Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Developing Dimensional Management For The Assembly of Aircraft Structures

Download as pdf or txt
Download as pdf or txt
You are on page 1of 300

Loughborough University

Institutional Repository

Developing dimensional
management for the
assembly of aircraft
structures
This item was submitted to Loughborough University's Institutional Repository
by the/an author.

Additional Information:

• A Doctoral Thesis. Submitted in partial fullment of the requirements for


the award of Doctor of Philosophy of Loughborough University

Metadata Record: https://dspace.lboro.ac.uk/2134/13683

Publisher: c David Jereys


Please cite the published version.


This item was submitted to Loughborough University as a PhD thesis by the
author and is made available in the Institutional Repository
(https://dspace.lboro.ac.uk/) under the following Creative Commons Licence
conditions.

For the full text of this licence, please go to:


http://creativecommons.org/licenses/by-nc-nd/2.5/
• • Lo~;~ghbprough
.,umvemty
University Library

Author/F!ltng T1tle •oo ,;.f.J;.IT.fl!;.i~f· ooD ~ 00 oooooooooooo

oo "'""""'" ''''" ''"' oo " o-o "'' ooo • •• ooo •••• ••• •••• ooo

Class Mark .. 00 0000 00 oo• 001, ooOOOO • oooo• oooooooo oooo• 00 00

Please note that fines are charged on ALL


overdue items.

0403411165

11111111111111111 11111 11 I 111111 I 111111 11111


DEVELOPING DIMENSIONAL MANAGEMENT FOR THE ASSEMBLY
OF AIRCRAFT STRUCTURES.

by

DAVID JEFFREYS

A Doctoral Thesis
Subnutted m partial fulfilment of the
reqmrement for the award of
Doctor of Philosophy
Loughborough Umversity

Loughborough Umverslty
Wolfson School of Mechamcal and Manufactunng Engmeenng
November 2006

© David Jeffreys 2006


-= Loughborough
Univcr~ity
Pllkmgton Ltbrary

'Jl\~
Date
leg
Class
I
~~~ ~0?>4--11 \ b5
ACKNOWLEDGEMENTS

The author wishes to thank

Academzc contacts
Dr Paul Leaney for his guidance and superviSion of this research proJect
Professor David WIIhams for his professiOnal guidance
Jagpal Smgh for all his pracl!cal support during the expenmentatwn

lndustrzal contacts
Graham Wood of Airbus
Kevm Fowler ofBAE SYSTEMS
Mark Lewis ofBAE SYSTEMS
Tom Kalkman ofUGS PLM SolutiOns.

To my wife Isabel and my two children Bethan and Thomas for their extended patience, love and
support.

lll
Abstract

This research IS concerned with the d1sciphne of dimensiOnal management (DM) and It's potenl!al for
meetmg the development and producl!on needs of future aircraft structures. The prenuse is that next
generation m1htary and clVll aircraft development may not take full advantage of an integrated
approach to DM resulting m a disconnected approach and generatmg silos of activity In response,
the aims of this research IS to propose and develop a lugh level DM methodology to address aircraft
structure vanat10n management, develop and review vanahon analysis tools to be used in key stages
of aircraft structure validatiOn, design and conduct an experiment focusmg on the effect of
d1menswnal vanab1hty on the assembly of carbon fibre composite panels to alunuruum substructures.

A literature review and an mdustnal study were undertaken This work h1ghhghts the need for a more
structured and mtegrated approach to DM WJtlun the aerospace sector. Particular product and process
attnbutes were 1denl!fied and the effect of uncontrolled vanatwn highlighted An mdustnal study of
key US and European aerospace and automol!ve compames Jdenllfied the nature of DM and a matnx
of workmg practices and analysis tecluJ1ques. were established

The proposed physical architecture robustness engmeenng system (PARES) methodology has been
developed m response. The key steps of the methodology have been defined and mapped onto a
genenc aerospace mtegrated development process. Based on the PARES framework, key
architectural chunks have also been defined and mapped onto a software solutiOn framework. A key
actlVlty of PARES mcludes the development of tolerance analysis tools and techniques at key
development stages

The PARES methodology outlines the need for Im!Jal front loaded assembly analysis at the early
stage of development before CAD geometry IS available followed later by more detailed analysis
based on 3D CAD geometry In response a parametnc vanatwn analysis (PVA) tool was designed
and developed for early tolerance evaluatton The one dimensional PVA tool demonstrates the
opportumty for early tolerance design on a wmg structure and how the analysiS results 1denllfy the
key obJecl!ves for down stream three d1menswnal geometnc tolerance studies Three geometnc
vanatwn analysiS (GVA) case studies were conducted on an A1rbus commercial aircraft wmg box
structure The case matenal evaluates wmg box structure vanatwn usmg a commercial three
d1menswnal tolerance analysis software tool, VSA

IV
The Impact of advanced mrcraft structure matenals and their manufactunng process were mvesl!gated
through a number of expenments based on the design, manufacture and assembly of composite panels
to alummmm substructures Key characteristics associated with the productiOn process of panel and
substructure was physically and d1g1tally mo-delled

The conclusiOns outhne the case for the development of a DM methodology, supported by analysiS
tools and techniques, to be incorporated as part of the mtegrated product development process for
aucraft structure development. The PVA tool w1ll provide an evalual!on of early design mtent before
maJor geometry design and assembly spec1ficatwn has been cmmrutted Th1s front loaded analysis
will help detenrune the key areas for further detailed study which will be performed by the GVA tool
when 3D CAD, manufactunng, and assembly process have been defined Further work is Iden!Jfied
addressmg the opportumty to extend the DM methodology, to develop the capab1hty and scope of the
analysis tools and techniques, and to position and mtegrate these tools agamst ex1stmg engmeenng
dJSCiplmes. One particular area Jdenllfied for exploratiOn IS the mtegration of fimte element and
geometnc dimensional vanatwn modelhng techniques Tlus mtegratwn would present an opportumty
to better evaluate the effect of confonrung (flexible) parts and subassembhes as part of the digital
mock-up actlVlty

V
Table of contents

Title page
DeclaratiOn 1l
Acknowledgements 11l
Abstract IV
Table of Contents VI
List of figures X
Acronyms Xll

1 Introduction ................................................................................................................................ 1
1.1 Background .. . . . . ... .. .I
I2 Aims and obJectives . .... .. . .. . ... .4
I3 A guide to the thesis and the research approach .................................. . .................. 4
2 Literature review........................................................................................................................ 6
2I Manufactunng methodologies chronology ...... . .6
2.1. I Systems engmeenng . .. . . . . . . ... .. .9
2 I. I I Systems engmeenng defimtions 10
2.1.1 2 Defimtwn of a system II
2.1.1.3 Systems versus component thinking. II
2.1. I 4 Systems engmeenng process . . .... .. .. 11
2.1.2 Concurrent engmeenng . .... . . . . 13
2.1 3 Quality functiOn deployment . ... I6
2 I .4 Design for manufacture and assembly . .. I8
2 I .5 Failure mode and effect analysis ... ......... I9
2.1.6 Robust design and quality engmeenng . 20
2I6I Robust design 20
2.I 6 2 Quality engmeenng 21
2 I63 Statistical process control .. .. 22
2 2 Manufactunng orgamsatwnal Issues. . . 22
2 2. I Lean manufactunng. ... . .. .. . . 23
222 Integrated product and process development ... 24
2.2.3 Ag1le and flexible manufactunng. .. . . 26
23 Development of aircraft structures. . . 28
2.3.1 Matenals in aircraft structures ... . 29
2.3 2 Aircraft assembly technologies .... 3I
24 DimensiOnal management discipline .. . .3I
2 4.I Product features.. . .. .. ........ . . ... .34
2.4 2 Key charactenstics . . ........ . . ... .35
2.4.3 Tolerance specificatiOn. . .......... . .37
243I Tolerance standards . .38
24 32 Tolerance specification . . .40
2433 Tolerance analysis and synthesis. .. . .. .4I
244 Computer aided tolerance software tools and process ... . . . . .. . .42
25 Manufactunng methodology cnhque . . . .. ... . . . .43
2.5 I DM cntique . . .. ............ . .44
2 6 Concludmg comments . .. . ... .. . ... . 45
3 Assessing current best practice in dimensional management ............................................. .46
3 I USA study tour of leadmg aerospace and automotive compames . . . .. .. .. .. . . . . .46

VI
3 1.1 Chrysler Corporation. . ..... .... 47
3.1 2 Ford Motor Company . ....... 49
3.!.3 McDonnell Douglas Aerospace .... ....... 51
3I4 Raytheon Aircraft. 52
3 .!.5 N orthrop Grumman . .. . . . .. . ....... 53
3.! 6 The Boemg Company 55
32 DimensiOnal management WIIhm European aerospace compames . .. 56
3.2.1 BAE SYSTEMS (1998). 56
3 2.2 Swedish Aeroplane AB (1997). .. . ..... .. 57
3.2.3 Aubus (1998) ... . .. . .. . .. . .. 58
33 DimensiOnal management case examples 58
34 Dimenswanl Management best pracl!se cnl!que. 61
35 Concludmg comments. .. . . 62
4 The need for dimensional management ...........................................•..................................... 63
4I The need for a dimensiOnal management methodology. . ....... 63
42 Manufactunng challenges for future aircraft . ........... 64
42I Military fixed-wmg aircraft . . ... 65
4 2.2 Commercial fixed-wmg mrcraft .. . ........ 68
4.3 Supportmg acllvities/techniques to aid the DM Process . . ........ 69
5 Physical architecture robustness eugiueeriug system (PARES) .......................................... 73
5.! The PARES scope .. 73
5.2 PARES methodology process, tools and techniques. 74
5.3 Methodology arclutecture . 82
54 PARES benefits. . .. . . 86
6 Development of a parametric variation analysis (PVA) tool for wing box assembly ........ 88
6.1 PVA tool development background . . 88
62 Reqmrements specification and scope . 90
63 Design and development methodology 91
64 PVA tool functiOnal overview. 92
64 I Analysis configuratiOn .. 93
642 Parameter feature defimtwn 94
643 Tolerance specificatiOn .. .. 94
644 SimulatiOn parameter defimt10n 95
645 Analysis output.. .. 97
646 Knowledge capture 99
65 EvaluatiOn of the PVA tool. .. . .. .. 99
66 Benefits and limitations of the PVA tool .... 103
6. 7 Further development work .. . .. 104
68 Concludmg remarks.. . . . ... .. . . .. .. . .... 105
7 Geometric variation analysis (GVA) on proposed Airbus wing box assembly ...•......•.•••. 106
7I Background to Airbus A3XX case studies . . . . . 106
72 Case study analysis methodology .. 107
73 DimensiOnal analysis for nb to complex skm assembly. 108
73I Analysis mms . .......... . ..... 108
732 Analysis objectives . ... . ... . ... 109
733 List assumpl!ons . ... . .. . ... 110
734 Design of assembly components and fixture. . ... 110
735 Idenl!fy key features and tolerances .... 111
7.3.6 Create CAD assembly model .. . ..... 115
7 3. 7 Define assembly sequence ...... . . ... 115
7.3.8 Idenllfy assembly moves .... . . .... 115

Vll
7.3 9 Define analysis measurements .116
7 3 I 0 Perform assembly sJmulatwns. .117
7.3 11 SimulatiOn results and analysis concluswns .. .. .117
74 D1menswnal analysis for Simple nb to skm assembly ..... 118
7.4 I Analysis mms. . . . . .... . 118
742 Analysis obJeclives. .. . ..... . 118
743 Define assembly sequence .. . . 119
744 SimulatiOn results and analys1s concluswns 119
75 D1menswnal analys1s for C spar eqmppmg assembly ... 120
75I Analysis a1m . . . . . . . .. ... ... . . . . . .. . . .......... 120
752 Analysis objeclives . . .. ... . . .. .... . .. . . . . . ... . . ............. 120
7 5.3 Define assembly sequence ......... . .. . .... .. 121
7 5.4 SimulatiOn results and analysis conclusiOnS .............. 122
76 Overall concluswns . . .... . ........ ····· .. 122
8 Assembly of carbon fibre composite panels to aluminium substructures ........................ 124
8I A1ms of the expenment . . .. 124
82 ObjeCtives ofthe expenmentatwn. . .. 125
83 Des1gn of the expenment . .. .... . . . 126
8 4 Des1gn of panel and substructure .... . .. .. . . .. . . . .. . . . ... . . . ...... .127
85 ldenlificalion of matenals and manufactunng process ..... 128
85I CFC panel manufactunng process 129
852 Alunumum substructure manufactunng process. . ........... 130
8.6 Des1gn of the assembly process .... . .. ... .. .. 131
8.7 Detenmnatwn of assembly key charactenstics .. 132
8 8 Detenmnatwn of component des1gn tolerances 135
89 ldenlify quahty measures for mvesligatwn 138
8 I 0 Design of mspeclion methods . ... . . . ..... .139
8 I 0.1 Controlled environment mspectwn methods .............. .140
8.10 1.1 Surface profile error.. . . . . . .. . . . 140
8 I 0 1.2 Step and gap error. . . . .. ......... . 140
8 I 0 1.3 Vntual hole cond1lion error . . ....... . 141
8 I 0.1 4 Multiple temperature environment mspectwn method ... .. 142
810.1.5 CMM mspectwnusmgVALISYS VS 3 I software . 143
8.I I VSA tolerance analysis .. . ..... . ... 144
8.12 Expenmental results . .. . .. . . . . ..145
8 I 2.1 Geometnc surface form error of panels and substructure. ..... .145
8.13 Step and gap mspectwn results.. .... . !50
8 14 Hole conditiOn of panels and substructure . . . ....... !55
8.15 Assembly dowel pmmspeclion results . .. . . .... . . 160
8.15 I Multiple temperature environment mspectwn results .. . .. .. . ... . 163
8 16 VSA tolerance analys1s results .. ... . . .. 167
8.16 I Analys1s objeclives .. ......... .167
8.16 2 Modellmg assumptiOn . .. . . .. ............ 168
8.16 3 Assembly sJmulatwn .. . .. ..... . .168
8.16 4 Component datum and tolerance .. . ...168
8 16 5 Analysis measurements .. 169
8.16 6 Analysis results .... ............. 170
8.17 Expenment dJscusswn . 171
8.18 Concludmg remarks. . . . ... . ... 175
9 Discussion .................................................................................... - ....•.............................•....•. 177
9I Rev1ew of d1mensJonal management .... .177
92 The need and nature of d1menswnal management . . ................ .178
93 PARES; proposed methodology.. .... .. . .. .. . . . . . 179

vm
94 Development of a PVA assembly analysis tool 181
95 Case study for GVA assembly analysis tool . . 182
96 CFC panel to Alummmm substructure expenment .... 183
10 Conclusions and further work .............................................................................................. 185
10 I Conclusions .. .. . . .. . . .. .. ... . . . ..... .. . .. 185
I 0 I I The need and nature of dimenswnal management . . . 186
10 1.2 Proposed dimensiOnal management methodology, PARES .. 187
I 0 I 3 PVA tool development for early wmg box analysiS .. . 188
I 0 1.4 GVA case studies on wmg box structure . . . ..... . .. . 190
I 0 1.5 CFC panel to alummmm substructure expenmentatwn.. . 191
I 0 2 Further work . . . . . .. . .. . .... . 192
10 2.1 DM methodology and process 192
10 2 2 DM tools and techniques. . 193
10 2 3 DM expenmentahon ..194

References 195

A endices 203
Tools and techniques for SPC 204
2 Computer aided tolerancmg (CAT) software tools and process 208
3 Analysmg tolerance accumulation 222
4 A dimenswnal control methodology for the vanatwn analysis of aircraft 225
structure assembly (Pubhcahon by author)
5 Dimenswnal control as an mtegral part of next generatiOn mrcraft 232
development (Pubhcat10n by author)
6 Stage I - VSA panel to substructure tolerance analysis results 239
7 VSA process reports 245
8 Different methods of constructmg tolerance zones m CAD systems 249
9 Wmg Box Assembly Analysis Tool- User manual 252
10 Pubhca!lon Review, Tolerance design A handbook for developmg op!lmal 257
specifications
11 Stage I - Vahsys measurement report 258

ix
List of figures
Figure I Systems engmeenng process overview (IEEE 1995) . . . . . ... .. .. ... . ..... .. 12
Figure 2 Sequential versus concurrent engmeenng methods (Prasad 1996) . .. 14
Figure 3 Lean and mtegrated system (LIS) concept (Rossetto and Franceschm1 1995).... 15
Figure 4 Common CE tools for product development (Syan and Menon 1994).. . 16
Figure 5 House of quahty matnx schematic (Syan and Menon 1994) ... . 17
Figure 6. The HOQ matnx cascade (Molloy, T!lley and Warmam 1998) .. . . 17
Figure 7. An example FMEA ratmg chart (Fox 1993) .. 19
Figure 8 The quadratic loss functiOn (Taguclu 1993). .. 21
F1gure 9. H1gh level IPPD process (JGA, lnc 1999) ..... . 25
Figure I 0 Classification of different flexJbihtles (Raouf and AnJuml995) 27
Figure 11 Case examples ofDM m the aerospace mdustry ... 59
Figure 12 Case examples ofDM m the automotive mdustry 59
Figure 13 Advanced hghtwe1ght aircraft fuselage structure (ALAFS 1997) 67
Figure 14 PARES high level DM program .......................... . .. 74
Figure 15 IPD/DM team allocatiOn agamst genenc airframe zones ..... 75
Figure 16 PVA tool mterface . . ... . . .. . ...... . .. 76
Figure 17 Feature speclfical!on taxonomy ... . . . 76
F1gure 18 KC assembly flow through ... 77
F1gure 19 PKC vanatwn outputs.. . ...... . . 78
Figure 20 Component SPC database mcludmg Cp and Cpk . 79
Figure 21 Post FAI SPC data defimhon and collectiOn. . .. ..81
Figure 22 PARES top level architecture chunks 82
F1gure 23. Overall IDEFO representatiOn of the PARES process 85
Figure 24 Fmt level IDEFO represenatwn of the PARES process. 86
Figure 25. Example wmg box structure w1th top skm removed, courtesy of Airbus .89
Figure 26 Components w1thm the genenc wmg box .. .... 91
Figure 27 PV A process .. . . 92
Figure 28. Wmg box assembly process defined for this analysis . . 94
Figure 29 Data dnver table, D nose to front spar assembly example .. 95
Figure 30 Example frequency chart forD nose .. . .. 98
Figure 31. Example sensihvity chart forD nose .. .... 98
Figure 32. Example statJstJcal chart forD nose .. . . 99
F1gure 33. S1mple 4 part assembly ... lOO
Figure 34. V1sVSA process report .. 100
Figure 35. V1sVSA contnbutor chart .. .. 101
Figure 36. VisVSA statJsl!cs table .. .. 101
Figure 37. PVA process chart . . . . 101
Figure 38 PVA sensitiVIty chart . ..... 102
F1gure 39 Statistical chart . . . . . . ... ...... . 102
Figure 40 Contnbutor report companson . . 103
Figure 41 Example wmg box assembly w1th D nose and A frame, courtesy of Airbus . 107
Figure 42 Rib, complex lower skin and assembly fixture . .. . . .. .. . . ..... 108
Figure 43. Complex nb (dark shaded) to lower skin m genenc wmg box ... . . . . .109
Figure 44. R1b design schemahc . . . . . .. . ...... . . .110
Figure 45. Fixture design schematic .... . ... Ill
Figure 46. R1b cntlcal features and tolerances schematic ... .. . .......... 112
Figure 47. Complex skm cntlcal features and tolerances schematic ..... 112
Figure 48. Fixture cntlcal features and tolerances schematic . 114
Figure 49. Assembly sequence dmgram ..... . ... . .115
Figure 50 Measurement strategy for gap condition . 117
Figure 51. Assembly sequence diagram ... ........... 119
Figure 52. Measurement pomts for C spar eqmppmg .. . . 121

X
Figure 53 Assembly sequence diagram ............. . 121
Figure 54 Panel and substructure design- Stage I. .. . . . .. . ..... .... 128
Figure 55 Panel and substructure design- Stage II . . ....................... . 128
Figure 56 Lay-up and cure tooling plates- Stage I and satge II CFC panels. . . . . ........... . . 129
Figure 57 Sabre 400H and Wadkm V4-6 machine centres . 131
Figure 58 Panel cntJcal features- Stage I.. . ..... .133
F1gure 59. Substructure cnl!cal features- Stage I .. . . .134
Figure 60. Panel cntJcal features- Stage II ........ . .134
Figure 61 Substructure cnl!cal features- Stage II . . . ... .. . .. . ..... 135
Figure 62 Feature tolerance sepecification for stage I and stage II ................ . .... 138
Figure 63 The VALISYS procedure ............ . . .......... 143
Figure 64 Stage I- panel I and panel2 form error. . . .. . ..... . . .. .. . ...... 145
Figure 65 Stage I - Talysurf Senes li wavmess mspecl!on results. . ...... 146
Figure 66 Stage I - panel I and panel 2 thickness mspectwn results ........ 147
Figure 67 Stage II -panel I and panel2 thickness mspectJOn results ... . ....... 147
Figure 68 Stage I- panel I and panel2 form mspectJOn results .. . ..... . ..... 148
Figure 69 Stage I- substructure form mspection results . . . . . . . . . .. .. .. . ...... . . ... ..149
Figure 70 Stage I - panel I, panel 2, and substructure form mspectwn results .................... . .... 150
Figure 71 Stage I- panel I step conditiOn mspectwn results ..... . ... 151
Figure 72 Stage II- panel I step condJI!on mspecl!on results . . . . ...... 151
Figure 73 Stage II - panel2 step condition mspecl!on results .... . ................ 152
Figure 74 Stage I- panel I gap conditiOn mspecl!on results . ... . ... .153
Figure 75 Stage I- panel2 to substructure gap conditiOn mspectwn results... . . ..... 153
Figure 76 Stage II- panel! to substructure gap mspectwn results ... ... . .. . ..... .154
Figure 77 Stage II- panel 2 to substructure gap mspection results . . .. . . . ....... 154
Figure 78 Stage I- panel I, panel2, and substructure hole form mspectwn results ... . . 155
Figure 79 Stage 11- panel I, panel2, and substructure hole form mspectwn results . . .. .156
Figure 80 Stage I- panel I hole p1tch devJal!on inspecl!on results ..157
Figure 81 Stage I- panel 2 hole pitch deviatwn mspectwn results .. 157
Figure 82 Stage I- substructure hole p1tch deVJatJOn mspectwn results . . ... 158
Figure 83 Stage li- substructure hole pitch deviatJon mspecl!on results . . .. 158
Figure 84 Stage II- panel I hole pitch devJal!on mspection results .. . ..158
Figure 85 Stage II- pane12 hole pitch deVJal!on mspectwn results ..................... . ... 159
Figure 86 Stage I- panel I, panel2, and substructure hole diameter mspectwn results . . . .. 159
Figure 87 Stage 11- panel I, panel2, and substructure hole diameter mspectwn results .. . 160
Figure 88 Dowel pm mspecl!on results.. . . . . . .. .. ... . ... . .. . ...... . . 160
Figure 89 Stage I- panel I to substructure gap mspectwn conditiOn. ... . . 161
Figure 90 Stage I- panel I to substructure step cond!lwn mspectwn results ..162
Figure 91 Stage I- panel2 to substructure gap mspectwn results . . ... . . . .......... . ..162
Figure 92 Stage I- panel I to substructure step conditiOn results at different temperature .. ..163
Figure 93 Stage I- panel I to substructure gap condition results at different temperature ... ..164
Figure 94 Stage I- panel I and panel2 perpendJCulanty inspecl!on results.. .. ..164
Figure 95 Stage I- substructure lmear mspecl!on results. 165
Figure 96 Stage I- substructure perpend1culanty mspectwn results .. .. . ..165
Figure 97 Stage I- panel I hole diameter error at different temperature .... ..165
Figure 98 Stage I- substructure hole diameter error at different temperature .. . 166
Figure 99 Stage I- panel to substructure pm assembly condJtJon at different temp . 166
Figure I 00. Substructure datum and tolerance . . ... . .. . . . . . . .. . 168
F1gure I 0 I Panel datum and tolerance. . .. . .. ........ . 169
Figure I 02 Pm datum and tolerance . . . . . . . . . ... .... . 169
Figure I 03. Step, gap, and virtual hole condJI!ons to be measured. 170
Figure I 04. Gap condJI!on at locatiOn plane 4 substructure . . . 170
Figure 105 Step conditiOn at locatiOn plane 4 substructure..... . . 171
Figure I 06. Hole 2 VJrtual hole conditiOn. . . .. . . . . . .. 171

XI
Acronyms

A3XX Airbus future commercial large aircraft


AMT Advanced manufactunng technology
AI Alummium
ADCATS Association for the Development of Computer Aided Tolerancmg
AKC Assembly process key charactenstic
ANSI Amencan NatiOnal Standards Institute
ASME Amen can Society of Mechamcal Engmeers
AVA Assembly vanation analysis
BS Bntish Standard
BIW Body-m-wlute
BoM Bill of matenals
CAD Computer aided design
CAE Computer aided engmeenng
CAM Computer mded manufacturing
CAT Computer aided tolerancing
CE Concurrent engmeenng
CEM Complex electro mechamcal
CFC Carbon fibre composite
CIM Computer mtegrated manufactunng
CIRP Intematwnal InstitutiOn for ProductiOn Engmeenng Research
CMM Coordinate measunng maclune
CNC Computer numencal control
CPL Computer profile !me
DC DimensiOnal control
DCS DimensiOnal control systems
DDC Design for dimensiOnal control
DFA Design for assembly
DFM Design for manufacture
DFMA Design for manufacture and assembly
DFV Design for vanatwn
DM DimensiOnal management
DMU Digital mock-up
DOD Department Of Defence

Xll
DOF Degrees of freedom
DSA Damsh Standards Assoctallon
FD&T FunctiOnal destgn and tolerancmg
FEA Ftmte element analysis
FMEA Failure mode effects and cnticahty analysts
FMECA Fmlure mode and effect cnllcal analysts
FMS Flextble manufactunng systems
GD&T Geometnc dtmenswmng and tolerancmg
GM General Motors
GPS Geometnc product specificatiOn
GRP Glass remforced plastic
GVA Geometnc vanatwn analysts
HOQ House of quahty
ICAF Industnal College for the Anned Forces
ICY Interchangeabthty
ILS Integrated system support
IML lnstde mould !me
IPD Integrated product destgn
IPT Integrated product team
IPPD Integrated product and process development
ISO Intematwnal standards orgamsatwn
JGA J ames Gregory Assoctates
JIT Just-m-llme
JSF Jomt stnke fighter
KC Key charactensllc
LAI Lean mrcraft mtttallve
LEM Lean enterpnse model
LIS Lean and mtegrated system
LOF Lme of fltght
MTG Master toohng gauge
MKC Manufactunng process key charactensllc
NC Numencal control
NGMS Next generallon manufactunng system
OML Outstde mould !me
PARES Physical arclutecture robustness engmeenng system
PD&T Parametnc dtmenstoning and tolerancmg
PDM Product data management

Xlll
PDP Product development process
PKC Product key charactens!Ic
PLM Product hfecycle management
PVA Parametnc vana!Jon analysis
RCS Radar cross sectiOn
RMS Root mean square
RSS Root sum square
SE Systems engmeering
SPC Statistical process control
StatKC Statistical key charactens!Ics
STEP STandard for the Exchange of Product mformatwn
Ti Titanmm
TPS Toyota productiOn system
QAP Quality assurance program
QE Quality engmeenng
QFD Quality functiOn deployment
UGS Umgraphics Solutions
VM Vanatwn management
VR Vanatwn reductiOn
VPM VIrtual product manager
VSA Vana!Ion systems analysis
WC Worst case
WIP Work m progress

XIV
Chapter 1

1 Introduction

Objectives: Th1s chapter w11l provtde:

• A background to the work addressed in thts thests


• The atms and obJecttves of the thes1s.
• A gmde to the research approach and the structure of the thests

1.1 Background

The dtsctphne of dtmenswnal management (DM) was concetved wtthm the automotive mdustry
m order to better manage the effects of vanatton dunng destgn, manufacture and assembly
(Leaney and Marshall 200 I) through the use of various tools and techmques The mam dnver for
the tmplementatlon of DM ts a nght-first-ttme assembly of vehtcles from parts that vary m stze
One of the key areas for analysts has been the velucle structure known m the automotlve mdustry
as the body-m-whtte (BIW) More recently, dtmenswnal analysis has been extended to other
areas of the vehtcle mcludmg dnve tram, mtenor and extenor tnm, and suspensiOn. From tts
automohve ongms, DM has contmued to gam momentum m supportmg the concurrent
engmeenng actlvtty and has been adopted by a number of product sectors mcludmg electncal
whtte goods, nuhtary vehtcles, medical devtces, and heavy plant eqmpment.

Recently, the aerospace sector has begun to review and adopt some of the DM techmques in
support of aerostructure destgn and development. Both the commerctal and nuhtary atrcraft
sectors of the aerospace mdustry are facmg extenstve competltwn from a raptdly emergmg global
market when attempting to wm new and ongomg busmess contracts (Shumaker and Thomas 1998,
Marguet and Mathieu 1997) Thts has resulted m orgamsatwns mergmg mto large aerospace
conglomerates extendmg theu technical capabthty, thetr economy of scale effictency and thetr
busmess market share potenhal An example oftlus achvtty has been the amalgamatiOn ofBoeing
and McDonnell Douglas creatmg a substanttal company wtth a constderable share of the market
(Bradley and Pme 1997) Desptte these advantages aerospace manufactunng orgamsations have
recogntzed the need to mcrease engmeenng effictency m thetr current and future workmg
prachces. Thts has been achieved m part through the adoption of a number of mature automohve
best practice techniques such as concurrent engmeenng, JUSt-m-lime and lean production. Some
of the core obJectives for these 1mlia1Ives w1th respect to the aerospace mdustry are

• ReductiOn m aircraft development and production lead-limes


• M1mnuze a1rcraft development and production costs
• Develop busmess practices that are more agrle to market demands
• Improvements m design to reduce aucraft ownership costs.
• Organisation wide product and process quahty Improvement.

In response, some maJor aerospace compames have recogn1sed the success of DM m the
automottve mdustnes to meet these objecttves and ach1eve nght-first-ltme assembhes. They have
mtroduced a DM process to support a1rcraft development programs, BAE SYSTEMS on the
Eurofighter Typhoon for example (Barrow 1997)

Whtlst 1t is poss1ble to apply the same technologies developed for the analysts of BIW structures
as a means of DM for the assembly of aucraft structures fundamentally there are two maJor
differences m the two mdustnes that have to be considered: (I) The automobt!e mdustry IS h1gh
volume and the aerospace industry IS by companson low volume; (2) The trend m aerospace IS for
reduced part-counts ach1eved through w1der mtroduclion of large structural parts moulded from
compos1te matenals. Consequently the atm of DM m the automottve sector IS rap1d productiOn at
low cost (usually by automatwn) whereas m the aerospace sector 1t IS the productiOn ofhtgh value
products to a considerably greater degree of prec1s10n. In add11ton, due to the htgh cost of atrcraft
parts customers are requinng mterchangeabthty of structural parts placmg a greater demand on
DM m aerospace that does not ex1st m the automobile sector

D1menswnal vanat1on exists m all a1rcraft physiCal archttecture hardware (manufactured


components, subassembhes, assembhes) and cannot be ehnunated It IS mhented from des1gn,
manufactunng, assembly and the mspectwn processes (Cra1g 1992). Management of vanat1on IS
achteved through the spec1ficatton of tolerance to nommal d1menswns of size and form. An
important part of the des1gn process should be the detenmnatton of nonunal d1menswns and the
apphcatton of an effecttve tolerance type and metnc These need to be adnumstered such that
each md1v1dual component w1thm a structure w1ll meet the overall performance mtent of the
des1gn Many of these 1ssues are covered by 'blanket' company standards for the specification of
component tolerance wh1ch are then left to vanous mterpretattons by the down stream aci!VJI!es of
manufactunng, toohng, assembly and mspect10n The end result IS shop floor assembly fitters
'adjustmg' out vanat10n problems where ever posstble, leadmg to wasted time and money, as well
as compronused product quahty

2
This prachce IS possible where pnmary and secondary aircraft structures are manufactured
predonunantly from alunumum matenals However, the next generation of commercial and
m1htary aircraft programs are mcorporatmg more advanced matenals and associated
manufactunng processes such as carbon fibre composites for example A charactenshc of this
new matenal1s Its ab1hty to demonstrate a level of geometnc mstab1hty resulting from a complex
manufactunng process In addi!ion, due to their mherent properties these new matenals will not
be sympathehc to trad1!1onal assembly 'adjustment' techniques These Issues coupled with the
demand for higher mterchangeab1hty and mamtenance spec1fica!lons will require marked
Improvements m component, subassembly and assembly dimensiOnal quahty management. These
Improvements will best be achieved through the deployment of a concurrent and mtegrated DM
methodology

This thesis IS concerned w1th the development of methodology for DM m the concurrent design
for manufacture and assembly of aircraft structures High-level methodologies exist m the
automotive sector but such a methodology does not yet exist for the requirements of the aerospace
sector It will mves!Igate the potenhal of DM m meetmg the design, manufactunng and assembly
needs for the successful development and productwmsatwn of robust aircraft structures Both the
automotive and aerospace mdustry sectors will be reviewed to 1denhfy key DM elements and their
effect on product development The opportumty IS to be explored by proposmg a DM
methodology which begms at the early stages of design and contmues through to senes
productiOn. The methodology will be supported through the design and development of a
parametnc vanatwn analysis tools designed to vahdate aerostructure design for manufacture and
assembly at the early stage of design In addi!ion, an expenment focusmg on the effect of
d1menswnal vanab1hty on the assembly of carbon fibre composite panels to alununium
substructures will be undertaken to idenhfy the key Issues associated w1th the use of new
matenals and manufactunng processes proposed for future aircraft structures.

Th1s research work has two hypotheses:

1. An extended DM methodology for aerospace assembly manufacture will proVIde significant


advantage to support a systemahc approach for management of complex aircraft structures
n. A DM methodology can be enhanced through the use of supportmg software tools and
techniques

3
1.2 Aims and objectives

The atms of the research are

1 To mveshgate the use of a DM methodology to support the manufacture and assembly of


complex aerospace structures
n To explore the use of software tools to reahse the DM methodology and tllustrate tts
apphcatwn.

The maJor objechves of the work are·

• Propose, develop, compare and test a htgh level DM methodology to support the
development of atrcraft structures.
• Design, develop, and test a ID parametnc vanatwn analysts tool tatlored to be used
dunng the early stages of product destgn to atd the allocatiOn of tolerance metncs m
atrcraft wmg structure manufacture and assembly Undertake a number of geometnc
vanatwn analysts case studtes on atrcraft wmg assembhes usmg a commercial 3D
tolerance analysts software tool based on dtgttal geometry
• Undertake an expenmental mvest1gat10n mto the fundamental nature of dtmenswnal
vanatwn on the assembly of CFC panels to alurrumum structures of planar and cyhndncal
geometnc form

1.3 A guide to the thesis and the research approach

The nature of vanatwn makes any DM dtsctphne both complex and diverse It encroaches upon,
and can naturally be mtegrated wtth, a number of formal manufactunng techmques rangmg from
quahty function deployment at the early stage of destgn nght through to statishcal process control
at atrcraft senes productiOn Much of the work m thts thests has been formed through extenstve
mvolvement wtth the aerospace mdustry and ts reflected m the research approach The approach
conststs of:

Literature review: Identlfy past, current and planned areas of work bemg undertaken m the area
of DM relatmg to atrcraft structure development The boundary hrnits of DM may also be defined
wtth respect to the thests

4
Dimensional management best practice bench marking: A number of automo!ive and aerospace
compames m the USA who were perceived as bemg Ieadmg participants of DM were VISited and
bench marked Orgamsatwns were tested for DM Implementa!ion and scope.
Assessing the need for dimensional management: Based on the findmgs from acadenuc review
and mdustrial current best prac!ise, the need for DM to form part of an mtegral aJrcraft structure
development program IS mves!igated
Methodology and variation analysis tools development: Design and development of a
methodology and supportmg vanation analysis tools to smte the particular needs of the aerospace
mdustry
Industrial case studies: An industnal placement and a number of mdustnal based case studies
have provided evidence for the need for a structured and comprehensive approach to the DM
dJsciphne for aJrcraft structure analysis The aerostructure case study mforrna!ion feedback has
also provided an opportumty to mfluence the DM related practices w1thm the aerospace mdustnal
orgarusatwns
Experimentation to investigate the assembly of carbon fibre composite panels to aluminium
substructures: Conduct expenmentatwn to h1ghhght Issues resultmg from the assembly of
carbon fibre composite panels to alunumum substructures, a particular charactenst1c of some
future aerostructure designs. The mves!igatwn will mclude the effect of temperature change on
both matenals and their resultant assembleab1hty

The thesis has been structured to reflect the approach taken above In the followmg sec!ion the
literature review provides a background to the commercial and m1htary aJrcraft mdustry,
1denlifymg manufactunng methodology and orgamsatwnal Issues, and prov1dmg an overview of
the DM d!SC!phne

5
Chapter 2

2 Literature review

Objectives: This chapter revtews the hterature m the domam relatmg to DM wttlun the context
of aerospace and automottve destgn for manufacture The revtew mcludes

• Manufactunng methodologtes chronology


• Manufactunng orgamsatwnaltssues
• DM dtsctplme

The aims of tlus thests to develop a htgh level DM methodology and assoctated analysts tools to
support the development of aerostructures To first understand what ts reqmred a reVJew of
current manufactunng methodologtes and thetr background ts presented m the followmg secttons

2.1 Manufacturing methodologies chronology

Between the 1950's and the 1970's, nuhtary aerospace orgamsatwns based wtthm Europe and the
US have tradttionally had one customer, thts bemg thetr own respecltve governments (Bradley and
Pme 1997) Conversely, the governments of the day tradttwnally only constdered one suppher
whtch was thetr domesttc aerospace mdustry Thts mutual monopohsttc sttuatwn resulted m httle
mcenltve amongst the aerospace compames to tmprove busmess effictency and reduce nuhtary
atrcraft development and produclton costs

Dunng the 1980's the end of the cold war and a down turn m the world economy signalled a
change wtthm the aerospace manufactunng sector. In order to reduce defence expendtture,
western governments started to look outstde thetr domesltc mdustry. Aerospace sector
manufactunng orgamsal!ons now had to respond to the threat of global market compel!!ton. Thts
threat, coupled wtth increasmgly aggresstve product demands, tnggered a number of aerospace
development and productiOn imltaltves. As a result, manufactunng process technology began to
mcrease m complextty across manufactunng mdustry (Leaney and Marshall 200 I)

In the early 1970's destgn for manufacture (DFM) and destgn for assembly (DFA) became drawn
to the forefront of manufactunng (Leaney and Wtttenberg 1992). The techntques emerged m

6
order to address the trend towards the automation of small complex products requmng manual
assembly As concepts such as flexible manufactunng systems, flexible manufactunng
technology grew and With the impendmg mtegration of robotic assembly the mterest in design for
manufactunng and assembly (DFMA) mcreased It became clear that the assembly design
philosophy of the day would not transfer well mto an automated Situation and a new des1gn for
assembly (DFA) technique would be a way of addressmg th1s As DFA became more accepted,
several techn1ques were developed of wluch the best known are:

Assembly evaluatiOn method (AEM), by H1tach1.


" Boothroyd method, by Prof G Boothroyd and Dr. Dewhurst (Boothroyd, Dewhurst and
Knight 1994).
m. Lucas method, by the Lucas orgamzatwn and the Umvers1ty of Hull

C1rca 1978-1983 concepts of large scale flexible automatiOn began to dormnate future
manufactunng strategies 11us d1rectly led to ph•losoph1es such advanced manufactunng
technology (AMT) and flex1ble manufactunng systems (FMS) The 1mpendmg dnve for flexible
automatiOn led to research mto the use of robotics as a resource for the cheap and reliable
assembly and manufactunng of products Dunng the later part of this penod 1t was realized that
this scenano would never be realized on the scale first suggested mostly due to the lim1tatwn of
robolics mtelligence systems The declimng trend towards AMS and FMS was also due to the
mtroductwn of Toyotas Just-In-Time (JIT) philosophy. The JIT approach paved the way for
'lean' produclion wh1ch encouraged the !denlificatwn and reductwn of non-value addmg enlilies
such as matenal storage, excess1ve work-m-progress (WIP), product scrap and rework (Womack,
Jones and Roos 1990) The emphasts was to reduce all manufactunng buffer levels to a rmmmum
m order that quality related defects would become more VlSlble. Thts could only be achieved
through improvements m batch set up and changeover performance whtch in turn would tmprove
manufactunng flextbtlity (Mclntosh, Culley, Mtleham and Owen 2001) To reduce buffer stocks
and to ensure limtted manufactunng dtsruption, the need for process management became
apparent and thts became addressed by stalislical process control (SPC).

The techntque of DFMA, FMA, and AMT began to htghlight the potential of constdenng the
destgn, manufactunng and assembly processes simultaneously at the early stage of product
development. Thts led to the development of stmultaneous engmeenng which developed further
mto what has become to be known as concurrent engmeenng Concurrent engineenng (CE) was
seen as a key to achtevmg compelittve advantage through the development of htgh-quality, lughly
functiOnal products produced effictently through the synergy of mtegrated product and process
destgn, whtle also constdenng multtple life cycle factors such as funclionality, servtceabtlity,

7
manufacturabtiity, marketab!ltty and recyclabthty (Molina et a/ 1995). CE focused on the needs
of corporate structure and encouraged the busmess enterpnse to dtssolve departmental boundanes
allowmg greater cross funchonal commumcatwn

The CE paradigm constdered product development through the integrat10n of technologtcal based
computer tools such as computer atded destgn (CAD, computer atded manufacture (CAM),
computer atded produchon plannmg (CAPP), and computer atded analysts (CAE), (Bedworth,
Henderson and Wolfe 1991) These elements help generate the framework for a number of
enterpnse modelling techntques and thetr fus10n mto a central database ts percetved to be the key
to successful computer mtegratwn. Thts vtsion has mostly been ptctured as computer mtegrated
manufactunng (CIM) As computer and software technology develops and mtegrates wtth
manufactunng methodologtes, increasmgly the mdustry finds ttself destgmng, developmg and
testmg products m a 'vtrtual' envtronrnent One of the fundamental problems of vtrtual product
development ts their current mabtlity to address some real life problems, for example dtmensional
vanat10n related tssues Generally, destgn orgamsations only validate theu work m a nonunal
(perfect geometnc form) condthon wtth httle constderatwn gtven to the many sources of
manufactunng and assembly vanatwns actmg on that nommal A developmg techntque destgned
to address thts tssue ts DM A number of DM software tools have been developed wtthm the
CAE domam for a vanety of analysts mcludmg tolerance and datum reference evaluat10n on
product and process, assembly sequencmg and constramt methodology, and quality standards
va1Idat10n, predtcted aerostructure external step and gap dtmenswnal vanat10n (Whttney 2003).

Concurrent plulosophtes led the move towards an mtegrated product team (!PT) onentated
approach to product development and mtegrated product destgn (IPD) techntques began to
emerge IPD concepts, also known as mtegrated product and process development (IPPD), were
dnven out of complex product development and first emerged m establishments provtdmg
products to mtiitary based customers m the USA (Shumaker and Thomas 1998) Currently the
Defence Advanced Research ProJects Agency (DARPA) ts developmg the raptd destgn
explorahon and ophnuzahon (RaDeo) program, (formally MADE), (Whttney 1997) Thts
program wtll focus on problems of destgn and manufacture of complex electro-mechamcal (CEM)
ttems, an area that does not have a smgle techntcal focus or well developed engmeenng models
accordmg to Whttney (1997) CEM products, such as nuiitary aircraft, are tradttionally developed
vta a systems engmeenng (SE) approach where components and systems are destgned together
SE methodology reqmres a product rea1Izat10n process from a systems-onentated approach as
opposed to a component-onentated problem (Prasad 1996) The approach relies on the
decomposihon and classtficahon of the product m order to tdenhfy stakeholders wttlun an
orgamsat10n who become responstble for each product development sechon. The SE approach

8
exploits a transformatiOn system for product reahzahon design. The aim of the transformation
strategy IS to uncouple the system mto mutually separable transformatiOn states so each only
affects only one set of outputs, based on Suh's first axwm (Suh 1990) Each state IS then
modelled and reconstructed to give a systems defimhon from the aggregation of the defimhons of
Its constituent states.

Other less known methodologies that have emerged recently are·

Conformabzlzty analyszs a knowledge based design technique for the predzctwn of vanab1hty
nsks m component manufacture and assembly (Batchelor and Swzft 1996).
Correctzve actzon a response to problems encountered by manufactunng firms dunng fabncation
and assembly of the product, such as assembly of mechamcal structures hke automobiles and
mrcraft (Cunmngham 1997).

A number of aerospace manufactunng methodologies have matured and currently play a maJor
role m engineenng two of which have already been Identified as SE and CE These
methodologies are reviewed in more detail m the next sectwn

2.1.1 Systems engineering

The concept of SE IS not a new one. A number of books have referenced the concept back to as
far as the Second World War (Chase 1974; Goode and Macho! 1959) In Its origmal form SE
tended to be used by mathematicians, electncal and aerospace engmeers, and other scientific
diSCiplmes m order to define, analyse and manage complex problems Over the years
manufactunng engmeenng orgamsatlons have adopted the SE concept m vanous different forms
This has commonly occurred m orgamsatwns manufacturing CEM product types associated with
h1gh complexzty, cost of manufacture, and high nsk. The aerospace mdustry IS a good example of
this (Loure1ro 1999) Organisations reahsed that to manage the development of CEM products
they needed to mtegrate product development teams Methods of CE have focused on
simultaneously solvmg expensive design changes some way down the development cycle but falls
short in Its ab1hty to framework manage all aspects of complex design and development SE IS
deszgned to mtegrate With the CE methodology through a multidisciplme approach to the
defimtwn, analysis and venficatwn of a product design and development

In order to clanfy and mtroduce consistency to the SE concept in 1990 the mtematwnal council
on systems engmeenng (INCOSE) was formed The mam objective of INCOSE zs to develop a
better understandmg of SE through Improved defimtwn and clanficatwn of concepts and

9
techniques Their work partly conhnues through an annual symposmm held at dtfferent locatiOns
around the world There are currently two mam standards whtch descnbe a process for
engmeenng a system The first ts a US natiOnal standard EIA 632 developed by the electromc
mdustnes association (EIA 1997) and the second IS IEEE 1220 developed by the mshtute of
electromc and electncal engmeer (IEEE 1995). More recently the EINIS 731 have been
developed and are available m the pubhc domam Both the EIA 632 and IEEE 1220 standards are
based on MIL-STD-499B

2.1.1.1 Systems engineering definitions

Several defimhons of what SE IS, and what It does, exist Blanchard (1998), a well regarded
advocate of SE defines it as:

the effective apphcatwn of sctenttfic and engmeenng efforts to transform an operatiOnal need mto a
defmed systems configuratton through the top-down 1terattve process of reqmrement analysis,
functiOnal analysis and allocatiOn, synthesis, design opttm1zatwn, test and evaluatwn It mvolves
the des1gn engmeenng process of bnngmg a system mto bemg, With emphasis on an mtegrated top-
down hfe-cycle approach

Blanchard's (1998) defimhon ts based on the methodologies of the MIL-STD-499B. Current


literature on SE methodology and defimhon appears to be based on the IEEE 1220 standard

The EIA 632 ( 1997) standard defines SE as a systemahc approach m whtch to engmeer a system
The standard IS based upon a number of best prachces that have evolved smce Wold War Il. The
approach consists of two main elements

• DlVlsiOn of the problem mto multiple layers of manageable layers.


• The applicatiOn of a process set to each element.

Each of the defined processes are defined w1th regard to a number of gUidmg pnnctples, these
bemg the nght thmg should be done first hme, by the nght people.

The dtstmctwn between a system bemg 'hard' or 'soft' has been mtroduced by Checkland (1981)
The needs and objechves of a hard system can be well defined where as for a soft system they
cannot. An example of each concept IS g1ven by Parnaby ( 1981) The hard system elements refer
to hardware, processes, and I/0 stgnals The soft elements mclude soctal aspects such as human
mteractwn and are regarded outside the system domam.

10
2.1.1.2 Definition of a system

The word "system" has been denved from the Greek term "systema", whtch translated means an
orgamsed whole. The Colhns Enghsh Dtc!!onary (I 988) defmes the meamng as a "complex
whole; orgamsation" These basiC defim!!ons are vague and need to be expanded m order to fully
descnbe a system component and It's relevant to the SE concept.

There a number of defimhons and classtfica!!ons of what a system can be, Blanchard ( 1998},
Martm (1997), and Blanchard and Fabrycky (1990) These range from a CEM development
approach m support of rmhtary atrcraft destgn to naturally occuning phenomenon such as a river
system.

2.1.1.3 Systems versus component thinking

Smce the rmd-80s CE methodologies have developed m aerospace manufactunng related


orgamsatwns CE methods have supported what IS descnbed as the optirmsation of mdtvtdual
components m the behef that a product can be reduced to a collective number of self-supportmg
parts (Gormely and Maclsaac 1989) Organtsatwns pursumg CE methods began to constder hfe
cycle process reqUirements Components were bemg evaluated accordmg to CE formal methods
(DFA and DFM) to tmprove thetr mdtvtdual and collective manufactunng and assembly
suttabthty. As the complextty of products such as rmhtary atrcraft mcreased poorly structured
evolutionary development took place. CE fatlcd to provtde the framework management for the
destgn and development phases of mcrea smgly complex products The CE of mdtvtdual
components would help component evolutiOn, but only an mterdtsctphnary, collaborative
approach to denve, evolve and venfy a hfe-cycle balanced system can dehver better results that
meet customer expectations Thts approach is systems engineenng (IEEE 1995).

2.1.1.4 Systems engineering process

Martm ( 1997) suggests the SE process ts compnsed of three sub processes wtth mterfaces to the
design, mtegrated logts!!cs support (ILS}, productiOn and deployment processes

• SE management sub process.


• ReqUirements and architecture defim !!on sub process.
• System mtegra!!on and venficatwn (SI&V) sub process.

11
The SE process allows a product to be decomposed mto mdtvtdual component and element blocks
(chunks) at any level of the archttecture winch can then be passed on to a development team for
destgn. The SE concept recogntses each system block IS an mtegrated whole even though
composed of dtverse, spectahsed structures and sub functwns SE endeavours to optnruze the
balance of objecl!ves between each of the systems m order to achteve maxtmum compattbthty of
tts parts.

The systems engmeenng process atms to achteve the correct balance among operatiOnal
(performance), econonnc, and logtstical factors, m the evolutton of functiOnal detatl destgn
reqmrements (Bianchard 1998)

The SE process defines a complete system development begmmng wtth general requrrements and
endmg wtth a comphant product or process. A block dtagram of the SE process at tts htghest
levelts gtven below (IEEE 1995)

Process
Inputs-~ Rcvnanml

'""'-""'
ImP""

____, Requu·ements I
Analvsts Rtqllircmnt&
I con
Coa•tnlnt
mm
Requlrtments

-
~ Requuements TndeStudtl'$
Basebne &
Requlrt'ments Asus.sments
Baseltne
Valtdallon

,. -
Dffompo.ltiOftl
~Validated Requuements Ba.sebne Allonllot~
Tradr-oiTs & Systems
tmpafiS

Funcltonal
Dttomp<l!litl ... &
I
____, AnahsiS
llltquiNmtal
Funellonal
! Alkx'clotlllll

-
Functional Tr2de Studtes
·hit• tore &
Assessments
Functional
Venfh:atioQ

-
____,
l
Synthesis

J,
Venrted FunetJonal Arcbtttc111re

Phvstcal
I ""''"
Sol all Oil
Rrqllirnntalll
& Allft"llatlvn
""''"
Solatlo.
Tra&-dfs&.
lmparh
Analysts

Design Tnde
Studtl'S

- Pbulcal
Venficatlon
Archtte lure

IVerified Pbvstcal An:bilecture


Control i
&
Assrs.smmts

L Process
Outputs

Figure I Systems engmeenng process overview (IEEE 1995).

12
2.1.2 Concurrent engineering

The philosophy of CE IS not a novel one (Parsae1 and Sulhvan 1993) Its use can be traced back
to as far as World War II where product accomplishments such as the P-51 Mustang prototype
was designed and bmlt m only 102 days. The use of CE formal techniques such as DFM and
DFA are also recorded. Parae1 and Sulhvan (1993) compare the design and development of two
Word War II aircraft, the Supermanne Spitfire and the ME-109. The design of the Spitfire d1d not
consider any aspects of DFM and DFA resultmg m over 13,000 man hours of productiOn lime
Alternately the German ME-109's design d1d take these formal techmques mto cons1deratwn
resultmg m only 4,000 man hours reqmred to produce each aircraft.

Due to the recent globahzatwn of market forces most aerospace engmeenng orgamsatwns today
are faced With snmlar challenges, these bemg more demandmg customers, rapid advancements m
technology, environmental Issues, competihve pressure on quahty and cost, and shorter lead lime
to market w1th enhanced product features (Syan and Menon I 994) In the last decade western
manufactunng orgamsaiions have had to contnve better ways m winch to compete agamst
Japanese Imports which have mfiltrated and m some cases dommated the home product markets
Western manufactunng orgamsaiions began to develop technological solutiOn w1th the a1m of
recaptunng these markets with tools such as CADICAEICAMICIM It soon became apparent that
JUSt usmg a wholly technical approach to product manufacture would not be effechve The
ex1stmg engmeenng pracllces, frameworks, and product management methods were not
compallble With the new technology based strategies A new approach was reqmred
encompassmg all aspects of both product and process design/development which took m a much
broader v1ew of the eniire manufactunng actiVIty The CE methodology, also known as
simultaneous engmeenng, concurrent design, hfe-cycle engmeenng, has been developed to full-
fill this functiOn

The CE methodology IS descnbed by Kusmk (1993) as a pracllse of mcorporallng vanous life-


cycle values into the early stages of design. These values include a product's aesthetics,
manufacturability, assembly, serviceability, recyclab1hty, and not JUSt Its primary functiOns Life-
cycle design methods consider all the phases of a products bnngmg mto bemg through to disposal,
these phases bemg design, development, productiOn, d1stnbutwn, usage and disposal. CE IS
designed to overcome the disadvantages of operatmg a sequenllal product design and development
process, traditionally pracllced m the West (Prasad 1996). Sequenllal engmeenng mvolves the
diSJOinted efforts of successive engineering sectiOns completmg their functiOnal actiVIties and
then transfemng all product related mformation and hardware to the next stage. This approach IS

13
also known as senal engmeenng, tune-phased engineenng, and 'over the wall engmeenng' A
companson of sequential and concurrent engmeenng 1s g1ven below (Prasad 1996)

Information flow

-
~------------------

Errors changes and correct1ons

THE SEQUENTIAL ENGINEERING PROCESS

lowest overall life-cycle costs

Problem prevention mstead of -


problem solvmg and re-des1gn

THE CONCURRENT ENGINEERING PROCESS

F1gure 2 Sequential versus concurrent engmeenng methods (Prasad 1996)

A w1dely accepted defimtion of concurrent engmeenng IS g1ven by Wmner ( 1988).

a systematiC approach to the mtegrated, concurrent design of products and their related processes,
mcludmg manufacturrng and support Tlus approach IS mtended to cause the developers, from the
outset, to consider all elements of the product hfe cycle from conceptiOn through disposal, mcludmg
quahty, cost, schedule, and user reqmrements

CE w1ll only be effective through the establishment of a number of elements accordmg to


Dowlatshahi (1994):

• InformatiOn systems
• CADICAMICAE
• L1fe-cycle engmeenng
• DFM and DFA.
• Orgamsatwn and cultural changes

The d1versity of computer software, hardware, and operatmg platforms coupled w1th the demand
for data shanng has forced the development of mtegrated commumcatwn protocols (Mohna et a/
1995) Max1mum cross functiOnal data and mformat10n transfer could only be ach1eved through
the exchange of data v1a neutral formats (t.e, IGES and STEP) The development of seamless

14
data mterchange elements has been slower than anhctpated and th1s m part has delayed the full
technological potential of CE.

Currently there IS a move to comb1mng CE w1th lean manufactunng (lean manufactunng IS

outlined later m th1s rev1ew). This concept ts referred to as lean and zntegrated systems (LIS), a
model of wluch IS shown below (Rossetto and Francesch1m 1995).

Ftgure 3 Lean and mtegrated system (LIS) concept (Rossetto and Franceschmt 1995)

The mam ObJectives of CE are summansed by Syan and Menon (1994)

• Decrease product development lead-lime


• Improve profitabtlity.
• Greater competitiveness
• Greater control of des1gn and manufactunng costs
• Close integratwn between departments
• Enhance reputatiOn of the company and Jts products.
• hnproved product quality.
• Promotion of team spmt

The methodology of CE ts supported by a number of diverse formal tools. These tools ensure a
conststent approach to engmeenng des1gn and development actJVJtles. The correct selectiOn of a
tool and 1ts timely use 1s Important to the CE process F1gure 4, shown on the next page,
represents the typ1cal uses of common CE tools m the product development process (Syan and
Menon 1994)

15
QFD ~IReqUirements I -......
defin1bon , )

OFD ~~I_
Problem-solv1ng __..,Concept : ~
DeSign for manufacture 'development }

OFD ---i~
-----=---~~J J
Deta1led
~~~~ FMEA des1gn
Des1gn for manufacture ~
QFO r-- -~

Problem-solvmg ~ Manufactunng -...._


Process FMEA systems-' )

QFO
Process FMEA
~ Process 1 ~
Taguctn --~eSig~
Process capab1hty ~- )
Poka Yoke SPC _ ~~
Problem solvmg _ _ . . Manufactunng' ~

'-~I
Serv1ce andj
tl

~~~~

Figure 4 Common CE tools for product development (Syan and Menon 1994).

There are a number of formal manufactunng analysis tools and methods used m support of CE.
These are·

• Quahty functiOn deployment.


• Design for manufacture and assembly
• Failure mode and effect analysis.
• Robust design and quahty engmeenng
• Stal!sl!cal process control
• Value engmeenng
An explanatiOn of these IS g1ven m the next sectiOn

2.1.3 Quality function deployment

Quahty funcl!on deployment (QFD) originated m 1972 at the Mitsubishi's Kobe shipyard Site m
Japan and was later adopted by Western engineenng orgamsations dunng the nud-80s QFD IS
designed to capture and translate 'the vmce of the customer' mto a set of structured reqUirements,
and then evaluate each proposed product capability m terms of Its Impact on meetmg those needs
(Cohen 1995). This techruque allows pnmary product concerns to be traced throughout
development prov1dmg a common and integrated approach to engmeenng and manufacture
(Leaney and Marshall 1998). Once a product has been defined, QFD enables a design team to
focus on key customer reqmrements, those elements that are defined as bemg very Important to
the customer (Bossert 1991).

16
The QFD process IS mullifunclional and normally implemented through a nuxed-<liscipline destgn
team, whtch may mclude personnel form marketmg, destgn, manufactunng, logtsl!cs, mamtenance
and finance. The process starts through the development of a translatton matnces often called the
house of quahty (HOQ) whtch progresses customer reqmre agamst engmeering reqmrement.

Ftgure 5. House of quahty matnx schemalic (Syan and Menon 1994)

The classic HOQ matnx (see figure above) w1ll be made up of SIX mam elements (Syan and
Menon 1994), namely.

Customer pnonttsed reqmrements (What hst)


2 Engmecnng charactensttcs (How hst)
3 Engmeenng correlation matnx (HOQ roof)
4 Customer requuements mterrclatmnsh1p
5 Customer preference chart used to assess relative product competitiveness

6 Cost and techmcal assessment used to allocate resources

A number of smular mterrelated matnces are then developed dependmg on the complexity of the
product under constderatwn. In each matnx cell a symbol IS entered whtch mdtcatmg the
relatwnshtp between the mdivtdual elements of the 'what' and 'how' hsts The symbols can be
custonused to smte any praclittoner's reqUirement but usually at least three categones; weak,
medmm, and strong are tdenlified

The number of matnces generated can vary but the most common set are htghhghted m the table
below (Molloy, Ttlley and Warman 1998):

HOQ Matnx No. The ~what' hst The 'How' hst


One Customer reqmrements Engmeenng charactensttcs
Two Engmeenng charactens.ttcs Component destgn
Three Component destgn Process planmng

Four Process planmng Production plannmg

Figure 6 The HOQ matnx cascade (Molloy, T11ley and Wannam 1998)

17
The QFD process provides a well defined structured framework enabling customer reqUirements
to be captured and dissolved mto mdividual component feature spec1ficatwn, throughout the
product life cycle.

2.1.4 Design for manufacture and assembly

Design for manufacture and assembly (DFMA) can be descnbed as set of good engmeenng
pnnc1ples and techruques focusmg on the integration of manufactunng and assembly cnteria mto
the product design process (Molloy, T1lley and Warman 1998) It IS esl!mated that up to 85% of
product development costs are comnutted dunng the design and plannmg phase (Andreason,
Kahler and Swift 1988, Nevis and Wh1tney 1989) It must be therefore reasonable to assume that_
If these pnnc1ples and techniques are used at the design phase there IS a huge potenl!al to save
manufactunng and assembly later m the product development cycle

The DFMA process can be naturally spht into two separate analysis, design for manufacture
(DFM) and design for assembly (DFA). The techniques associated with DFA are relal!vely
mature where as the concepts for DFM have developed more recently and relies on a close
worlang relatiOnship between the product design and manufactunng acllvity with the mm of
1mprovmg manufactunng performance (Corbel! et al 1995) It rehes on analysis that degrades
individual components mto mdlVldual features and validates them agamst planned manufactunng
process DFM tools such as design for maclunmg and design for sheet metalworlang have been
developed by Boothroyd and Dewhurst (Boothroyd, Dewhurst and Knight 1994).

DFA techniques have been dommated by three evalual!on methods. These are·

• Assembly evaluatiOn method (AEM) developed by H1tach1 m 1967, (Leaney and


Wlltenberg 1992),
• Boothroyd method developed by Prof. Geoffrey Boothroyd and Dr Dewhurst in 1980,
(Boothroyd, Dewhurst and Knight 1994),
• Lucas method developed by the Lucas orgamzal!on and the Umvers1ty of Hull 1989,
(Leaney and W1ttenberg 1992)

All three techniques have a smular approach mvolvmg the analysis of an assembly design to
ascertam Its proJected cost The aim of all the techniques IS to reduce the cycle lime and umt cost
of an assembly through a number of Improvements such as the reducl!on of components for
example

18
DFA techniques were designed to address small scale products which can be man handled, and
are made m large volume. DFA IS not so effective on large scale low volume products such as
mrcraft for example. There IS potential scope for a new DFA methodology wluch could evaluate
large scale complex products With a primary focus on assembly quahty attnbutes and not JUSt on
assembly cycle time and cost

2.1.5 Failure mode and effect analysis

Failure mode and effect analysis (FMEA) IS a CE tool designed to help predict what failures may
occur, what the effect of such failures might be on the functiOnal operatiOn of a system or product,
and what steps might be taken to prevent the fmlure or Its effect (Fox 1993). The pnmary mm of
FMEA IS to ehnunate (or reduce) potential product or system fmlure- effects through the
IdentificatiOn of important charactenst1cs that have to be measured, controlled and momtored
Types of failure are charactensed as (Pllt 1994):

• Seventy, the seventy of the failure m terms of customer dissatisfactiOn


• Occurrence, the probab1hty of the occurrence of failure.
• Detectzon, the probab1hty of the design or development process detectmg the failure.

An example FMEA ratmg chart IS shown below (Fox 1993).

Ratmg Seventy Occurrence Deteet10n


I Exceeds specificatiOn but not not1ced by Never Very h•gh -programme des1gn process
customer wlll detect failure
2 Noticed by customer but does not effect Very occasionally H1gh - programme des1gn process 1s
the product function hkely to detect fa1lure
3 Noticed by customer, mmor effect on Very occasJonally H1gh - programme des1gn process 1s
product function, customer accepts hkely to detect failure
conditiOn

4 Customer dissatisfied w1th function of OccasiOnally Medmm - programme des1gn process


product may detect fmlure
5 S1gmficant effect on customer Occastonally Medmm - programme des1gn process
sat1sfact10n may detect failure
6 S1gmficant mconvemence to customer Frequently Low - programme des1gn process 1s
unhkely to detect failure
7 S1gmficant annoyance to customer Frequently Low - programme des1gn process 1s
unhkely to detect failure
8 Customer endangered Very frequently Zero - programme des1gn process w11l
not detect failure

Figure 7 An example FMEA ratmg chart (Fox 1993).

19
2.1.6 Robust design and quality engineering

Dunng the reconstructiOn of the Japans manufactunng mdustry followmg World War II, Dr
Gemcht Tagucht developed a new methodology for the development ofhtgh quahty products and
processes. From hts research m the 1950's and early 1960's he developed robust destgn, a
concept whtch underpms the methodology of quahty engmeenng

The pnnctples of robust destgn and quahty engmeenng form the corner stone of the DM
dtsctphne. Robust destgn, quahty engmeering, and DM are all concerned wtth the evaluatton and
management of product and process vanatton, tt's potenttal causes, and tt's resultant effects
measured as a 'quahty loss' on product functiOnal attnbutes The current practtce of DM m the
__ UK and USA manufactunng mdustry ts primanly concerned wtth product phystcal arclutecture _
quahty attnbutes, for example, car bodtes and engmes m automottve, and atrframe structure and
engme propulsiOn m aerospace. Robustness engmeering pnnctples has tradttwnally been apphed
to all product attnbute types and ts not confined to the phystcal architecture of a product Robust
destgn and quahty engmeenng are descnbed m the followmg secttons

2.1.6.1 Robust design

Robust destgn draws on many tdeas from stattsttcal expenmental destgn to plan expenments for
obtatmng dependable mformatwn about vanables mvolved m makmg engmeenng decisions It
addresses the followmg two reqmrements faced by product and process destgners (Phadke 1989)

• Econonucal reductiOn ofvanatwn of a product's functiOn m the customer's envtronment


• Ensure that decisiOns that are found to be opttmal dunng laboratory expenments wtll
prove to be so m manufactunng and m customer environments.

The mam atm of robust destgn ts to tmprove product quahty through numnuzmg the effect of
causes of vanatton wtthout ehnunatmg the causes Thts tS achteved through a process of
parameter destgn Robust destgn utthses two mam techmques The first ts stgnal-to-nOtse ratio
whtch ts used as a measure of quahty and the second ts orthogonal arrays whtch are used a study
multtple destgn parameters stmultaneously. A defimtton of robust destgn ts gtven by Phadke
(1989)

robust design IS an engmeenng methodology for 1mprovmg productiVIty dunng research and
development so that h1gh quahty products can be produced qmckly and at low cost

20
The above defirution mcorporates the word quality This word means different !lungs to different
people. Taguch1's descnpt10n IS we measure the qualzty of a product m terms of the total loss to
society due to functzonal variatzon and harmful szde effects This suggests that m the Ideal world
the quahty loss would be zero, and 1! follows that the greater the loss the lower the quahty.

2.1.6.2 Quality engineering

The pnnc1ples of quahty engmeenng (QE) can be descnbed as an interdiSCiplinary science


involvmg engmeenng design, manufactunng and econorrucs, and IS concerned w1th the reductiOn
of costs incurred pnor to, and after, the sale of a product.

QE IS synonymous with Dr Gemclu Taguchi (1993) who suggests that quahty engmeenng
pertams to the evaluatiOn and Improvement of the robustness of products, tolerances specification,
the design of engmeenng management processes, and the evaluatiOn of the econorruc loss as a
result of product functional vanatwn (Marshal! 1998) Taguclu (1993) suggests a three stage
methodology for des1gnmg quality mto products and processes The stages are:

• System design
• Parameter design
• Tolerance design.

Quahty loss m QE IS quantified m terms of cost and IS established through the quadratic loss
functiOn The functiOn quantifies the vanatwn m a process which can be used to evaluate the
effects of tolerance parameter reviSion for econorruc and quahty unprovements (Taguclu 1993)

LSL USL

2
L = k (Y· T)

loss loss
T y

Figure 8 The quadratic loss functiOn (Taguch1 1993).

The loss functiOn IS defined as L(y) ~ k(y- T)2 , where


k = quahty loss coefficient, a constant
y ~ quahty charactens!Ic of a product
T ~ target value for y

21
QE's main obJec!Ive IS to reduce both quahty loss, wh1ch IS the cost mcurred after the sale of a
product, and the umt manufactunng cost, the cost mcurred pnor to product sale.

2.1.6.3 Statistical process control

Fox (1993) suggests that sta!Is!Ical process control (SPC) 1s an aci!VIty that uses sta!Is!Jcal
methods to evaluate the process of making parts and enables the applicatiOn of control of that
process The techn1que allows engmeers to understand how a process behaves, enabling them to
take control of the process, and serves also to ass1st m the redes1gn of the process where
necessary

There are several tools whJCh support the SPC techmque These are

• Frequency h1stograms

• Check sheet.

• Pareto chart

• Cause-and-effect d1agrams

• Control charts

• Defect concentratiOn dJagram .

• CorrelatiOn d1agrams .

• Control chart .

An explanatiOn of each of the tools can be found m the append1x of this thes1s.

2.2 Manufacturing organisational issues

Prac!Jcing a CE plulosophy mvolves the mul!Ifunctwnal mvolvement of numerous departments


and busmess centres throughout the en!Ire hfe cycle of a product or system. It IS also essen!Ial to
attam a computer optmused manufacturing system wh1ch IS regarded as the next generatiOn of
computer integrated manufactunng (CIM) systems (Hanud and Sulhvan I 993) G1ven this,
orgamsatwnal 1ssues w1thm a manufacturing busmess have become ever more Important and a
number of plulosoph1es have been developed The most pertment of these w1th respect to aJrcraft
manufacture are (Shumaker and Thomas I 998).

22
• Lean manufactunng
• Integrated product and process development (IPPn
• Agile and flexible manufactunng

2.2.1 Lean manufacturing

Smce the m1d 1980s the JUSt-m-time (JIT) philosophy has been synonymous w1th manufactunng
engmeering through out the West. The JIT philosophy was ongmally denved from the Toyota
production system (TPS) developed by Mr Tanclu Ohno in Japan (Ohno 1988). The strategy of
the JIT productiOn system encourages, only the necessary products, at the nght time, m the
reqmred quantity are manufactured, and stock on hand IS held to a m1mmum. The basic
overlappmg elements of tile JIT philosophy are well documented as being (Harrison 1992, Leaney
and Marsha112001, Ohno 1988; Womack, Jones and Roos 1990).

• The ehmmatwn of waste


• Total quahty
• People development

The concept of lean manufactunng was mtroduced to the West m 1990 through a book entitled
"The machme that changed the world" by J. Womack, D. Jones and D Roos (1990) The book
records the findmgs of a five year study on the automotive mdustry m order to establish how the
Japanese manufactunng mdustry had become so compehtlve in companson to 1ts Europe and
USA counterparts Womack, Jones and Roos (1990) placed lean productiOn as the next
development along the path wh1ch had so far come from craft and mass production

The mrcraft manufactunng sector became aware that lean philosophy was bemg pursued by the
automohve (Womack and Jones 2003) It began to consider 1fthe lean manufactunng framework
may be a way to restructure and establish a new culture w1thm 1ts own mdustry (Shumaker and
Thomas 1998) As a result, m 1992 the lean a1rcraft 1mhat1ve (LAl) was launched m the US
funded by a number of maJor contractors and the US A1r force (MIT 2004)

The LAl contmued to gam momentmn through the lean enterpnse model (LEM) The LEM
addressed broader 1ssues such as factory operatiOns, suppher systems, orgarusation and hmnan
resources, and external envirorunent B1cheno (1994) suggests that there IS no smgle defimtion
for lean produchon, but most would agree that any aerospace lean enterpnse should include all
external actlVlty such as supply cham management, d1stnbutwn and des1gn B1cheno (1994), goes

23
on to suggests that the core of any lean enterpnse wtll be made up of three elements. These three
elements are.

• JIT
• Total quahty.
• Team mvolvement.

In general terms lean productiOn can be descnbed as a hohstlc manufacturing system compnsmg
many parts. Some of these are JIT, zero buffer stocks, total quahty control, and bmldmg quahty
m rather than post-bmld rectificatiOn, maximum delegation to direct workers, small lot
productiOn, contmuous improvement, quick set-up times, standardised work, total preventative
__mamtenance, VIsual control systems, and teamwork The term lean IS denved from the preiDise
that lean productiOn reqmres less of everythmg compared wtth a mass productiOn system, 1 e half
the mvestment m toohng, half the human effort, half the product development time, etc.,
(Womack, Jones and Roos 1990) Thts sentiment has been embraced by the LAI strategy to
reduce all elements of productiOn. less time, less mventory, fewer management layers, less
capital, and fewer supphers for example.

The LAI can be said to m elude all aspects of the JIT philosophy The baste dtfference between
JIT and lean production IS that JIT stops at the orgamsatwnal factory floor boundary JIT can be
thought of as operatmg wtthm the orgamsatlonal plant boundary, and lean productiOn extends
these boundanes both mternally to the orgamsatwn (1 e management, marketmg) and externally
(1 e. supphers and dtstnbutors)

Wtckens (1993) suggests that lean manufactunng IS the way forward, but must be managed by
people who are focused on people. Wtckens beheves that the future lays m a phrase corned by
Sengenberger (1992) 'Synthesis between h1gher ejfic1ency and h1gher quahty of work and jobs'
From thts, the term 'lean, people centred, volume productiOn' has been developed

2.2.2 Integrated product and process development

The concept of mtegrated product and process development (IPPD) was mtl!ated by the US
Department of Defence (DOD) wtth the atm of nnprovmg systems destgn and acqulSltlon (James
Gregory Associates 1999, Shumaker and Thomas 1998) The DOD has comprehensive
mandatory procedures for the acqmsitwn of CEM products such as IDihtary atrcraft and at the
centre of thts procedure lays the IPPD process

24
One of the key obJecllves of the IPPD process 1s to assess product and system affordability in
terms of performance, producllon, cost of ownership, and the1r assoc1ated nsk. The IPPD process
1s also supported by six sigma quahty tmllatlves to pred1ct the cost and risks assocmted w1th early
des1gn deciSions.

James Gregory Associates (JGA) are a consultancy company m the US that prov1de IPPD
software and service support to a number of mrcraft manufacturers. A lugh level IPPD process
flow dmgram developed by JGA can be seen below

The IPPD process 1s defined by Blanchard ( 1998) as

a management technique that simultaneously mtegrates all essential acqutsitlon actlvtties through
the use of multtdtsctphnary teams to opttmtze the destgn, manufactunng, and supportabthty
processes

--...__
• Cu5~
a c.-.
Dl.aklg
,.,.
Todu>olog...
.._...
Customer Technology Value Analysis
Requirements Alternatives

A...,.8 Perform
Alternatives Transition
CUstomer and Value
Requirements Approaches Analysis

i
, _____'(___ _
~ i Draft Technology lntertm & Final Technology

~
, Delivery/TranSition Technical Delivery/
Custo-mer 1 Plan w/Buslness Reports Transition Plan wl
Business Case

~
Requirements

Ftgure 9 Htgh level IPPD process (JGA, !ne 1999)

The IPPD process IS used to define cost and performance objectives for product des1gn through to
senes productwn, mcludmg field support. The core objectives of IPPD can be hsted as (Leaney
and Marshal 1998):

• Customer focus.
• Concurrent development of products and processes

25
• Early and contmuous hfe cycle plaruung
• Maxmuse flextbthty for optnmsatwn and use of contractor unique approaches
• Encourage robust des1gn and 1mproved process capabth!y.
• Event dnven schedulmg
• Mullldtsctphnary teamwork.
• Empowerment
• Seamless management tools.
• Proacllve tdenllficatwn and management nsk

The IPPD concept IS facthtated by multldtsctplme mtegrated product teams (IPT). Each team IS

asstgned to one or more destgn/development programs to ensure 1ts success management. The
- process is controlled by a number of matunty gates which control the nsk of proceedmg wtth
subsequent phases of the development and ensure that the destgn and development parameters are
constantly momtored for deviatiOn (Barrow 1997) IPTs' can constst of both mtemal and external
personnel (suppliers or dtstnbutors).

The pnnctples of IPPD are supporttve of both concurrent engmeenng and systems engmeenng
methodologtes It facthtates the core product destgn and development mfrastructure for destgn to
senes manufacture Leaney and Marshall ( 1998) suggest that IPPD should be developed as a core
competence, where core competence refers to the collecllve knowledge wtthm an orgamsat10n.

2.2.3 Agile and flexible manufacturing

Wtth the mtroductwn of component mterchangeabthty at the Ford Motor Company m the early
1900s, mass productiOn techmques began to develop Manufactunng managers, such as Henry
Ford beheved that producmg products and components m large batches would ulttmately be more
profitable. Thts trend was remforced by the mtroductwn of mechamsed processes destgned to
replace manual labour and lead to dedtcated productiOn systems (Raouf and Anjum 1995) Thts
concept worked effecllvely for a market place whtch could be tdentlfied wtth products of low
complextty, long market hfe expectancy, and !muted product chmce

The future global a1rcraft mdustry w1ll be tdenllfied wtth products of very htgh complextty,
relatively short llme to market, and mcreasing consumer chmce. Thts sttuatwn has led to the need
for development of a1rcraft manufactunng strategtes that facthtate htgh product destgn and
development effictency, combmed wtth flextbthty

26
Currently, there 1s no agreed defimtion of what manufactunng flexJb1hty or ag~hty may be and
where the1r respective boundaries lie W1th regard to flexJb1hty, 1t IS essential to understand the
type of flexJbihty m questiOn, the level at wh1ch 1t operates, and the context to wluch 1t IS apphed
(Bamett and Leaney 1995) Without knowmg these elements, 1t will become difficult to
understand what 1s meant by flexib1hty. Slack (1991) suggests that there are two elements to
flexib1hty.

• Range jlexzbzhty, how much a genenc process can be changed, 1 e. capab1hty to produce a
greater vanety of vanant products, JSF m1htary aJrcraft program for example
• Response jlexzbzlzty, how qu1ckly can a genenc process be changed, 1 e. capability to
respond to a rev1s10n of reqmrements

Slack (1991) suggests that dJstinchon needs to be made between the flexJbihly of the whole
operatiOn (system flexJbihly), and the flexJbJhty of an mdiv1dual resource (resource flexJbihty)
Total system flex1bil1ty can be v1suahsed by cons1denng the enhre business operation as a single
block process, and 1ts outputs the resource flex1bil1ty functwns

There are many suggested categones for flexJbJhty and a flexJbihty framework. One developed
by Raouf and AnJum (1995) suggests that a number of flexlblhty categories ex1st wluch can be
frame worked mto a genenc structure, see table below

Category Flex.Jbdlt)
Product flextbJhty Volume flextb!IJty EqUipment flexJbthty
ProductiOn flexibility Des1gn change tlexibJhty
ConfiguratiOn flexJbJhty New product flexJbJhty
Mo<hficatlon flex1bthty
Product m1x flexJbJhty Custom1zmg flexJbJhty Mnt chdnge flcx.Jbthty
ProductiOn flextb1hty Expans10n flexJblhty
Configuration flexJbJhty ModificatiOn flexibility
Process flextb1hty Machme flex.JbJhty Routmg flcxJb1hty
Operation flcxtbJIJty AdaptatiOn flexiblhty
Path flexJbJhty Sequencmg flexlbJhty
Quahty flexJbJhty Matenal handlmg flexJbihty
Environmental flexJb1hty Routmg flexibiltty Volume flextbJhty
Capacity flextbdtty Dehvery flextblhty
Matenal flextbJ!tty Sequencmg flex1blltty
Demand flextbl1tty Apphcatmn flexJbd1ty

Figure 10 Classification of different flexibihlles (Raouf and AnJuml995)

Four mam categones of flexJblhty have been identified each having several sub categones Th1s
gives some mdiCatiOn mto the difficulty mvolved m pnmanly defimng flexibility, and secondly

27
measunng tts effechveness Idenhfymg cntical cntena for the measurement of flextbthty can be
dtfficult. One way ts to constder flextbthty measurements ts to rate 11 m terms of performance
measures (BenJaafar and Ramaknshnan 1995). These mclude reductiOn capactty, volume rrux,
produchon cycle limes, operatwnal costs, and mvestment.

Manufactunng ag!ltty m tts broadest terms ts sirrular m nature to flextbthty Agtle manufactunng
embraces flextbthty concepts but has a greater assoctahon wtth time and hme onentated
constramts, i e , product lead-hme to market etc Iromcally, manufactunng orgamsatwns who
wtsh to deploy flextbthty concepts need to develop rules for flextbthty constramts Tlus IS

reqmred m order to define the boundanes m whtch the concepts can be apphed, and to what
extent. Wtthout these flextbthty boundaries in place, the manufactunng organisation would
attempt to become mfimtely flextble m ever context, resultmg m an unmanageable concept

Flextbthty concepts wtthm the aerospace mdustry are also beggmg to emerge A spectfic example
of thts can be found at BAE SYSTEMS (fonnally Bntish Aerospace Mthtary Atrcraft and Aero
structures). A number of large scale htgh performance maclumng cells have being developed for
dnlhng, countersinkmg, and edge routmg to facthtate the produchon of mterchangeable carbon
fibre compostte (CFC) panels (Barrow 1997) The machme cells have htghly automated
fixture/tool loadmg and removal capabthty. Each of the CFC panel 'kits' assoctated wtth an
mrcraft type can be preloaded onto a dedtcated 'ptcture frame' set up Thts concept allows the
raptd changeover of CFC panel 'kit' types and enables high product rrux flextbthty m the
manufacture of vanous product types The same pnnc1ple ts also bemg undertaken for hght
weight sub-structures

2.3 Development of aircraft structures

Thts sectiOn outlmes the mam tssues wtth regard to atrcraft structure development These have
been revtewed in two mam sechons, these bemg

• Matenals m aucraft structures


• Atrcraft assembly technology.

The areas are both extenstve and complex Due to thts, the scope of thts revtew IS hrruted to
htghhghtmg the mam tssues assoctated wtth each area m the context oftlus research project.

28
2.3.1 Materials in aircraft structures

Early a1rcraft structures were bmlt from s1mple and lightweight matenals such as wood, and
fabnc. As manufactunng techn1ques for metals 1mproved m the early 1900's, metallic
components began to replace wooden components m a1rcraft structures (McCracken 2004) Tlnn
metallic skins were introduced producmg a more robust structure less affected by the weather
Individual components such as nbs and spars were assembled through a nvetmg process. By
1930, the technology for alunnmum alloy became avmlable and wing structure components such
as nbs and spars were manufactured from alunnmum matenals (Dav1es 1996)

Compos1te matenals although perce1ved as a recent innovatiOn have been ID use m aircraft
___ structures smce the 1950's (Johnston _1997) Composite matenals _offer the_ advantage of _
lightweight and strong propert1es and are constructed from two or more orgamc or morgamc
elements One element serves as a "matnx" bmdmg the matenal together, and the other element
serves as remforcement, usually m the form of fibres winch are embedded in the matrix Until
recently, the most common matnx matenals were "thermosettmg" matenals such as epoxy,
bJsmaleJmJde, or polynrude These were used w1th glass fibre, boron fibre, carbon fibre, for
reiDforcement

The first composite to be used ID the commercml aerostructures was fiberglass used ID the BoeiDg
707 compnsmg about two percent of the structure By the 1960s, other composite matenals
became avmlable, m particular carbon fibre, boron fibre and graphite, embedded ID epoxy resms.
The first maJOr military productiOn use of boron fibre was for the honzontal stabilizers on the
Navy's F-14 Tomcat mterceptor. By the 1980's, the Bn!ish Aerospace-McDonnell Douglas AV-
8B Ham er flew w1th over 25 percent of 1ts structure made from composites

Incluswn of composite matenals ID mrframe manufacture has IDcreased Modem commerc1al


a1rcraft such as the Boemg 777 has composite matenal making up about ten percent of 1ts
structural we1ght. Modem military a1rcraft, such as the F-22, use composites for at least a third of
the1r structures. It IS predicted that future nnhtary a1rcraft programs will mcrease the use of
compos1te matenals to as much as seventy percent Tins trend may not be the same for the
commercial sector due to the complex maiDtaiDabJiity 1ssues associated With compos1te matenal.

Composites also have a number of disadvantages, some of these are

• Rela!ively expens1ve and complex to manufacture.


• Difficult to IDSpect for defects

29
• Complex process for reparr.
• Manufactured part feature accuracy and repeatab1hty can be poor
• Little opportumty for component rework
• Environmental and cost Issues ansmg from product disposal.

More recently, new composite manufactunng processes have been developed and are used m the
aerospace mdustry. These cannot overcome the mam disadvantages outhned above, but can
address Issues such part accuracy and repeatab1hty:

• Resm transfer mouldmg


• Resm film infusiOn

New concepts for composite materials

• Smart structures, structures that momtor themselves for stress and stram

Thennoplasl!cs are a relal!vely new matenal that IS replacmg thennosets as the matnx matenal for
Aluminum sl!ll remams a remarkably useful matenal for aircraft structures and metallurgists have
worked hard to develop better alummum alloys (a mixture of aluminum and other matenals) In
parl!cular, alummum-hthmm IS the most successful of these alloys It IS approximately ten percent
hghter than standard alununum Its adoption by commercial aircraft manufacturers has been
slower due to the expense of litluum and the greater difficulty ofusmg alununum-hthmm. But It
IS hkely that alummum-hthmm will eventually become a widely used matenal for both
commercial and nuhtary aircraft

In recent years, new metalhc manufactunng processes have been developed and are used m the
aerospace mdustry These are.

• Superplastic fonnmg
• D1ffuswn bondmg

The aerosapce mdustry contmues to research matenals technology trymg to Identify new strength
to weight advantages for aircraft structures This IS diverse and complex area of engmeenng and
although this reaserch acnowledges It's existence It has been considered beyond the scope of this
research prOJect. One area for future research could be to mvestigate the d1meswnal stab1hty of
these new matenals and their associated manufactunng process

30
2.3.2 Aircraft assembly technologies

Some of the largest costs assoctated Wtth atrcraft structure development and production can be
attnbuted to assembly toolmg and fixtunng systems, to enable product bmld (Thornton 1997) In
order to reduce thts cost the aerospace mdustry has millated a number of advanced assembly
technologtes, one of the most stgmficant being fixtureless assembly Thts technology has the
followmg charactensllcs·

• ReductiOn of expenstve 'hard' fixtunng and checkmg systems


• IntroductiOn of dtgttal ahgmnent systems back to CAD master.
• -IntroductiOn of dtgttal checkm-g and gatigirig techniques back to CAD master -
• Reqmrement to destgn part-to-part dtrect merfacmg features to facthtate !at type assembly
bmld
• Destgn of parts and assembhes to accomodate mterchangeabthty (ICY) spectficatwn and
classt ficatwn.
• Usually assoctated wtth part count reductwn program

Jtgless assembly 1s achteved through the apphcallon of new 'self toohng I location' pnnctples
Each component and subassembly wtll be produced wtth mherent self locatmg features and
toohng features, m addtllon to nommal functwnal arclutecture. These features wtll be used to
facthtate part to part locallon for assembly purpose Stgntficant attentwn ts reqmred to feature
type and tt's defimtwn, and to thetr control wtth respect to geomerttc tolerance control and metnc
The DM pht!osophy can stgntficantly atd thts process by usmg the assembly analysis tools
avatlable to predtct net assembly dtmenswnal capabthty.

Atrbus are currently adoptmg thts type of phtlosophy, avotding where posstble, large and
expenstve Jtgs, toohng and hard gaugmg systems (Coyne 2004) BAE SYSTEMS are also
developmg thts technology m conJunCI!on wtth flextble manufactunng techntques such as the
development of advanced machine tools centres for the manufacture of parts for Jtgless assembly,
automatic fastener mstallallon, and sealant apphcatwn (Johnston 1997).

2.4 Dimensional management discipline

As compellllon m the aerospace manufactunng market mcreases, orgamsallons have been forced
to revtew thetr busmess strategy and efficiency m order to survive As a consequence, a number

31
of dtfferent efforts are bemg pursued by compames to Improve the 'affordabthty' of customer
dnven product specificatiOn through the concurrent mvolvement of engmeenng function. DM is
becoming part of a growmg field of endeavour pertment to the design, manufacturing and
assembly of complex products aircraft manufacture bemg one example (Leaney I996).

Thts area of activity IS most commonly referred to as DM (DM). Other aheses are dtmenswnai
control (DC), design for vanatwn (DFV), vanatwn management (VM) design for vanatwn
analysts (DVA), and assembly variation analysts (AVA) For the purpose of this report, tlus area
of actlVlty will be referred to as DM

There are a number of defimtwns for DM (Leaney I996; Leaney and Marshall 200 I; L1gget
I993), but Leaney and Marshall (200 I) define 1t concisely as:

the dtmenswnal management dtsctphne refers to total product dimensiOnal control whtch recogmzes
and manages vanatiOn dunng the design, manufacture and assembly stages of development and
productiOn

D1menswnal vanat1on IS inherent wtthm any component part, subassembly, tooling and
manufactunng processes and cannot be elmunated However, vanation can be Identified,
quantified and analysed so therefore it can be managed G1ven tlus fact, there has been relaltvely
httle effort placed on optmuzmg tolerance allocatiOn and evaluatmg potential changes m the
des1gn that would allow for mcreased vanation m non-cnt1cal areas (Craig I992) Conversely,
cnltcal areas havmg a s1gn1ficant Impact on product specification need to be 1dent1fied and their
vanatwn controlled through the appropnate specificatiOn of tolerance. These cntical features and
tolerances should then be destgnated as product key charactenstics (KC) and managed throughout
the product hfecycle (Boemg). Further explanalton ofKCs' are g1ven in the next sectiOn.

The DM process has traditionally been undertaken to support and evaluate tolerance allocation.
Tolerance analysts software tools, for example VSA, eM-TolMate, etc, have been used to
evaluate the resultant assembly bmld parameters takmg mto account mformalton such as
component and fixture feature tolerance, component locator strategy, and assembly sequence (see
sectiOn enlttled 'Computer a1ded tolerancing software tools and process' for further mformatwn).
The analysts model1s 1m1tally based on des1gn tolerances but w1th the ava1lab1hty of prototype or
surrogate process capab1hty data there IS an opportumty to 'close the loop' and feed back
mformalton from manufactunng to des1gn (Leaney and Marshali200I)

32
In addition to the software tools there are a number of DM best practice techmques wh1ch gmde
engineenng, some of these are (UGS PLM Solutwns 2004)

• Gauge repeatab1hty and reproducibility techmques.


• Geometnc d1menswmng & tolerancmg
• Feature defimtion and assembly locallon strategy
• Quahty measurement strategy

Pertment research and industrial projects associated w1thm the area ofDM are:

ADCATS: Assocmllon of the Development of Computer Aided Tolerancmg software, (Chase


1~88) Consort.'U1Il of 12 member compames interested m tolerance analysis coordmated through ___ _
Dr K Chase at Bngham Young Umvers1ty (BYU).
http·//adcats et byu edulhome html
ANDREA: program aunmg to renewing and strengthenmg the research and education m
engmeenng design and engmeenng management m Sweden Projects areas mclude design for
manufacture, tolerancmg m the aircraft mdustry, Improved assembly quahty Research base was
Lmkopmg lnslltute of Technology (L!TII)
Thzs research group zs no longer actzve
Conformability Analysis: eCA technique IS a developmg tool for the analysis of quahty costs
associated With mechamcal designs and manufactunng processes using process capabihty md!Ces,
FMEA, and Cost Mappmg Most recent development IS the adaptatwn of the standard eCA
technique to evaluate the functwnal, manufactunng and test process capability of electronic
c1rcmts Projects based out of the Umversity of Hull
http://www hull ac uk/eca
CAM-I: The CAM-I Institute for Manufactunng and Automation Research (!MAR) Projects
mclude
• Quality Assurance Program (QAP): development of DimensiOnal Measuring Interface
Spec1ficatwn (DMIS) now as American Nallonal Standard and bemg progressed as
mtenhonal standard
• Next Generation Manufacturing System (NGMS): project to combme the worlds best
th1nlang on the next generation of manufactunng systems to gm de the ten year NGMS,
IMSR&D.
• Robust Quality Engineering: development of a system that w!ll enable companies to
reahze the benefits of Robust Design concepts when apphed to proven Quahty
Engmeenng pnnc1ples earher m the hfe cycle
http·//www cam-1 org/mdex html

33
CIRP, International Institution for Production Engineering Research: Orgamse the CIRP
internatiOnal semmars on computer mded tolerancmg, (CIRP 2001).
MADLab research group: Research group based at The Umverstty of Texas at Austm currently
mvolved wtth several proJects relatmg to DM
http //www me utexas edu/-madlab/
SPE, Japan Society for Precision Engineering: Orgamze and promote computer atded tolerance
systems.
http //www 1spe or m/english/
The International forum on design for manufacture and assembly: Orgamse the mternahonal
senunars on DFMA, (Boothroyd and Dewhurst 2003).
http·//www.dfina corn

In support of the DM process there are a number of analysts techmques, tools, and methods The
most stgmficant of these are

• Product features.
• Key charactenshcs
• Tolerance spectficatwn
• Computer aided tolerance (CAT) analysis software tools

2.4.1 Product features

The use of features ts constdered by many researchers as the key to the genume mtegratwn of
many aspects of destgn and the planmng of manufacture (Case and Gao 1993). Feature
defimllons can not only be used to express the dcstgn mtent and form the basts for destgn analysis
but can also provtde the baste geometry knowledge for the integratiOn of manufactunng, assembly
and the dtmenswnal mspect10n process

The term feature ts denved form the Latin word 'Factura' whtch means the 'act of making' or
'informatzon '. The defimhon of a feature has been mterpreted by research m many dtfferent
ways. The problem of definition ts compounded further due to the dtfferent feature
mterpretattons for destgn, manufacturing and analysts. Some defimtwn are presented highlightmg
the scope of the dtfferent meamng:

a feature ts a group of geometnc enhhes Wlth some meamng for the particular act!Vlty to
be performed wtth them (Herbert et a/1990)

34
a feature IS a regton of mterest on a surface of a part (Pratt and Wdson 1985).

a feature IS a portiOn of the work ptece generated by a certam mode of metal cuttmg (Chm,
Barash and Anderson 1984)

Currently there 1s no formal way of categonzing or spectfymg features, therefore there are an
infirute number available (Allada and Anand 1995) However, orgarusat10ns dtvtde features mto
related classes for easter access by destgn and manufacturing. These classes can be subdtvided
mto subclasses to form a hierarchy known as a feature taxonomy. Because of the hierachtal nature
of taxonomy, any attnbutes assoctated wtth a feature class wtll also be related to any sub-class.
The structure of any feature taxonomy wtll depend on the methodology used for the tdentlficatlon
and classtficatlon of features wtth respect to how they are gomg to be utilized wtthin an
orgarusat10n

2.4.2 Key characteristics

Most modern manufactunng product development techmques mclude a mecharusm to capture


customer attnbutes whtch form the basts for product functiOn such as QFD. These product
specifications are further broken down mto functiOnal features and attnbutes. Techmques to
ensure the IdentificatiOn and management of these attnbutes, tolerance for example, are bemg
developed One of these techmques used to ensure product quahty JS key charactenstlcs (KC)
The KC techmque ts accredited to MlT m the US and has gamed populanty wtth maJor aerospace
compames such as Boemg and McDonnell Douglas, both of whom currently use KC methods m
support of product development The KC techmque has been defines as (Lee and Thornton
1996)

KCs' are product features, manufactunng process parameters, and assembly features that
stgmficantly affect a product's performance, funchon and form.

Research work undertaken by Lee and Thornton (1996) suggest the classification of KC's mto
three categones The first ts a product KC (PKC), the second a Manufactunng KC (MKC) and
finally an assembly KC (AKC) These features are tdentlfied through a process of product
decomposttlon firstly to subassembly, then assembly, component, and finally to component
features and attnbutes. These elements are then bmlt back up mto the final assembled product.

Decomposition of the product down to part feature level allows the analysts to define what
manufactunng processes wtll be reqmred to meet the attnbute and therefore what process

35
capability many be anticipated from related statistical process control (SPC) data As the product
IS built up the assembly features and attnbutes will also become apparent allowmg an appraisal of
assembly toohng reqmrements The product development process needs to consider both aspects
of product decomposition and rebmld equally as Issues relatmg one will have an effect on the
other (Wlutney 2004)

Lee and Thornton (1996) suggest a forth category of KCs', namely the StatKC The StatKC IS a
subset of the three mam categones and IS defined as a KC whose deviatiOn from tolerance will
pose Sigmficant risk to the quahty of the final product It could be descnbed as an element of a
'cnllcal path' ofmanufactunng and assembly components and processes.

After the IdentificatiOn of KCs', they can be categonzed mto the following (Thornton 1997)

• Safety Issues and government regulatwns these apply to corporate as well as mternatwnal
safety.
• Customer product reqmrements these are reqmrements duectly relatmg to product
performance, function and form
• Internal corporate reqmrements these are associated With mternal corporate reqmrements
such as schedule reqmrements, cost controls.

The benefits of usmg KC techniques, Lee and Thornton ( 1996), are

• Better deternunatwn of detml design decisions Improved cornmumcatwn of cnllcal


design areas With respect to cost, rnanufactunng and assembly process selectiOn.
• Root cause analysis: IdentificatiOn of root causes by travelling up, down and across
product KC tree.
• Aid eqmpment deciSions manufactunng and assembly equipment may be Improved
based on KC data
• A1d mvestments and make-buy decisiOns KC data can be used to analyse new
manufactunng capab1hlles to help deternune make-buy deciswns.
• Improve product plannmg: areas of company weakness can be deternuned based on
customer requirement history agamst receded StatCKs

The key charactenst1cs technique has been highlighted above This technique IS well documented
m a number of pubhcatwns.

36
2.4.3 Tolerance specification

The manufacture of perfect form ts unrealistic, and even tf 11 were posstble we would never know
because 11 would never be posstble to venfy perfect form by phystcal measurement (Reqmcha
1993) Tolerance, m the context of product development, has been defined as (Jeang 1995):

the maximum deVJahon from a nommal value wtthm whtch the component ts ~till acceptab1e for Its
mtended purpose.

Dtmenstons are controlled through the allocation of a tolerance whtch allows a nommal value to
vary in some way by a spectfied amount Two mam types of tolerance groups, parametnc (lmear)
and geometnc, have been developed for mdustnal use (Voelcker 1993) A thtrd tolerance group
has been developed to control kinematic variation which occurs from small resultant adjustments
between matmg parts at the assembly stage (Chase, Magleby and Gao 1997)

Tolerance mformation has traditionally been placed onto engmeenng drawmgs through notation,
symbols, and the specification of 'blanket' company standards. However, the mtroduction of
CAD/CAM systems has resulted m much research and development wtth regard to how best
tolerances many be represented wtthm these systems Anstides Reqmcha (1993), a Professor of
Computer Sctence and Electncal Engmeenng at the Umverstty of Southern California has been
responsible for the development of the mathematical and algonthm theory behmd several
commercial solid modelling systems Hts research htghlights the two mam syntactical approaches
to tolerance defimtton, dtmenstonal linut spectficatton and geometnc tolerance specification for
solid modelling Hts current work ts wtth the programmable automation laboratory (PAL) wtth a
focus on computatiOnal geometry, robotics and arttfictal mtelligence

Some of the mam techmques used m computer solid model to represent tolerances are (Reqmcha
1993)·

• Pure parametnc tolerancing


• Tolerance zone semantics
• Offset zone theory.

The spectfical!on of tolerance m CAD/CAM systems needs to be comprehensively defined m


mathemal!cal terms to avoid ambtguity. Measunng eqmpment also needs clear geometnc

37
spectfical!ons to ensure correct mterpretatwn of computer reference tolerance zones Currently
there are dtfferent methods of constructmg tolerance zones m CAD systems, these being

• Parametnc dtmenswning and tolerancmg


• Geometncal dimenswrung and tolerancmg.
• Kmemal!c dtmenswnmg and tolerancmg.

An explanahon of each of these tolerance types ts avatlable m the rear of thts report

2.4.3.1 Tolerance standards

There -are -mimy -world wtde tolerance standards includmg Bnttsh, Japanese, Clunese, USA,
European, and Russtan The mternatwnal standards orgamsahon (ISO) ts attemptmg to
consohdate the stronger elements from predommantly western standards to form a number of tts
own. The followmg current standards are considered to be relevant to the scope of thts thests,

British Standard (BS); 113 mam standards based on 'Techmcal product document Spectfical!on
for defimng, specifymg, and graplucally representmg products':
BS8888 2002
Geometnc product spectfical!on (GPS)
International Standards Organisation (ISO); 8 mam standards
ISO!! 01, IS02692, IS05458, IS05459, IS07083, ISO! 0578, ISO! 660, and ISO-DD16792.
American Society ofMechanical Engineers (ASME); 2 mam consohdated geometnc standards
ASME Y14 5M 1994 and ASME YI4.4IM 2003

Tolerance standards generally destgned to control parametnc (hnear) or geometnc enl!ttes


Standards such as the BS 308 (1985), ASME 145M-1994, and the ISO (1995) ISOIR 1101,
ISO/RIIOI/11 and ISOIR1660 were dlVlded mto hnear (mcludmg angular, radms and dtameter)
and geometric tolerances (Cox, McMahon and Tannock 1995)

Geometnc controls of a tolerance standard are defined through geometnc dtmenstorung and
tolerancmg (GD&T) schema GD&T standards are charactensed through sets of symbols which
are used to define part features and thetr tolerance zones (Moh 1996) Parametnc controls are
defmed through boundary condil!ons and can be expressed as baste dtmenswns wtth a tolerance
apphed.

38
The most commonly used GD&T standard m the US IS the ASME Y14 5M-1994 TJus has
evolved out of a consolidatiOn of other standards, ANSI Y14 SM-1982, ANSI Y14.5-1973, USASI
Y14 5-1966, ASA Yl4 5-1957, SAE Automolive Aerospace Drawmg Standards (sectiOn A6/7 and
8-September 1963) and MIL-STD-8C, October 1963, (Foster 1994) The consolidatiOn of all
these standards was necessary and dnven by members of military, industnal and educalional
organisatiOns m the US The standard 1s reqmred by these orgarnsatwns to consolidate the
defirnlion of des1gn and drawmg reqmrements w1th respect to actual funclion and relalionsJup of
part features Further to th1s, the standard prov1des a functional manufactunng and quality
mspeclion commumcatwn technique wh1ch consistently relays back to the des1gn defimlion. The
ASME Yl4.5M standard has contmued to develop through the support of the ANSIIASME
B89 3.2 dimenswnal measurement method (Beckw1th and Parson 1994) TJus was developed to
ass1st the d1rect measurement of component d1mensw_n_ and tolerance m acc~rdance w1th _the
ASME Yl4 5M standard

The ISO are developmg a standard on geometncal product spec1ficat10ns Th1s IS bemg
undertaken by vanous ISO technical cornn:uttees who are 1denlifymg nussmg and mcomplete
sectwns on tolerance with the mtenlion of fulfilling them (Henzold 1995) The new concept of
geometncal product specJficalions standards have been used successfully m Dernnark for
tolerancmg and standard! sa lion. They have developed a system called chams of standards used on
engmeenng drawmgs and CAD systems to define the geometnc features of a component such as
Size, distance, form, onentatwn, locatiOn, and surface roughness (Bennmch 1994). The cham of
standards consists of SIX links

• Product documentatiOn md1cal!on- codificatiOn


• Defimtwn of tolerances - theoretical defimtwn and values
• Defimtwn of actual features - charactenslics or parameters
• Assessment of the dev1alion of the work piece - companson w1th tolerance linuts
• Measurement eqmpment requirements
• CalibratiOn reqmrements- calibration standard

The harmomzatwn of the two pnnc1ple GD&T standards ASME Yl4 5M and ISO IS beconung
more desirable m the west The aerospace mdustry has been consolidatmg by fonmng mulliple
alliances in order to ensure that the1r JOmt technologies and econonues of scale can sustam the
market share of the1r busmess. As a result, many orgamsatwns have to overcome the bamers
resultmg from the physical dislocatiOns of mulliple company Sites Potenlialissues can an se, for
example, between US and European based collaboratiOns The US consortmms may have defined
product tolerances usmg the ASME Yl4 51 usmg 1mpenal dimensions European companies may

39
favour the ISO tolerance and metnc dimensiOns The difference between these standards will
need to be well understood by both parties to avmd nusmterpretatwn of product and process
defimtwn The development of a smgle integrated GD&T standard would help tJus situatiOn by
providmg Improved and consistent engmeenng commumcatwn for the global development of
complex products

2.4.3.2 Tolerance specification

The most common tolerance specificatiOn problem encountered by engmeenng design IS tolerance
allocahon, which IS the d1stnbutwn of the specified assembly tolerance among the components of
the assembly (Chase et a/ 1990)

Designers are pnmanly concerned With functiOnal tolerances ensunng the assembled product as a
whole meets some d1menswnal quality spec1ficahons (Craig 1996) The high level functiOnal
tolerances will be made up from an accumulahon of mdividual component tolerances that make
up an assembly. Funchonal tolerances are dnven by product specificatiOn but component
tolerances are mostly denved from process capability data A potenhal concern for the design
orgamsatwn is how to dJstnbute the h1gh level functiOnal tolerances across component parts

Each component will be associated With a cost to the orgamsatwn The level of direct cost will
partly depend upon component complexity, Its geometnc features, and their associated design
tolerances One of the primary consideratiOns for design IS affordab1lity This bemg the case,
feature tolerance should be associated with a umt cost With this mformatwn the design engmeers
will be able to develop tolerance related quahty loss functiOn models and Will associate these with
design mtentwn to establish the most economic and affordable solutiOn (Crevelmg 1997).

There are currently several models for evaluatmg cost versus tolerance These are (Chase et a/
1990)

• Lagrange mulhphers
• Combmatonal process selectiOn method.
• Zero-one discrete search method
• Umvanate search method
• Nonlinear programmmg method

40
The research by Chase et a/ (1990) suggests that an exhaustive search method usmg a Lagrange
multtpher to allocate tolerance and combmatory to test all posstb!e process combmatwns. The
zero-one method, a totally combmatory based method, ts too meffictent to be of practtcal value.
The sequential quadra!tc program (SQP search algonthm) method, based on nonlmear
programmmg was shown to be capable of treatmg multtple loop assembly func!tons but could not
guarantee tdenttfymg the global mimmum The umvante search method was shown to be the
most effictent. Thts ts based on a Lagrange multtpher and a process to reduce the set of process
combma!tons tested Although t! could not guarantee findmg the global mtmmum, t! always
found t! for unconstramed problems and moderate-to-large constramed problems

2.4.3.3 Tolerance analysis and synthesis

Tolerance analysts and synthests modelhng ts performed to

• Detenmne tf the predtcted varia!ton levels expected from the mtended manufactunng,
toohng and assembly processes wtll achteve the dtmenswnal specifications of the
mtended destgn.
• Perform root cause analysts on current products to rmttgate functiOnal non-conformance

The analysts of tolerances can be grouped by dtmenswnal space and their related degrees of
freedom (DOF).

ID I transla!ton - I DOF
I 5D 2 translattons- 2 DOF
2D 2 transla!tons, I rotatiOn - 3 DOF
3D 3 translattons, 3 rotatwns - 6 DOF

There are four mam groups assocmted wtth analysmg tolerance accumulatiOn (Chase, Gao and
Magleby 1997, Ntgam and Turner 1995), (Turner and Gangm!t 1991), (Ltggett 1993):

• Worst case.
• Root-mean square (RMS) or root sum square (RSS).
• Sta!ts!tcal analysts
• Kmernattc analysts.

An overvtew of these 4 groups ts avatlable m the appendtces of tlus thests

41
2.4.4 Computer aided tolerance software tools and process

A number of computer a1ded tolerance (CAT) and quality software tools and processes have
emerged m support of the DM actlVIty Some of the CAT software tools available are·

• UGS PLM SolutiOns (UGS 2006). http //www ugs corn


• DimensiOnal Control Systems (DCS 2004) http //www 3dcs corn
• CATIA V5 Tolerance Analysis of Deformable Assemblies
http //www.3ds corn/en/home asp
• Varatech http.//www varatech corn
• Tecnomahx Technologies Ltd http //www tecnornat1x corn
• _CogrutJon Corporation http //www Cl corn
• CEITI 6 s1gma http·//www s1gmetnx corn
• Saltire Software http.//www saltJre corn
• Toltech

There are a number of specific benefits of usmg these software tools-

• Perform complex analysis to establish and Improve aircraft structure dimensional


capability
• Optmuze the design, manufacture and assembly as a smgle process
• Mmmuze and resolve producl!on development Issues by supportmg FMECA and root
cause analysis
• Reduce warranty, scrap and rework caused by dimensiOnal vanahon.
• Reuse dimensional quality informatiOn on future designs by creatmg a product and
process quality knowledge base
• Identify cntical product and process dimensiOns (KCs ') and manage vanatwn
• Aid the development of measurement process planrung
• Improve commumcatton and collaboratiOn of product quality mfonnahon throughout the
enterpnse and supply cham.
• Assess Interchangeability and mamtamab1lity specificatiOns
• Venfy design component manufactunng and assembly specificatiOn agamst actual
process capability
• Directly support and dnve mtegrated product development (IPD) Initiatives dunng
product quality engmeenng reviews
• Facilitates product and process development process capability knowledge capture.

42
A reVIew of the CAT software tools IS available m the appendices of this thesis.

2.5 Manufacturing methodology critique

The manufactunng methodologies cronology outhnes the key strategies adopted by western
engmeenng orgamsatwns over the past decades Between the 1950's and 1970's the aerospace
manufactunng industry operated m response to the demands of therr respective defence
procuennent orgamsation usually controlled by the govermnent of the day. Tins Situation
changed m Europe m the 1980's wtth the end of the cold war and the onset of a global econonuc
downturn _These factors tnggered the search for new tmltatives m the aerospace mdustry who
now for the first time had to compete m tt's own domestice and the global market place As a
result, aerospace manufactumg technology began to mcrease m complexity m the search for
competttive advantage over thetr new competttors.

Both m the aerospace and the wtder manufactunng mdstry a number of new techntques and
tmtiattves began to appear During the 1970's DFM and DFMA became drawn to the forefront of
manufactunng followmg the percteved demand for mcreased automatton for the assembly of
complex goods. Ctrca the late 1970's the concept for large scale flextble automation was also
bemg constdered by manufactunng orgamsatwns m order to accommodate customer demand for
product custonuzatwn. Tins dtrectly led to such phtlosophtes as AMT and FMS whtch were m
part based on the developmg Toyota JIT and lean productton system

The techniques of DFA, FMA and AMT began to htghhght some of the shortfalls of a design
process whtch dtd not take mto consideratiOn subsequent reqmrements for the down stream
manufactunng and assembly process It became apparent that the development and manufactunng
process of complex goods needed to be better mtegrated to provtde a more comprehensive and
complete product hfe-cycle management (PLM) process This mtegrated PLM approach lead to
the development of stmultamous engmeenng whtch later developed mto CE The development of
CE lead to an mcrease of computer based technology development and lead to the mtroduction of
CAD, CAE, CAM, and CAPP software A new and emergmg technology supportmg CAE in the
automottve mdustry was DM These computer technology elements were developed and
archttected under the overall banner of CIM although at thts time thetr mtegratton at system level
was poor.

43
From an orgamsatwnal perspective the aerospace mdustry began to adopt a new strategy called
IPPD supported by the IPT. Th1s strategy was 1mtially developed m the US m order to fully
support the SE approach for total product and process development already bemg adopted by the
aerospace mdustry The SE methodology has been adopted for the development and mannfacture
of CEM 1tems m industry sectors such as aerospace. It d1ffers from the CE approach m that 1t
reqmres a product realisatwn process based on a systems-onentated approach as apposed to a
component-orientated approach Tlus allows the reqmrements of a complex product w1th a lugh
level of electro mechamcal mterdependency to be des1gned and developed from an mtegrated
systems and not a p1ece part onentatwn

2.5.1 DM critique

DM was 1mt~ally developed m the automotive mdustry as an engmeenng d1sc1plme as a way of


managing product vanatwn dunng the des1gn, manufacture and assmbly process. The emergence
of DM followed the recognitiOn that s1gmficant productiOn costs were bemg mcurred due to the
poor understandmg of product and process variatiOn The aerospace industry has followed th1s
lead and DM IS now practised m a number oflarge a1rcraft manufactunng compan1es.

DM IS not umque to the auto and aero sectors. It currently ex1sts m a number of other mdustnes
each of wh1ch are developing the1r own set of techn1ques to manage assembly van all on m the1r
products Techniques such as KC, feature defimtwn, tolerance spec1ficatwn, and tolerance
analysis form some of the bmldmg blocks of their respoct1ve DM process

Product h1erarcy can be desolved to feature defimtion Features are bas1c elements wh1ch
descnbe product phys1cal attnbutes such as a component mterface for assembly constramt, bas1c
architecture for KC des1gnat10n, and p1ece part manufactunng arclutecture reqmrements Feature
management IS a fundamental part of the DM porcess and IS an area that should be developed

The KC process IS defined m literature as a des1gn to manufacturing and assembly mtegral!on


techn1que Th1s mtegratwn 'check' IS aclueved through the VIrtual decompos11Ion of the product
to assembly, part, and feature levels wh1ch are then rebmlt for overall product validation. The KC
process 1s currently not supported by any software tools relymg on md1v1dual compan1es to
document the1r own process One observatton made by the author IS that the KC flow d1agrams
could be Improved through a more standard and consistent generatiOn process

From the rev1ew, the author recogn1ses that most products are currently des1gned in 3D CAD
usmg sohd modelling techn1ques. These CAD models form the bas1s of the nommal design intent

44
------------------------

and are defined m terms of assemblies, components, and features Tolerance specification to each
of these features, and related components, describe how the CAD nommal component may vary
by a spec1fied amount Tolerance spec1ficatwn and 1ts applicatiOn IS comprehensive documented
m a number of tolerance standards most of wh1ch are based on GD&T promc1ples One area of
weakness 1denlifed by the author is that of tolerance cost modellmg. From the literatures rev1ew 11
has been idenlified that no applicatiOn IS available to perform such an analysis

A number of software tools exist to aid the DM process most of wh1ch mvolve the srmulatwn of
assembly bmld usmg 3D CAD environments These tools allow comprehensive modelling based
on 3D nonunal geometry, tolerance soft gauge applicatiOn, assembly constraint management, and
assembly sequence One potenlial shortfall for the DM process 1s that these tools are typically
used on matunng des1gn concepts because they require_ detmled 3D CAD geometry and
supportmg data, GD&T spec1ficatwn, for example This presents a potenlial problem m that by
the lime a comprehensive tolerance analysis and synthesis has been undertaken, a des1gn may be
to mature for an orgamsatwn to conduct product mod1ficat10n.

2.6 Concluding comments

Tills chapter comprehensively rev1ews the domam relatmg to DM. The rev1ewed literature m DM
has focused on manufactunng methodology, orgamsatwnal ISSues, a1rcraft structure development,
and the DM d!SC!plme. Although the literature research prov1des an m depth mvesligatwn mto
each of these areas 11 1s evident that no overall DM methodology ex1sts to support the des1gn and
development process This research aims to explore such a methodology enabling a
comprehensive and genenc approach to DM for the aerospace mdustry The author has
recogn1sed that key areas 1denlified in the literature need to be mcluded as part of a DM
methodology Spec1fic key areas mclude

• Feature defimlion
• Key characterislics
• Tolerance specification
• Tolerance modelling

In chapter 3 the thesis Will present a fact findmg assessment of current best praclice m DM m both
the automolive and aerospace mdustry sectwns The findmgs of chapter 3, together w1th the
literature rev1ew undertaken m tills chapter, form the bas1s for chapter 4 wh1ch outlines the need
forDM.

45
Chapter 3

3 Assessing current best practice in dimensional


management

Objectives: The literature reVIew m chapter 2 outlines the current thinking associated w1thm the
DM domam which IS dormnated by the automotive and aerospace sectors To understand how
these sectors currently make use of the DM diSCipline this chapter presents an overview of the
current best practices relatmg to both the aerospace and automotive mdustnes The findmgs are
presented m the followmg sections._

• Present the findmgs of a US study tour of leadmg aerospace and automotive


manufactunng compames practicing dimensional management
• Highlight some European aerospace compames practicmg dimensiOnal control activities
• Outline sample cases of dimensiOnal analysis in the aerospace and automotive mdustnes

3.1 USA study tour ofleading aerospace and automotive


companies

Dunng August 1996 the author arranged a study tour of selected aerospace and automotive
compames m the USA The author was accompamed on the tour by two engmeers from Bntish
Aerospace Military Aircraft and Aerostructures (BAe MA&A), Samlesbury. The mm of the study
tour was to assess current best practice w1thm selected aerospace and automotive companies With
regard to the use of DM techniques. The aim of the study tour was addressed through the
followmg obJectives·

• IdentifY companies operatmg DM techniques to be mcluded m the study tour.


• Develop a structured method of captunng the scope of DM and the related
ImplementatiOn Issues of each company visited.
• Document findmgs of study tour.

46
The following companies were selected for the study tour. Their selection was the result of
vanous commumcatwn With parties associated With DM at that time who included personnel at
Loughborough Umversity, BAe MA&A, VSA Inc and DCS Inc

• Chrysler Corporation

• Ford Motor Company .

• McDonnell Douglas Aerospace .

• Raytheon arrcraft .

• Northrop Grumman

• The Boemg Company.

A structured method was defined to capture the scope and ImplementatiOn of any DM process
operating w1thm a company The scope of a company was captured m the DM scope matrix and
Implementation Issues m an Implementation matnx. Both these matnx together with their
respective guidelmes can be seen m addendum 5 and 6 m the study tour report (Jeffreys 1996).

The findmgs from each company VISit are discussed m the followmg sectiOns

3.1.1 Chrysler Corporation

The Chrysler Corporation VISited IS based at the Chrysler Technology Centre, Auburn Hills, MI

The Chrysler CorporatiOn has been operatmg DM techniques smce 1990. Chrysler first became
mvolved after becommg aware of the vanation simulatiOn work undertaken at General Motors
(GM), a maJOr compel!tor. Clmms of success by GM led to a concern that may have develop a
competitive advantage forcing Chrysler to mvesl!gate the new technique. Chrysler IS perceived m
the automotive industry as bemg a leadmg DM practitiOner. Over the last seven years they have
contmued to develop a DM strategy Some of the strategy has successfully been reahzed with m
company through some of the latest vehicle programs such as the Neon and Cirrus

Currently, the DM process will typiCally focus on body m white structure and mtenor tnm
ahgmnent m truck and car productiOn. Typtcal consideratiOns of the vehicle program are quahty
Issues relating to:

• All closures to fixed panel fit and functiOn, 1 e , doors, hood, trunk to BIW.
• Power tram functiOnal and assembly analysis

47
• Selected mtemal and external packagmg assembly analysis, 1.e , mtenor and exterior tnm
components.

Chrysler also used DM on dnve tram assemblies but 11 was not possible to Jdenl!f}t what work had
been undertaken

Velncle program methodology captures and deploys customer target specificatiOn (1 e functiOnal
attnbutes) through a quahty functwn deployment (QFD) process Target vehicle speclfical!ons
are idenllfied and documented m the dimenswnal co-ordmation manual as part of the up front
design process These are established through comprehensive bench marlang exercises
mcorporatmg reverse engmeenng techniques undertaken at the Chrysler Vehicle Research Centre
Currently Chrysler perceives Toyota as best car manufacturerwllh respect to dJmellsional quality_
and they have therefore become the benchmark The QFD process forms part of a systems
engmeenng type methodology This mvolves decomposmg the whole vehicle mto maJOr
assemblies, subassemblies, component parts, features and attnbutes Chrysler places great
Importance on Idenllfical!on, documentatiOn and management of functiOnal features on both
component and assembly Jigs

Chrysler operates an mtegrated product development (IPD) strategy to help dnve Its concurrent
engmeenng philosophy. W1thm this strategy there exists a DM engmeenng group made up of
approximately I 25 engmeers Each IPD team will be assigned I or 2 engmeers who will become
responsible for the dimensiOnal quality of that product sectiOn. Important feature tolerance
attnbutes are Jdenl!fied and managed by a special team of 3-4 engmeers who apply all geometnc
dJmenswmng and tolerance call outs m accordance to the ASME YI4.5M, 1994 standard Only
approximately 25% of the DM engmeenng group are Chrysler employees The remammg 75%
are short to medium term contractors from vanous compames such as VSA, DCS and Tnkon
The core DM expertise IS held with the company and many of the contract workers funcl!on as a
flexible vanatwn modelling resource to the core DM group The DM engmeers have orgamzed a
DM techmcal club There IS no corporate reqmrement to attend meetmg and engmeers orgamze
their own agenda and schedule This appears to be smular m nature to the process of quality
Circles

Chrysler IS committed to DM techniques and all current and new engmeers receive in house
GD&T and DM basic trammg Engmeers considered for DM related posJI!ons are selected from
the followmg cntena:

• Perceived as havmg a good understandmg ofDM techniques.

48
• DM team leaders are reqmred to have constderable expenence m toohng destgn/bmld and
gaugtng
• Mmtmum acadenuc quahficatwn - engmeenng degree.

Clrrysler beheves that tt takes up to stx years of trainmg and expenence to become a competent
DM engtneer.

The dtmensional analysts software tools at Chrysler are VSA-3D and VALISYS Assembly
software for Tecnomal!x Both are heavtly mtegrated mto thetr pnmary CAD system, CATIA
The DCS software had also been mvesttgated but currently was not bemg used. The VALISYS
Programnung and Inspectwn modules were also bemg used for dtgttal mspeclton.

There has been one mam drawback of the DM process recogmzed at Chrysler The engtneenng
environment has percetved the VSA system as analyl!cally based, and not as a pracl!cal tool set
Thts has led to problems in terms of resistance to cross funcl!onal commumcal!on, a vttal resource
to the DM engtneers There ts also related umon based problems because of what they regard as
'exposure' of some engmeenng operatiOns to other parts of the orgamsatwn

Chrysler clatm that the introducl!on of DM techmques has Impacted on product quahty in two
areas Ftrstly, the reductwn of vehtcle program development lead-limes resultmg m cost savmgs
and dehvery lime to market. Secondly, dtmenswnal quahty Improvements m velucles have been
acknowledged.

3.1.2 Ford Motor Company

The Ford Motor Company vtstted is based at Vehtcle Operatwns, Oakwood Boulevard, Dearbom,
MI

Ford has been mvolved wtth DM techmques smce the early 1990's Ford, stnularly to Chrysler,
became mvolved wtth DM tecluuques after beconung aware of the clatmed successes at GM
They currently run a large DM program and are clearly comnutted to the techmques. Part of the
Ford strategy has been to set up and mamtam a DM department located at the body and assembly-
vehtcle operal!ons based at Dearbom Ford constders the whole vehtcle construction for variation
management mcludmg analysts on

• Door, trunk, hood to BIW fixed panel flush and gap condttwns
• Drive tram analysts - engine, gearbox, axle, etc

49
As appears to be the case with most automotive manufacturers, the pnmary Issues address by DM
techniques at Ford are aesthetics such as steps and gaps for example

As part of their product development strategy Ford are givmg more up front consideratiOn to
design for manufacture (DFM) Process capability reqUirements are agreed at the beginning of
the program based on histone SPC data and dimensiOnal quality specificatiOns for a vehicle
program are recorded m the LB-506 document, known m Ford as the 'Gray book'. The Sign off of
this document becomes an Important stage of the program It effecbvely comrmts many
engineenng sections to a product and process specificatiOn which becomes difficult to revise m
the later stages of the program Important functiOnal features are 1denhfied through the
_ _ Significant Charactensbc (SC) process which IS similar m nature to the Key Charactenstlcs (KC)
process discussed m the review.

Ford's productiOn pohcy currently does not use a geometnc d1menswmng and tolerancmg
standard (GD&T) Havmg spoken to members of the DM department the author IS of the opmwn
that there IS current pressure from scmor management at Ford to mtroduce a standard such as the
ASME Yl4.5M for use m body design Tolerance specificatiOns tend to be coordmate pomt
based which references a digital model created dunng the reverse engmeenng process of vehicle
clay mock-ups Manufactunng and assembly would hke design to define tolerance specificatiOns.
However, design would prefer If manufactunng and OM secbons became responsible The root to
the problem IS that currently only design can approve and release drawmgs and are therefore
accountable for their content Design would hke manufactunng and OM to provide mput to the
drawmgs but also become responsible for the content, somethmg manufacturmg appears reluctant
to undertake.

Ford have developed their own comprehensive common locatiOn standard which is being used on
all vehicle programs m the USA The common locatiOn scheme has helped standardizing locabon
features and their associated tolerances on both components and manufactunng/assembly tooling
Features of datum have also been mcluded m the locatiOn standard. It IS now policy for tooling
design to orientate fixture part planes normal to vehicle coordmate axis system m x, y, and z

Ford has pnmanly been usmg the VSA software for circa fourteen years. They have also been
usmg the DCS software for Circa seven years Currently the OM department have engineers who
can operate the computer analysis software, but some of the modelling IS undertaken off sile A
maJOr subcontractor to Ford IS Craftline Inc Crafthne manufacture assembly Jigs for vanous
vehicle programs and they use the DCS software for their own computer analysis. The bmld

50
analysis IS presented to Ford as part of Crafthne's quahty assurance documental! on Most velucle
design work 1s undertaken in the PDGS Ideas CAD system

The benefits clmmed by Ford as a result of pursumg a DM strategy are

• Reduction m vehicle development lime


• Reduced program development cost.
• ReductiOn m warranty claims
• Fewer problems associated with vehicle launch

3.1.3 McDonnell Douglas Aerospace

The McDonnell Douglas Aerospace visited 1s based at PO Box 516, St. Loms, MO 63166

McDonnell Douglas Aerospace is usmg DM techniques to address product design as part of a


broader mii!al!ve focusmg on the mtroduct10n of an mtegrated product development (IPD)
strategy. The mol!ve for the Jm!Jal mtroductwn of DM stems from 1ts perceived success m the
automol!ve mdustry. McDonnell Douglas beheved that these tecluuques may bnng benefits 1t
terms of product development efficiency A DM p1lot project was mtroduced m 1992 shortly
after which 11 was launched over several other projects It was used m conjunction with other
techniques such as design for assembly (DFA) and design for manufactunng (DFM) to aid m the
development of the FIS E and F vanants of the Hornet program The 1mllal project mvolved the
analysis of the sphce between the centre/aft fuselage because of the concerns idenllfied form
previOus mrcraft bmld The VSA (22) assembly analysis product was used to model the sources
of vanatwn McDonnell Douglas claim th1s has resulted m savmgs of both cost and time w1th
regard to product development and manufacture

The DM process IS mtegrated w1th their own IPD strategy The product IS spit mto several zones
each havmg an IPD team responsible for the development for that sectwn A DM engmeer IS
placed w1thm each IPD team They become responsible for ensunng target dimensiOnal quahty
specificatiOns are Jdenllfied, documented and managed for their product secl!on. DimensiOnal
quahty targets for product manufacture and assembly are developed form IPD data sheets wh1ch
contam functiOnal design mformal!on They are also responsible for the cross funcl!onal
commumcal!on and coordmatwn of common datum's and locatiOns between the product zones.
The OM engmeers m each IPD team make up a OM group who regularly meet and deal w1th
d1menswnal quahty related Issues

51
McDotu1ell Douglas suggest the mtroductwn of the geometnc dtmensiOrung and tolerancmg
standard Yl4 SM has made a considerable Impact to all sections of the orgamsation and has
htghltghted to destgn engmeers the effects of poor tolerance, location and datum philosophy. Thts
has been achieved by large scale trammg across all sections of the engineenng organisatiOn wtth
some I ,800 staff bemg trained m vanous DM related techmques The wtdespread use of GD&T
has led to the destgn mtent beconung better understood by the manufacturing, toolmg, mspectiOn
and assembly engmeers who now share a cormnon reference datum. Tlus had the effect of malang
the conunurucat10n of product development more clear and consistent and dtrectly asststs the
cross functiOnal ac!Ivtty reqmred by the IPTs'

The use of DM by McDotu1ell Douglas ts pnmanly for destgn validation and both the standalone
and mtegrated VSA assembly analysts software and process are currently bemg used on all destgn
releases.

3.1.4 Raytheon Aircraft

The Raytheon Atrcraft viSited IS based at Central PO Box 85, Wtchtta, Kansas 67201-0085.

Raytheon Aircraft employees a DM process ongmally mtroduced through small pilot projects
They became aware ofDM through the percetved success at the Chrysler CorporatiOn and dectded
to mvestigate the tools and techntques bemg used there. The ongmal ptlot project was use to md
an engmeenng analysts on a landmg gear umt. The analysts was undertaken usmg the VSA
software and the atm was to Identify and quantify sources of potenhal manufactunng and
assembly quahty concerns. Raytheon were further mterested in usmg DM to make Improvements
m the destgn engmeenng process and to gam a better understandmg of thetr current
manufactunng and assembly process capabthty

Raytheon are followmg an IPD strategy for product development and senes manufacture.
Customer reqmrements are captured up front vta a QFD type techntque wluch then tdentifies and
quantifymg the Important quahty attnbutes The product destgn ts then spht wtth each zone
becommg the responstbthty of a POT. A DM engrneer is asstgned to each team and they become
responstble for the management of dtmenswnal quahty issues. Each DM engmeer makes up a
DM group who regularly meet and dtscuss related tssues. The DM group make up part of the
quahty assurance and management department whtch ts a httle unusual as most other DM groups
etther operate mdependent of any department or are associated wtth manufactunng and assembly

52
Raytheon have mtroduced a GD&T standard mto the engmeenng orgamsation and currently have
two eng~neers responsible for the Jdenl!ficatwn and allocation of call outs on des1gns. They
beheve the placmg of GD&T call outs onto drawmgs will play a significant role m makmg the
des1gn mtent clearer A potenl!al concern for the company IS the level of understandmg of GD&T
through the engmeenng orgamsatwn The work undertaken by dedicated experts to place correct
GD&T call outs may become undone through the mcorrect mterpretation by untrained shop floor
and mspectwn person11el

One of the aims of the DM ImplementatiOn was to mvesiigate and analyze manufactunng process
capab1hty Raytheon have mtroduced comprehensive SPC techniques m order to help address the
1ssue of SPC ratmg schemes which are required to ensure the capture and mamtenance of defence
busmess contracts. Some of the data bemg generated from SPC forms the basis for tolerance
metncs to be used m assembly model analysis.

Raytheon clmm as a result of followmg a DM strategy the mam benefits have been to recognize
the weaker areas of their manufactunng process and Its resultant effect on product d1menswnal
quahty Due to a better understandmg of mspeciion techniques, they have developed smarter
ways of workmg which have led to a reduciion m lime spent on inspecl!on The OM
ImplementatiOn has been managed and mamtamed by the quahty control department Raytheon
WISh to promote a more cross functiOnal approach to product development whlch they hope will
lead to better commumcaiion and boundary demarcatiOn throughout the eng~neenng

orgamzatwns

3.1.5 Northrop Grumman

The Northrop Grumman VIsited IS based at One Northrop Avenue, 3855/63, Hawthorne,
California 90250-3277

Northrop Grumman does not have an official OM process m place but they have been operatmg a
vanation reductiOn (VR) group for some years The VR group was responsible for the
mtroductwn ofDM after becommg aware of the claimed success of the automotive mdustry. First
pilot proJeCt was undertaken on an aucraft air mtake whiCh was associated with hlgh assembly
cost and lead limes

Northrop develops aircraft usmg an IPD strategy but does not use a QFD approach to capture and
deploy attnbute data From top level requirements key charactensl!cs are 1denl!fied and

53
developed w1th the assocmted mformahon bemg recorded onto IPO data sheets. Tlus mformation
becomes the focus of all vanatwn reductiOn activihes

The VR group mtroduced the concept of manufactunng and assembly vanatwn and 1ts effects on
product d1mensJonal quahty Very httle assembly vanatwn modelhng 1s undertaken by the VR
group the remamder of which 1s contracted out to the VSA company as and when necessary. The
VR group has received trainmg on 1ssues surroundmg both the process of OM and the use of
assembly modelhng software from VSA. The1r computer analysis has been undertaken on a DOS
based VSA-30 product by they have recently moved to the VSA-30 Unigraph1cs mtegrated CAE
system. Northrop Grumman also makes us of the !GRIP software 1mbedded w1thm the
Umgraph1cs CAD system for assembly modellmg

Northrop Grumman has had an SPC program m place for several years Currently, they use small
portable computer umts known as data ffiltes to capture the ongomg SPC data. This data 1s then
transmitted v1a a company network cable to a central database Tlus data base IS mamly used for
process control but 11 add11lonally used by the VR group as the bas1s for model analysis data.

Northrop Grumman cla1ms that the VR group and the1r processes have benefited the company m
several ways. As a d1rect result of the1r work and followmg an analys1s by VSA, an estimated
90% assembly cost, and reductions m lead-limes, was recognized on the milia! mr mtake project.
Th1s was ach1eved by des1gn alteratiOns to both product and to the related lengthy sluffiffilng
process. Northrop also benefited from the type of engmeenng analys1s wluch often h1ghhghted
other quahty related 1ssues such as poor fixture des1gn. One concern of the VR group was the
perceptiOn of the1r acllv!lles by the engmeenng organ1sallon. Des1gn, manufacturing and
assembly engmeers perce1ved the VR group as bemg more akm to ded1cated computer
programmers who have become removed form the everyday challenges of a1rcraft engmeering.
This 1s an understandable observatiOn and is not exclus1ve to the Northrop CorporatiOn Th1s
problem occurs m many compames m the early stages and can be attnbuted to the a~nount of
emphas1s placed on the use of vanatwn analys1s software m an attempt to resolve an engmeering
concern. Th1s IS a misconceptiOn and the real value of DM wlll only be reahzed through the use of
much broader actlVlties coupled to a comprehensive company w1de OM process. Th1s problem of
percepllon has been exasperated some what by the difficult and user-unfnendly analys1s software,
wh1ch IS currently bemg addressed w1th 1ts mtegrallon mto CAD and VJsuahsatwn systems.

54
3.1.6 The Boeing Company

The Boemg Company vtstted ts based at the Commerctal Auplane Group, PO Box 3707, M/S SC-
01, Seattle, WA 98124-2207

Boeing clatm to be the first large aerospace company to mtroduce the techmques of DM mto thetr
product development strategy Theu mo!tve, as wtth McDonnell Douglas, was to inves!tgate the
success bemg clatmed by the automo!tve companies m vanatwn reduction and mtroduce these
techniques It mto thetr own organizatiOns

DM was first introduced into the orgamsatton m I 991 by the subcontracting of several VSA
analysts engmeers m house. The number of subcontractors have been steadily reduced over the
years each one bemg replaced by a fully tramed Boeing employee The strategy was to transfer
the expertise and techmcal know how from VSA mto the DM group of engmeers at Boeing,
therefore transfemng and holdmg the expertise m house Makmg the same changes as other
aerospace orgamza!tons, Boemg have partly dtssolved the functiOnal destgn organtsation and
replaced tt wtth mtegrated product destgn teams The teams of engmeers are made up from
dtfferent dtsctphnes, wtth each team bemg responsible for a section of the product. The role of
the DM group IS to support and work wtthm the IPD teams takmg overall responstbthty for all
aspects of dtmenswnal quahty. Boeing are usmg DM for more than product and process design
vahdatton purposes, that IS they do not undertake computer analysts on every drawing released as
pohcy. They have developed a techntque known as Key Charactensttcs (KC) as part of a
Hardware Vanabthty Control (HVC) program Thts techntque ts used to tdentifY product features
whose uncontrolled vanation wtll play a major factor m product fit, performance, and semce hfe
KC are tdenttfied and captured at the early stages of the destgn process. The process dtssolves the
defined customer reqmrements down to the system level detatled features and attnbutes through
the 'drawmg tree'. These are then bmlt vta the process of product assembly through 'but Id trees'.
These key charactens!tcs then become the focus for the DM process and are managed through all
the destgn, development and manufactunng stages

Boemg have also introduced geometnc dtmenswmng and tolerancmg (GD&T) to extenstve parts
of their engineenng orgamsatwn and have expenenced the same success as McDonnell Douglas
They have adopted the ASME Yl4 5M standard for all tolerance call outs on certain atrcraft
projects

A great deal of Boemgs products are subcontracted out and the vanatton management of these
products are a major tssue wtth regard to quahty Subcontractors are tramed and expected to

55
follow the KC techniques mto their level. Boemg suggests that the DM engmeers Within IPD
teams need have excellent conunurucatwn strategies Tins is because often the situation exists
where a subcontractor may be supplymg part of a product that will mclude more than one IPD
team

In addition to the US study tour, a number of European based aerospace manufacturers associated
with DM were visited by the author. The findings of these VISits are presented m the followmg
sectiOn

3.2 Dimensional management within European aerospace


companies

A further review of aerospace companies withm Europe perceived as practicmg DM was


undertaken. These compames were:

• BAE SYSTEMS.
• Saab Military Aircraft
• Airbus

3.2.1 BAE SYSTEMS (1998)


Site VISlt:
Warton Aerodrome, Warton, Preston, Lancashire PR4 lAX.
Samlesbury, Balderstone, Blackburn, Lancashire BB2 7LF

BAE SYSTEMS formerly know as Bntish Aerospace Military Aircraft and Aerostructures (BAe
MA&A), were operatmg some pilot schemes in DM with the first emergmg m approximately
1994 The motive for DM came from the assembly orgamsatiOn seekmg Improvements m
component and assembly vanabihty. This agam was to mvestigate claims by the automotive and
aerospace mdustnes m the successful reduction of manufacturmg and assembly vanation
followmg a DM strategy

It has pnmanly been used to evaluate variatiOn m airframe structure assembly and Its effect on
component mterchangeabihty on the Typhoon (Eurofighter) program. The main pilot proJects
have concentrated on the dimensional analysis of the front fuselage structures usmg the VSA and
VALISYS Assembly software product. Advances made m the Typhoon warplane design has

56
presented engmeers w1th new challenges w1th regard to d1menswnal accuracy. Unlike prevwus
aircraft, the Typhoon's outer panels are des1gned to carry a Jugher proportiOn of the dynarmc
loadmg and some are also reqmred to be mterchangeable This combmation of reqmrements
reqmred careful analysis to facilitate the allocatiOn of the correct tolerances In addition, 1t IS
bemg used to aid the development of a new advanced manufactunng process, and associated
toohng and mspectwn philosophy. BAe MA&A use the Dassault Systemes CATIA CAE system
and are usmg the integrated digital assembly and mspection modules of the Tecnomatix
Technologies VALISYS product to analyse potential variatiOn problems

BAe have mtroduced an IPD strategy mto their orgamsatwn and are movmg away from ng1d
departmental product development They are developmg their own DM process wh1ch is tightly
mtegrated mto their product development process The process places - emphasis on
- -
the
comprehensive capture of customer reqmrernents and the1r breakdown mto company wide steak
holders. The process follows through all stages of development and senes manufacture

BAe MA&A believe that DM should not only be used to a1d product des1gn but should
additionally be used to Simultaneously des1gn and develop manufactunng, assembly and tooling
processes. Th1s means both product and process become the subJeCt of consideratiOn and
analys1s. A comprehensive SPC program m now m place and tlus will be used to momtor, control
and evaluate manufactunng and assembly process capability

GD&T is gradually bemg mtroduced mto the orgamsation and a number of trammg programs are
underway. DAe MA&A w1sh to reduce the use of company standards for tolerance allocation and
focus on the comprehensive allocatiOn of tolerance call out w1th a clear datmn of reference. Th1s
mforrnation IS also mtended to be used by mspectwn m order to make the management of
consistent d1menswnal quality consistent.

3.2.2 Swedish Aeroplane AB (1997)


S1te VISit:
Warton Aerodrome, Warton, Preston, Lancashire PR4 lAX
Swedzsh Aeroplane AB (SAAB) are based at Lznkopzng, Sweden but were vzsited at BAE SYSTEMS
Warton.

SAAD started to mtroduce DM mto the orgamzation dunng 1996. Once agam, the perce1ved
success of other aerospace organizations in the reductiOn of product vanatwn and claims of
reduced development lead-times and cost provided the mohvation to mvestigate the tecluuques of
DM

57
SAAB remam at the early stages followmg a pilot project wluch began m the early part of 1997.
The prOJect addressed the area of tdenhfymg and controlling the 'key charactensttcs' of an
mterchangeable canopy assembly and how tlus relates to Issues of GD&T control. The mm was
to tdenhfy, quantify and manage these charactenshcs from the up front design stage through to the
senes manufacture The development of process and product data wtll be used for the purpose of
vanatwn trace abthty at any stage of product development As part of thts development Saab has
an ongomg SPC progrmn wluch is bemg used to tdenhfy levels of vanatwn associated wtth each
manufactunng process.

The mam purpose of the ptlot proJect was to develop an ImplementatiOn plan for a tolerance
management methodology The mm IS to reduce time and cost of product development, reduce
~ - -
levels of rework and Improve product dtmenswnal quahty.

3.2.3 Airbus (1998)


Stte vtstt:
Airbus, New Filton House, Ftlton, Bnstol BS99 7AR.
Airbus, Broughton, Chester, Clwyd CH4 ODR.

Both Atrbus sttes were VISited m order to identify and define the case study work to be undertaken
in support of tlus thests The work undertaken followmg these VISits are presented in chapters 6
and 7 of thts thests

A small number of European based aerospace manufacturers associated wtth DM were tdenttfied
and VISited The findmgs of each of these VISits have been documented and presented m thts
sectiOn of the report. The Airbus VISit was of particular mterest as these led to the IdentificatiOn
of a number of small proJects whtch have been use to support the findmgs of thts research proJect

3.3 Dimensional management case examples

The followmg tables htghhght a selectiOn of DM analysts cases It mcludes information such as
company name, problem case subject, and the computer analysts tools used
The case examples were Identified from the US study tour, European company actiVIty research,
and the authors own commercial knowledge. These analysts examples have been undertaken
between 1996 and 2004

58
Some aerospace case examples are gtven below:

Company Problem case Software tools


McDonnell Douglas Analysts of sphce between VSA, UG CAD mtegrated, and
Aerospace centre/aft fuselage sectwns on FI 8 Cogmtwn- Mechamcal
Advantage
Aerospattale Cockptt destgn- hnkage analysts Cogrutwn- Mechamcal
on Rafael Advantage
BAESYSTEMS Analysts of Typhoon (Eurofighter) VSA, CATIA V4 CAD
front fuselage as part of Proof Of mtegrated, and
Concept study VALISYS assembly
Raytheon Aircraft Analysts oflandmg gear assembly VSA,DCS Both standalone on
on Beechcraft PC
Swedtsh Aeroplane AB Prehmmary studtes on Gnpen Intendmg to use VSA, CATIA
(SAAB) canopy structure for CAD mtegrated
mterchangeabthty
Northrop CorporatiOn Analysts of arr mtake assembly VSA
(atrcraft type not known)
Boemg Analysts oflarge fuselage, VSA and CATIA V4 CAD
closures, wmgs and flymg surfaces mtegrated, V ALISYS
on many cornmerctal aerostructure Assembly, and Cogmtton-
destgns Mecharucal Advantage DCS
CATIA VS mtegrated
Arrbus Analysts of wmg box structure on eM-ToiMate DCSCATIA VS
A3XX concept arrcraft and mtegrated
A400M
Ftgure I I Case examples ofDM m the aerospace mdustry

Some automottve case examples are gtven below.

Company Problem case Software tools


Ford Motor Company Analysts ofBIW assembly, hght VtsVSA and DCS CATIA
clusters, suspensiOn, engme mtegrated
deckmg on Focus (Cl 70)
Analysts of engme deck, and
mstrument panel on Mondeo
(CDWI62), other
Chrysler Corporatton Analy<ts ofBIW assembly, dnve VtsVSA and DCS CATIA
tram assembly, and mtenor fit of mtegrated
Neon and a number of trucks
Jaguar Analysts of body assembly, bonnet VtsVSA and eM-ToiMate
and boot closures on XK8 (X I 00)
Analysts offuiJ body on XJ6
replacement (X200)
Analysts (FEAD) on AJ28 V8
engme, other
MG Rover Group Full BIW analysts on 600 senes VtsVSA and eM-ToiMate
replacement (RDI)
Analysts on FEAD ofRDI engme
SuspensiOn analysts of Rover 7 5
Full BIW analysts on MGTF
General Motors (GM) Analysts ofBIW, (velncles not VtsVSA, DCS standalone, and
known) CATIA V4/5 mtegrated
Analysts of dnve tram assembly,
(vehtcle not known)
Ftgure 12 Case examples ofDM m the automottve mdustry

59
These aerospace case studies provide a valuable ms1ght mto the application of DM The most
notable examples are BAE SYSTEMS, McDonald Douglas, and Boemg who used to DM
technique to deternune early product assembly bmld vanatwn as part of a design validatiOn
process. BAE SYSTEMS also used DM techniques to help Identify smtable manufactunng
process specificatiOn, for example, CNC machme tool capability reqUirements Some case
studies, namely Northrop Corporation and Raytheon Aircraft, were undertaken based on root
cause analysis at the full productiOn stage.

The followmg section md1cates the type of product development area for DM analysis w1thm
different mdustnal sectors. Tlus list has been produced by the author based on commercial
knowledge, but IS not exhaustive of all current DM applicatiOn case studies Tlus case study
matenal1s not avmlable in the pubhc domam. _

Automotive sector:
Power tram
• Mamfold assembly
• Beanng clearances
• T1mmg gear alignment
• Seallocatwn
• T!mmg vanatiOn (piston to valve clearances)
• CompressiOn ratiO.
• Fuel pipe alignment
• EGR p1pe mstallation
• Exhaust bracket design
• Accessory dnve belt ahgmnent
• Engine mounts and deckmg
• Gearbox functional analysis.
BIW
• Complete BIW assembly capab1hty
• BIW functiOnal fit
• lntenor/extenor packagmg
• Power tram packagmg.
• Toohng manufacture and assembly standards evaluation
• SuspensiOn kinematic analysis

Aerospace sector:
• Aircraft fuselage structure final alignment

60
• AI fuselage structure/CFC panel mterchangeab11Ity functiOn analysiS.
• Aucraft structure fuselage/wmg redesign evaluation
• Aircraft structure mam closures fit analysis
• Wmg assembly process capability studies
• Aero engme mstallatwn and alignment
• Jet engme reverse thrust shell kmematic function
• Jet engine nozzle kinematic analysis.
• Undercamage assembly.
• Manufactunng process capability study
• Tooling manufacture and assembly standards evaluation
• Aircraft cockpit linkage kmematic analysis.
• Cockpit canopy assembly analysis ---

Medical equipment sector:


• Inhaler mechamsm validatiOn
• Drug d1spensmg machmes.

Electrical goods sector:


• Thermal sw1tchmg kinematics analysis
• Electncal switch gear assembly and function analysis.
• Mobile phone assembly validatiOn
• Photocopy assembly and functiOn
• Computer pnnter assembly and functiOn

Heavy plant sector:


• Engme alignment analysis
• Engme essential eqmpment alignment
• Armoured vehicle hull assembly validatiOn

3.4 Dimensioanl Management best practise critique

In order to establish DM best practise a study tour of key compames was planned and executed
The tour mcluded a number of leadmg automotive and aerospace compames and gave an ms1ght
to how DM methodologies were bemg used by these two mdustnes The tour mcluded VISiting

61
two automol!ve and four aerospace comparues m the US, and three aerospace companies m
Europe.

The tour revealed that the two US automol!ve companies, Chrysler and Ford, have been and usmg
DM techniques for a longer period of lime compared to the aerospace compames Both
companies clmm to have a fully documented DM methodology as part of their overall PLM
strategy and this IS used m both a prevental!ve and trouble shooting capacity. Tirree of US
aerospace compames, McDonnell Douglas Aerospace, Raytheon Aircraft, and Northrop
Grumman, only use DM techniques for the purpose of nsk nul!gatiOn on hnuted areas of the
a1rframe Idenllfied as bemg a potenllal problem The areas of concern are Idenllfied either from
past bmld expenence (carry over geometry) or from new and complex geometry systems None of
the three US aerospace comparues have an integrated approach to the use of DM _with respect !O
an overall PLM process. The Boemg Company d1d however have an mtegrated DM methodology
to support aircraft development and manufacture throughout the PLM hfe cycle. Tlus
methodology IS called HVC and mcludes the extensive use of KC feature Jdenllfiers for product
defimllon and validatiOn.

The DM study tour was extended to mclude three European aerospace compames, these bemg
BAE SYSTEMS, Saab M1htary Aircraft, and Airbus As IS the case with the majonty of the US
aerospace compames, all three European compames only used DM techniques for the purpose of
nsk nul!gatiOn on potenllal problem areas and d1d not have an mtegrated methodology. The
possible excepl!on to this was Airbus Airbus do not operate a comprehensive DM methodology
as part of an overall PLM process, but they do undertake complete zone analysis An example of
tJus type of zone analysis could be the dimensiOnal mvestigatiOn of a complete wmg structure

3.5 Concluding comments

The mdustnal review of DM pracl!ces m automotive and aerospace has complemented the
literature review by exposmg the mdustnal and commerctal best pracl!ce. The literature and
mdustnal reviews m chapter 2 and chapter 3 provtde the basis mformat1on from which the DM
need can be drawn This need IS outlmed m chapter 4

62
Chapter4

4 The need for dimensional management

Objectives: The mam objecttve of this chapter IS to outline the need for a DM methodology
w1thm the aerospace mdustry based on evidence presented m the literature and mdustnal reviews
presented m chapter 2 and chapter 3 respectively This chapter further outlines the case for the
DM methodology and It mcludes

• Outlme the relattve Importance of vanatwn control m mrcraft manufactunng


• Highlight future challenges m aircraft structure manufacture
• Make the case for DM m the development of next generatiOn aircraft

4.1 The need for a dimensional management methodology

As previOusly h1ghhghted, Withm the current compettttve aucraft mdustry enVIronment, new
design, manufactunng and assembly methods are bemg developed to address the Issues of
affordab1lity and to provide a more consistent product (Muske 1997) One area of mrcraft design
and development could be far better explOited m order to achieve these aims, this area IS DM.

Trad1ttonally, design engmeers are pnmanly concerned With Issues of functiOnal reqmrement,
structural mtegnty, and style The commumcatton of tolerance, datum and locatiOn schemes of a
product then becomes a secondary concern (Crmg 1996) An Important part of the design process
reqmres the determinatiOn of nommal d1menswns and the applicatiOn of tolerances These need
to be adnunistered such that each md!Vldual part will meet the performance mtent of the design
and specific functiOnal reqmrements, mterchangeab1lity for example DimensiOns therefore not
only specify the SIZe and shape of a product, but they mamtam the design mtegnty of the parts
dunng manufactunng and assembly (Henzold 1995; Spotts 1983) Many of these Issues are
covered by blanket company and mternatwnal standards and are then left to vanous
mterpretatwns by the down streain ac!tvittes of rnanufactunng, toolmg and assembly bmldmg m
unnecessary cost mto some cases This has led to the existence of the 'hidden factory' (Leaney
1996), where even at the late stages of productton operators and fitters are "adjustmg" out
vanatton problems The result IS productiOn waste affectmg domesttc affordabdity, as well as
comprormsmg product quality.

63
Sources of waste due to vanatwn {dtmenswnal notse) wtll need to be idenltfied and managed tf
atrcraft manufactunng orgamsatwns wtsh to become more compeltltve Vanalton waste can be
found m manufactured components and subassembhes, assembly fixtures and tools, the
destgn/manufactunng/assembly/mspeclton processes, and human mterventton (McCmslton 1994)

Wtth the advancement of CAD/CAM/CAE systems mtegratwn and a strong reqmrement for
enterpnse collaboralton there ts an opportumty to further evaluate the 'dtgttal' product and
process at the early hfecycle stage to venfY destgn, manufacture, and assembly against customer
spectficatwn There are however, potenttal drawbacks. Most destgn and analysts work ts
undertaken m the dtgttal environment whtch can become detached from the actual manufactunng
world where product and process IS fraught with vanalton. Withm thts vtrtual and prectse
envtronment the consequence of actual manufactunng and assembly variatiOn cannot be foreseen
posmg an engmeenng development nsk The DM process can address thts by perfonmng nsk
mtltgatwn on product and process destgn by evaluatmg thetr robustness to actual upstream
vanation. Thts evalualton may be done usmg vanatton analysts software tools These software
tools asstst in the venficatwn of destgn, manufacture, and assembly tolerance spectficatton
agamst product attnbutes such as flymg surface step and gap

The mtroduclton of dtmenstonal vanation from upstream manufactunng processes m addtlton to


poorly defined locatiOn and datum schemes may produce major problems for an assembly
orgamsatwn, (Jeffreys, Leaney and Wood 1998) Gtven that mdustry experts suggest that
approximately up to 60% of senes productiOn waste ts attnbuted to assembly process problems
there seems to be JUsltficalton to mcorporate DM techmques m order to reduce this metnc.

4.2 Manufacturing challenges for future aircraft

The lndustnal College for the Armed Forces (ICAF) defines four categones of atrcraft
manufacture, these bemg (ICAF 1998)

• Commerctal fixed-wmg atrcraft (commerctal atrcraft).


• Mthtary fixed-wmg atrcraft {truhtary atrcraft).
• Rotorcraft
• Jet atrcraft gas turbme engines.

64
The scope of this thesis is confined to the commercial and rmhtary fixed-wmg sectors, although
the broader pnnc1ples relatmg to DM could equally be apphed to the rema1mng sechons.

The next generatiOn of both rmlitary and commercral aircraft will be designed usmg novel
concepts, new matenals, and revised manufactunng and assembly strategies This new generatiOn
can be partly cancatunsed by (Jeffreys, Leaney and Wood 1998)

• Integrahon of digital design and test env1romnents - virtual product development.


• New airframe structure design concepts
• Introduchon of new matenal types
• Development of Improved aircraft mamtenance systems to reduce owner-operatmg costs.
• MigratiOn from master toohng gauge systems to digital mspectwn
• IntroductiOn of complex component and assembly mspechon analysis techniques
• Investment reduchons m manufacturing and assembly toohng and fixtures.
• Part to part assembly strategy (nummum fixture toohng concepts)
• Much greater emphasis on product and process affordab1hty.

4.2.1 Military fixed-wing aircraft

In the nuhtary sector the design and development of programs such the Eurofighter Typhoon and
the JOint strike fighter (JSF) has heralded a departure to conventiOnal aircraft structure design.
The maJor component of dynarmc load m traditiOnal aircraft IS earned by the a1rframe structure
With the extenor skms prov1dmg the aerodynarmc shape. This can be hkened to a car w1th a
chassis With the outer body panels proVIdmg functwnal shape and aerodynarmcs. TradJI!onal
aircraft structures and extenor skins, such as those m the BAe MA&A Tornado GR4 and Hawk
T45 for example, are mamly assembled usmg large numbers of mechamcal fasteners such as
nvets and bolts. Structure and skin components are dnlled usmg 'hard' (fixed) Jig/fixture tooling
templates and the holes fac1htate both down stream locatiOn and fastener msertwn access features
Adequate levels of process repeatab1hty are achievable with this type of toolmg regime However,
It becomes difficult to make toohng adJustments m response to component or subassembly
vanatwn created m earher upstream manufactunng and assembly actJVJI!es A further
disadvantage w1th tradJI!onal Jig and fixture tooling IS that It IS produced from a master toohng
gauge or master med1a wh1ch are physical representatiOns of the aircraft subassembly design
These are used to create, and transfer product standards throughout a toohng farruly (Fowler
1997) There IS also a significant cost associated w1th the constructiOn, mamtenance and
mod1ficahon of these tradJI!onal toolmg methods.

65
A number of mihtary aerospace manufacturers have movmg away from th1s traditional
manufactunng and assembly methodology reahsmg 1t would not be capable of dehvenng a
product wh1ch would fulfil (or exceed) customer techmcal reqmrements g1ven the mtended future
a~rcraft des1gns

The des1gn of future nuhtary a1rcraft structures w1ll be characterized by a monocoque concept. In
th1s concept the majonty of the dynaJruc Joadmg of the a1rcraft IS carried in the outer panels wluch
also function as aerodynamic flymg surfaces. The panels are then assembled to a s1mphfied ultra
hghtwe1ght metalhc and compos1te substructure Tlus can be hkened to a modem high
performance racmg car where the structure and outer shape are combmed mto a monocoque
configuration. One of the pnmary objectives of all nuhtary a1rcraft des1gn 1s to mcrease the
dynanuc fl1ght envelope and tlus can mamly be achieved through the reduction of a1rcraft mass.
Current and medmm term programs will ach1eve th1s through the introduction of new matenals
and manufactunng processes producmg components w1th an mcreased strength to we1ght ratio.
Compos1tes such as CFC and glass remforced plastic (GRP), and advanced metalhc' such as AI
and T! are an example of matenals Some compos1te materials have been use m a more hnuted
way for a number of years m the aucraft mdustry but generally for pnmary structure. The
mtroductwn of major CFC panel components throughout the a1rcraft has allowed the monocoque
des1gn to be reahsed given 1ts excellent strength/stJffness/we1ght ratios

Aggressive customer demands place further reqmrements to be absorbed mto concept des1gns.
An example of th1s would be the mtroductwn of component mterchangeab1hty (ICY) on some of
the pnmary structure CFC panels A number of the CFC panels on current and future a1rcraft w1ll
need to be removable to allow access for the mamtenance and repair of a1rcraft dunng down time
Tlus has required a trade off m des1gn spec1ficat10n between the CFC panels and correspondmg
substructures An example of th1s IS for CFC panel and substructure fastener hole poSihonal
tolerance to be large enough to accommodate ICY, but also be capable of accommodatmg the
stnngent stress reqmrements w1th regard to the hole/fastener pos1tJonal mterface

66
Uruhsed comiOOlhl dorsalassy to
nnp-ove fabncahon eff'lCJency
Reduce part count, andkllWrO&S
Flat upJ=er skm to nnprove fibre cost
placement effic~ency

FJbre steered bwersklll to


Improve fah effiCJencyand
we1ght

I Fibre placed caibon epoxy


(] RTMor RFI calbon-glass e}X)xy Stuches 1n-work..
[]Fibre plc'd&. rnan Lay-upcalbon - Sparfab &assy
- W1ng fuselage spliCe
0Alunuruurnmachmmgs - K1ck rib contmurty
I T118lUurn rnaclunmgs - Pylon nb de Slgn

Figure 13 Advanced lightweight a!Tcraft fuselage structure (ALAFS 1997)

In the constructiOn of tradihonal mrcraft the maJonty of structural and external skm components
are made from alurrumum/htaruum (AVT1). These matenals are duchle and may be stretched or
flexed to smte These AI and T! components therefore can be regarded as 'conformable' m
nature This allows the vanatiOn of a component to be absorbed through 'finessmg' during the
assembly process. Future a1rcraft monocoque des1gn concepts w1ll reduce the opporturuty for
component and assembly finessmg The incorporahon of CFC m both the outer panel surface and
substructure w1ll be substanhal and by the nature of the matenal w1ll not 'conform' The concept
also mcorporates a reductiOn of structural components Where AVTI substructure assemblies are
to be used their part count will be reduced compared to tradihonal aircraft, resultmg m structural
component bemg more complex, substanhally stronger, and more ngid

An exaJnple of the next generation a1rcraft des1gn technology has been tested on a section of the
F/A-18 ElF Hornet IDJhtary a1rcraft The purpose of the advanced hghtwe1ght aucraft fuselage
structure proJeCt was to explore and develop new design methodologies by conductmg a 'clean
sheet' design on a sectiOn of the aircraft. This involved combmmg the centre fuselage and the
Inner wmg sectiOn mto an mtegral airframe assembly A diagram of the concept design can be
seen m the figure above.

The mam dnvers for rruhtary aerostructures are

• Low observab1hty, reqmres the aerostructure properties to have spec1fic control with
regard to step and flush specification.

67
• Increased fltght envelope, reqmres an mcrease in aerostructure mass to dynamtc strength
ratiO achieved through new materials and manufactunng techmques.
• Variant flextbthty, destgn of a core common aerostructure which, wtth the use of
modulanty, may be used as the basts of several denvahves
• Reduction m aircraft umt purchase cost; achieved through the mtroduction of lean and
agtle manufactunng process and modular destgn pnnctples.
• Htgher maintamabthty standards; achieved in the mam by destgnmg customer ICY
spectficatlons.
• Target market; customers are domestic and foretgn governments facmg spectfic defence
threat.

4.2.2 Commercial fixed-wing aircraft

Currently, only two maJor commercial atrcraft manufactunng orgamsatlons extst, Boemg and
Atrbus. Both compames are considenng future super JUmbo destgn concepts, but Atrbus has
taken the mtttatlve wtth the development of the A380

It ts hkely that the commerctal sector wtll follow the lead of the rruhtary sector atmmg for destgn
stmphficatlon, part count reduction, and the further mtegration of compostte wtth metalhc
matenals for pnmary and secondary structural assembhes, parttcularly m wmg design and
development. Composttes such as CFC and GRP have been an estabhshed technology for
structures such as honzontal and verttcal stabthsers (tat! sectwn) and wmg box stnngers for some
years. However, tt has never been used extenstvely as a pnmary structure matenal, for example,
for an enttre wmg box constructwn Thts may change for future atrcraft as destgners pursue
vanous concepts for wetght reductiOn advantages.

One of the critical elements of any large atrcraft destgn tS the wmg constructiOn Aubus tS
explonng the posstbthty of producing a wmg box constructed predommantly from compostte
matenal This ts a stgntficant departure to tradttlonal wmg construction that conststs of mostly
Alfft matenals. Current wmg box assembly mvolves the use of techntques known as 'fetthng'.
The fetthng process reqmres destgnated interface surfaces on wmg structure components to be
dehberately manufactured wtth excess matenal which ts then removed to facthtate assembly to a
spectfic cntenon. Conversely, where not enough matenal extsts a 'shtmmtng' process is
mcorporated whtch involves the bmldmg up of a material surface at a subassembly mterface
These techniques have evolved because certam wmg components and subassembly mterface
features cannot be controlled well enough to achteve the target specification An example of thts
where excess material ts left on the mterface features of wmg nb and wmg skm components

68
known as 'vanal!on float'. During the assembly process the reqmred matenal thickness on the nb
mterface surface is calculated by measuring the correspondmg skm thickness The fetthng
operatiOn IS used to remove excess alurmmum matenal usmg an assembly Jig mounted routmg
tools. Th1s technique IS necessary due to an assembly plulosophy that subrmts to not bemg able to
management component accuracy such that no mterventwn IS reqmred These techn1ques lead to.

• Extra cost mcurred directly attnbuted to the assembly process.


• Lead-lime of wing box assembly is longer than necessary
• Each wmg box constructiOn IS geometncally umque resultmg m an mcreased nsk w1th
regard to subsequent subassembly bmld opera lions
• Opportumties for component and subassembly interchangeability at the Jrulial assembly
stages and at subsequent mamtenance stages are 1mpa1red
• Concern for environmental control on the assembly hne; the fetthng process results m
small particles bemg produced and depos1ted on the product and toohng

The introductiOn of CFC matenals Will greatly reduce the vJab1hty of fetthng techn1ques due to
the constrammg nature of the matenal Much greater control of component to subassembly
mterface features will be reqmred and this can m part be ass1sted through the DM process

Commercial aerostructures mam dnvers:

• Cost of ownership, reqmres fl1ght efficiency wluch IS dependent on structure ahgmnent,


stnngent step and flush specificatiOn to avmd turbulence
• H1gher mamtamab1hty and operatiOn flex1b1hty; reqmres customer levellCY.
• Safety, reqmres matenal structural mtegnty.

4.3 Supporting activities/techniques to aid the DM Process

The prevwus secl!on lughhghted some of the new des1gn concepts anlic1pated m the next
generatiOn of military and commercial a1rcraft TradJI!onal a1rcraft engmeenng praclices will
need to be evaluated for their smtab1hty m accommodatmg these new des1gn concepts In
response to the need, a number of engmeenng techniques and pracl!ces have been identified by
the author based on the literature and industnal rev1ew These aclivJI!es should be supported by a
comprehensive DM program

The DM techniques and pracl!ces are highlighted on the followmg page

69
Develop new, or revise existing, manufacturing and assembly business processes; the
manufacture of components from composite matenals such as CFC and GRP is an expensive and
spec1ahst activity and is a marked departure from AIIT1 related processes Large AIIT1 structures
elements will be geometncally complex and will need to be dimensionally robust

Adequately and effectively characterise manufacturing and assembly processes; the


mtroductwn or evaluatiOn of current statistical process capab1hty data for all manufactunng and
assembly busmess processes Tins mformatiOn should be openly avmlable to all engmeenng
sectiOns of the orgamsatwn and should be presented m a clear and unambiguous format.

Reduce manufacturing and assembly jig!Jtxture 'hard' tooling requirements; a number of


manufacturers are now usmg large precision (5 axis) machmmg tools. These~machmes are bemg _
used to dnll/countersmk and net profile CFC panel components m preparatiOn for mechamcal
fasteners These precisiOn operations allow CFC aircraft removable components to be
manufactured with Interchangeable charactenstics. Automated assembly cells are also bemg
considered for the dnlhng/countersmkmg and fastener msertwn processes

Elimination of MTG and master media; the mtroductwn of precision and flexible measunng
equipment will reduce or ehmmate the requirement for MTG and master media. These systems
are based on laser mterferometer technology (SMART 310 system from LexiCa, for example) With
dynanuc trackmg capab1hty. These systems can be used in conJunCtiOn with master digital
models created w1thm the organisatiOns CAD/CAM/CAE environments Orgamsatwns need to
move away from gauging and towards a measunng philosophy.

Develop digital inspection techniques; tlus will provide a process for component, assembly and
fixture toohng dimensiOnal quahty venfication The master digital data stored w1thm the
CAD/CAM/CAE env1romnent can be utihsed as the nommal product media onto wluch
mathematical tolerance zones known as soft gauges may be overlaid The mspection process data
may be fed back mto the digital CAD media to allow powerful mspectwn analysis Some
CAD/CAE mtegrated systems have the capab1hty to senu automatically develop off !me
mspectwn programnung for equipment such as a co-ordmate measunng machme (CMM). Tins
technique can also be used m conJunctiOn With measurement eqmpment such as mobile CMM
arms and VISIOn based mspectiOn systems.

Introduce geometric dimensioning and tolerancing (GD& T) concepts; m order to create clear
and unainbiguous component and assembly feature tolerance data a GD&T standard will need to
be Implemented and understood throughout the orgamsatwn. This will allow the senunatwn of

70
dimensiOnal quality reqUirements to engmeenng and will help m the charactensation of
manufacturing and assembly process capability md1ces The applicatiOn of feature and tolerance
specificatiOns should also be established m the dig~tal CAD/CAM/CAE component and assembly
database.

Use of variation analysis tools and techniques throughout the development process; to evaluate
product parameter charactenstics at the early stages of design, ID and I 5D vanat10n analysis
should be performed pnor to detail CAD geometry bemg available. The use of parametnc
vanation analysis (PVA) tools and techmques will validate early target product specificatiOn
agamst manufactunng and assembly capability. The analysis findmgs provide the foundatiOn for
more detailed specificatiOns to accompany master 3D CAD geometry. Comprehensive 3D
synthetic geometnc vanatwn analysis (GVA) studies should be performed usmg full digital mock-
up geometry on all aircraft zones to venfY target quality product specification. These studieS may
later be reused to investigate an assembly process concern, for the validation of process capability
levels, and for tolerance types and metncs

A key feature for future aero,pace orgamsatwns will be the effective utilisatiOn of mtegrated
product development (IPD) strategies This Will be used in an mtegrated product environment by
mtegrated product teams (IPD), (Barrow 1997) Part of the DM process reqmres the product and
process mformatwn from a number of cross-functiOnal and multidJscipiine engmeenng sechons
such as design, manufactunng, assembly and mspectwn. In this respect the DM activities directly
promote IPD activity where each engineenng sectiOn or diSCipline IS mv1ted to mput their
particular expertise ensunng that all engmeenng concerns and/or opportumties for Improvement
are taken mto account

Due to the reductiOn of substructure component parts m the nuiitary and commercial aircraft
sectors, new advanced assembly plulosoplues are emergmg. The concept mms to eiinunate a
significant proportiOn of assembly fixture tooling and Its ex1stmg cost through the applicatiOn of
new 'self tooling/location' pnnc1ples Each component and subassembly will be produced With
mherent self locatmg features and tooling features, m addition to Its nommal functional
arclutecture. These features Will be used to facilitate component to component location for
assembly purposes. Significant attention will need to given to the feature types and their control
with respect to tolerance metncs If tlus new process IS to be successful DM techniques can again
play a maJor part in the design of mdiVIdual components and nummal toohng, their features and
their gauges The process could be effechvely used to evaluate the nsk of product none
conformance and the engmeenng of alternative strategy.

71
In chapter 3 the findmgs of a study tour of USA and European aerospace compames by the author
lughhghted a number of Issues relatmg to future mrframe development programs. These Issues
relate to the design, manufacture, mspection and assembly processes of next generatiOn rmhtary
and commercial aircraft The product and process charactenstics associated With these aircraft
are more radical and demandmg compared to current designs, partly tnggered by the mtroducl!on
of new structure matenals and advanced manufactunng processes. Such developments present
challenges, and opportuml!es, to engmeers across the whole spectrum of design to series
productiOn

Tlus thesis has already h1ghhghted the development of an mcreasmgly competitive mrcraft
mdustry, which demands new design, manufactunng and assembly methods to be developed at
lower costs while simultaneously prov1dmg a more consistent and affordable product. In additiOn,
there are new demands on a1rfrmne specification, for example, higher specification for
mterchangeab1hty One of the maJor challenges engmeenng orgamsations are facmg IS how to
ensure the control of product dimensional vanation m order to achieve such aggressive
specificatiOn for mterchangeabii1ty as well as specific controls such as flymg surface steps and
gaps wh1ch fundamentally effect m service performance. This research proJect m part addresses
the challenge by

o Developmg a proposed DM methodology for use m the aerospace mdustry. The


methodology, called PARES, IS presented m chapter 5
o As part of the PARES methodology, design and develop a basehne parametnc vanatwn
analysis (PVA) tool m support of early aerostructure wmg box development. This is
presented m chapter 6
o The PARES methodology IS agam supported through the geometnc variatiOn analysis
(GVA) of a development wmg box structure dunng the early detmled design phase when
CAD geometry becomes available. The analysis IS conducted usmg advanced 3D
tolerance analysis software and IS progressed through a senes of DM case analysis
examples which are presented in chapter 7

72
Chapter 5

5 Physical architecture robustness engineering


system (PARES)

Objectives: Thts chapter outlines a proposed Ingh level DM methodology m response to the need
tden!Jfied m cheaper 4 The physteal arclntecture robustness engmeenng system (PARES) has
been developed by the author m order to structure and mtegrate DM actiVIty so tt may
complement current aerospace development tm!Ja!Jves such as the IPD process The approach
respects the genenc reqUirements of the aerostructure manufactunng organisations taking into
account the types of analysts and engmeenng prac!Jces pursued The PARES methodology ts
presented under the followmg sectiOns

• The PARES scope


• PARES methodology process, tools and techmques
• Methodology archttecture.
• PARES benefits.

5.1 The PARES scope

The PARES methodology was developed by the author m response to the expenence of industnal
surveys and placements The process ts designed to tdenllfy, analyse and manage complex
product vana!Jon (product nmse) from the early concept design stage through to full senes
produc!Jon It wtll endeavour to manage product robustness, dtrectly hnkmg target spectfica!Jon to
quality reqUirements management. The methodology challenges are:

• Evaluate dtmenstonal reqUirements agamst mtended destgn, manufacturing, assembly, and


mspectton process
• Maxnruse the use of dtgttal product development technology therefore reducmg phystcal
prototypes to a nnmmum.
• Destgn and develop the overall quality management process to complement the product
knowledge capture process
• Span the whole development and productiOn hfecycle

73
5.2 PARES methodology process, tools and techniques

The proposed PARES methodology ts made up of a number of concurrent actlVltles and wtll be
present through out the entire aircraft productiOn phase The tumng of each actlv!ly wtll need to
be syncbromsed wtth the overall systems development and productiOn method, for example the
IPD process. The actlVltles assoctated wtth PARES have been Illustrated agamst a genenc IPD
process for mthtary atrcraft production (Barrow 1997) These actlvttles can be seen m the figure
below. The top bar mdtcate the maJOr development phases and the pnnciple matunty gates, while
the nuddle bar deptcts the phase codes

Usmg the IPD phases as a genenc baselme the pnnctple PARES tasks have been developed over
the product hfecycle timeline. The task bars- md!Cate the PARES methodology phases starting
wtth mnovation and endmg wtth productiOn The PARES actlVlty bars are set m the nuddle
section. The axts on this sectiOn stgnifies the development time !me across the bottom and cost of
change nsmg m relation The shaded areas at the bottom the Illustration mdtcate the effort
reqmred at the early stages of development m order to save the peak of product and process
change further down the time hne

;/ ,..
!11
KCtSreclf'x:4boau.iGD&T ~/r ,
P?A \DflDStllo:bes 11 CMUIIJOVA30SyttMI¥:Tol.mlnA~ (}'~ ,'

r-~~~~~~~~=·=-m~~~='~~~~~l'l,
Mllllui'>s1un!&:As=rblyF~Illllys» 11 ].
._
.><V

--
Figure 14 PARES high level DM program

74
The PARES development diagram can be seen m figure 14 The process tasks are descnbed m
the followmg sectwn.

IPDffiM team: The first acttvtty involves the appointment of an IPD DM team leader The
!PO/TEAM leader wtll need to appomt a number of DM analysts engmeer to act concurrently
wtth other IPD members on spectfic phystcal zone of the atrframe. An example team structure
has been proposed by the author and has been assoctatwn wtth a genenc set of airframe zones as
shown in the figure below

1
IT ... dTI«L~b t ~-
I
[

I
c.m-« 'J'Kifttad•

Ah"S,.t-

I
Aa-Veblck
--
I

l
- - - . - -

l~T!""'-a,•J
IPD/DM ERcinHI' ~H
rl
·""-
Fenruoll'oulace J I
Air Vthfde IYSUIIII$

I I
Wota.,.,..i)"Jte- I

IPDIDME•clna:r ~H CI'JitnFP$da~;e
I I I I
IPDIDME.,p._. ~H AlfFmela~;e
I I I I
IPDIDM E•Jfrl_. H ""'" I I I I
~H I I I
·-~
IPDIDM EPcin_.

IPDIDM E•clnea> ~ H EoSr;eiCtDCre Swi'aus j

F1gure 15 IPD/DM team allocation agamst genenc atrframe zones

Competitor assessment: Thts acl!vtty involves assessmg the capabthty of other nval
orgamsal!ons Thts would mvolve a number of benchmarkmg acttvtty to evaluate compel!tor
product and process capabthty

Set dimensional targets: Define the dtmenswnal targets for htgh level product attnbutes on the
vanous defined zones of the fuselage, wmgs, empennages, and honzontal/verttcal stab1hsers. The
definmg of these zones must also be established. Product dtmenswnal targets may mclude:

• Lme of fltght (LOF) profile tolerance for surface flushness (steps), gap and groves,
wavmess.
• Interchangeability standards for removable closures for steps, gaps, seals condtl!ons
• Radar cross sechon spectficatwn for m to wmd steps, gaps, fastener proJecl!on for
fuselage, wmgs, engme air mtake geometry

PVA 1D/2D studies: Undertake trul!al parametnc vanatton analysts on available parameter
design mformatwn Thts mvolves the use of stmple generic tolerance models to help vahdate the

75
target concept design dunng the detail option development phase. An example of how tlus type of
analysis may be developed IS presented as a case study m chapter 6 of this thesis

__ ... _... __
,...
.1--::.•~.:. ~"' ~ t.• I ,_.(I~•)
'><; ~ ...... ' ~ • -' ' J

Figure 16 PVA tool mterface

The PVA analysis Will Identify potenl!al problem areas which become the focus for followmg
detailed DMU geometnc vanatwn analysis performed m full 3D.

KC specification, GD&T, and feature identification: The PV A studies will established a first
pass KC Idenl!ficatwn The detailed KC Idenl!ficatwn and documentatiOn will need to be further
developed by dissolvmg the product structure to piece part usmg 3D digital geometry to review
manufactunng process Parts are then bmlt back up to Idenl!fY the assembly process From tlus
acl!vity all PKCs', MKCs', AKCs', and StatKCs' can be Idenl!fied. Component and assembly
features can be grouped and categonsed into feature taxonomy The followmg table outhnes such
a taxonomy developed by the author as an example

Type Domain
M Manufacturing feature: A feature used to facthtate the manufacture a component
which may be subsequently removed dunng later process operatiOns
F Functional feature: A detail on a component that will fundamentally effect the bmld
quahty of the aircraft
I Interface feature: Part/assembly feature which will mterface Wlth other
part/a.sembly
D Datum feature: Component or assembly datum reference feature
L Location feature: A feature on a component used for locatiOn to an assembly fixture

Ftgure 17 Feature specificatiOn taxonomy

76
These features can themselves be further defined down to a mathematical defirnt10n. For
example, a set of assembly locatwn features on a component des1gnated as (L) can be further
defined as:

Pl-z Primary surface pomt I, m Z d1rect1on


P2-z Pnmary surface pomt 2, m Z d1rect1on
PJ-z Pnmary surface pomt 3, m Z dJrectwn
p4-z Pnmary aux1hary pomt 4, m Z direction
H-xy Secondary hole I,
S-x Tertmry slot I

The KCs need to be defined at all assembly levels m add1t1on to md1v1dual components. Th1s can
be achieved by breaking the product down level by level focussing on manufactunng and
assembly attnbutes

Assembled product (PKC)

Assembly tree (AKC)

Manufactured parts (MKC)

F1gure 18 KC assembly flow through

Once the KCs have been defined the1r relevant GD&T tolerance specification can be spec1fied as
a first pass based on manufactunng and assembly process capab1hty data. Manufactunng
capability data nJaY be categonsed by component fanuly types, matenal types, an advanced
manufactunng types Product GD&T apphcatwn IS essentially done by bmldmg the product back

77
up level by level usmg the dtgttal model where available to virtually assemble the product Part-
to-part assembly tolerance allocation must be adrrumstered such that bonus tolerance systems are
used m conJunction wtth fixed and floatmg fastener methodology Tins wtll ensure
mterchangeabthty and functiOnal assembly can be aclueved at each parent I stbhng I chtld part
mterface

Digital mock-up (DMU) to geometric variation analysis (GVA) 3D: Once the KCs' and
associated GD&T tolerance metncs have been apphed a comprehensive 3D synthetic tolerance
analysts study can be performed This IS undertaken m a 3D CADICAE/DMU envuonment and
wtll analyse the entire assembly process for hardware vanation output. It will take mto account

• Nommal clash detection and proxmuty analysts for all geometnc sets or mdtvtdual
components
• Component and sub assembly sequences.
• InteractiOn of fixture toohng to facthtate assembly
• Intermediate stages of mspectmn and subsequent process alteratiOn
• All component and assembly GD&T datum reference and tolerance metncs
• Defimtmn of assembly move type.
• Static and kmematic assembly conditions
• Process and component assembly over constramt deformation based on FEA analysts.
• Temperature effects on assembly

Figure 19 PKC vanatwn outputs

78
The analysts can be tteratlve wtth the mam product study bemg broken down mto tts major
arclutecture zones, for example the forward, central, and aft fuselages. The analysts, to an extent,
must follow the progress of 3D CAD geometry creatlon and the assembly process plannmg both
of whtch are reqmred to undertake the zone study. An example of thts type of analysis IS
presented m chapter 7 of tlus thests. It mcludes a number of wmg box structure case studtes

Manufacture and assembly fixture analysis: Tlus act!Vlty is an extensiOn of DMU product 3D
synthetlc tolerance analysts It has been separated m the PARES process due to the effort
reqmred to vahdate manufactunng and assembly fixture tooling Many aerospace orgamsattons
now outsource thts actlvtty The quahty of the design and manufacture of assembly fixture
toohng can have a huge Impact on product quahty. Products mhent much of thetr dtmenstonal
quahty form the fixture system used to place and secure them m 3D space Traditionally, the
destgn and manufacture of assembly fixture systems have been controlled by blanket coordmate
tolerance standards wluch result in less than effectlve accuracy There IS an opportumty to could
be improved assembly fixture productiOn through

• Spectficatton of fixture KCs'.


• Apphcatton of complete GD&T geometry control.
• Evaluation of component locator strategy for correct mathematlcal constramt (6 DOF
check)
• Integration of effectlve fixture feature adjustment to accommodate m servtce cahbratton

Cp and Cpk predictions and measurement analysis: In support of the DMU product 3D
synthet!c tolerance analysts actlvtty outhned above, the manufactunng and assembly busmess
centres need to tdentlfy domestlc and outsourced capabthty agamst target product spectficatton.
An example mterface IS gtven m the figure below

r..to.... M.r.-"~ •r.r•


'lill.-».o. :.Lr..lila1, .,-

()AI I
;; ..... ~
t:.~~.~-:r~ .• ~' ~~-~ ~-

i::k
__
!~::::::.-~·;:_~:-::-:_:_~--~:_:=:_::- ;:"_f.~,') -
••••••••oua•••••l••••••• '

Ftgure 20 Component SPC database mcludmg Cp and Cpk

79
The synthetic tolerance analysis is undertaken mttially wtth design tolerances based on normal
distnbuttons and target spectfication lmuts. These then become the mput to the Monte Carlo
stmulation engme whtch processes thts informatton to determme the synthehc stack outputs

Analytical model correlation: The next stage ts to convert the tolerance analysts model mto a
synthests model whtch combmes all avatlable manufactunng, assembly, and fixture toohng
process capabthty data to the nommal model The synthests model, where data ts avatlable, can
then stmulate the butld charactensttcs based on correlated actual measurement data provtding a
more accurate predtctton It IS therefore very tmportant to butld and mamtam a measurement
knowledge database (as htghhghted later) of all domestic, and where posstble, suppher
manufacturing and assembly process capabihty Where tlus mformatton ts not at all avatlable, tt
- - - -

is posstble to acqutre tt from 3"' party software and productton capabthty standards, ISO for
example. It ts tmperahve that orgamsations butld and retam knowledge of thetr own process
capabthties to be used m future development programs. Fathng to do thts wtll result m
knowledge waste- one of the key wastes of the lean productton phtlosophy

Measurement process plans: Dunng the last stages, all KC and GD&T defimtion were defined
and vahdated through the CAD/CAE/DMU synthetic tolerance model These steps wtll vahdate
the htgh level destgn specification KCs' m a cnttcal parameters brochure, agamst manufacturmg,
assembly, and mspection process capabthty metncs When all the tolerance analysts cntena are
met, measurement process plan can be developed. The core mformatton wtll be avatlable form
the last stages such as KCs and GD&T. Where appropriate, off hne measurement programmmg
may commence based out of the 3D CAD/CAE/DMU envtronments and these can later be proved
out when prototype parts become avatlable Dunng this phase, the management of development
and senes productton quahty data should be constdered Issues such as who wtll be the future
customers for thts data, what data wtll they requtre, and m what format wtll they need tt needs to
be addressed The planmng of a central quahty database for all measurement data types plus the
type of customer mterface wtll need to be constdered

Gauge R&R: Gauge repeatabthty and reproductbthty (R&R) spectficatton ts requtred to estabhsh
the appropnate medta for mspection plans defined m the last stage. The gauge R&R techmque
assesses the process capabthty of a system to perform measurements defined by an mspectton
plan. Repeatabthty IS an evaluation of how capable a measunng mstrument may be to m terms of
consistent accuracy. Reproductbthty ts an evaluation of the measurement system process
mcludmg the man/equtpment mterface. When KC and GD&T mformatton ts avatlable for
component or assembly mspectton, then gauge R&R techniques are used to define whtch

80
measunng systems are the most appropnate. Thts may be based on extstmg gauge R&R
mformation from scheduled cahbratwn routmes.

Statistical process control: At the D7 final arl!cle defimllon (FA!) development stage,
preparatiOn for stal!sl!cal process control (SPC) dunng senes producl!on should commence
Product and process SPC undertaken at vanous levels of product matunty provtdes a constant
mdtcal!on of process capab!ltty and stabthty. It can also act as an early wammg system flaggmg
when a process parameter has becomes unacceptably out of control and m need of attentiOn
Again, the key mformatwn such as KC and GD&T IS avatlable and due to the gauge R&R
mformation, the appropnate measunng systems have been defined. So the 'what' and 'how' IS

defined, all that remams IS to spectfy the 'when' There are a number of SPC standards to
calculate the appropnate sample stze and measurement frequency After evalual!on, the most
appropnate standard should be adopted An example interface ts giVen in the figure below .

- ,,....,
._,........

--
-- .
~==;7-·
"' .., __, . _,
~-

-
Ftgure 21 Post FAI SPC data defimtwn and collectiOn

Prototype and production gauge design: Dunng the prototype and senes productiOn phases
there wtll be a reqmrement to validate the dtfferent atrframe components and sub assembhes for
geometric comphance Thts IS tradttJOnally undertaken by physiCally companng the component
or assembly to a htgh spectfical!on master toolmg gauge (MTG) and other master medta
hardware. Thts means that mull!ple MTGs' are reqmred for the many stages of assembly bmld to
vahdate component and assembly resultant geometry Due to their htgh spectficatwn MTGs' are
expensive to both produce (manufactunng cost) and mamtam (environmental spectfic storage).
These costs could be radtcally reduced through the mtroduct10n of precisiOn and flextble
measunng systems. One such system ts based on laser mterferometer technology, 1 e, SMART
310, which has very htgh gauge R&R capabthty over large volumes These measunng systems
can be used m conJunctiOn wtth master dtgttal models created wtthm the orgamsal!on's
CAD/CAE/DMU environments These dtgttal masters act as a vtrtual MTG and the reqmrement

81
for physical MTGs' can be dramatically reduced or even elimmated Orgarusatwns need to move
away from gaugmg and towards a measuring philosophy.

Dimensional validation: The final stage is to repeat the update the tolerance synthesis model
agam with senes productiOn measurement data and analyse the correlatiOn between the virtual and
the physical product. This synthesis model may now be used to predict overall product
performance based on up loaded actual data This IS useful If there IS a long lead time associated
With major assemblies from a supplier If the outsource orgamsatlon has actual data on the
assembly they are about to ship, this data can be uploaded mto the synthesis model to vahdate the
vanatlon effect on bmld specificatiOn If there IS a problem, It can be highlighted before the
assembly IS shipped and a collaborative review could take place to resolve

5.3 Methodology architecture

The top level PARES architecture, developed by the author, IS Illustrated m the figure below The
architecture IS made up of SIX mam chunks each bemg hnked to the remammg. An explanatiOn of
each chunk and how It relates to the activities m the PARES process is given below

Logistics
Database

QMU
CoUa~qr~tion

Quality Data Web Based


Visualisation Knowledge

Figure 22. PARES top level architecture chunks

CAD and BoM links: Is responsible for the hnk to product and process digital mforrnation. Any
DM analysis will be undertaken based on a digital data snapshot of the overall development

82
schedule Updates and changes to digital data will need to be flagged to the IPD/DM analysis
engmeers who my then manage the analysis model The CAD and BoM links include:

• CAD defimtwn of product dimensiOnal control zones for external surface controls, steps,
gaps, wavmess, surface fimsh
• Product defimtwn assembly sequence, product vanant data, fixture tool mterac!Jon, and
mspectwn process stages
• CAD feature defim!Jon: fully defined KC and cn!Jcal parameters brochure features,
datum's', tolerances, GD&T Reqmred for both product and assembly fixture systems.
• Supplier CAD data for evaluatiOn of maJor mterfaces for Iandmg gear, engme eqmppmg,
weapon systems, etc.

Logistics database: The IogJs!Jcs database IS a central quality data depository which should
mclude all hardware dJmenswnal mformatwn The PARES aciiVIIIes relatmg to tlus chunk are

• Quality mformatwn cache: all CMM measurements mcludmg laser tracker, mobile arm
(Faro, Romer), h1gh level product specificatiOns such as steps, gaps, wavmess metncs.
• Recordmg and stonng of component or assembly pomt cloud format data to record
detailed form measurements.
• Manufactunng and assembly process SPC· manufactunng and assembly mspectwn data,
fixture calibratiOn and dimensiOnal mamtenance records
• Manufactunng group or set knowledge capture Record Cp and Cpk values for
manufactunng process sets based on group technology.
• Interface to off !me CMM mspectwn process and generate base programs
• Message alertmg of product non conformance based on aircraft zone ownerslup

DMU collaboration: The tolerance synthesis model may be used to evaluate resultant product
charactens!Jcs based on real measurement data mput from mtemal and external 3'd par!Jes. In
addJtion, contnbutor reports may be run to aid root cause analysis where specificatiOns are
predicted as not bemg met

• Use of closed loop tolerance synthesis models updated with actual measurement
mformallon supplied from a 3'd party SPC database to predict effects on overall product
specificatiOns.
• Development of digital MTG model to evaluate component or assembly pnor to supplier
sign off.

83
• Use of dtgttal MTG to evaluate batch SPC on long lead time suppher component and
assembhes

Tolerance analysis and synthesis: The tolerance analysts model may be used to vahdate
manufactunng and assembly process capabthty agamst htgh level product spectficatton.

• Create I D/20 PVA studies to evaluate early structure concepts for dtmenswnal
capabthty.
• Creatmg full 3D synthettc tolerance analysts based on detatled CAD of product and
process Analysts to be vahdated agamst product dtmenstonal spectficattons
• Spectfic nsk assessment on suspect design areas Use and mtegratwn ofkmemallc, FEA,
and custom models to perform nskrntllgation
• Enable root cause analysts where dtmenstonal spectficallons are not bemg met through
the detatled contnbutor reports

Quality data visualisation: The abihty to integrate and mme all avatlable quahty data, then
recover tt m a conststent and useful format to support engmeenng design. Currently organisatwns
hold quahty data m multtple raw formats wtth no conststent reportmg mterfaces makmg tlns
acllvtty ttme consummg and meffictent Product high level quahty data should be web pubhshed
and made accesstble to all engmeenng functtons The pubhcatton of quahty data should be
undertaken and released to a stnct schedule Spectfic ttems are

• The development and deployment of a web based quahty reportmg system made avatlable
to the orgamsatwn Thts ts to outhne product htgh level scores through the use of a
dtmenstonal quahty dashboard.
• Advanced mteracttve quahty vtewmg for GD&T and pomt cloud spectfic domesttc and
suppher data trends on components and assembhes Thts should be avatlable to all
IPD/DM analysts engmeers and personnel responstble for quahty management.

Web based knowledge: Revtew suppher hardware quahty performance through mteracttve
collaborallon When quahty data ts recetved and tt tS subject to tssue, a detatled review may be
undertaken to present the overall effect on product spectficatton usmg the root cause analysts
tolerance model. Mam tssues are:

• Store and manage all measurement data types


• Portal to quahty commumcatton ptpehne between domesttc and thtrd party supphers and
customers.

84
• Storage of all supplied actual measurement data for an SPC database.
• Collaboration to review supplied quality data, and resolve issues

The operatiOnal structure of the PARES process has been modelled usmg the IDEFO (Integrated
Defimtwn) modelling methodology. An overall representatiOn and a first level decompositiOn of
the PARES process can be seen respec!Jvely m figure 23 and figure 24.

Five mam process blocks have been iden!Jfied to m the IDEFO first level representatiOn, these are:

• Create PVA modeL

• Create GVA modeL

• Create mspectwn plan _

• GaugeR&R .

• InspectiOn process

.,.,.
Manufacturmg

~-\
(Jla~!y
StarJdards defrn lion

Part~trv
Key ctlaractens!Jcs
TQieraro:::e spe<:•~cabon
GO&TSPf:'Qficabon
Fudla'e geot"IW!y
Pall locator strategy
Part teatu•e del\rMnoo

Furtae toleraoce speafi<.abon PARES Output


Gauge R&R plan
Ueasuremert mspecnoo plan
Measurement data
CHM mspectlon program
Assembly process
Measuremerf mspecllOn data
Competrllve assessme/11

Kmema~cal models

PDM PVA GVA Qualty Po 111 SPC


logisttcs tloOO analySis
database analysts

FEA CAD Pi~ Requore-nents OOOn


piannmg del1nrlOn lme
tnspectiOn

Figure 23 Overall IDEFO representatiOn of the PARES process

85
- --

I II
CreatapVA I
""""" A1 --~~11 I PVAana.,_

~,.,.--

··-... -
j-~
~~---::--::-:a
"' Q~;~l-
- - -
-
- '
ll~
• ·~

----
Esl- ____.---
~"""oei'Y
-) r -- i •••
] Me ........

l~ •
lfll::l
I;
I
::=---- ,I
II
'
' --~--·
ru~<n­
-- I ''
........
oped&t&Doo> - - -
I 'I
I'

·--
-K-- '

,_
) '
~-~
'
-:-..~ ;1-------.,

-
J
I
I ,I ..__.,. Ga.goR&~

I I
I

.~:~
o.,.,.~y CA0 fEA
I ~ocal ~

,_-
lollr.>Q """'..• e..r"''"""

Ftgure 24 Ftrst leve!IDEFO represenatton of the PARES process


--
-
P<>"'l Otl-()n
u.
S!'C
............

5.4 PARES benefits

The PARES methodology wtll potentially bnng the followmg benefits

• Lead to a more structured approach to product hfecycle DM, the PARES will need to be
mtegrated mto an extstmg product hfecycle management process clearly definmg each
process element and tt's matunty gate
• Produce a more robust product, PARES dnves product and process destgn vahdal!on to
ensure vanalton is tdenllfied, quanllfied, and managed through out the design and
development stages ..
• Promotes the tdea of destgnmg quahty mto the product and not mspectmg tt out, the
process gmdes as orgamsatwn to constder the effects of vanalton at all stages of the
destgn and development process It parttcularly promotes product and process destgn
evaluatiOn at the early dtgttal mock-up stage and not wattmg unttl toohng and prototype
product hardware are avatlable for quahty check
• Promotes a measurement culture, not a gaugmg culture; promotes the use of a closed loop
measurement strategy based on dynanuc process checkmg to ensure manufactunng and
assembly activtty stays wtthm reqmred control hnuts

86
• Reduce production defects, tmprovmg scrap and rework rates, scheduhng flow, and
dehvery times; Improved product and process control ensures a more predictable and
controllable manufactunng environment.
• Makes effective use of destgn, manufactunng, assembly, and mspection process
knowledge from legacy programs or credtble 3ro party organisatiOns. Ensure all relevant
current producl!on process data IS recorded as reusable knowledge Reduces knowledge
waste resultmg from quahty data bemg etther not avatlable, or in an unusable format.

The proposed PARES methodology, the process activities, and top level architecture have been
presented m thts chapter. Two stgntficant sect tons of the methodology have been identified for
further inveshgal!on as part ofthts research proJect, these bemg

• Develop a PVA tool for wing box assembly analysis: Destgn and develop a spectfic
PV A tool to analyse the key parameters of an atrcraft wmg box structure. The wmg box
assembly analysts tool wtll be destgned to perform I D/2D tolerance analysts to validate
the target concept destgn for capabthty. The PVA wtll be developed based on a number
of reqUirements tdentified by Atrbus together with wmg structure mformation The PVA
development process IS presented m chapter 6 of thts thesis
• Digital mock-up (DMU) to geometric variation analysis (GVA): Three case studtes
were tdentified and undertaken on an Atrbus wmg box structure mvestigatmg the
assembly fit of nbs, spars, slans, and D nose parts Tlus mvolved the use of a 3D
CAD/CAE/DMU environment to make use of product mformatton such as tdenttfied
KCs' and associated GD&T tolerance metncs. The three case studtes of 3D synthetic
tolerance analysts are presented m chapter 7

The two sectiOns tdentified from the PARES methodology for further work wtll be mvesttgated m
the followmg chapters

87
Chapter 6

6 Development of a parametric variation


analysis (PVA) tool for wing box assembly

Objectives: This chapter outlines the design and development of a PVA wmg box assembly
analysis tool m response to a specific need Identified m chapter 4. This IS remforced m Chapter 5
which positions the requirement for a number of PVA tools w1thm the proposed DM process
PARES This specific PVA tool is designed to evaluate assembly process agamst parametnc
_ target capability, for example, wmg box structure steps and gaps. The approach embodies the
reqUirements established at Airbus talang into account the analysis needs, worlang practices, and
wmg structure informatiOn The PVA tool Is presented under the followmg sections:

• PVA tool development background.


• Requirements specificatiOn and scope.
• Design and development methodology
• PVA tool functional overview
• Benefits and IInutatwns of the PVA tool.
• Potential further development work

6.1 PVA tool development background

A number of PVA analysis proJects were Identified m support of tlus research The reqmrements
for these proJects were based on specific assembly scenanos Identified by the composite wmg
development umt at Airbus, F1lton An Imhal search mto smtable PVA analysis tools capable of
perfonmng early parametnc studies With sufficient flexibility failed to Identify a smtable
candidate It was therefore agreed that a smtable PVA tool be developed to support the planned
analysis work The development of the PVA tools was to be undertaken by the author and
subnutted as part of this research proJect

The PVA analysis proJects were managed under a smgle proJeCt imtiatJve entitled the 'Vanahon
analysis of aircraft carbon fibre composite wmg box structure assembly' (Jeffreys 1998a) which
mcluded.

88
• Design and development of a PVA tool to analysis wing box assembly: the PVA tool
for wmg box analysis IS presented m tins chapter
• Conduct case studies to review DMU 3D geometric variation analysis on wing box
structures: the case studtes are presented m chapter 7.

Atrbus defined the reqmrement for the PVA u!ihty as bemg capable of tden!Ifymg baselme
estimates of parametnc dtmenswnal variatiOn wtthm aircraft wmg box features dunng the early
stages of design. The reqUirement was documented and an agreement was set m place for the
author to destgn and develop a demonstratiOn PVA tool

A wmg box provtdes the pnmary and core structural elements of a wmg system It is physically
located m the central part of the wmg and factlita(es the base platfcmn for other wmg elements to
be assembled, such as leading and trmlmg edges, engine mountmg pylons, main landmg gear, ai!d
so on Further more, a large portiOn of the wmg box structure m many large military and
commercial aircraft also functiOns as the mam fuel cell In baste terms, the wmg box IS
genencally made up from a number of nbs set penodtcally across the length of the structure, a
front and rear spare typtcally of I, C, or Z configuratiOn, and an upper and lower skin to complete
the closed box structure (see figure below) Traditionally, these structures have been designed
and manufactured from metallic matenals and associated processes Recently, both military and
commerctal aucraft comparues are mves!Igatmg the further mcorporatwn of composite matenals,
CFC for example, wttlnn wmg box primary structure destgn Dnvmg thts reqmrement IS the
destre to produce a lighter wmg structure whtch wtll result m an Improvement for m-servtce
affordabthty.

Ftgure 25 Example wmg box structure wtth top skm removed, courtesy of Atrbus

89
6.2 Requirements specification and scope

The reqUirement was to destgn and develop a parametnc dtmenswnal vanatwn analysts
apphcatwn capable of modellmg a number of tdenllfied key features wtthm the atrcraft wmg box
structure. The mam reqmrements were developed and vahdated wtth Atrbus, these were
pnmanly.

• Vanatwn analysis for I and I .50 configuratiOn


• Perform analysis on a genenc wmg box structure to mclude nb, front/rear C spars,
top/bottom skins, D nose, and A frame components.
• Idenl!fy key parameters for analysts, for example, tdenllfied key charactensttcs such as
steps and gaps
• Apply vanal!on to each parameter based on selected dtstnbutwn.
• Capable of worst case, root sum square (RSS), and stal!sl!cal analysts
• Apply vanatwn based on sample manufactunng measurement data
• Stonng analysts results for recall.
• Pnnt out stmple analysts reports
• Operate on low spectficatwn PC

From the primary reqUirements a functional and techmcal spectficatwns for the PVA tool was
defined by the author m support of thts research The scope of the PVA tool was hmtted to a
smgle genenc wmg box structure configuratiOn, based on a set assembly bmld philosophy (see
figure 26) This was defined as·

i. Locate front and rear spar sub-assemblies, D nose to front spar and A frame to rear spar.
n Locate nb; nb upper flange to be flush wtth upper front and rear spar flanges
111 Fettle/slum nb tf out of tolerance on step control
tv Locate upper skin
v. Locate lower skin.

90
Rear spar

Ftgure 26 Components Withm the genenc Wing box

6.3 Design and development methodology

The core development was undertaken wtthin the MS Excel apphcatwn malang use of available
math and statistic functions and macro commands. In addttion, the DeclSloneenng Crystal Ball
Professwnal (Decisioneenng 2000) software module was mtegrated to the MS Excel application.
Thts has allowed the use of Monte Carlo stmulatwn to be performed wllhm the base PVA ullhty.
The pnmary development blocks for the PVA tool were

• Review and define generic wing box structure components, parameters, assembly
sequence to be include in development scope; define all wmg box structure product
details for mcluswn m the analysts model and venfy these wtth Atrbus
• Define analysis model process flow; outhne the blocks, structure and process flow of the
analysts tool.
• Develop main wing box component gallery; define all component and parameter
configuratiOn back to central wmg box gallery
• Generate component and feature parameter tolerance chains for each assembly stage;
develop feature data tables for each component and define the tolerance stack cham
• Build the set viewing table; define the data set viewmg table to enable storage and
recovery of spectfic configuratiOns mcludmg nommal data and parametnc feature lmuts.
• Consolidate and test all the above activity and integrate Crystal Ball Pro software
module; mtegrate all analysts blocks wtth Crystal Ball Pro software where appropriate
and test.
• Perform analysis tool validation; generate test data and validate output agamst 3"' party
analysts engme.
• Develop user instruction; generate instructiOn and help section

91
• Present the analysis tool back to Airbus; revtew analysts tools wtth Atrbus.

The development blocks were generated and the resultant functwnal elements are presented m the
next sechon.

6.4 PVA tool functional overview

The PV A functwnal process ts defined m figure 27 below There are stx steps in the PVA
process, these are

Analysts configuratiOn
11 Parameter features defirut10n
m Tolerance spectficatwn
IV Stmulahon parameters defimtwn
v Analysts of results
VI. Knowledge capture

Parametric Variation analysis

.. '

Ftgure 27. PVA process

92
The seventh optiOnal step is to rerun the analysiS w1th revised are previOusly saved parameter
data An explanatiOn of the six steps IS g1ven below m sectwns 6 4 I to 6.4 6.

6.4.1 Analysis configuration

The wmg box assembly configurahon can be viewed m the 'Wmg box component gallery'. The
gallery IS made up of eight sechons, one for each maJor component. At the centre IS a schemahc
of the complete assembly wmg box w1th reference pomters to all mdividual components, these
bemg:

• D nose configuratiOn

• A frame configuratiOn

• Front spar configuratiOn

• Rear spar configuratiOn

• R1b configuratiOn

• Upper skm configuratiOn .

• Lower skm configuratiOn

To a1d navigatiOn, the Images w1thm the component gallery are hyperhnked and selechon of a
specific assembly area them would lead the user to the appropnate component analysis dnver
table.

The wmg box assembly process, 1denhfied for th1s analysis, IS Illustrated m figure 28

Key

q Flush allgnmenl

q FeHie process

• Locauon plllltlole

stage 1 J1 Stage 2

Q~:
Locate 0 nose 1n Locate Aframe 1n fixture
fiXtUre Posrtlon rear spar and
Posruon front spar fettle A frame bracket to
and fettle D nose to remove step condrl!on
remO\fe step
condrt:Jon

93
Stage 3
Locate D nose and A
frame 1n fiXtUre

n_
Jl Stage 4

0
.0
0
0
,
,,

2
Pos•tJon upper spar
flange and flush to
adjacent nb edge
Onve vanat10n to lower

1T
nb area

1f
Stage5
Pas ,a on upper skm to
nb and spar surface
Position lower skin to
nb and spar Interface
surface

Figure 28 Wmg box assembly process defined for this analysis

6.4.2 Parameter feature definition

For each assembly stage mcluded w1thm an analysiS the parameter detmls for each component
mcluded will need to be defined. This IS specified Within the data dnver tables which reqmre
nommal informatiOn on all parametnc elements makmg up a dimensiOnal cham. The component
parameters represented m these studies are consistent with a standard aerospace methodology for
a genenc wing box assembly process.

6.4.3 Tolerance specification

At each assembly stage, component parameter tolerances need to be defined. Th1s IS agam
specified m the data dnver tables The tables contam a number of available cells mto which data
relatmg to a tolerance cham feature may be mput

94
Data analysiS drive tabl• 1 D noS(! aoo tront spar assem~

0
0

oo
0

F1gure 29 Data dnver table, D nose to front spar assembly example

The reqmred parameter tolerance data can be categonsed as:

• 3 Std. Dev.; the 3 standard devtatwn tolerance apphed.


• Distribution; the parameter feature vanat10n d1stnbut10n type
• LSL; the lower specificatiOn hnut of tolerance
• USL; the upper spectficatwn hm1t of tolerance

6.4.4 Simulation parameter definition

When all parameter feature nommal and tolerance mformatwn has been completed the s1mulat1on
cntenon must be defined The analysis tool can perform three types of s1mulat1on, these are

• Worst-case stmulatwn
• Root sum square stmulatwn
• Statistical stmulatlon; Monte Carlo and Latm Hypercube SimulatiOn

Worst-case simulation: Worst-case (WC) SimulatiOn techmque represents a conservative


approach for assembly analysts It assumes that all vanatwn levels w1ll be at the1r maxtmum (or
nummum) levels. WC vanatlon levels are calculated by the simple hnear add1t1on of all
mdiv1dual values. WC SimulatiOn does not take into account the dtstnbutlon of feature s1ze and
they do not exceed theu respective spectficatwns.

WC T total= tl(+ or-) t2(+ or-) t3(+ or-) . (+or -)In

95
The tolerance vanables wtll take a value at one of tls hrruts m such a way to yteld an extreme
condttton at a pomt of analysts, 1 e. the maxtmum expected variatton at that pomt WC does not
take mto account the laws of probabthty.
Root sum square simulation: Root sum square (RSS), or hnear stack, predtcls vanallon at a
gtven pomt whtch ts not expected to be exceeded more than I m 370 assembhes (99 73%) Tlus
analysts was developed as a more accurate representalton of vanance when considering a large
number of tolerances One assumpllon made ts that the analysts represents features that are
manufactured about the mean pomt of the tolerance, 1 e an equal btlateral tolerance ts assumed.
The RSS method, ts used when an assembly response functton can be expressed as a hnear
functton of the component parameters. The math functton for RSS tS

RSS T total= (11 2 + t2 2 + t3 2 + ...+ tn2 ) 05

Where tn" represents the mdtvtdual tolerances and T total equals the total predtcted assembly
tolerance (equal btlateral vanatton) usmg the RSS model
Statistical simulation: Stallsllcal analysts employs a stallsllcal dtstnbullon whtch represents
more accurately the expected process vanallon of each parameter dtmenston The analysts tool
allows the use of actual manufactunng process data or an approxtmation of a distnbullon curve
and hnuts to gtve a better representatiOn of actual or calculated vanallon data Thts can be
established by dtrectly samphng available data and attaclung tl' s charactensllcs to a spectfic
parameter feature

Further explanahon on techmques such as Monte Carlo stmulallon ts avatlable m the appendtces
of thts thests

Several standard dtstnbutton types are avatlable to represent destred vanallon characterisllcs. The
followmg staltshcal dtslnbutwns are supported by the analysts tool:

• Normal dtstnbutwn .
• Umform dtstnbutwn
• Tnangular distnbutwn .

• Bmonual dtstnbutwn
• Potsson dtstnbutton
• Geometric dtstnbullon .
• Hyper geometnc dtstnbullon .

• Lognormal dtslnbullon

• Exponential dtstnbutwn

96
• Waybill d1stnbutwn.
• Beta d1stnbution

These were all available VJa a drop down menu from each Distribution Type entry cell. Uruque
groups of manufacturing sample data established from development expenmentat10n at Airbus
were available under descnptive headmgs such as

• Specific cured CFC panel vanance


• Specific debulked CFC panel vanance

Other specific manufactunng process capability based on mdustnal best practice was also
available for samplmg These mcluded

• Alurmmum h1gh speed machmmg.


• Resin film mfuswn
• Resm transfer mouldmg
• Specific Pre-pregnated CFC.

Once the parameter feature d1stnbutwns are established, the SimulatiOn nm IS defined The
SimulatiOn engme IS a Monte Carlo random number generator For each simulatiOn nm the
number of limes the assembly is to be bmlt will need to be defined. To ensure good confidence
levels from the output, current best pracllce would typically spec1fy a count of 3,000 simulatwns
or more, but this has to be established dunng the analys1s

6.4.5 Analysis output

The simulation outputs for WC and RSS could be established as soon as the parameter feature
normnal and tolerance lirmts were specified These outputs were represented as a smgle metnc
output made available m the data dnver table. The stallsllcal based analysis results can be run as
a secondary srmulat10n ulllising the Crystal Ball Pro utility. The analysis results are ultimately
assessed agamst to !me-of-flight wmg specifications for the profile tolerance of movmg surfaces,
fixed surfaces, and closure surfaces These mclude specification for overall profile, surface
flushness (steps), gaps and grooves, and surface wavmess These spec1ficat10ns d1ffer dependmg
on which defined zone of the wing IS bemg considered The Simulation outputs are presented m
the followmg charts

97
Frequency chart; simulallon outputs are d1splayed m a lustogram InformatiOn such as output
dtstnbution type and fit are reported, and certamty levels may be mampulated for further analysts
An example of a frequency chart for the D nose analysts IS given below

0 Fmec.ul Step 1!!!1[!1£1

500 Trials Frequency Chart 3 Outllers


036 18

115

9
il'
'll
45 i
~

Certainty ]i1Ht81HI " •J•lnfmlty


11F-,.JU

Ftgure 30 Example frequency chart forD nose

Sensitivity chart; evaluates the sources of vanance and determmes what contnbutwn each
element has made to the total measured vanance The senstllvtty chart dtsplays how much each
cham element affects the output the step measurement based on rank correlatiOn An example
chart is gtven below

C) SensiiiVIIJ' Chilli l!!l(il Et


Tnrget Forecast S1ep
;.]
Pt4 Pt5
Pt:J FH
Ptl Pt2
""'
,.,.
225>

Pt2 Pl3 1714


PIS Pt1
""'

0% 25': 50% 75% 100':


Me~ed by Conllbh:wltoV~1!61Ce

Ftgure 31 Example sensttlVlty chart forD nose

An example of a sensti!VIty chart for the D nose analysts ts gtven below. It tlus example, the Pt4-
Pt5 parameter feature IS the btggest contnbutor to vanatwn output

Statistical chart; reports standard stallstlcal measures includmg the mean, medtan, mode,
standard devtallon, vanance, skewness, kurtosts, and so on. An example of a stallsllcal chart for
the D nose analysts IS given in figure 32 on the next page

98
0 Ferec•:rt: Step l!lli!EI
.E~ fte!~lme':$ YJeW A~t~ Help
Cell BK252 Slatistlcs
v....
""""'
'""'
"'~
M .a.,
500
~00
000

'
'
'
Mode
Sta'ldardDev~a~~on
v~.
Skewness
K"'=
Coelt oiVo!lfl!lbiiy
..,
-
000
001
154
34963
Ra'IQO Mnrun ~21
Range Maxm.rn 021
Aange\1/dh
Mea-~StdEnor
os•
000

Figure 32 Example statiStical chart forD nose

6.4.6 Knowledge capture

An analysis sessiOn can be captured m two steps. The assembly mstance mformatwn such as
feature nommal, USL and LSL, d1stnbutwn, etc could be saved to the set data table held w1thm
the applicatiOn. Th1s could then be saved and reused for future analysis. The stal!sl!cal
simulatiOn output could add!l!onally be saved to a report file which may be pnnted

6.5 Evaluation of the PVA tool

The reqmrements for a PV A analysis tool were based on the literature survey, Identified mdustnal
reqmrements, and the PARES methodology together w1th mputs from Airbus and case studies to
vahdate the research The PVA apphcatwn was developed followmg the process outlmed m the
'design and development methodology' sectiOn. The pnmary mm was to design and develop the
PV A ullhty, to vahdate the mam functiOns, and handover to Airbus

Followmg development, the PVA ullhty was validated by the author. This was undertaken
through the followmg steps

• A simple 4 component ID assembly stack was defined usmg MS Excel (see figure
below).
• The model was then used to produce an analysis model usmg the PVA tool.
• In order to vahdate the Monte Carlo SimulatiOn outputs includmg the frequency,
sensitiVIty, and stahsl!cal charts, a basic ID model was generated m V1sVSA to represent
the same analysis conditions and outputs.
• Outputs form both the PVA and V1sVSA models were analysed

99
The PVA and V1sVSA analys1s tools demonstrated the followmg correlation against the
vahdatwn model (see figures 33-39)

Nom Tol D1st Tol/3 Nom +Tol Nom -Tol


Part A 115000 0 100 Un1 0033 115100 114900
Part B ~ 50,000, 0005 Nor 0002 50005 49995
Part C 12,000 0100 Un1 0 033 12100 11 900
Part D 1At\Nln 0030 Nor 0 010 180 030 179 970

Ga

PartD
180000
-+0 030/-0 030

F1gure 33 S1mple 4 part assembly

V ISVSA analys1s results:

---

Nonllllal 30000 Low« Spec I.Jrnt 2 9000 Cp 04957


Mean 30031 Uppe.r Spec lfflt 3 1000 Cpk- 04803
Sld DeYiahoo 0 0833 O.stnbt.obon Tested Inverse Stlb
,..,.
12 7333
Esbmale
12 0036
SarT'4)Ie
260€1
Estmate•
28013
134667 t3 6«6 3 2234 32049
262000
'"""' 0~174
•+! J~Range
04036
99~

F1gure 34. V IS VSA process report

100
""' Nomillal 30[)00
Process V=ce D~069

...
HLMVen~ DOOZl

"'--
PartC-Poont2
>jSPF jll200j
PartA-Po1nt
>jSPF jll200j
'~'
0 2000

0 2000
~E'I"'SIIM

1 0000

10000
'~

-9%
Part0-Pom2 00500 1 0000
>jSPF j0060j I 430%

9988"4
1 &ddl contnbulor(•) ( 1 00% eodl 0 1l"4

Figure 35. V1sVSA contnbutor chart

Nomnal 300 ,,.., 300


Stl Cev 008 lolt.'ef" llnut 290
Uppe.-"""' 310 Os- lnv Stub
Cp aw Cpk 048
•'-<S~ 12 73 ~Out of Spec 2620
Est %<Spec 1200 Est %>Spec
Est •:w Out of Spec 2585 low ""'
281
042
Hegh
Est t.a.v
322
200
"-"
Estt·hg"t 320
Est Range 040 Nom +1- Rarge 30J+i-!l20
Nom +1- 3S'3"Tla 300+1-025

Figure 36 VIsVSA 'tallsllcs table

The PVA applicaton results

gap
ou~_cc_~~~~-O-=~~~_c_c~~~~~~~-=• ._

.,,,_______ _ ______,_.,.
·~'-------------- ------·-----11 ""
_ _ _ _ _ _ _..;1' ""

~
.., _____ , ..

-=ooo
!,'". ___
j__ ---- --
_

.."'
2BIO 2853 21l9ll 2940 2963 3026 30!:19 3113 3156 31911
"

Figure 37 PVA process chart

101
S&nstt!Vity: gap
-500% -400% -30 0% -20 0% -100% 0 0%
! I·~·
r-" I" "
"'~
~

• ' -51.0%

c ,, ... "
"" .~

b _,
""

Ftgure 38 PVA senstllvtty chart

Statistics Forecast values


Tnals 3,000
Mean 2 999
Medtan 2 999
Mode ---
Standard
Devtatlon 0 081
Vanance 0007
Skewness 0 02269
Kurtosts 2 47
Coeff of
Vanabthtv 0 02698
Mmtmum 2 797
Maximum 3 208
RanQe Wtdth 0 412
Mean Std. Error 0 001
Ftgure 39 Statistical chart

From the results 1t can be seen that pndected gap range was 0 4174mm and 0 412mm from the
V1sVSA and PVA tools respecttvely. This represents a d1fference of only 0 0054mm. The
contnbutor reports have some subtle diffemecs (see figure below) The both analysis models rank
the contnbutors m the same order w1th the exception that VisVSA ranks the first 2 parts, part A
and part C, as JOint first where as the PVA tools prod1cts Part A as the first contnbutor followed
by m second place by part C There IS a further d1screpency in that the amout both analys1s
models rank the tlnrd contnbutor V1sVSA predicts that Part D contnbutwn will be 4.30% where
as the PVA tools proed1cts 1.50%

102
Part VtsVSA PVA
Part A 4779% 5100%
Part C 4779% 4750%
Part D 4.30% I 50%

F1gure 40 Contnbutor report companson

The main elements for th1s companson are the predicted range of varaiattOn, and the order and
magnitude of the contnbutor elements. Although there are some mmor discrepancies in the %
contnbutwn and their rankmg, a companson of the other stal!sl!cal and vanatwn range outputs
suggest good correlatiOn

6.6 Benefits and limitations of the PVA tool

The pnmary mtenl!on of the PVA tool was to demonstrate to Airbus how a !muted, simple but
fast tolerance analysis tool could be made available to all personnel at the early stages of design
The benefits and hnutatwns of the tool m Its current from are descnbed below.

Benefits:
• Fast and simple to use, no formal trammg reqmred
• Does not reqmre detailed geometry to be available, can be used on genenc wmg box
assembly at early design stage An example of how this may be used for D nose to C spar
assembly fit up IS shown m figure 28.
• No software cost mcurred. Only reqmres a common desk top apphcatwn such as MS
Excel to operate The DecJsJoneenng Crystal Ball ProfessiOnal software module IS only
reqmred If statJsl!cal analysis is to be performed.
• Can mcorporate actual manufactunng measurement data or available mdustnal process
capability mto model If actual sample data IS available on mdlVldual features, the PVA
tool can be configured to run a SimulatiOn usmg tills data for each related model mput
replacmg the standard dJstnbutwns selected from the d1stnbutwn gallery With Crystal
Ball Pro Each mdiVldual sample data can also be represented by a s1mphfied statistical
model which mcludes values for mean shift, standard deVJallon, skewness and kurtos1s

103
Lnnitatwns
• Only capable of ID and I 50 stack analysts
• Can only constder pre defined assembly sequence.
• Cannot readtly accommodate all elements ofGD&T tolerance cntena
• Luruted to modellmg of parameters and not mdtvtdual features.

6.7 Further development work

The PVA tooits baste demonstrator but tt could be stgmficantly developed to perform much more
complex tolerance analysts and synthests tasks Areas for posstble consohdatton and
development could mclude:

• Material properties: add matenal properttes agamst each component


• Model temperature variation: allow the mput of thermal expanston vanatiOn based on
product workmg envelope versus matenal property
• ISO limits and fits: to be readily apphed to hole and pm based features
• Individual feature definition: add defimtton to baste features such as holes, pms, and
pomts, and not JUSt assembly parameters
• Individual features template: to allow the defimtton ofGD&T.
• GD&T paper gauging techniques: mcorporate paper gaugmg techmques m order to
accommodate GD&T matenal modtfiers and bonus tolerances
• Relational database integration: set up to store all analysts configuratiOns and results
Th1s could also be extended become a platform to store all domesttc manufactunng
measurement data, standard ISO process capab1hty data, and where avatlable suppher
measurement data apphcable to th1s type of study
• Include other aircraft structure zones: expand the analysts to other atrcraft zones, for
example, fuselage, honzontal and vert1cal stab1hzers
• Cylindrical 2D analysis to accommodate fuselage modelling: prov1de capabthty to
analyse cyhndncal 20 geometry based assembly 1ssues Model the potenttal effect of
skm th1ckness, part-to-part hole postllon nnsahgurnent, fastener capab1hty, and thetr
effect on curved geometry
• Limited 3D tolerance analysis: produce hnnted 3D analysis by mcorporatmg matnx
transformatiOn to component to component assembhes Th1s could be done by usmg a
number of pomt based 321 moves to each component m sequence unttl a measurement
pomt coordmate on the last component IS resolved to an x, y, and z co-ordmate. Tlus type
of development would be complex but 1s technically poss1ble

104
6.8 Concluding remarks

The development of the PVA utility has demonstrated the benefits and lnmtatwns of perfornung
Simple I D/2D tolerance analysis at the early design stage Th1s type of analys1s IS essentially
based on a genenc structure configuratiOn and w1ll prov1de an early ms1ght to assembly parameter
capability and can mfluence des1gn deCISIOns before detmled 3D geometry IS generated. The
analysis can also help m the product des1gn selectiOn process

However, later m the des1gn development cycle when deta1led 3D des1gn geometry IS avmlable a
second degree of tolerance analysis and synthesis IS reqmred In order to meet these demandmg
analys1s reqmrements a d1fferent approach IS reqmred. Such as approach IS outlined m chapter 7
and IS demonstrated on three A1rbus wmg structure case stud1es. The case stud1es prov1de the
core effort behmd the consolidated A1rbus proJect entitled 'Vanatwn analysis of a1rcraft carbon
fibre composite wmg box structure assembly' (Jeffreys I 998a).

105
Chapter 7

7 Geometric variation analysis (GVA) on


proposed Airbus wing box assembly

Objectives: This chapter outlines a number of geometnc variatiOn analysis case studies
performed on a development wmg box structure The case studies have been undertaken m
response to the need outlined m chapter 4 for synthetic geometnc vanatwn analysis to be
undertaken when detmled 3D CAD geometry IS avmlable Furthermore, the proposed PARES DM
process outlined m Chapter 5 reqmres that the I D/2D PVA analysis IS extended to a more detmled
3D geometnc analysis models The approach accommodates a number of specific analysis needs
for the Airbus A3XX wing box assembly. The case studies are presented under the followmg
sectiOns

• Background to Aubus A3XX case studies


• Case study analysis methodology.
• DimensiOnal analysis for Simple nb to skm assembly case study
• DimensiOnal analysis for nb to complex skm assembly case study
• DimensiOnal analysis for spar eqmppmg assembly case study
• Highlight results and conclusiOns

7.1 Background to Airbus A3XX case studies

Dunng the wmg development phase of the Airbus A3XX commercial aircraft there was a
reqmrement to mvesl!gate how capable the current manufacturing and assembly tooling would be
to facilitate an enhanced wing assembly JOirung process. This enhanced process involved the use
of film and paste bondmg technologies for the sealing and JOimng of CFC mterfacmg surfaces to
be used m conJuncl!on with conventional mechamcal fastemng methods A diagram of a genenc
wmg box with D nose and A frame brackets attached IS given m figure 41.

106
F1gure 41 Example wmg box assembly w1th D nose and A frame, courtesy of A1rbus

The reqmrement was partly addressed by the IdentificatiOn of a number of analysis tasks by
Airbus These tasks have been addressed through a number of case studies undertaken by the
author

These case studies represent the work undertaken m addressmg the 'Vanation analysis of aircraft
carbon fibre composite wing box structure assembly' (Jeffreys 1998a) project. The pnmary a1m
of the project was to simulate the propagation of wmg components mto assembly fixture in order
to predicted key mterface conditions which m turn may be analysed to estabhsh If the cond1t10ns
will meet defined assembly specificatiOns. These specifications mclude the maximum and
mimmum gap conditions penmtted between components for mechamcal fastener, film bondmg
and paste bondmg assembly processes The followmg three case studies were identified A
descnption of the analysis methodology for each IS given m the next sectiOn

• Case 1: DimensiOnal analysiS for nb to complex skm assembly


• Case 2: DimensiOnal analysiS for C spar eqmppmg assembly

7.2 Case study analysis methodology

Essentially, the case studies were undertaken followmg the same analysis methodology. The
methodology consisted of the followmg stages·

a Define study aims and objectives


b. List assumptiOns
c Design assembly components and toohng features
d. Identify key features on components and toohng
e Define tolerances from process capab1hty data for all key features

107
f Create assembly modeL
g. Define assembly sequence.
h Identify assembly moves
i. Define analysts measurements
j Perform assembly stmulations
k. Assess results

Each of the above stages wtll be demonstrated through the presentation of a single case study.
The example case study wtll be the 'Dtmens1onal analysts for nb to complex sk:tn assembly. case
2' The core atms and obJectives, and conclusiOns for Case I and case 3 are also presented.

7.3 Dimensional analysis for rib to complex skin assembly

Thts case study represents part of the work undertaken m support of the 'Vanat10n analysts of
atrcraft carbon fibre compostte wmg box structure assembly' proJect (Jeffreys 1998a). The study
atms and objecllves are gtven below

7.3.1 Analysis aims

The atm ofthts task ts to stmulate the assembly of complex skm and nb components mto a settmg
fixture (see figure below) so that a cnl!cal mterface gap may be analysed The gap spec1ficat10n
mcludes the maxtmum and nummum condttion permttted between components at tdenl!fied pomts
for the mechamcal fastener, film bonding, at1d paste bondmg assembly processes

Ftgure 42 Rtb, complex lower sk:tn and assembly fiXture

108
The analysts model wtll represent the resultant assembly complex mterface gap condihon between
the lower nb surface and lower skm upper surface. Thts mterface condthon wtthm a genenc wmg
box ts shown m the dtagram below

Figure 43 Complex nb (dark shaded) to lower skm m genenc wmg box

7.3.2 Analysis objectives

The a1ms of the analysts Will be addressed through the followmg objechves.

• Create CAD solid geometry models for complex rib, skin, and fixture: a base
geometnc 3D surface model of the complex skm (supphed by Atrbus) wtll be converted
mto a sohd model Generated nb model usmg an offset surface 'splash' from the
mterface area of the skm upper surface When skm and nb models are completed the
fixture sohd model can be generated All additiOnal features for assembly can then be
added All geometry was created using the Dassault Systemes CATIA V4 I 8 CAD
software.
• Create the assembly analysis model: the Vanatwn Stmulatwn Analysts software VSA-
3D, mtegrated mto CATIA, wtll be used to create the assembly model The FD&T
tolerance spectficatwn module w1th CATIA w11l be used generate feature and datum
mformatwn.
• Perform a number of 'what-if' scenarios: to explore the effect of dtfferent key feature
tolerances and destgned nonunal gap condthons for the followmg spectficatwns
o Assembly through use of mecharucal fasteners.
o Ftlm bondmg assembly techmques.
o Paste bondmg assembly techmques
• Document conclusions and recommendations: document findmgs and draw
conclusiOns.

109
7.3.3 List assumptions

The analysis assumptiOns are as follows:

• All component and tooling features m the analysis are absolute ng1d entities, 1 e ,
mechamcal conditions such as mass deflectiOn or component deformatiOn (caused by the
resultmg assembly configuratiOn) will not be taken mto account.
• Toolmg geometry will be mimnuzed to key assembly features and structures for
simplicity
• The complex nb and skm components tolerance specificatiOns will be based on a CFC
manufactunng process capability.-
• The fixture tolerance specificatiOns will be based on steel fabncation process capab1hty

7.3.4 Design of assembly components and fixture

The design of the nb and fixture toohng can be seen m the followmg figures The design details
for the complex skm are not shown for confiden!Iahty reasons The nonunal geometry and
dimensiOns for the nb and fixture were based on the complex skm model.

14554 82mm
1- - --------------

79mn>_-~~---~~~=i-l --
1

I !Smm
_. __ ~--- 1254
-~

~- Radms 5,980mm, top edge


-0-l 0'
E
E
0
0
9:1
'
'
O' - Radms 5,990-5,970mm bottom edge -0-
108 48mm
1 •

4x6 OOOmm hole


VIEW INBOARD FROM STARBOARD pattern VIEW AFT

Figure 44 Rib des1gu schematic

The fixture features will be represented through basic locatiOn and piCkup en!I!Ies reqmred to
onentate and secure the nb and complex skin components for assembly. Components and toohng
for this analysiS have been considered for the starboard side wing sectiOn, I.e. hole (4 DOF) at
wmg structure root

110
The nommal design for the fixture can be seen below.

360mm .,I
,.. __
1
-
360mm 117999mm 117999mm 160mm

6mm d1a
4 way hole SOmm
-- I ..:.~omm
20mm ® __,

1
somm

6mm d1a
4 way hole

PLAN VIEW END VIEW

Ftgure 45 Ftxture destgn schemattc

7.3.5 Identify key features and tolerances

R1b critical features and tolerances (see figure 45 below)


Pnmary datum A Flatness offonn of +/-0 125mm
Secondary datum B Hole perpendtcularly of dta. +/-0 080mrn on 4 way hole at MMC to
datum A, diametnc tolerance ofH9 (+0.030/-0 OOOmm)
Tertiary datum C. Hole posttton accuracy of dta +/-0 080mm on 2 way hole at MMC to
datum A, datum Bat MMC, and diametnc tolerance ofH9
(+0 030/-0 OOOmm)
Bottom edge feature Surface profile of +/-0 I OOmm to datum A, datum Bat MMC and
datum Cat MMC
Hole I and 2: Hole posttton accuracy of dta +/-0 080mm at MMC to datum A,
datum Bat MMC, datum Cat MMC, and dtametnc tolerance of HI!
(+0.075/-0 OOOmrn) on 6 OOOmm dia

Ill
DAlUM_
PER I D!A 0 160(M) I A
DIA 6 000 H9

-o-
DATUM_A
FLTj0250
I
-B- -6-
' •
BTM_SRFI-7 I
SRFIO 200 lA I B(M) I VIEW INBOARD FROM STARBOARD
DATIJM_C
POS I DIA 0 I60(M) IA I B(M) VIEW AFT
DIA 6000H9

Figure 46 Rib cntical features and tolerances schematic

Complex skin critical features and tolerances (See figure below)


Pnmary datum A: Planar theoretical datum proJected from bottom skm surface
Secondary datum B· Hole perpendicularly of dia +/-0 180mm on 4 way hole at MMC to
datum A, d1ametnc tolerance ofH9 (+0 030/-0.000mm)
Terttary datum C Hole posJI!on accumcy of dia +/-0.180mm on 2 way hole at MMC to
datum A, datum B at MMC, and dmmetnc tolerance ofH9
(+0 030/-0 OOOmm).
IML surface features: lmllal surface profile of +/-0 500 to datum A, datum Bat MMC and
datum Cat MMC (Bottom surface features).
OML surface features Surface profile of +1-0 500 to datum A, datum Bat MMC and datum C
at MMC (Top surface features)

DATUM_A
BTM_SRF Thtortl•cal plane
DATUM_B
SRF 11 000 I A I B(M) I C(M)
PER I DIA 0 360(M) I A
DIA 6000 H9
i
IML

I OML I

f'OS I DIAO ~60(M) A I B{M)


TOP_SRFI7 DIA 6 000 H9
SRF 11 0001 A I B(M) IC(M)

PI.AN VIJ-W VII- W INIJOi\Rf)


FROM nARJJ()ARD

Figure 47 Complex skm cntical features and tolerances schematic

112
Fixture critical features and tolerances
Rib associated:
Pnmary feature. Posttion accuracy of +1-0 080mm to toolmg datum A, datum Bat
MMC and Datum Cat MMC.
Secondary feature Hole perpendtcularly of dta +/-0 180mm on 4 way hole at MMC to
toolmg datum A, datum Bat MMC, datum Cat MMC, and dtametnc
tolerance of H9 (+0 030/-0 OOOmm)
Terttary feature· Hole pos1t10n accuracy of dta +/-0.180mm on 4 way hole at MMC to
toohng datum A, datum B at MMC, datum Cat MMC, and dtametric
tolerance ofH9 (+0.030/-0 OOOmm)
Assembly toohng pms to be used wtth assembly hole features are dta. 5 996mm nommal *

Complex skin associated:


Pnmary datum A: Ptck-up surface on top of main toohng structure forming a planar
datum feature Thts feature has a flatness of form tolerance of
+/-0.180mm
Secondary datum B Hole perpendtcularly of dta +/-0 180mm on 4 way hole at MMC to
datum A, dtametnc tolerance ofH9 (+0 030/-0 OOOmm)
Terttary datum C Hole pos1t10n accuracy of dta +/-0 400mm on 2 way hole at MMC to
datum A, datum Bat MMC, and dtametnc tolerance ofH9
(+0 030/-0 OOOmm)
Surface ptck-up 1-3· Surface profile of +/-0 180mm to datum A, datum B at MMC and
datum C at MMC

Assembly toohng pms used to secure nb and skm components mto the fixtures are dta. 5 996mm
nomma1, see below

Pm/hole tolerance
The standard assembly fixture tolerance wtth regard to pms and holes are

Pm= 5.990mm dta at h7 (+0 000/-0 012) for assembly fixtures


Hole= 6 OOOmm dta at H7 (+0 012/-0 000) for assembly fixtures and components.

113
SRF_l SRF_2
SRFj0360!A! B(M)jC{M) SRF l 0 360 I A I B(M) I C(M)

® ®

• •
DATUM_C DATLIM_B
POS IDIA 0 SOO(M) 1A I B(M) IC(M) PERIDIA 0360(M)IAIB(M)IC(M)
DIA 6 000 H9
PLAN VIEW
DIA 6000H9
I ENDVIEW

RIB_TER
POSjDIA 0360(M)IAIB(M)IC(M)
DIA 6 000 U9

f1gure 48 f1xture cnt1cal features and tolerances schematiC

There IS no facthty wtth the VSA software to model the hole/pm/hole tolerance relatiOnship
Therefore, for thts stmulal!on the pm has been specified as a nommal dtameter and tls tolerance
(float) has been added to the fixture hole tolerance Therefore:

Pm nommal & tolerance Hole nommal & tolerance CLEARANCE


5 990rnm (+0 000/-0 012) 6 OOOrnm (+0 012/-0 000)
Max cond1t1on ~ 5 990rnm Mm cond1t1on ~ 6 OOOrnm OOIOmm
Mm cond1t1on ~ 5 978mm Max condition~ 6 012mm 0 034mm

Total clearance (float) 1s 0 034mm- 0 OIOmm ~ 0 024mm

It IS not posstble to spectfy umlateral tolerance condtl!ons m Catha's functiOnal & dtmenswnal
tolerancmg (FD&T) module other than the use of ISO hmtts and fits. Therefore to sattsfy the
above spectfical!on make the component and fixture holes H9 (+0 030/-0.000) and make the
fixture pm a nommal 5 996mm dtameter. Thts creates a float spectfical!on on hole/pm/hole of
+0 024/-0 OOOrnm as reqmred

The tolerances specified m the above secllon represent the mtttal destgn spectfical!on The
complex skm OML and IML feature tolerances were changed as part ofthe "what-tf' scenanos m
order to represent dtfferent panel thickness vanal!on. In addtl!on the destgned nommal gap
conditiOn for the dtfferent assembly processes were also revtsed to smte certam "what-tf'

114
scenanos. These tolerance and nom mal gap reviSions are clearly stated on each of the different
s1mulatwn results

7.3.6 Create CAD assembly model

The components and fixtures Will been generated from the skm demonstratiOn panel surface
modeL The complex skm and nb geometry w1ll been constructed as sets of arcs, surfaces and
faces m the CATIA CAE system. The fixture model w1ll be constructed usmg exact sohd
geometry All cntical features and tolerances have been generated w1thm the CATJA functiOnal
d1menswnmg and tolerancmg (FD&T) apphcahon This data can then imported in the VSA-3D
tolerance analysis module and w1ll be used to create the necessary feature and tolerance
mfonnatwn reqmred for the simulation system

7.3.7 Define assembly sequence

The assembly sequence w1ll be as md1cated m figure 49 below. The complex skm Will first be
mstalled to the assembly tooling followed by the nb sectiOn

FINAL ASSEMBLY
FIXTIJRE

F1gure 49 Assembly sequence diagram.

7.3.8 Identify assembly moves

The first and second moves are to locate the complex skin component the nb component mto the
fixture respectively.

115
Complex skin to fu:ture move
The skm component ts located m three steps
• The first step ts to onentate the skm such that tts calculated profile !me (BTM_ SRF), is
laying approximately m the x and y planar wtth the OML facmg m the -z dtrechon It is
then placed onto the correspondmg fixture ptckup faces (SRF_I, 2 & 3) of the assembly
tooling and approximately ahgned m the x and y onentahon agamst the secondary and
terttary toohng locatiOn hole features.
• The secondary hole feature of the skm (DATUM_B) and secondary hole feature of the
toolmg (DATUM_B) are to be ahgned and the pm mserted
• The tertiary hole features of the skm (DATUM_C) and terttary hole feature of the toohng
(DATUM_C) are to be ahgned and the pm mserted

The above move wtll be stmulated m VSA-3D through a feature based 321 operatiOn

Rib to fu:ture move


The nb ts also located m three steps
• Datum A planar (DATUM_A) of the nb ts onentated m the x and z planar and offered up
to the correspondmg flat ptckup planar feature (RIB_PRJ) of the assembly toohng.
• Hole datum B (DATUM_B) on the nb ts ahgned wtth the correspondmg toohng hole
(RIB_ SEC) and an assembly pm ts mserted
• Hole datum C (DATUM_C) on the nb IS ahgned wtth the correspondmg toolmg hole
(RIB_TER) and an assembly pm ts mserted

The above move wtll be stmulated m VSA-3D through a feature based operahon.

7.3.9 Define analysis measurements

Measurement strategy for gap condition


For the stmulatwn analysts the obJechve ts to measure the vanatton at a gap condthon usmg a
normal vector between the bottom edge of nb 6 and the complex skm IML surface. Simulatton
measurement wtll be facthtated at 17 pomts between the correspondmg surfaces of each
component (see figure below) The pomt to pomt measurements wtll be performed usmg cntena
of the shortest dtstance between surfaces through a normal vector The measurement pomts are
located at the complex skm's surface/surface mterface, mtdway between these and at the skm
edges. The VSA gap measurement template wtll be used to perform the analysts.

116
Measurement pomts taken at
surfacelsmface mtersecllon and
m1dway between

Complex skm
top surface {IML)

.1 17 Measurement pomts
along nb 6 centerlme

Ftgure 50 Measurement strategy for gap condttJOn

7.3.10 Perform assembly simulations

Each stmulatwn wtll be performed usmg 2,000 assembly butlds This number of butlds wtll be
used to ensure a statistically stable set of results. All vanat10n stmulatwn wtll be of a normal
dtstnbutwn type

The stmulatwn wtll be undertaken m two stages

• Monte Carlo stmulatwn


• Htgh-Low-Medmn (HLM) stmulation

An explanation of the above stmulatwn techntques can be fond m the appendtx of thts thests

7.3.11 Simulation results and analysis conclusions

Some ofthe conclustons from thts work are listed on the next page.

117
• It IS ev1dent that each of the three assembly processes w1ll reqmre a d1fferent nominal gap
cond1t10n to ensure Cp and Cpk process capab1hty outputs are opt1m1sed
• The complex skm out of assembly tolerance cases were of a magmtude greater than those
associated w1th the s1mple skin profile (1 e Case I).
• For the s1mple skm (Case I) llghtening the surface tolerances had a much greater 1mpact
on out of assembly spec1ficat1on cases than for the complex skm
• For the film assembly process the optlmum gap cond1tion has changed for both the s1mple
skm and the complex skm
• The tolerance specificatiOn for a paste JOmt 1s sufficiently large to ensure that vutually all
assemblies, even With the complex skm, can meet the assembly gap specification
• W1th regard to the complex skm, top surface profile and the secondary datum were by far
the most s1gmficaut contnbutors to assembly vanatwn.
• The s1mulatwn results suggest that a reductwn could be accommodated by the paste
bondmg assembly process w1th httle or no effect on Cp or Cpk process capab1hty

7.4 Dimensional analysis for simple rib to skin assembly

The overv1ew of the s1mple nb to skm assembly IS presented below.

7.4.1 Analysis aims

The aim of this task IS to simulate the assembly of skm and a nb components mto a fixture so that
a pred1cted mterface gap cond1t10n may be analysed m order to establish If the cond11Ion will meet
defined spec1ficatwns These specificatiOns mclude the maximum and mmimum gap cond11Ions
penmtted between components for mechamcal fastener, film bondmg and paste bonding assembly
processes. The resultant VIrtual hole conditiOn between components to fac1htate mechamcal
fasteners w11l also be s1mulated

7.4.2 Analysis objectives

The aims of the analysis will be addressed through the followmg objectives

• Create nommal component and toohng feature geometry m CATIA V4 I 8 software The
geometnc models of the components and fixture will be cyhndncal m nature and w1ll not
d1rectly represent any actual Aubus wmg design,

118
• Undertake analysts m the Vanatwn Stmulahon Analysts VSA-30 software;
• Perform a number of 'what-If' scenarios to explore the effect of dtfferent key feature
tolerances and destgned nommal gap conditions for the followmg specificatiOns
o Assembly through use of mechamcal fasteners.
o Film bondmg assembly techmques
o Paste bondmg assembly techmques
• From the results document conclusions and recommendations

7.4.3 Define assembly sequence

The assembly sequence wtll be as mdtcated m figure 51 below The skm wtll first be mstalled to
the assembly toohng followed by the nb section.

RIB

FINAL ASSEMBLY

Figure 51 Assembly sequence diagram

7.4.4 Simulation results and analysis conclusions

Some of the conclusiOns from thts work are.

• Skm thickness vanatwn (surface profile) m the maJonty of cases has proven to be the
btggest contnbutor to out of specificatiOn condttwns m assembhes Simulated.
• Improvements m skin surface (thickness) control wtll stgmficantly Improve assembly
dtmenswnal quahty.

119
• Revision of the nommal gap condiiion for a given jomt assembly process can contnbute
to the Improvement or detenoral!on of their Cp and Cpk
• Where nommal gap condiiions are revised at JOIDI mterfaces, the adhesive manufactunng
compames should be challenged to Improve their process specificatiOns
• W Ilh regard to the fixture and Its tolerance, some rmprovements could be made but m
general these would have httle effect with regard to 1mprovmg the d1menswnal quality of
the assembly given the current tolerance specification.
• There IS an opportumty With machine finished CFC component features to Improve their
surface control with reference to their datum system This would result m better
component and therefore assembly quahty control

7.5 Dimensional analysis for C spar equipping assembly

An overview of the C spar eqmppmg assembly IS presented below

7.5.1 Analysis aim

The aim of this task IS to simulate the assembly of complex skm and spar components into a
fixture so that a predicted mterface gap conditiOn may be analysed m order to estabhsh If the
condition will meet defined specificatiOns These specificatiOns mclude the maximum and
m1mmum gap conditions perm!lted between components for mechamcal fastener, film bondmg
and paste bondmg assembly processes.

7.5.2 Analysis objectives

The aims of the analysis will be addressed through the followmg objectives:

• Define analysis methodology


• Create nommal front spar component and toolmg feature geometry m CATIA V 4 I 8
software The model of the complex skm will be reused and from the base geometry for
the C spar design.
• Undertake analysis m the Vanallon Systems Analysis VSA-30 software
• Perform a number of 'what-if scenanos to explore the effect of different key feature
tolerances and designed nommal gap conditiOns for the followmg specificatiOns
o Assembly through use of mechamcal fasteners

120
o Film bondmg assembly techmques
o Paste bondmg assembly techniques
• From the results document conclusions and recommendatiOns

Complex skm
top surface (IML)

r:~~t(
I - - / ~~ Front spar outhnc

\\(-,'-,,//
' ,
-----,jI Measu:rement pomts taken at .1 14 Measurement pomts
~' , ~ ___ " - /// surface surface mtersed1011 and along spar lo"'er flange
/ m1d\'aybctv.ecn m1dpom1

.__-~---

Figure 52 Measurement pomts for C spar equ1ppmg,

7.5.3 Define assembly sequence

The assembly sequence w1ll be as md1cated in figure 53 below, The complex skm will first be
mstalled to the assembly toohng followed by the C spar section

L
f1XTURE f1NAL ASSEMBLY

figure 53 Assembly sequence diagram

121
7.5.4 Simulation results and analysis conclusions

Some of the conclusiOns from th1s work are

• Improvements in assembly capability would be ach1eved through settmg a umque nominal


gap cond!hon for each of the assembly process.
• From the assembly s1mulatwn results 1t can be seen that the most difficult ophon to
control IS the mechamcal fastener process, followed by film bondmg and finally the most
controlled was the paste bonding assembly processes.
• For the paste bondmg assembly process, the s1mulatwn results suggest that a HL
specJficatwn reductiOn could be tolerated w1th little or no effect on assembly capab1lity
• The predommate variahon contnbutors for th1s assembly when the skm top and bottom
surfaces were set at non optimum tolerance were the actual skm surfaces themselves
• The tolerance range for the paste bondmg assembly gap cond!hon appears to be
sufficiently large to accommodate all assemblies However, even w1th a reVIsed nommal
gap condil!On the process mean lies to the left of centre of the LL and HL specificatiOns
causmg a number of assemblies to remam below the LL.
• The film bondmg assembly process With skm surface tolerances set to an oplimum and a
specified nommal gap condil!on resulted m the smallest number of out of specificatiOn
assemblies

7.6 Overall conclusions

The two case studies performed on the proposed compos1te wmg box assembly structure have
been presented m this chapter The defined analysis methodology has demonstrated the benefits
of performmg assembly tolerance analysis at a matunng des1gn stage where some or all detaJled
3D CAD product geometry IS available This type of analys1s should overlap the early ongoing
PVA tool analys1s, examples of wh1ch are outlined in chapter 6 Th1s later stage 3D analysis Will
prov1de a more powerful validation of proposed manufactunng and assembly capability and will
challenge des1gn for productiOn decJswns before 1t reaches the des1gn freeze matunty gate. The
methodology used m the analys1s forms a cnlical component of the proposed PARES
methodology outlined m chapter 5. It also directly addresses the need 1denhfied m chapter 4 for
the comprehensive 3D geometnc tolerance analysis of structures to be performed when detailed
geometry and key feature defimtion IS avaJlable.

122
An additiOnal need Identified m chapter 4 IS to develop new, or revise ex1stmg, manufactunng and
assembly busmess processes Some a1rframe manufacturers are mcorporatmg components made
from composite matenals such as CFC and GRP m additiOn to the traditional metallic materials
such as AVfi related processes Both composite and metallic components may be geometncally
complex and Will need to be dimensiOnally robust to ensure both their assembly and ach1evmg
final product quahty key charactenstics

In order to gam an understandmg of the Issues a number of practical expenments were designed.
These expenments explore the design, manufacture, assembly, and mspectlon of composite and
metalhc fuselage components for different geometnc form The key Issues for mvest1gatwn are·

• Possible effect on overall assembly charactenshcs at different operatmg temperatures


when usmg components made from materials with different thermal properties, for
example, alummmm and carbon fibre composite
• The potential effect of advanced and complex manufactunng process on piece part
dimensiOnal stab1hty, for example, the manufacture of carbon fibre composite.
• The effect of geometric shape complexity on overall assembly charactenstlcs

These conclusiOns have been formed from extensive case study material whiCh are confidential
from Airbus and due to a none disclosure agreement between the author and the company they
cannot be exposed

The expenmentatwn IS presented m chapter 8

123
Chapter 8

8 Assembly of carbon fibre composite panels to


aluminium substructures

Objectives: Thts chapter outlmes a number of expenmental tasks destgned m response to the
need tdenl!fied m chapter 4, section 4.3 for the development of new, or revise extstmg,
manufactunng and assembly busmess processes. Thts need anse due to future commerctal and
mthtary aerostructure destgns incorporatmg a mtx of compostte and metalhc matenals and
assoctated advanced manufactunng processes. The wmg box assembly PVA tool and geometnc
analysts case matenal presented m chapters 6 and 7 respectively outlme the type of analysts that
may be conducted. These types of analysts however wtll only be effective tf the geometnc
vanabthty of targeted matenals and manufactunng processes, CFC for example, IS well
understood Thts expenmentatwn explores some of these tssues first hand by conductmg a
number of practtcal destgn, manufacture, assembly, and mspectwn tasks of CFC panels to
alummmm substructures. The expenmentatwn IS presented under the followmg sectwns:

• Atms of the expenment


• Objectives destgned to meet the aims
• Define the expenmental method
• Htghlight the results and conclusiOns
• IdenttfY posstble future expenmental work

8.1 Aims of the experiment

The experimentatiOn presented m thts chapter has been developed m response to some of the
needs tdenl!fied m chapter 4 These needs translate to the reqmrement for a better understandmg
of geometnc vanatwn charactensl!cs on advanced matenals and associated manufactunng process
currently being constdered for some estabhshed and future aerostructure development programs
The mvestigatwn wtll generate matenal and process knowledge whtch may be mcorporated mto
assembly analysts such as the PVA tool and later wtthm a full synthetic geometnc study usmg
advanced tolerance analysts tools.

124
The mm of thts expenment ts to mvesttgate geometnc dtmenswnal vanatton and tts effect on the
abthty of eFe panels to be assembled to alummtum substructures Panel and substructure
assembhes are somettmes destgned wtth the functional reqmrement of interchaneabthty, i e ,
capable of bemg mstalled, removed and replaced wtthout adjustment or alteratton to panel or
matmg structure Thts functtonal reqmrement can become Impeded by vanous forms of geometric
vanation and thts tmpedance ts anttctpated to mcrease as panel and substructure geometnc form
becomes more complex The mam mm of thts work ts to address these tssues dtrectly through a
senes of practtcal expenments mcludmg the manufacture, assembly and mspectton of
representattve panel and substructure enttttes. In addttton, the expenment wtll mvesttgate mto
techniques designed to compensate for variabthty leadmg to a higher certamty of part to part
mterchangeabthty.

8.2 Objectives of the experimentation

The objecttves of the expenment wtll address the atms through the mvesttgatton of two geometnc
types of panel and substructure assembly, these bemg·

Stage I - Planar geometric profile: Stage I mvolved the design, manufacture, assembly and
mvesttgatJOn of two Stmtlar efe panels onto a smgle alummtum substructure.

Stage 11 - Cylindrical geometric profile: Stage 11 wtll mvolve the detatl destgn, manufacture,
assembly and mvesttgatwn of a number of eFe panels onto an alummmm substructure. The
concept destgn of the panel and substructure wtll be of cyhndncal profile

The expenmentatwn was destgned to mvesttgate the mdtvtdual component, assembly, and
mterchangeabthty elements of a planar and cyhndncal form of eFe panel to a correspondmg
alummtum substructure The followmg activt!tes were performed m support of stage I and stage
11 of the mvesttgatton

• Destgn of the expenment


• Destgn of panel and substructure.
• IdentificatiOn of matenal and manufactunng process
• Destgn of assembly process
• Deterrmnatton of assembly key charactensttcs
• Define destgn tolerances
• Destgn of assembly process

125
• Idenl!fy the quahty measures to be IDVesttgated
• Destgn of expenmental methods
• Perform measurement analysts usiDg a range of metrology equtpment, the Brown &
Sharpe Mtcro Measure IV and VALISYS software.
• Perform a tolerance analysts study on the panel and substructure assembly usiDg the VSA
and VALISYS analysts software
• Expenmental results and dtscusston.

8.3 Design of the experiment

The expenmentatton ts centred on the destgn, manufacture, assembly, and IDSpection of planar
and cyhndrical CFC panels to correspondiDg alumiDtum substructures. Panel and substructure
were designed and modelled m the Dassault Systemes CATIA system and a number of spectal
features were constructed and used m the analysts These component datum reference frame,
features, and tolerances were also generated ID CATIA's advanced 3D functtonal dtmenstoniDg
and toleranciDg (FD&T) module Thts features based IDformatton would be used for mspecl!on
and assembly stmulatton analysts

Both panels are assembled to the substructure vta the 20 dowel pms representmg a close tolerance
fastener stmtlar to those used ID advanced compostte aerostructures For expenmental purposes
the panel and substructure assembly mvesttgatton ts cantered about tts abthty to demonstrate
IDterchangeabthty, one of the key functtons of advanced structures. To demonstrate
IDterchangeabthty ID thts expenment the panel and substructure needed to:

• Assemble uml!ziDg all 20 dowel pms


• Demonstrate panel/substructure gap spectfical!on
• Demonstrate panel/substructure step spectfication

The expenment IDvesl!gated the panel and substructure enl!ttes ID a controlled envtronment usiDg
a Brown and Sharpe co-ordinate measunng machine (CMM) and the Mtcro Measure N software
These IDVesl!gal!ons outhned destgn and manufacture process vanal!on such as hole postl!on and
form, surface and edge form of both panel and substructure

The panel and substructure assembly was IDVes!tgated usmg preciston hard gauges over a range of
temperatures Thts was to demonstrate the effects of thermal expanston on the dtfferent matenal
used for the panel and substructure

126
The CFC panel has a coefficient of expansiOn of 0 004mm!metref C and the alummium
substructure of 0 026mmlmetref C This different coefficient values will have an effect on
certam panel and substructure feature reqUirements for assembly such as hole/hole ahgnment, and
step/gap conditions

These imllal results will then be used as a bas1s for further mvesllgallon usmg Tecnomahx
VALISYS software 1mbedded within the CATIA system This will demonstrate automated
mspechon programs generated off-hne directly w1thm the CATIA CAE env1romnent The
mspechon data can be mvesllgated usmg different analysis modes which will demonstrate
different result conditions The V ALISYS software also uhhzes functiOnal mspect10n analysis
which allows the use of a bonus tolerance system makmg use of the specified matenal condition
modifiers associated With some tolerances The use of the different mspect10n modes will
demonstrate different result conditions.

The panel and substructure will also be subject to an assembly SimulatiOn This will be
undertaken by the vanat10n simulatiOn analysiS (VSA-3D) software also integrated into CATIA
The SimulatiOn will be undertaken to mvestlgate predicted vanat10n results agamst known
measured inspectiOn data

8.4 Design of panel and substructure

The geometry design for stage I and stage 11 panel and substructure can be seen m figure 54 and
55 respecllvely The panels and substructures were created as sohd model enhlles allowmg the
des1gnallon of detailed component features which are central to analysis

All features were modelled m sohd geometry to provide a platform for the attachment of datum
reference frames and tolerance metncs These will be uhhsed for assembly and the measurement
mspectlon processes.

The design of the substructure pnmary face was lowered by a nommal !mm This was to allow a
controlled gap between the panel and substructure in their assembled state m case sh1mmmg
techniques were required Sh1mmmg m this way m order to control the relationship between
panel outer surface steps and gaps IS common m the aerospace mdustry

127
F1gure 54 Panel and substructure des1gn- Stage I.

F1gure 55 Panel and substructure des1gn- Stage I!

The panel and substructure 3D geometry IS avatlable for deta1led mterrogatton, 1 e, d1menswns,
angles, etc.

8.5 Identification of materials and manufacturing process

The matenals 1denttfied for th1s experiment were chosen to represent the characteristtcs and
properties of matenals currently used for this apphcatton m the aerospace mdustry. The same
matenal spec1ficat1on, where ever poss1ble, was used throughout th1s expenment to a1d
consistency. The panel matenal was a resm pre-pregnated (prepreg) CFC. The resm is a type
8552 epoxy matrix and the carbon fibre IS an IM7 grade contmuous weave des1gned for the
prepreg process The carbon fibre IS prepregnated with resin before dehvery to the customer

128
These matenal grades are currently bemg used m the aerospace mdustry for the manufacture of
Interchangeable panels Data sheets for resm and CFC matenals are avmlable from the author
upon request The productiOn of CFC components from th1s type of matenal reqmres a complex
lay-up and cunng process The deta1ls of this process are cons1dered to be beyond the scope of
thiS theSIS.

8.5.1 CFC panel manufacturing process

The stage I panels were la1d up and cured at BAE SYSTEMS The stage 11 panels were lmd up
and cured at Airbus The productiOn of net edges and holes was undertaken at Loughborough
Umvers1ty usmg the Wadkm V4-6 CNC and the Cmcmnal! Sabre 400H machmmg centres The
subsl!uctures were machmed from a sohd billet of 608276 grade alummmm This is bas1c grade
alummmm w1th s1m1lar charactensl!cs to a1rcraft grade matenals

Design and manufacture of tooling plates: to enable the stage I and stage 11 panel lay-up and
cure purposes, EN3B steel tooling plates had to be manufactured for each stage The plates
determme the form shape for each panel generatmg a CPL at the panel and toohng mterface. The
toolmg plates are shown m the figure below w1th the1r respecl!ve panels The plates were
des1gned and manufactured at Loughborough Umvers1ty m support of th1s expenmental!on.

F1gure 56 Lay-up and cure toohng plates- Stage I and satge 11 CFC panels

Lay-up and cure of CFC panels: this was undertaken followmg a BAE SYSTEMS process for
lay-up and cure. Th1s process was undertaken for stage I and stage II panels Data sheets for resm
and CFC matenals lay-up and cure are ava1lable from the author upon request

Routing of net panel edges and drilling of assembly holes: for stage I the production of holes
and net panel edges required the adaptatiOn of the toohng plate once the lay-up and cure process
was completed In preparatiOn of panel dnlhng and routmg, 9x6mm toohng holes were dnlled

129
into both stage I panels using the Wadlan machme tool m a manual operatiOn These toohng
holes were then used to secure both stage I panels onto a toohng plate vta bolts The first panel
(panel 2) was secured onto the toohng plate and then placed mto the machmmg centre and
clamped for operation. Panel I was prepared dtfferently to panel 2 due to the resultmg
delammatmg damage dunng the edge routmg process To reduce thts damage, panel I was
sandwtched by a two layers of 4mm thtck PVC plaslic sheet, one on the top face and one on the
bottom The plastic sheets were prepared by cuttmg a square profile to panel gross stze and
dnlling 9x6 5mm holes in accordance wtth toohng plate hole pattern Panel I was then
sandwtched between the plaslic sheets An additiOnal alummmm pressure plate produced to the
same geometry as for the plaslic sheets was placed on top of the CFC/PVC sheet sandwich The
whole assembly was then bolted together producing a compressed sandwich whtch mcluded the
toolmg plate, the CFC/PVC sandwtch, and the alummmm pressure plate. Thts was then secured
onto the machme centre VIa clamps and the holes and net edges were machined The mam
pmpose of the plaslic sheets was to ensure the cutter moved through sohd sacnfictal matenal
above and below the panel edge surface to reduce the nsk of dehmttatwn of the CFC panel The
edges of panel 2 were routed m one pass usmg an 8mm dtameter sohd carbtde routmg tool
runmng at 8,000rpm and feed rate of I 5 m/mm
The holes were dnlled normal to the surface profile m a smgle operatiOn at 8 000 rpm and a 0 314
m/mm feed rate The edges of panel I were routed m one pass usmg a 12mm dtameter sohd
carbtde routmg tool runmng at 8,000rpm and a feed rate of 0 4m/mm. Holes were dnlled m a
smgle operalion at 8,000rpm and a feed rate of 0 314m/mm A two flute sohd carbtde dnll was
used for producmg the holes m both panels. An off-hne CNC program was wntten to undertake
the dnlhng and routmg operatiOns Stage II panel manufacture was undertaken on the Sabre 400H
machme to accommodate the offset edges and hole axts. Essenlially the same feeds, speeds and
cutter details were used at each stage. The mam dtfference JS that the panels were routed usmg
ctbatool sacnfictal toohng matenal and no PVC sandwich technique was mcorporated Thts, as
for stage I panel productiOn, lead to some delammatmg damage dunng the edge routmg process

8.5.2 Aluminium substructure manufacturing process

Preparation of billet: the substructures were machmed from two sohd pteces of 608276 grade
alummmm The tmlial btllets was reduced to the reqmred net base and edge dimensiOns through a
manual mtlhng operatiOn usmg a 120mm dtameter face cutter.

130
Figure 57. Sabre 400H and Wadkin V4-6 machine centres.

Machining of net substructure shape and drilling of boles: stage l final planar and hole
geometry was then produced by the writing and executing of an off line CNC program. All
features were produced on the Wadkin V4-6 machine centre (see figure 57 above). A 25mm
diameter milling tool was used Lo great all planar geometry and initially ho les were drilled with a
5.5mm diameter drill and finished with an adjustable radius boring tool. The final hole diameter
was determined by offering the dowel pin to the ho le until a fit condition was establi shed. Stage IT
was undertaken on the Cincinnati Sabre 400H (see figure 57 above). The process was similar for
stage I except but a 4 axis mac hine was required to generate the cylindrica l geometry section and
hole axis offsets. A 14mm diameter radial nose tool was used on the cylindrical surfaces finishing
with a 6 mm diamete r bal l nose cutter.

8.6 Design of the assembly process

T he assembly process involved the lit of stage 1 and stage ll panels to their respective
substructures. ln both phases the pane ls will be secured via 20 steel dowel pins which are used to
represent a close tolerance fastener fit typically used in the aerospace industry. The panels were
secured in all 6 degrees o f freedom (DOF) in the following way:

1. Panel pla ne 6 > substructure plane 3 = 3 DOF in Tz, Rx, Ry.


11. Panel hole I > substructure ho le I = 2 DOF in Tx, Ty.
111. Panel hole ll > substructur e hole 11 = I DOF in Rz.

Only two dowel pins are required to secure an orientation and a repeatable assembly. However,
an additiona l 18 dowel pins have been used in this experiment to represent current milita1y
aircraft design which includes many panel to substructure assemblies with multiple fastener to
cage alignments.

13 1
Assembly definitions have mostly been derived by the vanahon stmulahon software companies
such as UGS PLM SolutiOns and Tecnomatlx Technologies The actwn oftakiDg one component
to another IS known as an assembly move The component bemg moved IS referred to as the
object enhty and the recetviDg component IS known as the target entity.

8.7 Determination of assembly key characteristics

The cntlcal features of any assembly are tmtlally determined from a detailed target product
spectficatwn. The spectficahon for thts assembly is genenc and represents a typtcal mthtary
atrcraft weapon systems spectficahon It ~Deludes.

Panels need to be fully interchangeable: thts means being capable of beiDg mstalled, removed
or replaced without adjustment Tradttwnally, mthtary atrcraft removable flyiDg surfaces are not
manufactured to an IDterchangeabthty spectficatwn and are therefore geometncally umque Thts
result ID htgh maiDtenance costs associated wtth component replacement Current and future
atrcraft destgn wtll demand more components of an IDterchangeable spectficatwn ID order to dnve
down maiDtenance costs All fasteners (20 dowel pms) must be used ID the assembly operahon

Panel to substructure assembly must be achieved through a range of environmental


conditions: assembly need to be achtevable ID a range of condttlons, 1.e panel removal for
mamtenance of flying systems at dtfferent world locatiOns and temperatures Mthtary and
commercial atrcraft have extenstve temperature envelopes m whtch they must function. Atrcraft
destgn now mcorporated a greater mtxture of matenal types wtth wtdely varyiDg thermal stabthty
and coefficients of expanswn

The resultant external surface profile of the panel must be within an overall calculated
profile line (CPL): all external flying surfaces are subject to a relahvely close tolerance profile
tolerance Currently thts IS not well understood by aerospace manufactunng compames due to the
complexity of measunng a phystcal enhty and then makmg a comparison to master dtgttal data
(MOD)

The resultant step and gap conditions between the panel and substructure: the assembled
panel must he wtthiD an ID to, and out of, wmd conditiOn spectficahon. Into wmd condttJOn refers
to geometry wtth an onentatwn across lammar atrflow and tf uncontrolled wtll IDter fear wtth
lamiDar atrflow Out of WIDd cond1t10n refers to geometry of onentahon coplanar to lamiDar
atrflow and therefore has a reduced effect on airflow over flyiDg surfaces. Dtfferent

132
specificatiOns are generated for mto and out of wmd component geometry's wtth respect to step
and gap condttton These are Important for two mam reasons. The first relates to mthtary atrcraft
whose destgn reqUirements demand low radar cross section specificatiOn to mimmise
observabthty The second related to both mthtary and commerctal who reqmre low drag
coefficients relatmg to flymg surfaces m order to max dynamtc performance, i e , lower fuel
consumptiOn, htgher erose speeds, and htgher maxtmum speeds for example

The expenment wtll allow the above to be explored Aspects of mterchangeabthty wtll be
invesl!gated through the fit and funcl!on tests at nommal and step ranges of temperature The fit
conditiOn of the 20 dowel pms wtll mdtcate the assembly capabthty at temperature and the
measurement of step and gap conditions wtll also be recorded

There are a number of features on both the panel and substructure whtch are cnl!cal to the
performance of the assembly wtth regard to the above. The key features of the panels and the
substructures can be seen m the followmg figures

)"
Ftgure 58 Panel cnl!cal features- Stage I

133
Figure 59 Substructure en Ileal features- Stage I

i(.

Figure 60. Panel cnhcal features- Stage 11.

134
Figure 61 Substructure cnl!cal features- Stage 11

8.8 Determination of component design tolerances

Part of the design process reqmred the detennmation of a tolerance type and metnc to control
each of the cntlcal features The tolerances were defined to the ASME GD&T Y14.5M 1994
standard. The detenmnatwn of the type of tolerance call out was directly related to the type of
control reqmred of the geometnc entitles, 1 e., hole position for the fastener hole pattern and
flatness of form for the interface landing surfaces The detennmatwn of appropnate designed
tolerance metncs was achieved by extractmg the process capab1ht1es of the manufactunng
eqmpment used for the productiOn of the cntlcal features, 1 e., the Wadkm V4-6 and the Sabre
400H This was achieved by investlgatmg machme capab1hty data from cahbratwn and past
process operatiOns

The tolerance type and metnc value for the manufacture of the panel and substructure are given
below.

Component Feature Tolerance


Panel -stage I Pnmary datum- plane I FLTI02
Secondary datum- hole I PERIDIA 0 025IA, DIA +/-0 09
Ternary datum- hole 11 POSIDIA 0 025(M)IAIB(M), DIA +/-0.09
Hole2 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 09

135
Component Feature Tolerance
Hole 3 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 09
Hole4 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 09
Hole 5 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 09
Hole6 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 09
Hole 7 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 09
HoleS POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 09
Hole9 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 09
Hole 10 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 09
Hole 12 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 09
Hole 13 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 09
Hole 14 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 09
Hole 15 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 09
Hole 16 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 09
Hole 17 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 09
Hole 18 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 09
Plane 2 PERI002IA
Plane 3 PER10021A
Plane4 PERI0021A
Plane 5 PERI002IA
Substructure- stage I Pnmary datum -plane I FLTIO 2
Secondary datum- hole I PERIDIA 0 025IA, DIA +/-0 015
Tertiary datum- hole 11 POSIDIA 0 025(M)IAIB(M), DIA +/-0 015
Hole 2 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole 3 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole4 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole 5 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole6 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole? POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole 8 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole9 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole 10 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole 12 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole 13 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole 14 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole 15 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole 16 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole 17 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole 18 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Plane 8 FLTI004

136
Cl!mponent Feature Tolerance -
Plane 9 FLT[004
Plane 10 FLT[004
Plane 11 FLT[004
Panel- stage 11 Pnmary datum- surface I NIA
Secondary datum- hole I NIA
Terttary datum- hole 11 POS[DIA 0 025(M)[A[B(M), DIA +/-0 09
Hole 2 POS[DIA 0 025(M)IAIB(M)[C(M), DIA +/-0 09
Hole 3 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 09
Hole4 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 09
HoleS POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 09
Hole6 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 09
Hole? POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 09
Hole 8 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 09
Hole 9 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 09
Hole 10 POS[DIA 0 025(M)IAIB(M)[C(M), DIA +/-0 09
Hole 12 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 09
Hole 13 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 09
Hole 14 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 09
Hole 15 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 09
Hole 16 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 09
Hole 17 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 09
Hole 18 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 09
Plane 2 SUR[O 02[A[B(M)[C(M)
Plane 3 SUR[O 02[A[B(M)[C(M)
Plane 4 SUR[O 02[A[B(M)[C(M)
Plane 5 SUR[O 02[A[B(M)[C(M)
Substructure- stage 11 Pnmary datum - plane I NIA
Secondary datum- hole I NIA
Terttary datum- hole 11 POS[DIA 0 025(M)[A[B(M), DIA +/-0 09
Hole2 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 015
Hole 3 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 015
Hole4 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 015
Hole 5 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 015
Hole6 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 015
Hole? POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 015
HoleS POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 015
Hole9 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 015
Hole 10 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 015
Hole 12 POS[DIA 0 025(M)[A[B(M)[C(M), DIA +/-0 015

137
Component Feature Tolerance
Hole 13 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole 14 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole 15 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole 16 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole 17 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Hole 18 POSIDIA 0 025(M)IAIB(M)IC(M), DIA +/-0 015
Plane 8 SURIO 021AIB(M)IC(M)
Plane 9 SURIO 02IAIB(M)IC(M)
Plane 10 SURIO 021AIB(M)IC(M)
Plane 11 SURIO 02IAIB(M)IC(M)

Ftgure 62 Feature tolerance sepectficatwn for stage I and stage 11

8.9 Identify quality measures for investigation

The dtmenswnal quahty measures to be mvesttgated for the panel to substructure assembly
components are

• Interchangeable assembly capabthty m both controlled and uncontrolled envuonments

• Panel and substructure component detatl mspechon

lnterchangeabthty, for the purpose of thts expenment, was defined by the author as

The abthty of any of the manufactured CFC panels to be assembled onto, and removed from, Its
correspondmg alummmm substructure utthzmg all of the dowel pm~. wtthout modlficatton to panel,
pms or substructure, and wtthout the use of 'undue' force

The abthty of the panel and substructure assembly to demonstrate mterchangeablity wtll be
affected by a number of sources of dtmenstonal vanabthty, some of these are

Surface profile error: may lead to devtation from the nommal destgn surface known as the
computer profile hne (CPL}. Devtatwn of the panel surface away from CPL may be the result of:

• Surface wavmess.

• Surface form

• Surface roughness

138
Step and gap condition error: a step error can be descnbed as any dtsplacement between a
surface destgned to blend between an adpcent panel and substructure to form a dtscontmued
condtl!on The gap error can be descnbed as any dtsplacement from the nommal dtstance
designed to extst across panel and substructure edge surfaces.

Hole and pin clearance error: hole and pin VIrtual clearance error between panel and
substructure could result from:

• Vector (axtal) error.


• Centre to centre (pttch) error
• Dtametnc error.

Environmental (temperature) variation: there ts a constderable mtsmatch between the


coefficients of expanston rates between the two matenals used resultmg m features bemg
dtsplaced m space as a result of the matenal mvolved movmg at dtfferent rates

Dtmenswnal mspectwn wtll take place usmg the followmg eqmpment.

• Brown and Sharpe Tessa 3D CMM usmg the MM4 software modules
• Brown and Sharpe Tessa 3D CMM usmg the VALISYS software modules
• Prectswn pm gauges
• Prectswn shp gauges
• Mtcrometer 25mm gauge
• Hetght gauge wtth dtal mdtcator.

All dtmenswnal mspectwn, except temperature related vanatwn, take place m the metrology
laboratory. Thts has a controlled envtronment wtth temperature at 20° C, +/-1° C, and humtdtty at
50'\+/- 5%

8.10 Design of inspection methods

The expenmental mspecl!on methods were destgned to address the quahty measures defined m
the prevtous sectiOn Detat!ed ptece part mspectwn was earned out tmtially m a controlled
environment Further mspection was undertaken at defined temperatures m order to investtgate

139
-I
I

the effects of thermal expanswn and contractton on the assembly An outhne of these processes IS
presented below

8.10.1 Controlled environment inspection methods

The followmg mspectton methods were undertaken for the controlled environment mvesl!gatton.
The environment parameters were temperature 20° C, +I- I ° C, hum1d1ty at 50%, +I- 5%. All
controlled environment analysis was undertaken m the metrology laboratory at the Wolfson
School ofMechamcal and Manufactunng Engmeenng

8.10.1.1 Surface profile error

Surface waviness error: detected through the use of surface fimsh and surface form mspectton
on CMM The surface fimsh detector was hm1ted to a range of 5.5cm m a hnear mspecl!on. The
CMM probed the enl!re panel surface recordmg data at 20-25 equally spaced pomts
Surface form error: detected by comprehensive CMM mspecttons for form error on each surface
and edge enl!l!es Surface relationships were also mvesl!gated via the CMM for squareness and
pelJlendiculanty.
Surface roughness error: detected through the use of surface fimsh analysis
Thickness error (panel only): detected through the use of a standard 25mm micrometer. A panel
thickness measurement was taken at the prox1m1ty of each assembly hole.

8.10.1.2 Step and gap error

Step condition: determmed by usmg a d1al gauge and square base The assembly and the dial
gauge were placed onto a gramte block surface, the stylus set on the panel upper surface and the
gauge was set to zero. The stylus and gauge was then manoeuvred onto the step surface of the
substructure and the value recorded Due to the form error of the panel a number of recordmgs
were made for each of the four step condii!ons, from these a mean was calculated In add!I! on the
CMM was also used for an mspecl!on using the MM4 and V ALISYS software modules

Gap condition: determmed by usmg shp gauges The mvesl!gatton was undertaken by fittmg the
different shp block gauges m the gap cond!l!on at approximately four equal posillons The largest
shp gauge that could generally be pressed mto largest gap was Idenllfied and recorded. The
CMM was also used ulllizmg the MM4 and V ALISYS software modules.

140
----------------------- --

In order to Simulate a larger number of panels to substructure assemblies each panel was rotated
through 90 degrees four ltmes and step and gap cond11tons were mvest1gated The panels were
then turned over and rotated through 90 degrees four ltmes and mvesltgated further Due to the 2
panels, this techmque Simulated the fit of 16 panels onto the substructure

8.10.1.3 Virtual hole condition error

The followmg methods were used to invesltgate panel and substructure hole form, hole p1tch and
hole d1ametnc error. The hole errors for each component were determmed from a smgle
measurement routme on the CMM. Each component was g1ven a datum as highhghted m figures
11 and 12 The MM4 routine incorporated an eight point mspect1on techmque at each hole Four
of the pomts were taken eqmd1stant at the extreme top of the hole and the remammg 4 pomts
equ1d1stant at the extreme bottom of the hole. The VALISYS mspeclton software uses the same
techmque but mcludes a th1rd set of four pomts m1dway between the first two g1vmg a total of
twelve measurements Th1s was done to ensure the most accurate mterpretatwn of hole geometry,
especially about 1ts ax1s F1gure 16 below h1ghhghts the MM4 techmque The routme mcluded.

Panel and substructure hole form error: the form (cyhndnc1ty) error was recorded for each
hole Cyhndnc1ty IS calculated by the creatwn of two coax1al cyhnders created about the 1m1tal
best fit cyhnder The coax1al cyhnders are created at the w1dest and narrowest mspectwn pomts
and the rad1al d1fference between the two IS recorded as the cyhndncal form error Th1s IS
denoted as 'f' on an mspeclton output

Panel and substructure hole centre to centre (pitch) error: hole to hole (p1tch) error IS
calculated by recordmg the x and y coordmates of each hole m1d pomt pos11ton and resolvmg for
d1stance. The centre of gravtty pOSIIton IS denved from the best fit cyhnder for the 8 inspeclton
points undertaken for each hole mspected

Panel and substructure hole diametric error: hole d1ameter was taken form the best fit cylinder
form the 8 pomt mspectwn techmque The error was calculated by calculatmg the d1fference
between the nommal and inspectwn d1ameter metncs

Inspection of assembly dowel pins: the assembly of the panel to the substructure was v1a 20
dowel pms These dowel pms were bought m and mspected on a CMM v1a a magnetiC V block.
The techmque was s1m1lar to the above hole mspeclton usmg an 8 pomt process, four at each end

141
of the pms. The data from the 8 points was then used to calculate a best fit cylmder around the
pm from which a diameter was recorded. The form of the pm was also recorded by the same
process as for Panel and substructure hole form error given above. The dowel pm specification
was diameter 6 mm, +0 004/+0 009 (based on suppher data)

8.10.1.4 Multiple temperature environment inspection method

The followmg mspectwn methods were undertaken for a range of environmental temperatures m
order to demonstrate Its effect on geometnc variation There IS a considerable mismatch between
the panel and substructure matenals with respect to their coefficients of thermal expansion rates
These expenments were designed to investigate the effects of temperature chance on the assembly
of panel to substructure and on some of 1ts cntical features The temperature range chosen for tlus
expenment reflect a typical m1htary and commercial aircraft operatmg reqUirements specification
M1htary aircraft have particularly demandmg environmental envelopes With not only a wide range
of operatmg temperatures to endure but also a rap1d transition from one end of a range to the
other. Three parameters were investigated, these were;

• Step and gap conditions of panel to substructure.


• Panel and substructure hole d~ametnc error
• Assembly performance of panel to substructure v1a dowel pms

The above mvestigatwns were undertaken m the same temperature batch sessiOn and m the order
given The analysis was performed usmg a controlled oven booth for all temperatures above 0
degrees, and a controlled refngerator umt for below 0 degrees. The environmental temperatures
were momtored at all times dunng the mspectwns usmg digital readout thermocouples.

Step and gap conditions of panel1 and substructure: the panel and substructure were 'soaked'
m the appropnate temperature environment for a penod of time m order to allow the components
to stab1hze at the reqmred temperature Once the components had stab1hzed, they were removed
from their environment and mspected rapidly m close proximity m order to reduce thermal change
(rapid heatmg and coohng) error. After each mvest1gatwn, the components were replaced for a
penod of time until they became thermally stable once agam This process was repeated until all
the relevant measurements had been completed.

Hole diametric error of panel 1 and substructure: the technique for mducmg the correct
thermal environment was the same as above for the step and gap mspect10n The hole d1ametnc

142
error of the panel and substructure was deterrmned mdependently through the use of precisiOn pm
gauges The gauges used were of 25 mtcron dtametnc steps

Assembly performance of panel to substructure via dowel pins: the teclmique for inducmg the
correct thermal envuonment was the same as above for the mspectwn cnterion The assembly
performance of the panel and substructure was determmed by the fit of the 20 dowel pms Thts IS

essentially a hard gauge test It was conducted by placmg the panel onto the substructure such
that the pnmary holes of each were ahgned 1 e , panel hole I to substructure hole I, and a dowel
pm mserted into both at hole I. The assembly performance was then measured by msertmg dowel
pms equally m the x and y dtrectwns of the panel! e, hole 2, hole 20, hole 3, hole 19 etc until a
no fit sttuahon was encountered. The fit condthons at each hole were recorded.

8.10.1.5 CMM inspection using VALISYS V5.3.1 software

The VALISYS system IS made up from a number of modules A detatled descnphon of the
VALISYS process and modules are available m the appendtces of thts thests. The procedure
hsted below (figure 63) was used to undertake the CMM mspechon usmg the V ALISYS software.

Step No. Activity Function Module


Step I Specify nommal dimensions of part - (CATIA)
Step 2 Create CAD model of part - .
Step 3 Define features to apply_tolerances and datum's FEATURES Design
.
Step 4
Step 5
Apply tolerances to features
Venfy and build tolerance schemes
GAUGE
CHECK ..
Step 6 Programmmg
Step 7
Create mspecuon path and processes for features
Define mspectmn probe (Remshaw PH9)
PATH I
.
Venfy/modify mspectiOn process
PROBE
.
Step 8
Step 9 Create calibratiOn process
PATH2
PATH3 .
Step 10 Export all Valisys data to system VUTILITY Design
Step 11
Step 12
Import export files to CMM controller
Acllvate the part
FILES
.
ln•pectmn

Step 13 Establish link between the system and CMI\1


FILES
.
.
INSPECT
.
Step 14 Calibrate the probe
. ..
Step 15 Onentate the part
. .
Step 16
Step I 7
Run mspectmn Process
" .
Step 18
Disconnect the link
Analyse the measured vanation vs tolerance QUALIFY .
Step 19 Analyse the measured data ANALYZE Analyse
Ftgure 63 The VALISYS procedure

The structure of the reports produced from VALISYS IS explamed m the followmg section.

143
1. Analysis mode summary, bnefly these are.

Cylinder ax1s method LEAST SQUARES (LS)


Limits of s1ze RMSAVERAGE (AVE)
Ax1s extrapolal!on YES (YES)
Datum method BEST FIT (BF)
Degrees of freedom All to GAUGE NIA

n L1st of feature variation and tolerance detmls and the1r post mspection status i e. PAS SED
or FAILED
ll1 Inspected part details and mspectwn process number
IV Summary of gauge analys1s

8.11 VSA tolerance analysis

As part of the expenment a computer tolerance analys1s study has been perfonned ulilizing the
VSA software The tolerance analys1s study w1ll analyse the effects of geometnc vanatwn on key
features w1thm the panel and substructure assemblies

The methodology used for th1s analys1s Will be Similar to that outlined m chapter 7. It Will mclude
of the followmg stages

a Define study aims and obJeCI!ves.


b List assumptwns
c. Des1gn assembly components and tooling features
d Idenlify key features on components and tooling.
e Define tolerances from process capability data for all key features.
f Create assembly model.
g Define assembly sequence.
h Idenlify assembly moves
Define analys1s measurements.
J Perfonn assembly s1mulatwns.
k Assess results.

The results of the tolerance analysis study are presented later in th1s chapter

144
8.12 Experimental results

The expenmental methods descnbed above produced the followmg results.

8.12.1 Geometric surface form error of panels and substructure

The following results were determmed for geometric surface form error in the panels and
substructures

Panel surface waviness- stage I


Panel surface wavmess IS the cychc deviation of an external surface form from nommal. In panel
I and panel 2, wavmess is probably caused by the crossmg of the carbon fibre bundles during the
lay-up procedure and resultant deformatiOn dunng cunng processes Surface wavmess IS difficult
to measure directly For this experiment wavmess was Identified using a Form Talysurf Senes JI.
This eqmpment IS designed to measure surface form and texture over a hnear range of !20mm
The panels were taken to Taylor Hobson at Leicester where they were measured The CMM data
of the panels, taken on the Tessa 3D, also gives an mdJcatJOn of form error (see figure 64)

Panel I Pane12
plane I -0 0575mm plane I - 0 0523mm
plane 6-0 2120mm plane 6-0 125lmm

figure 64 Stage 1- panel! and panel2 form error

Each panel plane was mspected with a cloud of 30 measurement pomts. Takmg any
measurements usmg a cloud of pomts philosophy basically mvolves takmg as many pomts as
deemed necessary (usually a h1gh number) on a surface, evenly dJstnbuted over an area, m order
to gam an understandmg of 1ts topographic defimtwn

From the results plane 6 on both panels displayed a much greater error of form compared to plane
I of each This IS probably due to each plane I bemg on the side of the laid-up toohng. Plane 6
possibly mhent their form defimtion from plane I plus all the subsequent CFC layers placed on
top dunng lay-up and cure.

The Talysurf Senes II wavmess and form mspectwn results are presented in figure 65 on the
followmg page (all measurements m miCrons).

145
Surface inspected Wa Wv Sm Wp Wt
Panel I, plane I m x 5 62 16 04 473 15 93 31 97
Panel!, plane 6 m x 14 76 45 36 4 81 61 2 106 56
Panel 2, plane I m x 14 34 42 37 564 54 55 9692
Panel 2, plane 6 m x 65 16 03 3 32 I5 69 3I 72
Panel 2, plane I my 4 91 12 24 4 I 15 76 28
Figure 65 Stage I- TalysurfSenes 11 wavmess mspectwn results.

Key·
Wa Correspondmg parameters for wavmess and primary profile; amplitude
Wv Correspondmg parameters for wavmess and pnmary profile; amplitude
Sm Correspondmg parameters for wavmess and primary profile; spacmg
Wp Correspondmg parameters for wavmess and pnmary profile, amplitude
Wt Correspondmg parameters for wavmess and pnmary profile, amplitude

From the results It can be seen that the wave forms m the panel I, plane I and panel 2, plane I are
more umform and of smaller magnitude when compared to panel I, plane 6 and panel 2, plane 6
These results are consistent w1th those from the CMM form mspectwn data Both mdiCate a
wavmess error present w1th the lower panel faces (plane 6) demonstratmg a greater magnitude
than that of panel upper faces (plane I). The last measurement on panel 2 was taken m the y
directiOn, where all previous measurements were m the x direction. This resulted m a different
pattern producmg a bolder wave form output but wtth stmtlar amplitude.

Surface wavmess has the effect of mcreasmg panel thickness and therefore mtroduces variatiOn
In a planar shape panel as used m stage I, the effects of panel wavmess are mmimal However,
the effects of excessive wavmess on a cylindncal or complex curve may present a bigger problem
Thts may lead to hole misalignment and out of specification step and gap conditions

Panel thickness variation - Stage I and 11


Panel thickness vanation IS mamly due to the CFC lay-up and cure process The process to lay-up
and cure CFC panels IS complex and mcludes a large manual mput The panels manufactured m
stage I and 2 are composed of more than 40 layers of prepreg weave matenal that has to be
vacuumed down between approx. every 4 or 5 layers to expel any mr mcluswns. Th1s type of lay-
up process mduces net panel thickness vanatwn When this IS combmed wtth the vanatwn
caused by wavmess, panel to panel and batch to batch thickness variatiOn can vary as much as 5%
The results of the panel thtckness measurements can be seen m the figures below.

146
Stage I -panel 1 and panel 2 thickness inspection results

6 200
6 150
6100
6050
~ 6000
" 5 950
" 5 900
~ 5850
5800
5 750
5700
5650
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Hole number prox1m1ty

: --Panel 1 -<>-Panel 2

F1gure 66. Stage I- panel! and panel2 th1ckness mspectwn results

The results show that for Stage I there ts an approximate mean wtdth dtfference of 0 2mrn for the
nommal 6mrn panels Thts vanat10n wtll affect assembly performance m respect to the resultant
step condtl!ons between panel and substructure

Stage 11 -panel I and 2 thickness mspect10n results

62
6 15
-5 61
:a 605

"""" 5956
0..

59
585
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Hole number prox1m1ty

1--Panel I -<>-Panel 2 J
F1gure 67 Stage li- panel I and pane12 th1ckness mspectwn results

For Stage 11 the results are less stable. There IS an approximate mean wtdth dtfference rangmg
from Omm to 0.15mm for the nommal 6mrn panels As for stage I, thts vanatton wtll affect
assembly performance m respect to the resultant step condttwns between panel and substructure
for stage 11

147
Panel geometric error- Stage I
The panel geometnc error represents the overall error of each planar surface. Form errors are
measured mdependently of any datum reference and present a fitted plane and deviations form
that plane. Form error generally dtffers from wavmess error. Form error descnbes devtatton
away from a nominal plane but tt does not generally used to descnbe a type pattern lt represents
a gradual increase or decrees of form devtatwn such as a twtst or a bow whtch could also be
descnbed as very long wavelength error. This may be cause by an element of the manufactunng
process such as the dtstortion of a steel CFC cure mould tool The figure below graphtcally
htghhghts the results.

Stage I -panel 1 and panel 2 form inspection results

L
~
0200
s
s
~
~
0100

/
0
t:
§" 0 000
0

"""
{0 100)
P1 P2 P3 P4 P5 P6
Plane

1-- Panel 1 -<>- Panel 2 I


Ftgure 68 Stage I- panel I and panel2 form mspect1on results

As can be seen from the results the form error on the top and bottom {planes I and 6) are of a
dtfferent magmtude. In panel I, plane I ts of much smaller magmtude than plane 6. Thts ts
probably due to the onentatton of the panel! ay-up Plane I was probably the stde m contact wtth
the lay-up tool whtch accounts for tts low form error. Plane 6 represents the final surface of the
lay-up process and mhents from bmlt up from each layer depostted

The form error of planes 2-5 represent the net panel edges. These were machmed as part of the
manufactunng process and it can be seen that thetr form error in considerably smaller when
compared to planes I and 6. The error seen m panel 2 ts generally larger than panel I Thts ts
due to a problem wtth the routmg process that occurred on panel 2. The feed rate was to fast
whtch resulted m the panel edge becommg delammated and damaged. CFC is a demandmg

I48
material to cut and the correct cuttmg speed and feed rates are crihcal m order to gam a good net
fimsh.

Substructure geometric error


Substructure geometric form vanation will mamly be caused by machme process capability The
form error can be seen m the figure below The mam contnbutors to machme process capabihty
are:

• Tool accuracy; effected by tool wear, tool damage


• Machme positiOn accuracy, effected by mherent specification, machme wear
• Environmental conditions; effected by temperature and humidity.
• Efficient programmmg, effected by programmmg techmques, feed and speeds.

Stage I - substructure form mspection results

0 200

E
5
~ 0100
g
"E ...... .....
...
0
0 000

(0 100)
P2 P3 P4 P5 P6 P7 P8 pg P10 P11
Plane

1-- Substructure I
Figure 69 Stage I- substructure form mspec!Ion results

As can be seen from the results m the figure above, the largest form error IS on the panel and
substructure assembly mterface, plane 3 Th1s IS due to the proporhonally larger surface area of
plane 3 in companson to the other planes

149
Stage I - panel!, panel2, and substructure furm mspection resuhs

0200
a'
~
a A
~
0100
g
~

0 0000
V-------"' "' ,,
A
A
1:.' "' "'
"'
(0 100)
P1 P2 P3 P4 P5 P6 P7 PS pg P10 P11
Plane

1-- Panel 1 -+- Panel 2 A Substructure I


Figure 70 Stage I- panel!, panel 2, and substructure form mspectwn results

There are no planar form measurements possible on stage II due to the geometry bemg cyhndncal

8.13 Step and gap inspection results

The followmg results were determmed for the step and gap conditions m panel and substructure
assembly

Step condition of assembly- Stage I and 11


The step conditiOns of the assembly represent the distance from the panel surface to the upper
substructure step surface, measured normal to each_ The step condition on an aircraft has an
effect on the Jammar a1rflow about the outside profile If allowed to dnft out of specification It
can affect an a1rcraft m two ways-

• Reduction of aerodynamic effic1ency leadmg to loss of dynamic performance 1 e , top


speed, cruse speed, range, etc,
• Increase of radar cross sectiOnal area Jeadmg to potential mcrease of observability,

For the purpose of stage I expenmental!on, the step conditions were recorded and no companson
was made to a reqUirement specification The step condition results record a value for the
difference from the substructure to the panel step height Therefore a positive step recordmg
represents the panel surface bemg relatively higher than that of the substructure, The step
condition results can be seen m figures 71, 72, and 73

150
Stage I -panel! step condrt10n rnspect10n resuhs

0 250

~ 0200

g 0150
~

] 0100
"
B" 0 050
VJ

0000
2 3 4

Onentat1on

1-- Plane 4 --+-- Plane 5 __...._ Plane 6 o Plane 71

Ftgure 71 Stage I- panel l step condtt!On mspectwn results

Due to the relative consistency of results, the error IS probably due to the panel th1ckness vanat10n
m the regiOn about plane 4 and 5, 1 e, the panel w1dth at these reg1ons are below nommal. The
step conditiOn of panel 2 could not be measured due to the damage resultmg from the routmg
process reqmred to produce the net edge.

For the stage II expenmental!on, the step condtltons were recorded and no companson was made
to a reqmrement specification. The step cond1lton results record a value for the difference from
the substructure to the panel step he1ght Therefore a pos1ltve step recordmg represents the panel
surface bemg relatJvely htgher than that of the substructure The results can be seen m the figures
below.

Stage 11 -panel! to substructure step condition


inspection resuhs

Plane 5 Plane 7
Substructure plane nurrberreference

I-+- Normal --Reverse I


Ftgure 72 Stage 11 -panel I step condttton mgpectton results

151
The step measurement error is agam probably due to the panel thickness vanatwn m the regwn
about plane 5 and 7, 1 e, the panel width at these regions are below nommal The relative
difference between the two onental!ons, which IS approximately 0 I mm, suggests uneven panel
thickness with some possible cylmdncal fonn error. The same observatiOns can be attnbuted to
stage !I, panel 2, step conditiOn results

Stage II- panel2 to substructure step condition


msp ection resuhs

Plane 5 Plane 7
Substructure plane number reference

I-+- Nonnal __.,_ Re-.erse I


Figure 73 Stage 11- panel2 step conditiOn mspecbon results

Gap condition of assembly- stage I and II


The gap conditiOns of the assembly represent the distance from the panel edge to the substructure
mtemal planar edge, measured normal to each plane The gap conditiOn on an aircraft can also
have an effect on the lammar airflow about the outside profile. If allowed to dnft out of
specifical!on It can effect aerodynamic effictency leadmg to loss of dynamic performance 1 e , top
speed, cruse speed, range, etc For the purpose of stage I expenmental!on, the gap condiiions
were recorded and no companson was made to a genenc reqUirement specifical!on The gap
conditiOn results can be seen below

!52
Stage I - pane! I gap condition inspection results

0000
E' (o o2o)
5 (0 040)
g (0 060)
.g (0 080)
g<.> (0 100)
g. (0 120)
" (0140)
(0 160) ' - - - - - - - - - - - - - - - - - - - - - - - - - '
2 3 4

Onntat10n

1-- Plane 4 -+--Plane 5 -A- Plane 6 o Plane 71


Ftgure 74 Stage I- panel I gap condt!Ion mspectton results

As can be seen form the results the mean devtal!on of the gap condtl!on IS consistently below
nommal, 1e the gap conditiOn IS smaller than nommal This suggests that the manufactunng
process used to produce etther the substructure or the panel was cuttmg under stze. Thts IS

consistent wtth a set of results measunng the dtstance between the datum hole and planar edge for
each panel (see panel I and 2 datum to edge dtstance inspectiOn results). These results
demonstrate that the CFC panels machmed consistently oversize whtch would result m a reduced
gap condtl!on after assembly to the substructure.

Stage I- panel 2 to substructure gap condttton mspectton


results

0400

I 0 300
0200
0 100
~ 0000
-g (0 100)
8 (0 200)
g. (0 300)
" (0 400)
(0 500)
2 3 4
Onentat10n

1-- Plane 4 -+-Plane 5 --...-Plane 6 - o- Plane 71


Ftgure 75 Stage I- panel2 to substructure gap condt!Ion mspectton results.

153
Stage 11 - panel I to substructure gap
inspection results
Or-----~-----r-----,-----,
"
0
~ '2 -200
il 0
r; ~ -400
., El
g.~ -600
0 -800 L __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __J

Plane4 Plane 5 Plane 6 Plane 7


Substructure plane number reference
I-+- Normal -11- Reverse I
Ftgure 76 Stage I!- pane! I to substructure gap mspectton results.

From the stage 11 panel I and panel 2 to substructure results tt can be seen that the gap conditiOns
were also understze compared to nommal, see figures above and below

The magmtude of error changes wtth the edge type. Plane 4 and 6 for panel I and 2 were
conststently measured as bemg approxtmately 0 6mm understze, where as planes 5 and 7 were
measured at only 0 2mm undefSlze Thts dtscrepancy may be attnbutable to a CNC machmmg
program error on etther the panel or substructure.

Stage 11- panel2 to substructure gap mspectlon


resuhs.
o~---~---~------~------~
"52
~'2
-200
- 0
~ b -400
.g E
§- ~ -600
0
-800~----------------------------~

Plane4 Plane 5 Plane 6 Plane 7


Substructure plane number reference

I-+- Normal - - Re-.erse I


Ftgure 77. Stage I!- panel 2 to substructure gap mspectwn results

154
8.14 Hole condition of panels and substructure

The followmg results were determined for the hole condtltons m the panel and substructure
components. The 20 hole patterns m panel I, panel 2 and the substructure were mspected The
measurements are dtvtded mto three categones, these bemg

• Hole form (cyhndnctty) error.


• Hole centre to centre (pttch) error.
• Hole dtametnc error

Stage I- panel I, panel 2 and substructure hole form inspection results


The followmg results represent the measured hole form error in panel I, panel 2 and the
substructure for both stage I and stage ll The hole form error of panel I, panel 2 and substructure
for stage I can be seen below

Stage I -panel I, panel 2, and substructure hole form mspectton


results

~ 0 010
E ooo8
§ '? 0006
<8 ..§, 0 004
0" 0002
:r: 0000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Hole No

1-- Substructure --+- Panel1 -....-Panel 2 I


F1gure 78 Stage I -panel I, panel 2, and substructure hole form mspect10n results

As can be seen form the results for stage I the hole form vanes relattvely httle Both panel I and
2 were m approxtmately the same range (between 0 007 and 0 00 I) wtth a mean of 0 003

155
Stage I! -panel 1, panel 2, and substructure hole form
mspectlon results

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Hole number

1-+- Substructure - - Panel 1 - t;. Panel 2 I


Figure 79 Stage 11- panel I, panel2, and substructure hole form mspect10n results

For stage I!, the results md1cate that hole form vanes relallvely httle. The average results for the
substructure, panel I, and panel 2 were 0.004mm, 0 0 14mm, and 0 009mm respecllvely. The
results of stage I! are less consistent w1th those from stage I due to an mcrease m geometnc
complex1ty The range for each of the outputs has also mcrease compared to the stage I results

For both stages I and I! the substructure holes were fimshed w1th a bonng bar where as the panels
were dnlled m a one h1t cycle usmg a two flute tw1st dn!L The form of the panel holes m stage I
are generally better than that of the substructure, wh1ch was unexpected. Th1s can be explamed
partly by the different depth of holes The depth of hole m the substructure IS more than tw1ce that
of the panel Th1s deeper hole m the substructure, although theoretically more accurate, recorded
a b1gger error m stage I Th1s suggests that the depth at wh1ch an mspecl!on process takes place IS
cruc1al to the output results, for example, 1f the substructure holes were to be mspected over the
same depth range as that set for the panel, then the author would expect a better form result from
the substructure holes.

Stage I and 11- panell, panel 2 and substructure hole centre to centre inspection results
The followmg results represent the measured hole centre to centre error m panel!, panel 2 and the
substructure. Panel I, panel2 and substructure p1th error results can be see m figures 81-85.

!56
Stage I - panal 1 hole pitch deviation inspection results

s 0020
5 0 010

~" 0000
5: {0 010)
"
-o
{0 020)
""B
0: {0 030)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Hole No

1-- X co-ord1nate - - Y co-ordmate I


Ftgure 80 Stage I- panel! hole pttch devtatton mspectton results

As can be seen from the results panel 1, panel 2 and the substructure hole mspectwn demonstrates
consistent and accurate hole post!ton. Hole posttion accuracy ts dependent upon machme tool
accuracy All components were dnlled on the Wadkin V4-6 which when last calibrated tndicated
a position accuracy of+/- 25micron over the full operatmg envelope. Due to the close proximity
and the planar axtal onentatwn of all the holes It can be seen from the results that all the hole
post!tons are well withm that specificatwn

Stage I -panel 2 hole pttch devtation inspectiOn results

ss 0005
0000
';:;' {0 005)
:2 {0 010)
l'i {0 015)
i) {0 020)
:::: {0 025)
B
0: {0 030)
{0035) L __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ J

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Hole No

1-- X co-ord1nate - - Y co-ordmate j


Ftgure 81. Stage I- panel2 hole pttch devtatton mspectton results.

157
Stage I- substructure hole pitch deviation inspection results

eE 0
0
020
015
~

c 0 010
0
:;: 0 005
> 0 000
"'" (0
..c (0
005)
B 010)
0:: (0 015)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Hole No

1--X co-ordtnate ...._ Y co-ordrnate I


F1gure 82 Stage I - substructure hole pttch devtatwn mspect10n results

Stage 11- substructure hole pitch deviation mspection results

I
·---· • ----·
0300
c
g
0200
0100 ••••••
0000
:0 -0 100
>
"'"
..c
-0200
-0300 -rt: !; !;
"
!;
!; ll. !; !; !; -n-~
B
0::
-0400
-0500 " !;
ll.
!; !;

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Hole No

1-- X co-ordrnate -+- Y co-ordrnate ll. Z co-ordrnate I


F1gure 83 Stage 11- substructure hole p1tch dev1at10n mspectwn results

Stage I!- panel I hole pitch deviation mspectwn results

eE 0 600

~...-;::: :::::.~ ~A
...... ~
~
0400
c 0 200

~:::: ~· -
g 0 000
:0 -0 200
> -0400
"
"'..c -0 600
-0 800
B
0:: -1 000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Hole No

1-- X co-ordrnate ...._ Y co-ordrnate ll. Z co-ordrnate I


F1gure 84 Stage 11- panel I hole p1tch dev1at10n mspect10n results

!58
Stage 11 - panel2 hole pitch deviahon inspectiOn results

8 0600
8 0400 ... -
~
0200 •• 0 0
• 0 ---~-A
jj- -;-a_- ll-- fl_--A~
"g 0000
"-"
A" '.

~/ "
~
,.
!:! -0200 {l-ll- {l--"
-0400
"' -0 600
'"0
" "
-=
B -0 800
,;:; -1 000 ·-
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Hole No

1--- X co-ord1nate ~ Y co-ordmate Z co-ordinate I


"
Ftgure 85. Stage 11- panel2 hole pttch devtatwn mspect10n results

Stage I- panel!, panel 2 and substructure hole diametric inspection results


The followmg results represent the measured hole dtametnc error m panel I, panel 2 and the
substructure The results can be seen m the figures 86 and 87 below.

Stage I -panel I, 2, and substructure hole dtameter mspectwn


results

.. • • .. • • .. .. .. a
~ 6100
a a • t • v• .. .. ..
*
';:::' 6 080

.;;
0"'
6060
8 6040
"' 6020
:r: 6000
• • • • ..------. • • • • • • • • • • • ....
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
.
Hole No

1--- Substructure ~ Panel 1 -.-Pane12j

Ftgure 86 Stage I- panel I, panel 2, and substructure hole dtameter mspec!ton results

The Panel holes were produced by a smgle cycle dnllmg operatton wtth a sohd carbtde two flute
twtst dnll of 6 I mm diameter. A smtable dnll dmmeter was tdenttfied to compensate for the
charactensttcs of CFC matenals. These charactensttcs cause a resultant dnlled hole dtameter to
shnnk back to below the nominal cutter stze This occurs because the fibre bundles become
forced apart dunng the dnlhng operatton which subsequently contract when the dnllts Withdrawn
resultmg m a dmmetnc reductton

159
The substructure holes were dnlled using a 6nun diameter twist dnll and then fimshed with a
bonng bar. It can be seen from the results that the substructure hole diameters are smaller than
those of the panels The panel holes demonstrate that they have consistently shrunk back below
dnll diameter by approximately 14 microns, less than anticipated The hole diameters were
expected to shnnk back to approximately 6 050nun or less but this did not occur Both panels
gave a consistent measurement which suggests that the resultmg shnnk back from a CFC dnlhng
operatiOn may be predicted accurately.

Stage 11 - panel I, panel 2, and substructure hole diameter


mspection results
~

ss 6 1
~(J.~~~~·i!~
*s"'
-:: 6 08
6.06

-
604
.;;
0"'
:r:
602 . ._
6
1 2
-
3 4 5 6
/
----
7 8 9 10 11 12 13 14 15 16 17 18 19 20
Hole No

1-- Substructure --a-- Panel 1 ·A· Panel21


Figure 87 Stage 11- panel!, panel2, and substructure hole d1ameter mspectwn results

8.15 Assembly dowel pin inspection results

The dowel pms used for the assembly of the panels to the substructure were mspected The
mspecllon results can be seen below

Dowel No. Diameter (mm) Form(mm) Pass/fail


I 6009 0003 Pass
2 6007 0002 Pass
3 6009 0 003 Pass
4 6010 0002 Fail
5 6009 0 001 Pass

F1gure 88 Dowel pm mspectwn results

Twenty dowel pms (6mm diameter, +0 004/+0 009) were purchase for the assembly of the panels
to the substructure A sample of 5 dowel pins (25% sample) was mspected on the CMM It can

160
be seen from the results that the dowel pin d1ameters were cons1stently on the upper hm1t of their
tolerance. One dowel pm was recorded as bemg I m1cron over tolerance (25% fa1lure rate), but
due to the s1ze of error bemg very small 1t was mcluded m the assembly
The dowel pms were also mspected for form for mformat10n only

Gap condition of pan ell to substructure


The gap cond1hon mspect10n results are presented m the figure below

Stage I - panel! to substructure gap mspection results.

20
'C
e 0
g" -20
"
0 -40
~;; -60
~ -80 ~~----+-------o
g. -100
0
-120
2 3 4 5 6 7 8
Onentation No

1-+- Plane 4 --..-Plane 5 o Plane 6 -<>-Plane 71


figure 89 Stage I- panel! to substructure gap mspectwn condliion

Step condition of panel 1 to substructure


The step cond1hon mspechon results are g!Ven m figure 90 The results md1cate the poor surface
cond11!on followmg the delammatmg problems dunng routmg

161
Stage I - panel I to substructure step condlhon mspectlon results

350
~
c: 250
2
"
_§_ 150
c: 50
0
~ -50 "'-
;;
.," -150
0.
.!l -250
V>
-350
2 3 4 5 6 7 8
Onntatton No

1-+-- Plane 4 --.-Plane 5 o Plane 6 - ~- Plane 71

Figure 90. Siage I- panel! to substructure step cond1t10n mspectwn results

Gap condition of panel 2 to substructure


The gap cond1t10n mspectwn results are given m the figure below The results mdicate the poor
edge condition of the panel followmg the delaminatmg problems during routmg

Stage I- panel2 to substructure gap mspection results

~
c: 20
2
~ -80
,§,
c: -180
~
il -280
.,"> -380
fir
'-' -480
2 3 4 5 6 7 8
Onentahon No

1-- Plane 4 -+- Plane 5 o - Plane 6 --Plane 7 I


Figure 91 Stage I- panel 2 to substructure gap mspectwn results

NB. No step condition has been recorded for panel 2 due to delammatmg damage dunng
manufacture

162
8.15.1 Multiple temperature environment inspection results

Step condition of panel 1 to snbstructnre


The step condition mspectiOn results for a range of different temperatures are presented m the
figure below.

S!age I - panel I to substructure step condrtxm mspecflon results at


different temp.

150
~
~
0
100
~
u
50
5
~

~ 0
g
.;; -50
~
0
u -100
0.
1l -150 D
Cll

-200
-30 0 20 40 60 80
Temperature (degrees C)

1-- Plane 4 -.--Plane 5 o Plane 6 -<>-Plane 71

Ftgure 92 Stage I- panel! to substructure step condttwn results at dtfferent temperature

It can be seen from the results that through the range of temperature planes 6 and 7 are behavmg
as predicted, 1 e substructure matenal IS expandmg at a faster rate than the panel and therefore
there IS an mcrease m step size Step condlllons at plane 4 and 5 however are behavmg as 1f there
IS a twist or bow occumng dunng the changes m temperature

Gap condition of panel 1 to substructure


The gap conditiOn mspectwn results for a range of dtfferent temperatures are presented m figure
93 below

163
Stage I -panel I to substructure gap condition inspection
results at dtfferent temp.

03
8
s
~
02
~

..
01
g" D
-a 0
5 0"~~

D~
~

"
0
<..>
-0 1
g. -0 2
0 -0 3 --
-30 0 20 40 60 80
Temperature (degrees C)

1-+- Plane 4 --Plane 5 o - Plane 6 ~ Plane 7 I


F1gure 93 Stage I- panel I to substructure gap condition results at d•fferent temperature

As can be seen form the graphic results. plane 6 and 7 are relahvely stable but as expected the gap
cond1t10n widens as temperature mcreases but thts ts only ev1dent at the higher temperatures
This would also suggest that there may be some tw1stmg or bendmg of the panel edges thought a
large temperature range The gap cond1t1on of plane 4 and 5 also wtdens as temperature mcreases
but With a more consistent rate of change

Stage I - panel plane-to-plane perpendicularity


The perpend1culanty relatwn between the different planes IS given below. The result suggests
that there were no maJor errors to be noted (see below)

Stage I - panel I and panel 2 perpendtculanty inspection results

0 200
8
,§,
k
g 0.100
"
~
~

" 0 000
i
~ (0100) L-----------------------------------------~
P2- P3 P3-P4 P4-P5 P5-P2
Plane-to-plane relatwnsh1p

1--- Panel1 -+-Panel 2 I


Figure 94 Stage I- panel I and panel 2 perpendtculanty mspectlon results

164
Stage I- substructure !mer Inspection results

006.-------------------------------------~

~
";::' 0 04
g
"'2 002
"'
.3
oL-----~-------L------~------L-----_j

P2-P3 P3-P4 P3-P5 P3-P6 P3-P7


Plane

--Plane-to-plane linear error

Ftgure 95 Stage I- substructure hnear mspect10n results

Stage I - substructure perpendtculanty mspection results

~004,-------------------------------------,
E
E
~
~

g 0 02
"
E
0

""
P3-P8 P3- P9 P3- P10 P3- P11
Plane

---Plane-to-plane perpendtculanty error

Ftgure 96 Stage I- substructure perpendtculanty mspectton results

Stage I- hole diametric error on panel! and substructure


The followmg results show hole dtametnc error for panel I on holes I, 6, I I and 16, whtch
°
represent the condtl!on at each corner of the panel All temperatures are C and all measurements
are In mm

Temp_ Hole I Hole6 Hole 11 Hole 16


-30 6095 6095 6095 6095
0 6095 6095 6095 6095
20 6095 6095 6095 6095
40 6 095 6095 6095 6095
60 6095 6095 6095 6095
80 6095 6095 6095 6095
Ftgure 97 Stage I -panel I hole dtameter error at dtfferent temperature

165
It can be seen from the above table that hole diametnc condii!on remained stable at all
temperatures Due to the coefficient of expansion of CFC bemg relatively low
(0 004mrnlmetref'C) these results were of no surpnse

Hole diametric error on substructure


The followmg results show hole diametnc error for panel I on holes I, 6, 11 and 16. . All
temperatures are ° C and all measurements are m mm.

Terno. Holel Hole6 Hole 11 Hole 16


-30 6 015 6 015 6 015 6 015
0 6 015 6 015 6 015 6 015
20 6 015 6 015 6 015 6 015
40 6 015 6 015 6 015 6 015
60 6 015 6 015 6 015 6 015
80 6 023 6 015 6023 6023
Ftgure 98 Stage I- substructure hole dtameter error at dtfferent temperature

As can be seen from the table above, the substructure holes demonstrated good stability through a
range of temperatures Due to the rate of coefficient of expanswn of the alummmm
(0 026mrn!metref'C) It was expected that the hole diametnc condii!On would mcrease Some
evidence of the hole diameter mcreasing was evident at 80° C but this was only a small mcrease
nght at the end of the upper temperature range.

Assembly performance of pan ell to substructure


The followmg results are a measure of panel I to substructure assembly performance

Stage I -panel to substructure pin assembly condition

u~
80

"~
bJ)
60

" 40
:::!-
a"
M
20
~
"sP. 0
"
1-< -30

0% 20% 40% 60% 80% 100%


%Assembly fit

I • Ptn gauge fit I


Ftgure 99 Stage I - panel to substructure pm assembly condttton at dtfferent temp

166
From the results it can be observed that an even d1stnbutwn of assembly fit has occurred, w1th
perhaps one exceplion The assembly performance at 0 degrees C IS somewhat better than may be
anl!c1pated by v1ewmg the other results A poss1ble cause for th1s may be a non lmear reactiOn to
the way the matenals expand and contract

8.16 VSA tolerance analysis results

The obJective of the analys1s was to pred1ct the probable d1stnbutwn and range of vanatwn for
step, gap and vutual hole cond11ion between the stage I substructure and panel assembly.

The VSA software was used to s1mulate the manufacture and assembly of the substructure and
panel assembly and analyse the vanation that could occur. The VSA software IS capable of

• The three d1menswnal geometry of the component parts and assembly toolmg.
• The sequence m wh1ch the parts are assembled and the means used to locate pos11ion and
secure the mating components
• The three d1menswnal vanatwn m the geometry of the components or fixtures and the
vanation m the locatiOn of matmg components
• The probability of d1menswnal vanatwn occurnng m the component parts and fixtures
• The overall combmed effect upon the final assembly due to the vanatwn m each separate
component/fixture and the vanatwn that occurs between matmg components.

The SJmulatwns will prov1de the stal!stJcal data to pred1ct the probable range, distnbul!on and root
causes of the vanalion m the local assembly

8.16.1 Analysis objectives

The obJecl!ve of the study was to create a dimensional model of the stage I substructure and panel
assembly. The model was used to determme the stalislical d1stnbutwn and range of vanatwn m
output, measured m terms of the followmg parameters:

• Vanal!on m step and gap between substructure and panel assembly


• Virtual hole to hole m1sahgnment due to manufactunng vanation.
• Compare the above to actual expenmental measured data.
• Rev1ew results and produce conclusiOns and d1scusswn.

167
8.16.2 Modelling assumption

Included w1thm the VSA model are the followmg basic assumptiOns:

• All parts will be considered ng1d


• All component tolerances will be represented by normal d1stnbutions with ±3cr range
centred on design nommal (I OCpk capab1hty)
• The components will be modelled at ambient temperature (the model will not consider
thermal effects)
• Dunng assembly SimulatiOn, part posii!ons will be allowed to float freely w1thm available
assembly clearances.

8.16.3 Assembly simulation

The VSA model was used to perform the followmg assembly sJmulatwns

• Nommal Simulation to confirm Imllal geometry


• Monte Carlo SimulatiOn to prediCt the mean and range for each measurement.
• HLM SimulatiOn to Identify and rank the leadmg contnbutmg tolerances to each
measurement

8.16.4 Component datum and tolerance

Figure 100 Substructure datum and tolerance

168
Figure 101 Panel datum and tolerance

I
I
l.
Figure 102 Pm datum and tolerance

8.16.5 Analysis measurements

The followmg areas were 1denhfied for mvestiga!ion. These mcluded step, gap, and Virtual hole
condJ!ion measurements The hm1ts for each of the measurement types were defined as·

• Steps +/-0.1 00
• Gaps +/-0 150
• Virtual holes +/-0.100

169
...

..,

Ftgure 103 Step, gap, and vtrtual hole condtttons to be measured

8.16.6 Analysis results

A number of output measurements were taken to established vanal!on levels for cnttcal features
assoctated wtth the stage I assembly mterchangeabthty Thts was established through a number of
measurements for step and gap around about the edges of the assembly. A sample ofvutual hole
condtl!ons were taken between the two common holes m the panel and substructure A sample of
the analysis results are presented below The full set of results from thts analysts ts avatlable m
the appendtces of thts thests.

--· -~
...... '"'"
"'"'-"'"""""'O.OC(l)
lilHV....,.collOO'll

~-
1 ..,..,..,li_PANti..)Uf'
'"""'JollliMI:I'<I
-.;,o,.
WO
-- -m
·= •·=
FW.S£1 SliiiST~UC'TliRf WO
~ IOn 0.(11~ ...._ 0(11~ I PO$ IDooO~II'IIAI~II

li'W.S£1 SLIBSTRUCTlJR[ ~~
·~
- - . I •on
>ISI'f iOO«<iAIB(Io!)IQI.!II

PIOOSEI_PAm:L_fllP--
., 11<~.00011 M"'OOOO IPIJS ID,.IlMI"liAIBP-111 I ~

f>\A.S[I_SUDSTRUCTUAE '"""""""'""""
., Wn 0(115 .... 0(115. lf'ER IOoollUI'liAI '" '" <D
PI\ASEl_IWIEL_I'lJP ...... ~
·~ ·=
>lll'f IOOWIA1a"'IIC11'11
F'>ll5fl5liBSIRUClliRI
>OAr I'.BIMJ:l><1 - - '~

·~

.... .._ ........


...... ~.
'l.<).fOIS,..C
-·=·=·- -·- -- -
·= ~041<

·~
•••.,s.g,..o_
-·-
l.ui.~•)( 1114-'>

Ftgure 104. Gap condttton at location plane 4 substructure

170
·-
..
-- -- _,,.,
-....
-'1-UOOJJ

....
lli.Wv...,..,. aoou
= ~~- H~loloo
\ f>w,(l PAH(l_fl.l' -~ o= ·~
>I Pf IU<WiAI8il<ll~ll

PI\A.S£1 SUBSTRUCTUII£ - -
>If< I 10<001 "~
,_
• 1 PHASEU'AN£Ln.P _,.....,
>IFliiO<OOI
,_ '~'
,_ ·~
!'H.>.S£1_Si.$SIRUCIURE """''
> ISPF IOO<OIAII»'IICI"II '""' •n

,. ...., .....
'J.<LCNtLW

'II.O..OIS<>oc
- -
01915
121&1
04<"1
~J--Wno:IO'O.
C&O:'e
11931
03<911

F1gure 105 Step cond1tton at Iocat1on plane 4 substructure

Ft<ASI'I_F't"'<H_fUP Idol -·~


> ""'009(lM..._009l ii'OSICI>o002Sflo!IIA19i'ti1QMII "' "'
l'tti>S£1-lfl.IP
>DOll' l'olti>OIQ... "' ~
I JO!jl);

l PHASEI_\.U9S1RUCIUR( ""'l ~
> ,.., OUI5 .,.,. 0015 IRl!.IO.OO!P!>(IoljiAIRflo!IIQiolll "" om
f'H.-S[I_F'»H£l_fLIP ......., _
> M"'-OlPlllMMOim II'OSiO .. OM(IoiJIAI8(1o!ll "' ""
l't<AS£1 SU9'S1RliCIUfl( - . -
> Hn 0015 M-. Ol~~ IHJ~ i0..00l5 ... 11AI~II "" ""
PH.>.SEI_I.Uf<IRJ(!l'R( ,.....,.,. _ _
> ""' 0015 ""'HI\ !PER ot.o0CI2SIAI "" ""

- - -
OHi>
~JS..A0<90
·~
01101
997300'10.

F1gure 106 Hole 2 VIrtual hole condt!ton

8.17 Experiment discussion

Manufacture of panel and substructure


The machmmg of net routed edge and dnlhng of holes m each panel produced some problems
Panel 2, the first panel, was machmed usmg an mappropnate feed rate of I 5m/mm for the
maximum rpm of 8,000 available on the Wadkm machmmg centre This led to panel2 sustammg
lammate damage about the edges Although It has been concluded that the excess1ve feed rate
was responsible for the edge damage, the CNC program was wntten such that the router was
directed to change duectwn of the cuttmg tool whilst Withm the panel. This may have also

171
contnbuted to panel 2s edge damage. The hole dnlling operatiOns for panel 2 went without
mc1dent apart from the expected burst through lammatwn damage local to the hole edges

Panel I, the second panel, was machmed usmg a different technique The panel was sandwiched
between two layers of PVC plastic The entire sandwich was then mounted onto the tooling plate
in the same manner as for panel 2 but with an additiOnal plate. The five entities were bolted
together m order to apply pressure to the CFC panel In addition, when the router made Its pass
along the edge It cut the sandwich of PVC plastic as well as the panel which probably helped stop
the lammate damage that occurred m panel 2 Other aspects of the cuttmg process also probably
helped reduce lammate damage These were the reductiOn of machmmg feed rate to 0 4m/mm
and d1rectmg the cutter path clear of the panel after each cut.

After an 1mtial face milling operatwn, the substructure was machmed m one operation This
occurred without mc1dent

Panel thinness variation


The planar geometry m stage I panel thickness vanatwn has had httle or no effected on hole to
hole assembly fit or panel/substructure gap condition However, the step condition has been
affected Panel width vanatwn will have a bigger effect on cyhndncal or complex panel
geometry lt will probably cause hole to hole misalignment, out of specd1catwn step conditions
and may even affect gap condition The result ofth1s will be a no fit condition for the assembly of
panel to substructure.

Manufacturing process observations


Assemblies that rely on component part specificatiOn and not toolmg 1 e !at bmld, the process
capability relatmg to the manufacture of each component will have a big effect on the resultmg
assembly performance. The manufactunng processes associated with carbon fibre composites
(CFC) are still m their mfancy Anomalies such as hole shnnkage after dnlhng and net edge
dimensions (after a machmmg operatiOn) appear to be to be left oversize are not well understood
although empmcal data IS available From the expenmental results the CFC manufactunng
processes effect

• Panel hole diameter and form, hole diameters suffer form shnnk back. Hole form can
also be effected by burst through causmg mternal hole profile damage.
• Panel hole to hole position; p1tch error recorded on the panels were very good Provided
the panel IS placed m the fixture correctly and machme accuracy IS high, hole to hole
pitch error IS not considered to be a problem

172
• Panel planar and surface form, the panels m Stage I demonstrated three types of form
error The first on the panel top side (plane I) which IS accumulated form the mould tool
surface The second on the bottom side of the panel (plane 6) which IS accumulated from
the mould tool surface and resultant lay-up process parameters The third form error,
found on the edges of the panel (planes 2, 3, 4 and 5) Imlially IS the result of the cork dam
lay-up geometry. But the final form error is attnbuted to the routmg process capability.
• Panel surface wavmess of form, this error IS generated dunng the CFC lay-up and cunng
processes The crossmg fibre bundles cause mlema! stress and stram in local areas of the
panel Ieadmg to waves appearing m the surface
• Panel edge form, lmlially from cork dam profile, but after routmg net edge relates to
routmg/machmmg capab1hty

Panel wavmess detected on both the panels form will have had two fundamental effects. The first
is to cause the panel to sit up on Its high pomts (lower peak of wavelength) contnbuting to greater
VIrtual thickness. The second is to change the virtual onentat10n of the panel m relatiOn to
substructure onentatwn Wavmess error is effeclively a subset of form error

A range of form errors were detected dunng the mspect1on Form error can be descnbed as the
deviatiOn of a geometnc surface, planar or other, form a defined datum. Contamed withm form
error will be a wavmess error which can be descnbed as a form error that displays amphtude and
wavelength From the surface measurements taken at Taylor Hobson 11 can be seen that waviness
error can themselves contam harmomc error. Harmomc error can be descnbed as a cyclic error
possessmg amphtude and wavelength which IS situated about wavmess error Fmally, harmomc
error Will conlam surface roughness error

Thickness vanat10n on the Stage I planar surface has httle effect on the assembly of panels to the
substructure. It has not caused m1sahgrunent of hole to hole axial pitch or gap conditions error,
but has affected the step conditiOn of the assembly

From the expenmentalion 11 was Idenlified that features with larger volumes tended to exh1b1t
higher or Similar geometnc error m companson to smaller volume features This IS despite the
fact that the larger volume features were produced by a more accurate manufacturing process

The manufacture of the substructure went without mc1dent and a post machming inspectiOn
md1cated no unpredictable conditiOns.

173
VALISYS inspection results
Pane12
It can be seen form the results that all four gap edges on the panel faded the mspectwn. The
perpendtcularly tolerance of 0 02 to datum A was to ttght for the manufactunng process used to
router the edges Datum C also fatled on posthon accuracy, but passed on diameter The 18 hole
pattern also faded on posttion accuracy but the report does not speedy whtch hole or holes were
responstble All other features measured m the process passed

Investigation ofpanel waviness and its effect on hole alignment on 2D panel and substructure
geometry
Excesstve panel wavmess on a 2D form wtll have two fundamental effects The first is to cause
the panel to stt up on tts htgh pomts makmg tls overall virtual thtckness greater than that detected
through dtscrete measurement The second is to change the vtrtual onentahon of the panel.
These effects upon mterchangeabthty specificatiOns are to be mvesttgated on 2D and 3D complex
surface parts.

Explore CFC panel geometry error taxonomy


The vanatwn of form tolerance can be reduced to several sub sets, these being

• Form error
• Wavmess error {eye he)
• Surface roughness {cychc)

Investigate panel thickness vanatwn and zts effects on a 2D panel and structure geometry
Thtckness vanatwn on a 2D and 3D complex surface wtll cause the mtsahgnment of hole to hole
axtal pttch In extreme, tt wtll affect both step and gap condttwn wtth respect to panel to
substructure fit.

Explore CFC manufacturing process


Anomahes such as hole shnnkage after dnlhng and net edge dtmensions {after a machmmg
operatton) appear to be to be left overstze From the expenmental results hole shnnk appears to
be conststent and therefore predtctable. Future work should mclude the development of a process
model denved from empmcal test data on aspects such as:

• Hole dtameter - dnllmg shnnk back.


• Panel edge profile - feeds and speeds for routmg.
• Panel lay-up -lay-up tool accuracy {tolerance chams)

174
• Panel cure - cure tool distortiOn

Investigate feature volume to conformance relationship


From the expenmentat10n It was Identified that features With larger volumes tended to exlub1t
higher or similar geometnc error m companson to smaller volume features Tins IS despite the
fact that the larger volume features were produced by a more accurate manufactunng process

Discussion drawn from the VSA simulation results


The analysis predicted the attnbutes of step, flush, and virtual hole conditiOns based on panel and
sub structure tolerance allocation and defined assembly process The modelling process was
essentially the same as for the GVA studies descnbed m chapter 7 The results predicted that m
the mam the panel to substructure assembly philosophy would be capable of achievmg the gap
specificatiOn and VIrtual hole conditiOn However, the model prediCted that up to 15% of the step
conditiOn would fall out of the specified limits

8.18 Concluding remarks


A number of practical expenments designed to explore the design, manufacture, assembly, and
mspectwn of composite panels to a metallic lightweight fuselage have been presented m this
chapter.

There are a number of over all conclusiOns that can be drawn from this chapter, these are

• The effect of different operatmg temperatures on components made from matenals With
different thermal properties, for example, alummmm and carbon fibre composite IS
s1gmficant. This was particularly noticeable m the expenmentation followmg the
assembly of a CFC panel to alummmm substructure when the environment temperature
approached the equivalent of aircraft extreme operatiOn limits,
• The manufacture of matenals such as CFC IS complex and lime consummg and demands
an advanced manufactunng process. In producmg the CFC panels some damage was
sustamed form the net routmg and the hole machmmg operations. The net routmg
damage was caused by mcorrect CNC machme tool spmdle speed m conJunction to travel
movement. This was unavOidable due to the available CNC machme tools not haVIng the
reqmred capacity to accommodate CFC machmmg operations Damage sustamed on the
hole as 'burst through' was caused by not usmg special dnllmg tools appropnate for CFC
due to unavailability No problems were encountered with the manufacture of the
metallic mould tooling plate or alummmm substructure components.

175
• Increasmg geometnc shape complexity from planar to cyhndncal form affects their
related overall assembly charactensl!cs. The measurement results of the planar and
cyhndncal panel and substructure assemblies after companson indicate that a higher
magnitude of dev1atwn IS present m the cyhndncal geometry. Measurements such as hole
posiiiOn, panel thickness were of a were higher resultmg in greater assembly no
conformance
• The GVA tool was used to model the planar components and predict their assembly
performance The key features 1denllfied for analysis were the step and gap condiiions
between the panel and the substructure mterface
o Gap conditiOn; the vanatwn range for the gap condiiion between the panel and
correspondmg substructure edge feature predJcl!on by the GVA tool correlated
well With the actual measured condiiion The measurements also detected some
panel distortiOn probably caused by waviness. The measured gap variatiOn can be
attnbuted to the machmed panel havmg a course surface fimsh caused by
machmmg process restnctwns Furthermore, the CFC panel consistently
machmed undersize (CFC charactensl!c} whiCh accounts for a small mean shift m
the measured data when compared to the GVA esl!mated data
o Step conditiOn, the vanal!on levels for the step condiiion between the panel top
surface and correspondmg substructure feature as predicted by the GVA tool was
higher by a magmtude of 2 compared to the actual measured data This can be
accounted for by the tolerance specifical!on given to the GVA analysis as bemg
high by about 50%
o Vutual hole conditiOn, the vanal!on levels for virtual hole condiiion predicted
from the GV A tool md1cated that available hole size for the pm would be reduced
by approximately 0 I mm per hole from the nommal 6.00mm. Given the precise
tolerance on the dowel pm, which measured at 6 009mm with vu1ually no range,
the GVA analysis tool essenl!ally concluded that there would be an interference
fit on a large number of holes. This, however, was not seen dunng the assembly
of the expenmental parts The reason for this is because the CFC holes were
dnlled usmg a 6 !mm dnll to allow for CFC shnnk back (CFC charactenstic)
The holes dnlled resulted on average to be 6 08mm resultmg m a predictable
0 02mm shnnk back. The resultant larger hole diameter allowed for more float
allowmg the fit up of all 20 pms between the panel and substructure iniiial
dnllmg of the CFC panel With a 6 OOmm dnll resulted m a hole of 5 08mm
dmmeter causmg an interference fit up between panel and substructure.

176
Chapter 9

9 Discussion

Objective: Thts chapter identtfies the mam issues for dtscussion and sets the foundatton for the
concludmg remarks presented m the next chapter. The dtscusston is based around the atms of thts
research project whtch includes the DM dtsciplme, tts principles, the assembly expenmentatton,
and the proposed DM methodology

9.1 Review of dimensional management

The domam of DM has been revtewed m chapters 2 and 3 from two perspecttves, these bemg a
hterature revtew of all pubhc domam matenal and a comprehenstve mdustnal revtew generated
from a study tour of leadmg automotive and aerospace compames and the author's commerctal
knowledge

The DM dtsctpltne has developed out of the automottve mdustry whtch has been stnvmg to
tmprove product qualtty whtle addressmg key generic busmess goals such as vehtcle and process
umt cost reductton, development cost reductton, and shorter ttme to launch schedules. It was first
to recogmse the potenttal commerctal advantages resultmg from DM practtse There ts evtdence
of the DM actlvtty m automottve smce the mtd 70's m the USA Many of the mfluenttal
manufactunng methodologtes currently used m the aerospace sector have m the first mstance been
developed and matured m the automottve mdustry As wtth other imttattves, the aerospace
mdustry has the opportumty to draw upon the expenence of the automottve sector wtth respect to
DM techmques and practtces. The aerospace sector could benefit from a potential technology
transfer from automotive wtth respect to DM practice

The hterature and mdustnal review highltght that aerospace manufactunng compames currently
face the btggest commerctal and techntcal challenges to date. More recently the aerospace sector
has revtsed the approach to development and productwn m response to aggresstve customer
dnven changes such as purchase cost reductton, cost of ownershtp reductton, tmproved operatton
performance for both dynamtc and mamtenance charactenstics. The aerospace mdustry
tradttionally has been assoctated 'craft' based product development producing one-off products
each bemg umque to the next. Thts has mfluenced the type of manufactunng methodology used

177
tradthonally m the mdustry The mtroduction of new manufactunng methodologies such as CE,
SE, and lean productiOn mto the aerospace sector has generated considerable focus on
development lead-hme reductiOn, produchon efficiency, waste reductiOn, and htgher quahty. In
dtrect support of these mihahves, DM teclmiques have been mtroduced mto the development and
productiOn methodology to better define and control key areas of engmeenng to accommodate
larger production runs where Issues such as modulanty, mterchangeab!ltty, and more efficient
assembly procedures ex1st

The aerospace mdustry IS begmnmg to recognise the potenhal benefits of applymg OM teclmtques
to vahdate aircraft design for manufactunng and assembly vanat10n robustness OM teclmtques
m the aerospace sector were tmhally used for aircraft structure (fuselage, wmgs, surface controls,
etc ) vahdat10n. These teclmtques are now extendmg, for example, to propulsiOn umts (Jet and
piston hydnde engmes), drive umts, and engme nacelle assembly, and undercamage assembly.
The area ofDM analysts IS still expandmg through research and commercial partnerships. Further
areas mclude inclusiOn of FEA analysts to model effects of process dtstortwn, mclusion of real
SPC data on tolerance analysis models, creation of full closed loop analysis

In support of the mdustnal revtew a study tour of selected aerospace and automohve companies
w1thm the USA was planned and undertaken. Each of the compames VISited were beheved to be
practitioners of DM Other compames beheved to be practii!oners of OM Within Europe were
also VISited In general, the maJor automohve compames VISited m the USA had been usmg DM
techmques and supportmg software tools for the longest penod of time compared to the USA
aerospace sector Furthermore, the automotive sector has been practising DM for longer penods
than their European counterparts For each of the automotive and aerospace compames visited a
set of DM case examples were Identified. The range of case examples for the automotive sector
suggested that the maJonty of DM actiVIty has been targeted at BIW quahty vahdahon. Wtth
respect to the aerospace sector the case examples indicated the maJonty of OM activity was on
atrframe structures

The key areas tdenhfied m the literature and mdustnal review have been presented. The
discussion now addressed the need for DM m support of aircraft development and production

9.2 The need and nature of dimensional management

The findmg of the review forms the basis for the need outhned m chapter 4 for Improved OM
achv1ty w1th a1rframe development and productiOn.

178
The next generatiOn of both m1htary and commerc1al a1rcraft will be des1gned usmg novel
concepts, new matenals, and reVJsed manufactunng and assembly strategies. It IS only through
these s1gmficant step changes m technology can today's challenging cost and weight targets of the
a1rcraft customers (both military and commercial) be fully met To be successful, aerospace
manufactunng organisatiOns need to fully understand the available technology and 1t's supportmg
tools and techn1ques by managmg the whole production des1gn to assembly cycle as a single
mtegrated process To achieve th1s 1! is 1mpera!Jve that the des1gn specificatiOn takes full account
of the down stream effects of hardware vanab1hty on the manufactunng, assembly, and mspechon
ac!ivJ!ies and their effect on ach1evmg aircraft customer and domes!Jc specJficahon Conversely,
the manufactunng and assembly busmess centres must capture all available process knowledge
and make this available back upstream to des1gn.

Aircraft manufactunng orgamsations currently fml to take the full advantage of tools, techniques
and methodologies for the management of product and process geometnc vanatwn Aerospace
manufactunng orgamsalions have the opportumty to mcrease engmeenng effic1ency m future
commercml proJects through the adoptiOn of a number of mature best prac!Jce techmques such as
CE, SE, and lean produclion The DM d!SC!phne embedded mto an IPD process would extend
and compliment the capab1hty of these Jm!ialives.

The reqmrements outlmed above could be met through the des1gn and development of a spec1fic
DM methodology Such a methodology IS has been proposed and IS discussed m the followmg
sectiOn.

9.3 PARES; proposed methodology

The proposed PARES methodology 1s presented m chapter 5 and has been proposed to address the
needs 1denhfied m chapter 4, section 4 3 The mam mm of PARES 1s promote an integrated
process for the Jdenlificalion, analys1s and management of complex product vanahon throughout
all the phases of a1rcraft structure development and productiOn

The PARES methodology structure has been developed against a genenc IPD process for a1rcraft
production prov1dmg a hmeline reference to spec1fic achvJ!ies and the1r synchromsatwn to the
IPD process backbone It mcludes a number of concurrent actJVJtJes des1gned to mtegrate the
overall DM process reqmred for product hardware vanab1lity controL The PARES process
mvolves a number of achvJ!Jes that need to be synchronised w1th an IPD backbone.

179
Aerospace compames m the USA and Europe have invested significantly m new and advanced
software tools such as CAD, CAM, and CAE, to gam competitive advantage Leadmg
engmeenng software development companies such as Dassault Systemes and UGS are currently
respondmg to this demand by developmg ever more integrated products under the banner of
product IIfecycle management (PLM) These PLM software solutions attempt to address the
needs of DM requirements through the development of specific software modules to support
actiVIties such as tolerance designation, tolerance analysis and synthesis, quality (measurement)
data analysis, key feature IdentificatiOn These modules however are deployed by manufactunng
companies m an ad hock manner With little consideratiOn to an end-to-end DM strategy.

The PARES methodology could become the dnver for a more structured approach to the
deployment of DM assocmted software tools and techniques. A software centnc architecture has
been developed to support this approach and IS made up of SIX mam chunks, each of which would
be supported by a specific software module The activities flow for DM has been developed to
position how the PARES methodology could be supported by current modules of commercml
software tools This IS presented m chapter 5

PARES IS a proposed methodology and therefore has not been fully evaluated. The development
of the PVA tool and the GVA case studies have, m part, evaluated some of the key actiVIties To
fully evaluate PARES the remainmg activities would have to be deployed and evaluated with m a
sUitable aerospace orgamsatwn. Deploymg PARES will reqmre a well defined Implementation
plan designed to challengmg the established process The deployment will need to consider the
followmg actiOns

• Establish a deployment team made up from appropnate members of the orgamsatwn


with the reqmred skills and mtemal position This should mclude an executive sponsor,
a project manager, and the mdlVldua] team members.
• Create a direct link of accountability to the orgamsatwn 's management
• Create extensive project plans With actiOns and timeiines
• Create a set of PARES related process performance metncs that the can be used to
measure where the orgamsatwn was at the beg1nnmg, and It progress path during and
after the ImplementatiOn

The proposed PARES methodology Identifies the need for more up front assembly analysis In
order to address this reqUirement a PVA tool was developed.

180
9.4 Development of a PVA assembly analysis tool

The need for a PVA tool to support mrcraft wmg box structure concept design pnor to detail 3D
geometry bemg avmlable was Identified by the mdustnal review There IS an opportumty to
mvestigate assembly process pnor to significant levels of CAD geometry being available After
this stage, design concepts are considered mature and the design orgamsatiOn would be reluctant
to make significant feature and geometry change based on any analysis undertaken at thiS stage.
The PVA tool can be used to review product architecture (parameter based) for baseline
functional assembly concepts m ID and I 5D to validate early target attnbutes

The PVA tool was designed and developed based on an Excel spreadsheet application but also
mcorporatmg the statistical analysis tool Crystal Ball Pro from DecisiOneenng. The mitial
development consisted of 1dentifymg the type of genenc wmg box components to be mcluded and
definmg their assembly sequence and process

W1thm the analysis model all s1gmficant ID tolerance chams were Identified based on component
features and assembly process Analysis tables were then developed w1thm the excel spreadsheet
to represent these features and tolerance chams The analysis tables were defined to mput
mformatwn reqmred to perform the analysis mcludmg parameter feature nommal distances,
applied tolerances, and d1stnbut10n type.

To perform statistical analysis, the crystal ball pro software was added to the application This
allowed the statistical simulation of parameters defined m the analysis dnver tables mcludmg
Monte Carlo and Latm Hyper Cube. Furthermore, root cause analysis could be performed
through rank correlation based on the mdiv1dual parameter chams The PV A tool IS capable of
producmg WC, RSS, and statistical predictions of measurements These statistical predictiOns
can be output mto a formatted report for pnntmg. Analysis sessiOns can be saved and recalled

PVA works well as a ID front loaded analysis tool It IS capable of giVIng a strong md1cation of
how capable an assembly press will be in terms of dehvenng target attnbutes, and where these are
not achievable, what elements are the mam contnbutors to vanatwn The mam limitation of the
PVA tool Is Its mab1hty to take mto account the effect of full 3D and geometnc vanation.

The opportumties to develop the functiOnality w1thm this type of simple parametnc tool are
considerable Some of these mclude temperate compensation, development of I 5D and 2D study
parameters, radial parameter analysiS, hm1ts and fits specificatiOn Other areas are identified m
the further work section of chapter I 0

181
When early PVA analysts has been completed 1t paves the way for more detatled GVA analysts
based on avatlable 3D CAD geometry.

9.5 Case study for GVA assembly analysis tool

Three case studtes were tdentified m conJunchon with Atrbus to mveshgate the use of a
commercial GVA tool called VSA. The case analysis was based on a set of wmg box
configuratiOns bemg used m support of the A3XX aircraft development program The main aims
were identified and agreed between the author and key Atrbus personnel. These were to predtct
the capabthty of a specified assembly operation mvolvmg a JOtmng process for mechamcal
fastener, paste bondmg, and film bonding techmques Pnnctple area for consideratiOn was the nb
to skm, and C spar to skm mterface

In conjunchon wtth the case studies, an analysts process was developed to dnve the GVA
software tool The process was made up of eleven steps and these have been defined m chapter 7.
Thts type of analysts normally requtres a stgmficant level of 3D CAD geometry to be avatlable
However, some CAD geometry was not available from Airbus. Thts resulted in the author havmg
to destgn many supportmg parts and fixture elements usmg the Dassault Systemes CATIA V4
CAD software The tolerance analysts was performed usmg the CAT3D-VSA software. Some
modelhng problems were encountered and all bugs and functwnaltssues were documented by the
user Thts hst was sent to the VSA developers m the USA for revtew

The analysts produced a large amount of predtcted measurement data based Thts data has been
sorted and collated mto a number of analysts proJect files From the results a number of
conclusiOns were draw The conclusiOns and analysts were data presented to back to Atrbus for
then evaluatiOn. The conclusiOns drawn were very specific and mdtcated predictiOns of assembly
capabthty based on the dtfferent part zones and thetr assembly usmg mechamcal fastener, paste
bondmg, and film bondmg process techniques The actual model data and VSA results are treated
as confidenhal by the author and cannot be released wtthout permtsswn from Airbus.

The GVA studtes provtded comprehensive analysts data on the dtfferent case assemblies based on
mechamcal fastener, paste and film bondmg feature interfaces These were based on destgu
tolerance and vanal!on metncs Further analysts can been undertaken by replacmg the destgn
tolerance values wtth actual measurement data and rerunmng the analysts model In addttton, the

182
use of FEA mtegratwn analysis to model component or tooling deflectwn may be undertaken to
support ex1stmg data models m VSA.

The GVA case analysis invesllgated a metallic assembly process to evaluate Its potenllal for use
in CFC assembly. Issues relatmg to the manufacture and assembly of alummmm and CFC
structures were mvesllgated as part of an expenment These are discussed m the followmg
secllon.

9.6 CFC panel to Aluminium substructure experiment

Advanced a1rframe design mcorporates a mixture of matenals mcludmg CFC and alummmm.
The mm of this expenment was to mvesllgate the potenllal Issues ansmg form the design,
manufacture, and assembly of CFC panels to alummium substructures and their effect on product
key attnbutes such as step and gap This was conducted through an expenment based on planar
and cylmdncal geometry configuratiOns

At the design stage all panel and substructure geometry were generated WIIhm the Unigraph1cs
CAD system usmg solid geometry functiOn From the panel solid models the accompanymg
mouldmg tool plates were designed agam m solid geometry These solid models then formed the
basis to produce the off-line CNC machmmg programs, the off-line CMM measurement programs,
and were at all limes used as the product master for all geometnc and d1menswnal reference.

The off-line CNC program was generated by the Umgraph1cs CAM module based on the sohd
geometry All cutter paths and tool selectwn were generated and then validated m the Vencut
simulatiOn software This software validates the process for tool I fixture I part clash detectiOn,
path pattern venficatwn, and resultant part net shape mtegnty. The same process was used to
generate the post process programs for the CFC mouldmg tools, the alummmm sub structures, and
the net machmmg of the CFC panels.

The machmmg routmg for net CFC panel edge and hole dnlling produced a number of
manufactunng problems The first problem resulted m all panels sustammg lammate damage
about the edges during the CNC machming operatiOn The second problem resulted in hole
damage resultmg from a condition called 'burst through' which caused lammatwn damage local to
the hole edges. Both these conditions were expected as they are well documented as
manufactunng Issues for CFC matenals The manufacture of the sub structure and CFC toolmg
occurred with no real problems

183
Followmg the manufacture ofthe md1vidual components, they were mspected for accuracy. The
mspect10ns of the CFC mouldmg tools ind1cated that the form ex1sted on both tools Th1s was
partlcularly ev1dent on the cylindrical (phase Il) mould tool and wh1ch displayed charactenst1c of
good form error but haVIng a poor s1ze d1mension.

The panel and substructure for stage I and stage 11 were mvestlgated for dJmensJOnal accuracy
usmg a CMM and advanced mspectlon software A number of tests were also performed to
evaluate the assembly capability of panel and substructure at nominal temperate Further analys1s
was undertaken to rev1ew assembly capability at different temperatures based on ranges typ1cally
demanded by m1litary and commercial mrcraft.

The VSA vanat10n analys1s tool was used to model the panel to substructure assembly process.
The analys1s mm was to pred1ct the attnbutes of step, flush, and Virtual hole conditiOns based on
panel and sub structure tolerance allocatiOn and defined assembly process The effects of
manufactunng and environmental (temperature) vanatlon on functiOnal assembly capability, for
example, Interchangeability of panel to substructure, were mvestlgated Due to the different
thermal propert1es of alumimum and CFC, dunng extreme temperatures, cond1t10ns of non
assembly were detected

The mam discussion for the pnnc1ple areas of th1s research have been presented above These
areas now form the bas1s of the conclusiOns presented m the next chapter.

184
Chapter 10

10 Conclusions and further work

Objective: This chapter presents a hst of conclusiOns drawn from the material presented m this
theSIS and idenllfies the opportuml!es for further work.

10.1 Conclusions

With reference to the a1ms and obJecl!ves outhned m chapter 1, an overview of the conclusions
are·

Need and nature: The hterature and mdustnal review provides the background and the
pos1tiomng of the subject area A tour of leadmg aerospace and automotive compames m the
USA and Europe has charactensed the need for the DM diSCipline as part of an mtegrated product
development process One clear message IS that DM needs to permeate through the whole
development process mcludmg aspects of design, development, manufacture, and assembly of
aJrcraft structures, covenng both m1htary and commercial sectors.

Methodology: An outline DM methodology has been presented to address the need for a Wider
scope of actlVlty m support of robust engmeenng at all stages of product development The
methodology outlmes the core acllvity at each stage of product hfecycle, mcludmg the need for
front load effort early m the process, and its Impact on the extended manufactunng enterpnse

Variation analysis tool: The need for a baselme vanatwn analysis tool to be used at the early
design stage has been Identified The design and development of such a tool has been undertaken
and Its scope, benefits, and hmitatwns have been h1ghhghted. The baselme tool was validated,
and used m support of a proJect at Airbus to mvesl!gate the capab1hty and smtab1hty of a metalhc
wmg assembly toolmg spec1fical!on for use With carbon fibre wing assembly.

Experimentation: This research work has 1denl!fied a number of potenl!al designs for
manufacture and assembly opportuml!es m both the commercial and military sectors for future
programs of aircraft manufacture These opportuml!es rely primanly on evaluatmg the assembly

185
fit and funchon capability of carbon fibre composite (CFC) pnmary structures to lightweight
alumimum (AI) sub structures takmg mto account product specification on interchangeability.

The conclusiOns outlmed above are expanded m the followmg sechons

10.1.1 The need and nature of dimensional management

The discipline of DM was reviewed to determine the best approach m whiCh to address the DM of
aircraft structures termed here as 'need and nature'. The review mcluded undertaking a literature
search, the planmng and undertakmg of mdustnal review of key compames, and 1denttfymg key
commercial mformation known to the author. The mam findmgs of the review are surmnansed as.

• The literature and mdustnal review revealed a number of pressures currently faced by the
aerospace mdustry underpmnmg the requirement for better product vanat10n
management These pressures can be broadly grouped as
o Lower cost of aircraft ownership, requires Improved flight efficiency which IS
dependent on structure ahgnment, stnngent step and flush specificatiOn to avoid
turbulence
o Higher mamtamab1hty and operational flexibihty, reqmres customer level ICY.
o Safety, reqmres new matenal test for structural integnty
• A number of engineenng techniques and practices have been 1denttfied by the author as a
means of addressmg the reqmrements outlined above, these are
o Develop new, or revise ex1stmg, manufactunng and assembly busmess process
o Adequately and effectively charactense manufactunng and assembly processes
o Reduce manufactunng and assembly Jig/fixture 'hard' toohng reqmrements
o Elimmatwn, as far as possible, of physical MTG and master media.
o Develop digital mspect10n techniques
o Introduce geometnc dimens10nmg and tolerancing concepts and standards
o Early and efficient use of vanat10n analysis CAD/CAE SimulatiOn tools.
• The DM process ongmated m the automottve sector and Its use contmues to grow to
present day DM techniques are gammg wider acceptance across a number of mdustnal
sectors and IS now emerging m the aerospace product development processes
• The mdustrial review Identified the need for early assembly tolerance analysiS to be
performed pnor to CAD availability. This parametnc analysis would be one or two
dimensiOnal and would IdentifY key areas for down stream three dimensiOnal geometric
analyses. Current tolerance analysis work m both the automohve and aerospace sectors IS

186
undertaken too late m the development cycle resultmg in httle or no opportumty to make
destgn change.
• The mdustnal revtew revealed that aerospace manufactunng companies are developmg
advanced matenals for the manufactunng and assembly of aircraft structures. These new
matenals and processes reqmre mvestigation to fully understand their dtmenswnal
capabthty.
• The case and industnal review also htghhghted the need for a DM methodology to be
developed m order to support the IPD process wtthm aerospace. The main attnbutes of
such a methodology can be broadly descnbed as.
o Supports a total vtew for product development and productiOn and underpins the
destgn to manufacture as a single process
o Atd alignment of cnlical product and process activities whtch are traditiOnally
conducted independently.
o Introduce real world product and process vanatwn to be analysed as part of the
dtgttal mock-up activity
o Dnve the defimtwn and selectiOn of particular analysts software tools and
techmques used to model vanatwn
o Ensure that, orgamsalion, product and process capabthty knowledge may be
captured and reused directly to set appropriate product quality standards.

10.1.2 Proposed dimensional management methodology; PARES

The PARES methodology was destgned and developed to support this research work and m
response to the needs outlined m chapter 4 It IS made up from a number of concurrent process
tasks whtch span a maJor segment of an aircraft product development hfecycle The mam
conclusiOns for this sectiOn are

• A proposed methodology for DM of aucraft structure development has been presented m


support of this research work. The methodology IS called phystcal architecture robustness
system (PARES).
• The PARES methodology key acttVItles have been mapped onto a genenc aerospace IPD
process and presented m chapter 5 The key phases of the PARES methodology are.
o InnovatiOn
o Selection
o Spectficatton
o Validation
o ProductiOn.

187
• The PARES methodology ts centred round the need for better mtegratwn of DM ac!tvt!tes
wtthm the full scope of the aerospace structure development process
• Some of the key areas addressed by PARES were.
o EvaluatiOn of key charactens!tcs against planned destgn, manufacture, assembly,
and mspec!ton process
o Maxtmtse the use of DMU analysts techmques for product vanatwn management.
o Capture product and process capabtlity knowledge for use m future product
development
o Develop a methodology for IPD process synchromsatwn to span the whole
product hfecycle.
• Where orgamsatwn, product and process capabihty knowledge data ts not avatlable the
methodology promotes tls mtroductwn providmg a closed loop quahty functwn.
• The top level architectural chunks of PARES have been outlined m chapter 5. These
chunks (hsted below) can be supported by a number of dtfferent software tools and
respecttve techmques These are
0 CADandBoM
0 SPC quahty data
0 Tolerance analysis and synthests
0 Root cause analysts
0 Quahty data vtewmg
0 Data collaboratiOn
• The effecttve destgn, development, and deployment of any DM methodology cannot be
undertaken wtthout first understandmg the extstmg busmess process spectfically
mcludmg software tools and techmques already commttted by an orgamsation and
PARES provtdes the framework that takes thts into account.

10.1.3 PVA tool development for early wing box analysis

The need for a PVA tool to support concept destgn pnor to detatl 3D geometry bemg avatlable
was tdenttfied by the mdustnal revtew. Part of the need stems from the destre to conduct front
loaded assembly analysts before a stgmficant level of CAD geometry is avatlable. By the lime
thts level of detatl IS avatlable many of the mam destgn concepts are constdered mature and
destgn orgamsatwn would be reluctant to make stgmficant geometry change based on any analysts
undertaken.

• The PVA tool func!tonal spectficatton was destgned Jomtly between key personnel at
Airbus and the author. The development of the PVA tool was undertaken by the author.

188
• The PVA tool IS a ID analysis tool developed m MS Excel nsmg the Deciswneenng
Crystal Ball module for stalistical simulalion and root cause analysis.
• The PVA tool was designed and developed to analyse the maJor assembly elements of a
genenc wmg box structure at the early {parametnc) des1gn stage The analysis tool
essenl!ally has a number ofpre-configured sub assembly study optwns wh1ch may be bmlt
up level by level The study output provides a measurement of vanatwn and Jts
contnbutors at 1denl!fied key features for the related assembly. Some of the mam wmg
box parameters were:
o D nose and front spar assembly
o A frame to aft spar assembly
o Rib to front spar and D nose assembly
o Rib to aft spar and A frame assembly
o Front and aft top and bottom skm assembly
o Splice plate assembly
• The base! me findmgs from any PV A study may be fed mto the down stream GVA
acl!vJty
• The development of the PVA tool presented m chapter 6 Jdenlifies some of the
advantages as bemg
o Relalively qmck and Simple tool to use once set up
o Does not reqmre detailed 3D CAD geometry to be avmlable.
o May be used at early design stage 1f genenc assembly data IS available.
o No complex software reqmred Only reqmres a common desk top applicatiOn
such as MS Excel to operate.
o The DeclSloneenng Crystal Ball ProfessiOnal software module may be mtegrated
1f stal!stical and root cause analysis 1s to be performed.
o Can mcorporate actual manufactunng measurement data or avmlable mdustnal
process capability mto the model
• The three GVA wmg box case studies each provided a number of concluswns, some of
wh1ch were
o Skin thickness vanatwn (surface profile) m the maJonty of cases has proven to be
the biggest contributor to out of speclficalion condJI!ons m assemblies simulated.
o Improvements m skm surface (thiCkness) control will sigmficantly Improve
assembly dimensional quality.
o With regard to the assembly fixture and 1ts feature tolerance, a general llghtemng
of specificatJon would have little effect with regard to 1mprovmg the functwnal
quality of the assembly.

189
o There may be an opportumty wtth machme fimshed CFC component features to
Improve theu surface position control wtth reference to thetr datum system Thts
would result m better component and therefore assembly quahty control.

10.1.4 GVA case studies on wing box structure

The PVA studtes were capable of evaluatmg the effect of parameter tolerances on overall
assembly charactensl!cs at the early stages of dcstgn These areas then become the focus of
attentiOn for later GVA analysts when more detailed geometry becomes avatlable Three tolerance
analysts case studtes were undertaken usmg a leadmg 3D GVA software tool called VSA Thts
type of approach forms part of the PARES methodology for the analysts ofassembhes when 3D
CAD data IS available. The case studtes gave a valuable mstght mto the effort and process
reqmred to undertake a full GVA study m 3D The mam conclusiOns from thts work are

• The three case studies were undertaken m conJunchon wtth Atrbus on early A3XX
concept geometry Atrbus tdenhfied the study areas, thetr scope, and the type of analysts
to be undertaken The areas for analysts were:
o Rtb to complex skm assembly
o Rtb to skm assembly
o C spar eqmppmg assembly
• The mam objechve of each case analysts was to predtct how capable an assembly system
may be based on the type of fixture techmque, the assembly sequence, the mdlVldual part
tolerance scheme, and predtct cnhcal AKC mterface measurements.
• Case study CAD data made avmlable from Atrbus and generated by the author allowed
the use of 3D GD&T to define each key feature, the datum system, and mdlVldual feature
tolerance Thts was defined usmg the CATIA V4 funchonal dtmenswning and
tolerancmg (FD&T) module whtch defined a funchonal tolerance soft gauge based on
GD&T spectficatwn as descnbed by the ASME or ISO standard
• The case study tolerance analysts was performed wtth the CAT3D-VSA software product
mtegrated mto CATIA V4 Thts analysts tool uses Monte Carlo stmulatwn and rank
correlatiOn to dnve stahshcal based outputs based on the measurement condthon defined.
Usmg the CAT3D-VSA software product revealed:
o Tolerance analysts usmg the CAT3D-VSA mtegrated module m CATIA V4 was
time consummg

• Spectfic conclusiOns have been defined based on the output results for each analysts
cases. These have been documented agamst thetr respective analysts case m chapter 7 of
the thests

190
10.1.5 CFC panel to aluminium substructure experimentation

The expenmentation provided an ms1ght to the potenl!al issues when an aerostructure IS


designed, manufactured and assembled from a range of different matenals and manufactunng
processes A two phased expenment mvestigated the potenl!al assembly problems when
fittmg CFC pnmary panels onto alummmm substructures.

• In support of the experimentallon, all components, manufactunng tooling, and assembly


fixtures, were designed as 3D solid models usmg the Umgraph1cs Vl8 and CATIA V4
CAD systems.
• Based on the 3D solid models created m the des1gn phase, all CNC machme tool program
generallon was developed through the Umgraph1cs CAM post processor.
• The manufacturing activity outlmed the potenl!al dJfficul!Jes associated With producmg
components from matenals such as CFC For some manufactunng operatwns, for
example, the edge routing and hole dnlhng of the CFC panels, the machming fac1hlles
were not Ideal. Th1s resulted m machmmg process problems producmg damaged features
on some component parts.
• Undertakmg all of the design and manufacturing acl!vates to produce the component parts
and associated toolmg I fixture gave the author first hand expenence of some the design
to manufacture difficulties and the potentmllead limes for the1r productiOn
• The VSA vanatwn analysis tool was used to model the panel to substructure assembly
process. The analysis a1m was to predict the resultant assembly quality attnbutes of step,
flush, and virtual hole condJI!ons based on panel and sub structure tolerance allocatiOn
and defined assembly process
• The modelling process for this analysis was the same as for the GVA studies descnbed m
chapter 7 No problems were encountered dunng the modelhng process
• The results predicted that the panel to substructure assembly philosophy would be capable
of ach1evmg the gap specificatiOn and virtual hole condJI!on However, the model
predicted that up to 15% of the step condJI!on would fall out of the specified hm1t (see
figure 94, sectiOn 8 16.8 Analysis results) Th1s predicted result also mdicted a
manufactunng capability mdex of Cpk 0.57 which would be unacceptable m the
aerospace mdustry as a deliverable process

191
10.2 Further work

The areas of research whtch were not posstble to address due to ltme constramts, or were beyond
the scope of thts theSIS, have been collated as further work The areas constdered for further work
have been dtvtded into the following categones:

• DM methodology and process


• DM tools and techmques
• DM expenmentatton

10.2.1 DM methodology and process

• Conduct further investtgation mto the pnnctple of DM wtthm the aerospace industry
Extend the process further mto a complete product hfecycle management process
Develop the PARES methodology to address the challenges.
• Develop a DM process audtt to evaluate the level of process mtegralton wtthm an
aerospace manufactunng orgamsatton
• Destgn a structured tmplementatton strategy for the current and further developed PARES
methodology.
• Development of a spectfic vanatton analysts model for mthtary atrcraft for the assessment
of observabthty charactensttcs takmg mto account the full destgn, manufactunng and
assembly product vanalton elements The low observabthty model could be developed
usmg a rule based technique to mclude both parametnc and geometnc destgn to model
resultant outer surface vanatton due to manufactunng and assembly process. Thts would
result m an evaluatton that gave the nommal radar cross section and the plus I mmus value
range based on vanalton parameters.
• Revtew the integratton of destgn for 6 stgma concepts to support the PARES
methodology
• Develop the PARES methodology to greater level of detatl based on systems engmeenng
framework.
• Introduce a cost model to the PARES methodology. Cost of product and process may be
based on reqmred product and process capabthty.
• Extend the DM analysts of products to tdenttfy the appropnate spectficalton for
manufactunng factlity destgn, 1 e , for CNC machine tool volumetnc accuracy.

192
10.2.2 DM tools and techniques

• Develop comprehensiVe feature taxonomy to address the feature design and management
as part of a component I assembly I product hterarchy This may be developed in
conjunction With the KC feature classtficatwn
• Constder the use of combmmg FEA and DM analysts techmques to model and predtct
structure deflectiOn and dtstortwn due to the assembly of components With dtfferent
stiffness properttes There may be Issues wtth the assembly of ngid none conforming
CFC panels to flextble conformmg alumimum substructures such that the alummmm part
will be deform and comply wtth the CFC matchmg mterface
• Constder the mcluston and mtegratwn of both tolerance and FEA based analysts tools to
effictently investigate the effect of vanat10n resultmg from thermal expansion, mass
deflectiOn, manufactunng process dtstortwn (1 e, heat related), restdual force dtstortwn
(reactive effect from door seals, pneumatic and hydraulic closure mechamsms) as part of
aircraft structures destgn vahdatwn. The effects of operatmg temperatures on products
such as atrframes could be modelled to understand the Impact of natural vanation
Temperature change propagation models based on FEA technology could be use to gauge
the effect on feature relatwnshtps and thetr potential dtsplacement The expenmentatwn
results from thts research prOJect suggest that temperature change propagation is not
hnear therefore further mvesttgatton ts reqmred to produce further analysts models
• The basehne PVA assembly analysts tools may be developed to take in to account thermal
vanatwn, Le, hole position dtsplacement, and tt's effect on ICY Thts may be used at the
early destgn stage for ID, I 50, and 20 parameter analysts when detailed destgn
geometry ts not avatlable The parametnc models should then be mtegrated mto the full
30 analysis tools such as VSA when 3D geometry ts defined
• Develop a comprehensive process capabthty hbrary based on geometnc dtmensionmg and
tolerance pnnctples, feature stze defimtwn, and manufactunng process famthes
• Explore the opportumty to develop a component cost model based on feature tolerance
allocation Features could be classified through a multidtmenswnal matrix whtch gave
consideration to manufactunng process capabthty, part matenal family, parametnc and
geometncal mathematical feature type and size Thts matnx database could provtde
valuable tools when constdenng component defimtwn and Its resultant cost.

193
10.2.3 DM experimentation

• Investigate the effect of surface fimsh elements on component vanahon Tlus may
mclude surface roughness (Ra) and surface wavmess (Wa) elements and the des1gn of a
variatiOn h1erarchy
• Further areas of DM expenmentatwn may be mveshgated, for example
o Sh1mmmg of panels and substructures to control th1ckness vanance.
o Evaluate CFC panel spnng back usmg FEA predJC!Ive tools; produce spnng back
models for different matenals and manufactunng process.
o The effect of mtroducmg steel grommets and mserts mto CFC panels w1th respect
to product vanatwn
o IntroductiOn of countersmk features on CFC panels
o Introduchon of mechanical fasteners
o Perfonn further tolerance analysis studies on larger mrcraft sub-assemblies.
• Investigate the use of DM techmques to support computahonal flmd dynam1c analys1s on
aircraft structure a1rflow The pred1cted vanatwn data output from DM stud1es could be
used as mput for the analySIS of flymg surface step I gap key charactenshcs perfonnance

194
References

ALAFS, 1997. Advanced L1ghtwe1ght A1rcrajt Fuselage Structure (URL:


http //www sawe org/docs/papers/categones/1996ABST htm1)
Allada, V. and S. Anand, 1995 Feature-based mode1hng approach for mtegrated manufacturing
state-of-the-art survey and future research directiOns. Intematwnal;oumal of computer mtegrated
manufacturmg, 8( 6), 411-440.
ADCATS: AssociatiOn of the Development of Computer Aided To1erancmg software.
Consortmm of 12 member compames mterested m tolerance analysis coordmated through Dr. K
Chase at Bngham Young Umvers1ty (BYU)
http I /adcats et bvu edu/home html
ANDREA: program mmmg to renewmg and strengthening the research and education m
engmeenng design and engmeenng management m Sweden Lmkopmg lnslitute ofTechno1ogy
(LITH)
Amlreason, M.M., T.L. Kahler and K. Swift, 1989. Deszgnfor assembly Kempston, UK. IFS
Pubhcatwns, Spnnger-Verlag
ASME Yl4.5M-1994 Dimensioning and to1erancing. New York Amencan Society of
Mechamca1 Engmeers (ASME)
Barrow, P ., 1997 Implementalion of new plannmg and manufactunng processes In The
manufacturzng challenge zn aerospace London !MechE semmar pubhcatwn
Barnett, L. and P. G. Leaney, 1995 Stnvmg for manufactunng flexibility m body constructiOn
Autotech 95 IMechE
Batchelor, M. A. and K. G. Swift, 1996 Conformabihty analysis m support of design for
quahty Proc lnst Mech. Engrs, Journal of engineenng manufacture, 210.
Bedworth, D. D., M. R. Henderson and P. M. Wolfe, 1991 Computer zntegrated deszgn and
manufacture New York McGraw-HIII, Inc.
Benjaafar, S. and R. Ramakrishan, 1995 Modelling, measurement and evaluatiOn of
sequencmg flexibility m manufactunng systems In The znternatwnal;ournal ofproductwn
research
Bennich, P., 1994 Chams of standards- a new concept m GPS standards. Amen can Soczety of
Mechamcal Engineenng (ASME) New York ASME
Beckwith, \V. A. and F. G. Parson, 1994 Measurement methods and the new standard B89.3 2
Manufacturzng revzew, 7(1)
Bicheno, J., I 994. The qualzty 50 a guzde to gurus, tools, wastes, technzques and systems
Buckmgham: PICSie books

195
Bjorke, 0., 1992 Computer-atded tolerancmg New York: Amencan soc1ety ofmechamcal
engmeers (ASME)
Blanchard, B. S. and W. J. Fabrycky, 1990. Systems engineermg and analysts New Jersey:
Prenlice-Hall, Inc.
Boeing. Advanced quailty system- Key characteristtcs D6-55596 TN RevisiOn A
Bossert, J. L., 1991. Quailty functwn deployment a practitioner's approach M1lwaukee,
W1sconsm: ASQC quality press
Boothroyd, G., P. Dewhurst, 2003 The 2003 mternationalforum on destgnfor manufacture and
assembly
Boothroyd, G., P. Dewhurst and W. A. Knight, 1994 Product destgn for manufacture and
assembly New York. Marcel Dekker, Inc
Bradley, R., and C. Pirie, 1997, Set the scene and outline the challenge. In The manufacturmg
challenge m aerospace London IMechE semmar publication
BS 308 'Engineering drawing practice', Part 2 - Dimensioning and tolerancing of size, Part
3 - Geometric tolerancing- Extracts. London Bnlish Standards lnslitute (BSI)
CAM-I: The CAM-I Inslitute for Manufactunng and AutomatiOn Research (!MAR).
http //www cam-i org/mdex html
Case, K. and J. Gao, 1993 Feature technology an overv1ew Jnternatwnal;ournal of computer
mtegrated manufacturmg (IJCIM), 6(1 & 2), 2-12
McCuistion P. J., 1994, The role ofdtmemtonal analysts wuhm concurrent engmeenng, OhiO
Umvers1ty, USA (Conference paper)
Chase, K. W., J. Gao and S. P. Magleby, 1997. Companson of assembly tolerance analys1s by
the direct lmeansat10n and mod1fied Monte Carlo SimulatiOn method Proceedmgs of the ASME
deszgn engmeermg techmcal conference, 82( I)
Chase, K. W. and W. Greenwood, 1988 Des1gn 1ssues m mechamcal tolerance analys1s.
Manufacturmg revzew, 1(I), 50-59
Chase, K. W., W. H. Greenwood, B. G. Loosli and L. F. Hauglund, 1990 Least cost tolerance
allocatiOn for mechamcal assemblies With automated process selectiOn Amencan Society of
Mechamcal Engmeers New York ASME
Chase, K. W., S. P. Maglehy and J. Gao, 1997 Tolerance analys1s of two- and three-
dimensional mechamcal assemblies With small kmemalic adJUStments In- Advanced tolerancmg
techmques Canada· John Wiley & Sons, Inc.
Chase, W. P., 1974. Management ofsystems engmeermg Malabar, FL Robert E. Kneger
publishmg Co , Inc
Checkland, P., 1981 Systems thmkmg systems practtce Chichester· John Wiley & Sons.
Choi, B. K., M. M. Barash and D. C. Anderson, 1984 Automatic recognitiOn of machined
surfaces from 3D solid model. Computer mded deszgn, 16(2), 81-86.

196
CIRP, International Institution for Production Engineering Research: Orgamse the CIRP
mternatwnal seminars on computer aided tolerancmg, (CAT).
Cognition Corporation, 2004 (URL http Ilwww Cl corn)
Cohen, L., 1995 Quahty functzon deployment how to make QFD work for you Readmg, Mass
Add1son-Wesley Publishing
Collins English Dictionary, 1988. England· Wm Collms Sons & Co.Ltd.
Conformability Analysis: eCA technique IS a developmg tool for the analysis of quality costs
associated with mechamcal designs and manufactunng Umvers1ty of Hull
http Ilwww hull ac uk/eca
Corbett, J., M. Dooner, J. Meleka and C. Pym, 1995. Deszgnfor manufacture strategzes,
prmczples and technologies Bath Addison-Wesley publishers Ltd
Coyne, B. 2004 PreslSlon assembly at Airbus Quahty today, June 2004, 11-12.
Cox, D. R., C. A. McMabon and J. D. T. Tannock, 1995 An approach to Imear and geometnc
tolerancmg for computer-aided design Journal ofdeszgn and manufacture, 5, 223-240
Craig, M. 1992. Controlling the variation. Manufactunng breakthrough, 1(6), 343-348.
Craig, M. 1996. Limits of tolerance. Manufactunng engmeer, 75(3), 139-143.
Creveling, C.M., 1997 Tolerance design A handbook for developzng optzmal speczficatwns
Massachusetts Add1son Wesley Longman, !ne
Cunningham, T. W., 1997 Mzgratable methods and tools for peiformmg correctzve actzons m
automotzve and azrcrafi assembly MSc report for MIT
Dassault Systemes, 2004, CATIA V5 Tolerance Analysis of Deformable Assemblies. (URL
http //www 3ds corn/en/home asp)
Davies, G.A.O., 1996 Aircraft structures Aeronautzcaljournal, Dec 1996,523-529
Decisioneering, Inc., 2000 Crystal Ball 2000, Professwnal Edztwn Product brochure
(URL' http //www crystalball corn)
Dimensional Control Systems, 2004, (URL http //www 3dcs corn)
Dowlatshahi, S., 1994 A companson of approaches to concurrent engmeenng Internatwnal
;ournal ofadvanced manufacturmg techmques, 9(1), I 06-113.
EIA 632-1 & 2, 1997 EIA 632 part I process characteristzcs, EIA 632 part 2 zmplementatzon
guzdance. G-47 systems engmeenng committee of electromc mdustnes associatiOn
Fowler, K. J., 1997 ApplicatiOn of tolerance analysiS techmques to aircraft structures lEE
manufacturmg dzvzswn colloquzum - control ofdzmenszonal varzatzon.
Foster, L. W., 1994. The me/ne apphcatzon ofgeometnc dzmenszomng and tolerancmg
techmques Massachusetts Add1son-Wesley publishmg company.
Fox, J., 1993. Quality through deszgn the key to successful product delzvery Berkshire, England·
McGraw-HII!mtematiOnal (UK) Ltd
Goode, H. H. and R. E. Macho!, 1959. Systems engineenng New York: McGraw Hill

197
Gormley, J. and D. A. Macisaac, 1989, The systems design approach (better late than not at all)
Vehzcular technology IEEE conference, San FranCISCO, 1, 401-412
Hamid, R. P. and W. G. Sullivan, 1993 Concurrent engmeermg London: Chapman & Hall.
Harrison, A., 1992 Just-in-tzme manufactunng m perspectzve Hemel Hempstead Prenl!ce Hall
mternatwnal
Henzold, G., 1995. Handbook ofgeometnca/ tolerancmg· design, manufacture and mspectwn
Chtchester: John Wtley & sons Ltd.
Herbert, P. J., C. J. Hinde, A. D. Bray, V. A. Landers, D. Round and D. M. Temple, 1990
Feature recognitiOn w1thm a truth mamtamed process planning system, Intematwna/ JOurnal of
computer mtegrated manufacturmg, 3(2), 121-132
ICAF, 1998 Azrcraft mdustry study report. (URL- http·//www ndu edu/nduhcafl!smr html)
IEEE-Std 1220-1994, 1995 IEEE tnal use standard for applzcatwn and management of the
systems engmeenng process New York The msl!tute of electromc and electncal engmeers, Inc.
INCOSE, 1998.People, teams, and systems Eighth annual mtematwnal symposmm of the
mternatwnal council on systems engmeenng
International forum on design for manufacture and assembly: Orgamse the mtemational
semmars on DFMA, (Boothroyd and Dewhurst 2003)
http //www dfma corn
International Organisation for Standards (ISO), 1995. Geometnc product specification (GPS)
-master plan, ISO/TR 14638 1995 (E) Geneva ISO
Jeang, A., 1995 Economic tolerance design for quality Qualzty and reliabzlzty mtematwnal, 11,
113-121. John W!ley & Sons, Ltd
James Gregory Associates, Inc., 1999 Integrated product & process development methods and
tools (URL http //www.JamesGregory corn)
Jeffreys, D. J., 1998. A dtmenstonal control methodology for the variallon analysis of aircraft
structure assembly In Advances m manufactunng technology - XII Suffolk: ProfessiOnal
Engmeenng Pubhca!Ions Ltd
Jeffreys, D. J., 1996 Assessmg current best practice m dtmenswnal management techntques.
IPPS, Loughborough Umvers1ty
Jeffreys, D. J., P. G. Leaney and G. Wood. A dzmenswnal control methodology for the vanatwn
analysts of azrcraft structure assembly Advances m Manufactunng Technology XII Proceedmgs
of the 14th Natwnal Conference on Manufactunng Research London: ProfessiOnal Engineenng
Publishing, 1998, pp 777-783.
Jeffreys, D., Assessmg current best practzce m d1menswnal management techmques for azrcraft
bwld. !PPS Report #96/3, Department ofManufactunng Engmeenng, Loughborough Umverstty,
England, Nov/96.

198
Jeffreys, D., 1998a. Work Proposal - Vanation Analysis of Aucraft CFC Wmg box Structure
Assembly !PPS Group Report 9810 Wolfson School of Mecharncal and Manufacturing
Engmeenng, Loughborough UniversJty.
Jeffreys, D., 1998b. D1menswnal Vanahon Analysis for a Sunple Rib to Skm Assembly- Task I.
!PPS Group Report 9811 Department ofManufactunng Engmeenng, Loughborough Umvers1ty
Jeffreys, D., 1998c Dimensional Vanatwn Analysis for a Rib to Complex Skm Assembly- Task
2. !PPS Group Report 9819. Department of Manufactunng Engineenng, Loughborough
Umvers1ty.
Jeffreys, D., l998d Dimensional Vanahon Analysis for C Spar Eqmppmg Assembly- Task 3
!PPS Group Report 98/10 Department of Manufactunng Engmeenng, Loughborough University
Jobnston, S., 1997. The mrframe manufactunng VISIOn to reahze the 21" century customer
dnvers In The manufactunng challenge m aerospace London IMechE semmar publication.
Kusiak, A., 1993. Concurrent engmeenng New York· John W1ley & Sons, Inc.
Leaney, P. G., 1996. Design for dimensiOnal control. In- Design for X- concurrent engmeenng
Imperatives London Chapman & Hall
Leaney, P. G. and R. Marshal!, 1998 Manufactunng challenges for automohve designers. In
Fundamentals of automotive design London: Chapman & Hall
Leaney, P. G. and R. Marshal!, 2001 The manufactunng challenges for automohve designers.
In. An mtroduction to modem vehicle design Oxford Butterworth-Hememann, 3, 57-92.
Leaney, P.G. and G. Wittenburg, 1992. Design for Assembhng The evaluatiOn methods of
H1tachi, Boothroyd and Lucas Assembly automatwn, 12(2), 8-17
Lee, D. J. and A. C. Thornton, 1996. The Jdenhficatwn and use of key charactenshcs m the
product development process MIT, ASME
Ligget, J. V., 1993 Dimenswnal vanatwn management handbook· a gwde for quality, design,
and manufacturmg engmeers New Jersey. PrentJce-Hall, !ne
Loureiro, G., 1999. A systems engmeermg and concurrent engmeermgframework for integrated
development of complex products PhD thesis, Loughborough University
Malloy, 0., S. Tilley and E. A. Warman, 1998 Design for manufactunng and assembly.
London Chapman & Hall.
Marguet, B. and L. Mathieu, 1997. Tolerancmg problems for mrcrafi mdustries Internahonal
msl!tutwn for productiOn engmeenng research Proceedmgs of the 5'" CIRP mternatwnal semmar
on computer mded tolerancmg Toronto, Ontario· CIRP
Marshal!, R., 1998. Design for modulanty A systems engmeering based methodology for
enhanced product reahsallon PhD Thesis, Loughborough Umvers1ty.
Martin, J. N., 1997 Systems engmeenng guidebook a process ofdeve/opmg systems and
products New York: CSC Press

199
MADLab research group: Research group based at The Umversity of Texas at Austm currently
mvolved With several proJects relatmg to dimensiOnal management
http //www me utexas edu/-madlab/
Mclntosh, R.I., S.J. Culley, A.R.l\lileham and G.W. Owen, 2001. Improvmg Changeover
Performance A strategy for becommg a lean, responsive manufacturer. Oxford. Butterworth
Hememann, Inc.
McCracken, M , The soc1ety of azrcraft matenals and process engmeers: The first fzfteen years
1944-1959, Sample Journal, Jan!Feb 2004,28-29.
MIL-STD-4998, I 992. Military standard, Systems engmeenng Department of Defence (DoD),
USA
MIT, 2004. Lean aircraft Imtiative (URL· http //web mit edu/ctpidlwww/lai)
Mob, S. C., I 996. The role of the coordmate measunng machines w1thm a d1mens1onal
management strategy M Se thesis Department of Manufactunng Engmeenng, Loughborough
Umversity.
Molina, A, A. H. AI-Ashaab, T. Ellis, R.I.M. Young and R. Bell, 1995 A review of computer
aided Simultaneous engmeenng systems, Research m engmeermg des1gn, London· Spnnger-
Verlag Limited.
Muske, S., I 997 Application of dimensiOnal management on 747 fuselage Proceedmgs of the 3'd
semmar on tolerancmg and assembly modelling
Nevins, J.L. & D.E. Whitney, 1989 Concurrent de;1gn ofproducts and processes A strategy for
the next generatwn m manufacturmg New York McGraw-Hiil, !ne
Nigam, S. W. and J. T. Turner, 1995 Review of statistical approach to tolerance analysis.
Computer-azded des1gn, 27( I), 6-15
Ohno, T. 1988. Toyota production system beyond large scale productwn TranslatiOn of 1978
Japanese editiOn 'Toyota seisan hoshiki', Pub. Productivity Press
Parnaby, J., 1981 Concept ofmanufactunng systems In Systems behav1our, 3'd edltwn
London· Paul Chapman, pp 131-141.
Parsaei, H. R. and W. G. Sullivan, I 993 Concurrent engmeermg contemporary 1ssues and
modem des1gn tools Cambndge· Chapman & Hall
Phadke, M. S., 1989 Quality engmeermg usmg robust des1gn New Jersey: Prentice-Hall
Pitt, H., 1994 SPC for the rest of us a personal path to stat1st1cal process control New York.
Addison-Wesley publishmg company, !ne
Prasad, B., 1996 Concurrent engmeermgfundamentals mtegrated product and process
orgamsatwn New Jersey. Prentice-Hall, !ne
Pratt, M. J. and P. R. Wilson, 1985 ReqUirements for support of form features m a solid
modelling system. CAM-I R-85-ASPP-01, Texas CAM-I, !ne

200
Raouf, A. and M. F. Anjum, 1995. Manufactunng systems flexibihty assessment. In. Flexible
manufactunng systems recent development Amsterdam: Elsev1er science B V
Requicha, A. A. G., 1993 Mathematical defimtion of tolerance specifications. Manufacturmg
review, 6(4).
Saltire Software, 2004. (URL: http://www.saltire corn).
Shumaker, G. C. and R. E. Thomas, 1998. Integrated processes m defence manufacturing, In.
Integrated product and process development New York· John W1ley & Sons, Inc.
Sigmetrix, 2004 CEITI6 s1gma (URL http.//www s1gmetnx corn).
Slack, N., 1991. The manufacturmg advantage Management Books 2000 Ltd, 77- 96
SPE, Japan Society for Precision Engineering: Organize and promote computer aided tolerance
systems.
http· Ilwww 1spe or 1p/enghsh/
Sports, M. F., 1983 Dzmenswnzng and toleranczngfor qualzty productwn New Jersey: Prentice-
Hall, Inc.
Sub, N. P., 1990 The prmczples of deszgn. New York Oxford Umvers1ty Press.
Syan, C. S. and U. Menon, 1994 Concurrent engzneenng concepts, zmplementatwn and
practice London· Chapman and Hall
Taguchi, G., 1993. Taguchz on robust technology development New York Amencan Society of
Mechamcal Engmeers.
Tecnomatix Technologies, 2004 (URL: http·//www tecnomatix corn)
Thornton, A., 1997 Usmg key charactenstics to balance cost and quahty dunng product
development. Cahfom1a: Proceedings of ASME DETCI97
UGS PLM Solutions, 2004, Teamcenter Qualzty Solutwn (URL http //www ugs corn).
Turner, J. U. and A. B. Gangoiti, 1991. Tolerance analysis approaches m commercial software.
Concurrent engmeermg, 1(2), 11-23.
Varatech, 2004 (URL http //www.varatech corn).
Voelcker, H., 1993. A current perspective on tolerancmg and metrology. Manufactunng revzew,
6(4), 258-268
Variation Systems Analysis, Inc. 1993 VSL users reference St. ClaJr Shores: VSA (Product
brochure).
Wickens, P. D., 1993 Lean productiOn and beyond the system, Its cntics and the future In
Human resources management journal, 3(4), 75-90
Winner, R. I., 1988 The role of concurrent engineenng m weapon systems acqwsztwn. IDA
report R-338 Institute for defence analysis Alexandna, VA
Whitney, D. A., 1997 Problems and Issues in design and manufacture of complex electro-
mechamcal systems Draper Laboratones, Inc. CSDL-R-2577.

201
Whitney, D. A., 2003. Proper constraint as a DFA pnnciple. In· The proceedmgs of the 2003
InternatiOnal forum on design for manufacture and assembly. Wakefield, I 01-I I 8.
Whitney, D. A., 2004. Mechanical assembiles New York Oxford Umversity Press
Womack, J. P., D. T. Jones and D. Roos, I990 The Machine that changed the world New
York: Rawson Associates.
Womack, J. P., D. T. Jones, 2003.

202
Appendices

I. Tools and techniques for SPC 204


2 Computer aJded tolerancing (CAT) software tools and process 208
3. Analysmg tolerance accumulation 222
4. A dimensiOnal control methodology for the vanat10n analysiS of aircraft structure 225
assembly (PublicatiOn by author)
5. DimensiOnal control as an mtegral part of next general!on aircraft development 232
(Pubhcat10n by author).
6. Stage I- VSA panel to substructure tolerance analysis results 239
7. VSA process reports 245
8 Different methods of constructmg tolerance zones m CAD systems 249
9 Wing Box Assembly Analysis Tool- User manual 252
10. Publication ReVIew; Tolerance design: A handbook for developmg ophmal 257
specificatiOns
11. Stage I - Vahsys measurement report 258

203
Appendix 1
Tools and techniques for SPC

There are a number of tools and techntques whtch support SPC These are

• Frequency histograms
• Check sheet
• Pareto chart
• Cause-and-effect dmgrams
• Control charts
• Defect concentration dtagram
• CorrelatiOn dtagrams
• Control chart

Frequency histogram
A frequency histogram ts used as a techmque to graphtcally represent process data It can Illustrate three
baste charactensttcs of the data

• Magmtude of the mean (average)


• Out! me the vanabthty of the data
• Indicate the pattern ofvanabthty (1 e the dtstnbutwn type)

The figure below ts an example htstogram of aircraft structure CFC panel thickness data

""
·~
' "'
• ""
'"'
""
..
0
u
"'
N ro
' "'
""'
"
"
"'
Ftgure L Example htstogram dtsplaymg dtstnbutwn curve and hmtts.

Chetk sheet
The check sheet IS used to collect htstoncal data for a process that may be under mvesttgatwn It IS destgned
to help tdenttcy the causes of process failures, whtch m turn wtll mdtcate potenttal product fat lures

The check sheet helps tdenttcy the sources of defects wtth respect to tune Thts may be stgruficant due to
faults that occur through dtfferent operator or machme use An example of a check sheet for morntonng a
cyhnder head IS gtven m figure 2

204
Check sheet for cylmder head defects Day! Day! Day2 Day2
Type of defect Sh1ft I Sh1ft2 Sh1ft I Sh1ft 2
Block mterface surface out of form 5 3 6 4
Castmg defects 2 I 0 I
Machined ports out of form 0 0 0 0
Block mterface surface fimsh out of 0 0 I I
specificatiOn
Damage to mternal threads I 0 I 0
Ftgure 2. Check sheet for cyhnder head defects

Pareto chart
The Pareto chart IS based on the research by Alfred Pareto whose law suggests that on average 80 percent of
problems stem from 20 percent of the causes In pnnctple Jt suggests that a small amount of problems
typically account for a large percentage of the total number of problems that occur, although Btcheno (1994)
suggests that thts IS probably more m the ratwn of90 percent to 10 percent

Pareto analysts charts are conunonly used m many dtsctplmes mcludmg mventory control (A, B and C
categones), forecastmg, marketmg, and personnel

40 . . . - - - - - - - - - - - - - - . . . ,
35
-30
;!'.
';. 25
g 20
g
g' 15
ti: 10
5
oLJ...JL.L.l._,LL..Ll-.LL..L...L_Ll.......l...l.....r::u
2 3 4 5 6 7 8 9
Defect type

Ftgure 3. Pareto analysts bar chart

Its truttal use was for the rankmg problems from the h1ghest to the lowest Once the btggest problem areas
have been tdent1fied then teams can work on a smtable solutwn The Pareto chart (bar type) presents all
defect types and thetr quanttttes These are presented as a percentage output, resolvmg each problem
element mto 1ts correct percentage proportiOn with the accumulated totals addmg up to 100 percent (see
figure above)

Cause and effect diagrams


The cause and effect d1agram, also known as the lsh1kawa and fishbone d1agram, IS used as a data framework
for the bramstormmg of possible contnbutmg causes of an tdent1fied problem or defect (Btcheno 1994) The
tcchmques are credited to Dr Kanoru Jslukawa and IS often utthsed followmg the development of a Pareto

205
analys1s chart (see above) Each category of ranked fault 1dentlfied can then be transferred to the spme of the
fishbone, W!th most effort bemg focused on the lughest ranked problems Each of the maJor problems can
then be subdtvtded mto secondary factors and these can also be subdtvtdcd tf necessary to tdentlfy the root
causes The cause and effect analysis conslltuent 'bones' cascade toward a smgle objective wh1ch IS usually
quahty onentated

Although they are not m themselves a stallst!Cal tool, they are generally mcluded m SPC programmes
because of the1r useful nature Wlth regard to V!Suahsmg sources of problems (P1tt 1994) They also
encourage multtfuncttonal team approaches to problem or defect solutions

Defect concentration diagram


The defect concentratmn dtagram, sometimes known as the measles chart, ts a htghly vtsual tool Defects or
problem areas are phys1cally plotted on an engmeenng draWing at the area of where the defects occur. The
accumulatwn of marks on the draWing g1ves a qmck unpresswn of where problems occur. This type of
s1mple representatiOn can be useful m understandmg the types of defects and the1r poss1ble related causes

Correlation diagram
The correlatwn d1agram IS used to 1denllfy relation charactenst1cs between two vanables It IS effecllvely a
scatter diagram wtth the defect level plotted on the 'x' axts and the expenmental vanable on the 'y' axts
Once a correlatwn d1agram has been developed 1t IS poss1ble to observe 1f there IS any relationship
(correlation) between the plotted data pomts Data analysts ts usually categonsed mto no correlation, clear
correlatiOn and posstble correlatwn Often tmtlal studtes cannot tdenttfy any correlatiOn and other factors

are expenmented With one at a ttme

Correlahon can be measured automahcally through computer apphcatwns and IS called the correlatwn
coefficient Stratification techmques can also be used This IS where groups of data are spht and plotted on
graphs Th1s techmque IS used to 1denllfy defects agamst llme For example, large groups of data may not
appear to have any correlation but If categonsed mto mdividual machmes or operators a pattern may become
apparent

Control charts
The control chart IS the mam weapon m the armoury of the SPC method Certam engmeenng processes are
momtored to ensure conforrn1ty w1th a target spec1ficat10n Th1s IS done through SPC control charts
(B1cheno 1994) There are two mam types of chart used, the vanable chart and the attnbute chart

Vanable charts can be descnbed as a chart that records a measured charactensttc withm a related scale
Examples are the measured diameter of a hole, the surface fmish on machmed component

The mam vanable chart records the average and range charactensllcs of a process sometunes known as the
'x' bar and 'r' chart respecllvely There are two elements to the chart, the flrst momtors the average
measurement of the sample, and the c;econd calculates the range, which ts the maxrmum mmus the mmunum

206
value Vanable charts explam process data m terms of both locatiOn (average) and spread (component-
component vanabihty) and are used m pairS

Attnbute charts some process charactensttcs cannot be recorded agamst a metnc scale Attnbute charts are
used m such sttuatmns where a logtcal judgement IS made and receded Examples are a pass or fatl cntena
on an electnc c1rcmt, a pressure test on a vessel

There are a number of attnbute control charts for different process applicatiOns, some of these are

The p chart sectiOn ofumts IS nonconfonnmg (sample size not necessanly constant)
The np chart number of umts IS nonconfonnmg (content size of sample)
The c chart number ofnonconfonm!Ies or defects (samples of constant size)
The u chart number of nonconfonnities per umt (sample size not necessanly constant)

The mam benefits of usmg control charts are

• Aids commumcatwn about a process so that mui!Ifunctwnal teams can quickly understand the
recorded charactenshcs
• M am statiStical based tool m support of the SPC methodology
• Help~ m predJCtiOn of future process events based on recordmg of the past and present trends
• Assists the detectiOn of assignable process defects and deviatiOns

REFERENCES

Bicheno, J., 1994 The quahty 50 a gmde to gurus, tools, wastes, techniques and systems
Buckmgham: PJcsJe books

Pitt, 11., 1994. SPC for the rest of us a personal path to statistical process control New York:
Add1son-Wesley pubhshmg company, I ne

207
Appendix 2
Computer Aided Tolerancing (CAT) software tools
and process

A number of Computer A1ded Tolerance (CAT) and quahty software tools and processes have emerged m
support of the dtmenstonal management acttvtty. Some of the maJor CAT software tools are avatlable are

• UGS PLM SolutiOns


• Qu1ck stack.
• Dtmenswnal Control Systems
• CATIA VS- Tolerance AnalysiS of Deformable Assemblies (TAA)
• Varatech
• Tecnomat1x Technologies Ltd
• Cogmtwn CorporatiOn
• CE/TI 6 s1gma
• Salllre Software
• Toltech

UGS PLM Solutions

UGS PLM SolutiOns (Umgraph1cs Product L1fecycle Management SolutiOns) IS a multmatwnal company
fanned as a result of a number of mergers and acqmstttons They are mvolved m a number commerctal areas
related to dtmensmnal management, these are

• Development and sales of dtmensmnal management analysts software tools known collectively as
Teamcenter Quahty SolutiOn (TQS)
• Dtmenswnal management process and deployment trammg
• Dtmenswnal management process tools, t e , geometncal dtmenstomng & tolerancmg, statistiCal
process control,
• Servtce based vanatton control consultancy support

VisVSA: The V1sVSA software was developed Imbally from the vanatwn SimulatiOn method ongmally
developed at General Motors at Detrmt know as VSM The VSA software tool has developed considerably
followmg 1ts migratiOn from the ongmal VSA company to Engmeenng Ammatwn (EAJ) m 1999 who
themselves were acqmred by Umgraph!Cs m 200 I Th1s product 1s a direct result of a VSA/EDS software
development team collaboratiOn, which has mamed the VSA core analysis software with Teamcenter
VIsuahsatwn tools to create an add-m component to run mth m the visuahsatwn software The V1sVSA
software IS now owned by UGS followmg Its separatiOn from EDS m 2004 The current capability of
V1sVSA 1s

• 3D assembly vanatwn analysis


• Monte Carlo, worst case, root sum square, plus any statJsttcally defined vanatwn can be
represented
• Support ANSI Y14 5M and ISO geometnc d1menswnmg and tolerancmg standard mth respect to
geometnc defirutwn, rules apphed to tolerance mth matenal condition and tolerance to datum
• Full graphics user mterface (GUI) mtegratwn mto V1sMocknp CAD VIsuahsatwn software
• GeneratiOn of analysiS data mcludmg process histograms and pareto ranlang of feature effects on
assembly vanatton

208
----------------------------------------------- ----

• Import all Jeadmg CAD geometry and supportmg meta data and product assembly structure via
translatiOn mto the Jt vtsuahsat10n format
• Import and export of all CAD neutral standards (IGES, STEP, etc )
• Ability to perform a limited graphical animatiOn of an assembly sequence Within the CAD
enVIronment The graplucal representatiOn ofvanatiOn 'sweep' can also be represented Withm the
CAD envrronment

VisMockUp: EAI ongmally created V1sMockUp as a digital prototypmg software tool to allow the VIewmg
and venficatwn of CAD data on an ordmary office PC or UNIX workstation This can be used to aid
concurrent engmeenng practices dunng product development to venfY assemblies for collisiOn and annotate
the CAD data with comments about corrective actiOn Where different areas of design may be based great
distance apart V1sMockUp can be used for onlme data collaboratiOn and product data management.
V1sMockUp provides further functiOnality such as the ability to ammate assembly processes and to take
measurements from the CAD geometry which can be useful m assembly !me situatiOns to replace engineenng
drawmgs as a means of commumcatlon

Vis Qualtty: A software tool used to graphically display, analyze and commurucate measurement data m a
high-end VIsuahzatmn enVIrorunent lt IS currently used m mdustry to troubleshoot assembly build problems
and share quahty mformatwn throughout the extended enterpnse. Graplucal reports combmmg measurement
mspectmn data with lightweight 3D geometry can be created Extremely large assemblies from multiple
CAD systems can be plugged directly mto the lightweight visualizatiOn enVIrorunent

VisPubltsl!: Provides the ability to produce dynamic and concurrent productiOn of Technical Documents
and PublicatiOns, With HTML reportmg capabilities An Assembly optiOn IS also available prov1dmg the
ability to display the assembly process of products throughout the life cycle

e-Vis: Allows direct visual collaboratiOn of complex, engmeered products m 3D Allows the data and tools
for mteractmg wtth that data m a secure, Internet-connected workspace

Competitive assessment
produet objKtlvt package
Bu1ld stntegy docum ant
Prtltm~nary GOT
Variation Simulation Model
Fin~ GOT
Measurement points
Me.uurernent plan
Chec:kin; lholdln; nxture
development .nd w.lldatlon /
CMM progrm~~ dewlop?t • ·---•..J •··"
;'
;'

-- ,,
-· _...

F1gure I UGS PLM dimenswnal management process (UGS PLM SolutiOns)

209
The UGSNSA d1menswnal management process IS defined m a SIX step process (see figure I), these bemg

1. Define product dunenswnal objeCtives


n Develop the design, manufacturmg and assembly process Evaluate to prove product IS capable and
robust
m Develop dunenswnal management product documentahon
IV Develop and vahdate the measurement process
v Ensure that the manufactunng process realises their design mtent
vt Productwn to design feedback loop

Quick stack

Th1s IS a legacy product capable ofundertakmg sunple ID-3D stack analysis and IS mtegrated directly mto
the UmgraphiCs CAD/CAM/CAE Software

Dimensional Control Systems

D1menswnal Control Systems (DCS) IS an engmeenng consultmg firm that have been mvolved m the field of
dunenstonal management, or dimensiOnal control as they constder, for some 15 years Mostly worlang m the
automotive mdustry they had htiie mvolvement m the aerospace mdustry but are workmg to penetrate this
market. They have undertaken some work With Raytheon but tlus has not been followed up Based m Troy,
they are well placed to serviCe the b1g three automotive compames namely Ford, Chrysler and General
Motors, and have undertaken d1menswnal control projects for all three

DCS has emerged from the Tnkon Corporatwn, Tnkon Design, Inc and therefore DCS and Tnkon products
traditionally displayed some simiianty, however the DCS product has now replaced Tnkon winch IS now
d1scontmued

The DCS' tools for vanation analysis are Similar to those used m V1sVSA m that they rely on geometnc part
mfonnatwn coupled With tolerance/GD&T data to create a 3D tolerance model This 3D model IS then
subject to two statistiCal tests, a Monte Carlo vanatwn analysis and a HLM sensitiVIty study

Improve
,Control
Analyze

Pnont1ze
Tunelme

Figure 2 DCS dimensiOnal control process (DCS)

210
Thts software study forms an mtegral part of DCS' ten-step dunenswnal control procedure (see figure 2),
whtch ts defmed as·

Identtty and document dunenswnal quahty goals


2 Team consensus and stgnatures
3 Develop strategtc plans to aclueve all dunenswnal quahty goals
4. Determme global tolerance and maJor datum for sub-assembhes
5. Generate tolerances and datum for all parts and assembhes, stattsttcal stmulatwn, work towards buy
m from all team members - thts ts the key engmeenng phase
6. Optuntse the destgn/process through 3-D analysts
7 Venty prototype tool and fixture destgns- vahdate gauge and fixture capabthty.
8. Evaluate prototype results
9 Venty productton tool and fixture destgns- vahdate gauge and fixture capabthty
I 0 Support dunng ptlot, launch and production.

InformatiOn gathered from steps 8 and I 0 acts as feedback to be put back m at step 6 to allow opttmtsatwn of
the product for dtmenswnal robustness

The software tools used operate on a standard PC workstatton platform easmg accesstbthty to non-CAD
operators The 3-D geometry ts tmported from CAD data vta an IGES file converter and there ts also the
optiOn to manually bmld stmple geometry Because of thts CAD-hke operatton there ts no manual
programmmg language accesstb1e to the user, whtch does mean that the ablltty to develop custom code is
lost The 3-D geometry can be vtsuahsed wtthm the dynamtc tolerance stmulatton package (3D-DTS)

DCS currently support 4 products types·

1-Dcs software solution: thts ts one dtmenswnal hnear tolerance stmulatwn software It runs on a pc
platform and ts a pomt based coordmate analysts system It can support a range of tolerance capabthty
mcludmg Worse-Case, RSSIRMS, or Monte Carlo wtth a range of staltsttcal dtstnbutwn types DCS has the
abthty to graphtcally mdtcate the 'sweep' of component/subassembly vanatwn DCS clatm that thts supports
engmeers when mterpretmg analysts results It has no capabthty of creatmg graphtcs but IGES files can be
tmported and exported DCS clatm that thetr ID product can be very qmck and effecttve for some types of
tolerance analysts DCS suggest use of full 3D ts not always necessary

3-Dcs software solution: as 1-DCS above, but has full three dtmenstonal capabthty These are to be run on
both pc and workstatton platforms One advantage of the 3-Dcs software ts tts abthty to graphtcally ammate
assembly sequencmg whtch mcludes the correct onentatwn and approach of all components mto as
assembly Thts ts a btg advantage when presentmg an assembly analysts to an engmeenng group

3-Dcs for mechanical desktop software solution: the product ts stmtlar to 3-CDS software soluttons but ts
fully mtegrated wtthm the AutoCAD mechamcal desktop product

GDM-3D: Used to graphtcally dtsplay measured data from a number of sources mcludmg raw coordmate
measunng machme (CMM) data, mobile arm measunng and Datamyte type devtces Measured data ts
formatted mto comparator or accumulator outputs and hard coptes can be produced m a standard report
Raw x, y and z data can be also dtsplayed

211
3-DCS for CAA- CATIA V5 integrated solution: Sumlar capabtiity to 3-DCS but dtrectly mtegrated mto
the CATIA V5 CAD envuonment Wtth a dtrect link to the CATIA FD&T, tolerances are automattcally
mcorporated mto the model helpmg to reduce model creation ttme

Model creatton ts atded by vtsual tolerance devtatwn of CATIA solids and vtsual ammatton of the assembly
sequence. In explormg the V5 archttecture DCS clatm the followmg key benefits

Faster mtegratwn by enforcmg an Object Onented Destgn wtth C++ and JAVA
Enablmg dtstnbuted computmg across platforms by uttltzmg CORBA/COM
Interfacmg the DCS applicatiOn wtth other knowledgeware tools m the CATIA V5 "hub"

CATIA VS- Tolerance Analysis of Deformable Assemblies (TAA)

The new P3 module has been developed for the Tolerance Analysts of Deformable Assemblies (TAA) Its
functiOn ts to 'assess the tmpact of the assembly process on flextble components' In real-world terms thts wtll
allow users to create a dtgttal stmulatton that follows the assembly of a parttcular set of components (such as
a car body panel) and gauge the effects that the vanous productton processes have on the tolerance of that
assembled product Whereas thts type of work may be posstble wtthm other htgh-end systems at a baste
level, what wtll make thts most mterestmg wtll be the abtiity to use the CATIA knowledgeware fimcttons to
expenment wtth dtfferent manufactunng processes (such as vanatwn m weldmg postttons and ttmmg) and
accurately gauge the effects they have on the tolerance of the final product All of thts allows users to
opltmtse, not only the form of the part, but also the productton of the part Key product features mclude
DeformatiOn and assembly process approaches- Thts product ts based on a mechamcal approach
that takes both deformatton and the assembly process mto account to predtct the tolerances for
welded (nveted, bolted, or glued) assemblies of sheet metal parts
2 Creatton of data and assembly process spectficatwns
• Commands for spectfYmg the attnbutes to be mcluded m the analysts
• Commands to spectfY the assembly process
3 Process venfication before stmulatton- thts function avoids simulatiOn deficiencies by allowmg
the user to make sure the specified process does not contam faults For mstance, tt can detect tf
there are two spot weld operatiOns for one pomt
4 Simulatton of the assembly to perform a set oftolerancmg analyses- the product provtdes
sensitiVIty analysis, detenmmst analysts, and statistical analysts that are based on the same conunon
computatton Integratton ofFtmte Element Analysts models the elasttc "deformabtiity" m the
assembly process and results m a finer and more reahstlc stmulatwn The user can get a senstttvtty
analysts to tdenttfY the key charactensttcs of the assembly
5 Easy re-computatiOn of the strnulatton- thts product avotds the use of ttme-consummg Monte
Carlo srmulatton Addtltonally, there ts no need to re-compute the stmulatton tf only the mput
vanattons are modtfied The type of stmulatwn used allows the user to do a quick update stmulatton
when the assembly process or few attnbutes need to be modtfied, added, or removed
6 Multt-dtsplay of the stmulatton
• Graplncal dtsplay of staltsltcal and determtmst analysts results are provtded through
dtsplacement presentatton (usmg FEA representatton) and pomt devtattOn (usmg arrow and
eliipsotdal representatiOn)
• Graphtcal dtsplay of senstltvtty analysts results ts provtded through the representation of input
devtalton contnbuttons (m percentage) of output devtalton
• Stattsttcal and determtmst analysts results are also avatlable through numencal dtsplay
The module mdtcates the potenttal for senously m-depth analysts of not only the behavwur of a product
dunng tls use, but dunng the manufactunng and productiOn process

212
Varatech

Varatech ts a tecluucal engmeenng consultmg firm based m Mtchtgan Thetr mam busmess m centred
around support for product development and assembly systems optimizatiOn They offer dunenswnal
management services utihsmg therr own software solutiOn and supported by the use of design for SIX s1gma
(DFSS) techniques

Varatech dunensmnal control engmeers offer expenence m the followmg areas.


• Design for assembly
• GD&T/Funct10nal datum structure
• 3D vanat10n modelling
• Functional gauge and fixture destgn

Key steps for the Varatech dunens10nal management process


I. Quahf'y and quantify an assembly's bwld objectives
2 Evaluate the geometnc sensttlvtttes of destgn concepts
3 Define and document locatmg schemes, datum structures, and GD&T
4 Gather and assess processmg capability
5 Evaluate assembly process

Varatech develops the S1gmund range of analysis software for assembly and piece-part tolerance analysis
The software product !me mcludes

Sigmund ID: S1gmund ID evaluates one-dimensiOnal tolerance stacks to obtam venfiable results usmg
mdustry accepted analysts tecluuques such as Worst Case and RSS analyMs

Sigmund 3D: S1gmund 3D evaluates, optim1zes, and validates the capab1htyofproposed designs and
assembly processes by creatmg virtual assemblies that emulate piece-part tolerances and assembly process
vanat10n

S1gmund 3D allows engmeers to bmld a large number of virtual assemblies for evaluatiOn With respect to
pre-defined bmld objeCtives It emulates manufactunng process vanat10n followmg the ASME Yl4 SM-
1994 GD&T standard as well as assembly process vanattons such as clampmg sequence, part onentation,
and gravity effects

Supported CAD/CAE
S1gmund 3D- Stand Alone
S1gmund 3D- SohdWorks Integrated
S1gmund 3D - Pro/Engmeer Integrated
S1gmund 3D- Sohd Edge Integrated

S1gmund ID- Stand Alone


S1gmund ID- SohdWorks Integrated
S1gmund ID- Pro/Engmeer Integrated
S1gmund ID- Sohd Edge Integrated
Tecnomatix Technologies Ltd

TecnomatiX was established m 1983 The Company went pubhc m 1993 (NASDAQ symbol TCNO) and
today has a Wide;pread, mternat10nal network of subsidiary offices and d1stnbutors m North Amenca,

213
Europe, the Pactfic Rlm and Latm Amenca. Tecnomattx are mvolved wtth comparues m the automotive,
aerospace, electromcs and heavy equ1pment mdustnes m the area of product analys1s and product hfe cycle
management

Tecnomat1x offers the followmg software analys1s tools·

eM-To!Mate enables the pred1ctwn and effect of tolerances and matmg operatwns on an assembly process
It Identifies key charactenst1cs of an assembly-such as the flushness of an a1rplane wmg, the gap between a
door and the fender of a car, or the clearance between the removable hard d1sk and the bay m a portable
PC-as well as the cnt1cal features of an assembly eM-To!Mate can calculate the vanatwns m these
charactenstlcs and the contnbutiOn of tolerances to these vanat10ns The stmu1ahon and analysts tools are
des1gned to help dehver optrmal tolerances for manufacturing and assembly

eM-Probe
eM-Probe enables the creatwn, optl£lllzal!on and venficatwn of off-hne mspectwn programs for CMM and
CNC machmes It prov1des colhswn detectwn capab1ht1es for complex parts and fixture set-ups With outputs
contammg DMIS programs for spec1fic CMM manufacturers

eM-Probe CAD
eM-Probe CAD enables the creatwn of off-hne mspectwn programs for CMM and CNC machmes Wlthm a
CAD envtronment, and provtdes the abthty to define measurement probes, and mspectwn processes
InspectiOn proce5'es created by eM-Probe can be run only by eM-Inspect and eM-Inspector

eM-Inspector
eM-Inspector enables readmg m DMIS files, and the1r executwn on NC Machme Tools and CMMs through
a dlfect bl-d~recllonal conrmumcatwns hnk It bnngs 3D CAD data to the shop floor w1th a Wmdows NT
based mspecllon system, and can dnve NC machme tools With DMIS 2 I and 3 0 commands eM-Inspector
provtdes consistent mathematical dtagnosttc analyses regardless of the measurement devtce

eM-Inspect
eM-Inspect enables the mspectwn and analys1s of parts on the shop floor It supports over 30 CMMs and NC
machme tools, performs onlme modtficatton of mspectwn processes, and revtews analysts results of
measured features m real tune
eM-lnspect-UG IS an embedded product m the UG CAD system.
cM-Inspect-SA IS a standalone product (not embedded m any CAD system) and IS ava1lable on UNIX
platforms only.

eM-Quahfy
eM-Quahfy enables the analys1s and evaluatiOn of measured data agamst nommal models, and proVIdes
conSIStent mathematical and graph1cal data to perform best-fit analyses and the venfical!on ofposs1ble

214
causes of failure

eM-Reverse
eM-Reverse enables the creatiOn of accurate 3D-CAD models from physzcal parts or master toolmg and
proVIdes best-fit analyses on measured mspectwn data It features adjustable accuracy levels and IS fully
embedded WJthm CATIA and UmgraphiCs

TecnomatJX Technologies produce a range of software products and servzces winch hnk the VIrtual
manufactunng envtronment to the real Thetr mam product hne ts

ROBCAD; for the deszgn, simulatiOn, optzmzsation and off-hoe programmmg of automated and manual
manufacturmg systems
EXALINE; for the deszgn, szmulatzon, optzmzsatwn and off-hne programnung of the pnnted cucuzt board
assembly process
PART; for the numencal control (NC) process planmng and programmmg
VALISYS; for the defuutwn, mspection and management of product tolerances throughout the
manufactunng process
DYNAMO; for the deszgn and venficatwn of assembly, packagmg and mamtenance processes

The V ALISYS product zs made up of a smt of modules

1 Deszgn, used to define 3D tolerance mformatzon m accordance Wlth ANSI Yl4 SM - 1982 GD&T
standard The tolerances are created dzrectly from CAD data Deszgn contams four functions:

Features, allows the selection of CAD features (faces, planes etc) to be controlled
Gauge, enables a 3D tolerance model (soft gauge) to be placed on selected features The 'soft gauge' IS a
computer representation of an allowable 3D geometnc tolerance zone assoctated wtth a feature bemg
controlled (1 e feature of flatness) When a statistical vanatwn Simulation IS operated the soft gauge
represents the hm1ts of that tolerance zone Gauge also performs a functiOnal check to venfY the relatwnsh1p
between the feature selected and the tolerance apphed are compat1ble
Check, used to validate d1menswns and tolerances are m accordance to the GD&T standard Yl4 SM
Vutzhty, allows the user to access, ed1t and save changes to the V ALISYS database

2 Programmmg, enables the off-lme programmmg of CMMs and NC machme tools wtth mspectwn
capabthttes Prograrnmmg contams five functiOns

Pathl, Path 2, Path 3, employs the features data form deSign to create a genenc mspectwn path for use by
NC mspectton equtpment
Generate, creates the VALISYS control language (VCL) wh1ch contams mstructwns for NC measunng
eqmpment Th1s may be to perform V ALISYS functions or to enable NC mstructwns to dnve mspectwn
eqmpment

215
Probe, allows VALISYS mfonnatwn to be attached to a CAD models of probes for the executiOn of
cahbratwn, onentatwn and mspechon processes

3 InspectiOn, allows the mspectwn program to be validated and mod1fied 1f necessary Also produces the
ana!ys1s data fonn the mspect10n program InspectiOn IS made up of four functwns,

F1les, enables the user to access, 1mport, delete and hst part files
Inspectwn, uses VALISYS programmmg mstruct10ns to dnve the mspect10n eqmpment, takes, retrieves and
stores measurement data
Qualify, creates a math model fonn mspect10n data and compares that agamst the related soft gauge math
model and creates a report companng des1gn and measured data
Control, used for definmg and executmg machme functions

4 Analyse, uses a graph1cal representatiOn to compare the measured agamst CAD data to 1dent1zy areas of
soft gauges m terms of pass, fa~ I and rework Contams only the analyse functiOn

5 Assemble, mcorporates the use of stattsttcal SimulatiOn techntques to predtct vanatwn stack up resultmg
form an assembly of a group of parts It also checks the assembly methods and Identifies cond1t1ons of over
constramt and under constramt Reports are produced identifying cnttcal part and assembly tolerances whtch
contnbute to vanatton There are SIX functions in assemble

Compmgr, enables component models to be managed to fac1htate assembly modelhng


Gauge, see Gauge m Des1gn sectton
Features, see Features m Destgn sectlon
Vutthty, see Vutthty m Destgn section
VSA-FCN, used to 1den11zy the mcomplete, mcorrect and mcons1stent apphcatwn of GD&T With regard to
the VSA s1mulahon software

6 Reverse, allows the construchon (d1g1hzahon) and mampulat10n of phys1cal features WJthm a CAD
system wh1ch may then be used m the constructiOn of complete component geometry Th1s module only
conststs of the Reverse function

7 Process control, destgned to track manufactunng process mformatwn to detenrune tf or when process ts
out of statiStical control Makes use of graplncal control charts and reports to allow the detectiOn of process
devtatwn and raptd correctiOn

Tecnomattx do not have a well documented drmenswnal control process However, they promote a
dunenswnal quahty process closely associated With the vanous V ALISYS software modules Currently th1s
1s the only software ava1lable wh1ch offers an mtegrated closed-loop path form tolerance allocatiOn to
dlffienswnal mspectton

216
Thetr mam busmess objectives are software sales, consultancy and trammg Thetr pnmary busmess area ts
m software sales wtth the consultancy and trammg elements run m dtrect support They are a multmatwnal
company and the1r headquarters 1s based m Herzthya, Israel

Advantages of the software

• Offers an mtegrated closed-loop dimensional quality process


• Can model and analyse complex 3D geometry
• Tolerance applicatiOn m accordance With GD&T ANSI Y14 5M standard
• Possible to check tolerance applicatiOn for mcorrect and Illegal call cuts
• Can be used with leadmg CAD systems- CATIA, Umgraph1cs, and IDEAS
• Promotes the use ofGD&T m an orgamsatwn
• CollisiOn free off-hne programmmg ofNC and CNC mspectwn eqUipment

Disadvantages of software

• Cost of software and hardware


• Software IS complex and therefore requrres extensive trammg.
• Software IS embedded Withm a CAD system therefore there IS a reqUirement to become familiar
with that CAD system

The author has used all the modules of the V ALISYS mtegrated software The software was used for the
dtgttal mspect10n of a number of aerospace and automotive component parts usmg a CMM m native mode
Assembly analysis has also been undertaken on the author's expenmental component part assemblies

Cognition Corporation

The Cogmtwn CorporatiOn IS a multmatwnal company who develops and markets the Advantage Senes of
software products These are a suite of engmeenng decisiOn support tools destgned to be used wtth an
evolvmg product development process focusmg on re engmeenng the design process

A maJor obJective of the company IS to achieve a 5 to 1 reduction m tune and cost ofbnngmg a product to
market while mcreasmg product quality and eiimmatmg product and toolmg changes caused by design error.
They aim to achieve this by the development of the VIrtual conceptual prototype method which IS to replace
the current dependency on physical prototypes to facilitate the design development process They stnve to
deliver a software solutiOn to the customer and make It operatiOnal very qUickly

MaJOr Objectives of Cogmtion are

• Dehver and tram customer on software solutiOns very qmckly.


• Enable customer to customize and modify software solutiOns very easily and qUickly.
• Product only released when fully developed and proofed
• Promote the modelling of the entire product/process development
• Promote the modellmg of the entire engmeermg program
• Conceptual model should dnve CAD detail design models

Cognition produce the folloWing products

1 Mechamcal Advantage I and II a feature-based vanatwn solid design system built for perfonnance
modellmg, destgn opttmtsatton, cnttcal parameter management, tolerance allocatiOn and analysts, and
kmemattcs and dynamtcs

217
Modules are made up of

• Tolerance modelling and analysis.


• Cnt1cal parameter analysis.
• Solidnote
• Sketch note
• Mathnote
• Program note
• Data capture and presentation (DCAP) note

2 Cost Advantage· a design analysiS tool that assists m cost assessment for both parts and assemblies The
software can be configured to allow the user to perform productiOn analysis, manufactunng alternative
studies, DFM and DFA The modules are made up of

• InJeCtiOn meldmg model


• Manual assembly model
• Electromc system assembly model
• Costhnk/PE

3 Des1gn Advantage an ANSIIISO standard drafhng system that can accept 2D/3D w1re frame models, 3D
surface models, and any ACIS solid model for the purpose of documentatiOn

Trammg focuses around use of software products and related procedures Cognition appears to have well
defined trammg courses for all1ts software products They Identify the duratiOn, prereqms1tes, and
obJectives of all courses All trammg based on real engmeenng problems and solutiOns Cognition focus
therr busmess on software sales rather than dedicated consultancy, therefore they reqmre thelf software to be
relatively fnendly for the engmeer to u'e Therefore the user slall requiTement to dnve the products Will be
low to medmm

A product development process IS currently bemg evolved Thelf process starts w1th a quality functiOn
deployment (QFD)approach wh1ch feeds reqmrements mto a system called SLATE produced by TD
Technologtes SLATE IS requtrements control software for design and manufactunng data on such Items as
time, resources, capaclttes, capabilities etc Each ttem IS then considered as an obJect, 1 e people, toohng,
processes, and IS modelled m an obJect related data base. CognitiOn argue that relatwnal data bases cannot
sufficiently handle large changes and tradeoffs' effectively and become difficult to manage

Thts process al!ows engmeenng to focus on the aspects of product parameter management dunng the
concept design phases m a 2D and 2 5D sketchmg environment VIrtual models are created of the product
and parametnc tolerancmg and srrnulatwn can be undertaken The virtual models can be lmked to data
tables and engmeers can model "what 1f' type equatiOns to explore parametnc vanat10n Cognition IS
developmg thelf system m order to tolerance all parameters and not JUSt geometnc

Thelf process focuses on 1dent1fymg and modelling all contnbutors to parametnc vanatwn and It is possible
to plot them graphically m order to analyse thelf range They promote product development by pnmanly
modellmg all parameters and then convertmg th1s data mto detailed 3D CAD geometry at a later stage
Cognition argne that organisatiOns commit to much trrne and money m detailed 3D design before all
parametnc concerns have been resolved If left unresolved these parametnc concerns 1fleft Will only force
destgn engmeers to make ttme consummg and expensive detailed redesigns at a later stage
CognitiOn suggests that engmeers need to undertake VIrtual prototypmg for all parametnc vanatwn not JUSt
geometnc tolerance analysts They recogmze three levels of tolerance analysts, these are

• Manual calculatiOns hand wntten solutiOns, use of spread sheets Problem ts reluctance of
engmeers to undertake th1s and when 1t 1s done many mistakes can occur due to length and
complexity of calculatiOns
• Cognition type- Tolerance Modelhng and Analysis can deal w1th 95% of tolerance problems m
engmeenng very quickly and eas1ly Problem IS the software cannot handle complex 3D analysis-
therefore 1t 1s limited
• VSA type can deal With almost all tolerance related modellmg and simulatiOn Problem IS product
complexity requmng lugh level of slall to operate and mterpret analysis results

218
The tolerance analysis module of the Mecharucal Advantage Sketch Note apphcal!on IS an mtegrated tool
that works w1thm normal design actlVlty ObservatiOns made of the tolerance modelling and analysis
software module

• Cannot handle 3D complex surface tolerance problems


• IntegratiOn to CATIA and UG via IGES format
• No support of STEP standard to date
• Cogmtmn views their role as fast tolerance analysis, ID and 2D which they claim can solve 95% of
tolerance engmeenng problems. They see themselves as the bndge between paper based
calculatiOns and full (VSA) 3D tolerance analysis.
• Tolerance data geometry to ANSI and ISO standard With regard to GD&T but VISually does not
appear the same
• Analysis models created m 2D and extruded to produce 2 3/3D
• DimensiOns on models can be "locked" 1f desired
• Cntical dtmenswns are not locked - remam flextble for vanatwn SimulatiOn
• IntroductiOn of feature based ACIS sohd model mtegrated With ANSI standard associallve 2D
drawmgs for parts and assemblies
• Capable senstttvity analysts, o/o contnbutwn analysts, 6 stgma and Monte Carlo analysts
• Directly compal!ble With ACIS based systems

Cognii!on does not have a well documented dimensiOnal management process The three mam areas of
Cogmtton are:

• Software development and sales


• Engmeenng consultancy
• Education and trammg on software and engmeenng processes

There are three mam areas that Cogmt10n market, these are m

• Software sales· This IS therr pnmary area ofbusmess Cogmtton would prefer to keep a lugh level
of software sales and trammg and keep engmeenng consultancy to a mimmum
• Trammg· Thts IS a maJOT supportmg part of the busmess to software sales
• Engineenng consultancy Cogniiion can provide a broad range of consultmg serviCes to the
manufactunng engmeenng mdustry They can provide serviCes relatmg to product support,
mcJudmg Implementation, planmng, trammg, consultatiOn and software. However, consultancy IS a
secondary market aim followmg software sales.

CEffl 6 sigma

Sigmetrix LLC IS a software development company focused on prov1dmg easy-to-use assembly tolerance
analysis and opllmizal!on software for mechamcal engmeers S1grnetnx IS a partnership between the ongmal
creators of the CE/TOL 6 S1gma tolerance analysis software and Rand Worldwide CE/TOL 6 Sigma has
been sh1ppmg to Pro/ENGINEER customers smce 1992 and IS d1stnbuted exclusively by Rand WorldWide
and Parametnc Technology Comoratton on a worldwtde basts

S1gmetnx IS contmumg to develop CE/TOL 6 S1gma as the premier tolerance opllnuzatwn software m the
world today Tlrrough the company's umque research relatiOnship With the top umvers111es developmg
complrmentary technolopes, hke Bngham Young Umvers1ty's ADCATS program, as well as the strategic
software development and distnbutiOn relatiOnship With Rand WorldWide and Parametnc Technology
Corporation,

219
Most of S1gmetnx key customers are the same mdustry leaders who are deploymg 6 S1gma methodologies on
a broad basis and reahzmg signtficant returns on mvestment and additional market penetration Sigmetnx
umque understandmg of 6 S1gma methodologies, Pro/ENGINEER mtegratwn, tolerance modelhng and
optimizatiOn methods has g1ven the company a demonstrated leadership pos1t10n m helpmg spearhead 6
S1gma programs m our customer base The partnership between S1gmetnx, Rand WorldWide and Parametnc
Technology Corporatwn ts allowmg these three premter cornpames to combme thetr collective talents m
prov1dmg mtegrated solutiOns that meet the demands of a ngorous 6 S1gma program

Saltire Software

Analytlx IS a sophisticated mechamsm des1gn-and-analys1s software package It IS essent~ally a parametnc


modeller where embedded kmemat1c equatwns are available for fast analysiS. It also has the flexiblhty to let
the user \\Tlte thetr own more formula for vanous type of analysis

It has the capab1hty of undertakmg statistical or maximmnlmimmum tolerance stack-up on designs An


mteractlve calculator ts also embedded that enables the user to develop therr own equations to analyse certam
destgn parameters

It operates form a PC m from Wlthm a MICrosoft Wmdows env1romnent. Its data exchange IS hmlted to the
data exchange format (DXF) or Wmdows Dynamic Data Exchange (DDE), cut and paste

Some of Its functions are

• Parametnc sketchmg -2D, patented vanat1on geometry engme, wtth auto-dunensiOnmg


• Mechamcal model mput, distances and angle values, velocities and acceleratiOn Tolerance on
mput dtstances and angles Apphed forces and moments Masses and moments of mertla
• Mechamcal model output, Geometnc position, velocity and acceleratiOn ReactiOn forces Shear
force and bendmg moments Stress and deflectiOn Tolerance stack, sensttivrty, trace, zones,
absolute or statistical
• Stmulatwn and analysts tools, Ammatlon, pomt trace, envelope EquatiOn calculator and
mteractlve equation solver Graphs and tables of output parameters

The product IS targeted at the up front concept stages of design development Relatively hnlc 1s known
about this product

Toltech

The Toltech computer a1ded tolerancmg system has been developed by The Norwegian Institute of
Technology, DIVISIOn of Production Engmeenng (BJorke) The system IS named TOLTECH, from
TOLerance TECHnology, and IS designed to mteractlvely perform tolerance related calculatiOns.

220
The TOLTECH system has Its own developed language m which the tolerance analysis must be wntten m a
senes of statements These statements contam four categones of mformatton·

Numbers, expressed as mtegers or reals


Symbols, formed as a senes ofletters (A-Z), digits (0-9), and characters" "
PunctuatiOn's, constst of one alphanumenc character and have a spectal meamng m the TOLTECH language
lgnorables, consist of one non-alphanumencal character

221
Appendix 3
Analysing tolerance accumulation

There are four mam groups assoctated W1th analysmg tolerance accumulatton, these are

• Worst case
• Root-mean square (RMS) or root sum square (RSS)
• Statistical analysts
• Kmemattc analysts

Worst case; does not take mto account drstnbutron of srze of feature and that they do not exceed therr
respective specrficatrons The tolerance vanables wrll take a value at one of rts hrmts m such a way to )'leld
an extreme condthon at a pomt of analysts, 1 e the maxtmum expected vanatwn at that pomt Thts analysts
rs only recommended for a hmrted number of drmensrons and where very hrgh assembly confidence IS
requtred

Root-mean square (RMS); predrcts vanatron at a g1ven pomt whrch IS not expected to be exceeded more
than 1 m 370 assembhes (99 73%) Thts analysts was developed as a more accurate representation of
vanance when constdenng a large number of tolerances One assumption made IS that the analysts
represents features that are manufactured about the mean pomt of the tolerance, 1 e an equal b!lateral
tolerance ts assumed

Statistical analysis; employs a stattsttcal dtstnbutwn whtch represents more accurately the expected process
vanatwn of each d1menswn It IS possrble to use actual manufactunng process data or an approxrmatwn of a
dtstnbutton curve and hm1ts to g1ve a better representation of actual or calculated vanatmn data Several
drfferent methods have been developed for statrstrcal analys1s

• Lmeanzatmn method
• Method of system moments
• Approxtmatmn by numencal mtegratton or quadrate techntque
• Monte Carlo StmulatiOn
• Rehab1hty mdex
• Taguchi method and modrfied Taguclu method
• Croft's method
• Extended Taylor senes approximatiOn.
• Hasofer- Lmd mdex method

Cunently m manufactunng there IS a mrgratwn towards 3D sohd modellmg for deta!l desrgn apphcahon
W1th the ava1lab1hty of the STEP mformat1on mterchange standards, the CAE systems wrll need to further
develop the ASMEIISO tolerance •tandards whrch w!ll mclude statistical tolerance analys1s. Current
commercral systems use two types of stahshcal SimulatiOn modelhng, namely hnear stack up and Monte
Carlo. Tins IS also the case wrth wrre frame and pomt based analys1s systems

222
The most commonly used methods by commercial tolerance analysis software IS linear stack up and Monte
Carlo simulation Lmear stack up, or root sum square (RSS) method, IS used when an assembly response
function can be expressed as a hnear functiOn of the component parameters The math functiOn for hnear
stack up 1s·

T total= (tl 2 + t2 2 + t3 2 + + tn2 ) 05

T total= Total tolerance


tn• = JndiVldual tolerances

Where T total= total predicted assembly tolerance (equal bilateral vanatwn) usmg the RSS model

Where tolerance d1stnbutwns are koown, a Monte Carlo sunulat10n (MCS) may be performed to proVIde an
accurate assessment of cumulative vanatmn effects The basts of the technique IS to Simulate samphng of
koown process distnbution For example, If the vanatwn of two elements A and B are considered, the
solutiOn requrred IS

P(A + B =A,+ B,)= PA,xPB,

The probability that the sum of elements A+ B IS equal to the sum of A,+ B, equals the probability of A,
times the probability ofB, If the d1stnbutwn charactcnst1cs for A and D are available a computer may be
used to select a random number for 1 and sample the two d1stnbutwns Sampling a frequency plot would
typically gtve two values for any gtven random number, therefore cumulative dtstnbutwns are used mstead,
see figure below

I '
I I I
L_._l_ _ _
-~

DimensiOn

a Frequency distnbution do not yield umque values

lllly
oil... lOO%-,/
'
50%
I

I
I

Dlmens10n D1mensJon

b Convert from frequency to cumulative distnbutiOn

PA, ---- [andl PB,


____ /
" " - - J . __ _ _ _ A
A,
c. Sample by random number

Figure I. Monte Carlo SimulatiOn process

The followmg tasks are performed by the computer system

Derive PA 1 from random number, sample the d1stnbutwn for A 1

223
u Denve PB 1 from another random number, sample the d1stnbut10n for B 1
m. Add A1 to B 1=I" output
IV. Multiply PA, With PB 1 = I • probability
v. Repeat n number of times
VI Add the probability of all values contamed Withm stated IInuts
VII Produce a histogram of the results

If a number of elements each dtsplaymg a normal statistical charactenstic are used m a vanance stmulauon
the RSS and MCS methods will produce essentmlly the same results However when elements are other than
nonnally distributed MCS Will Yield results una !tamable by the RSS method.

224
Appendix 4
A dimensional control methodology for the variation
analysis of aircraft structure assembly

D J JEFFREYS and P G LEANEY


Department ofManufactunng Engmeenng, Loughborough Umversity, LeiCestershire, UK
GWOOD
Bnllsh Aerospace Airbus Ltd, FIIton, Bnstol, UK
Advances m Manufactunng Technology XII Proceedmgs of the 14th Natwnal Conference on
Manufacturmg Research London ProfessiOnal Engmeenng Publishing, 1998, pp. 777-783

ABSTRACT

Thts paper descnbes the methodology used m a dtmen~wnal control analysts case study undertaken m
conJunction With Bnllsh Aerospace Airbus Ltd (BAAL) The methodology was apphed for the assembly
vanal!on modellmg of a carbon fibre composite (CFC) Wing box structure developed for future c1v1llarge
atrcraft proJects W1th the planned mtroductwn of large scale CFC components mto future atrcraft
structures, BAAL are usmg dtmenswnal control analysts to mvesttgate and evaluate new manufactunng and
assembly tecluuques and processes Focusmg on tolerancmg and assembly operations, both components and
fixture have been mvest1gated to IdentifY areas where d1menstonal quahty could be Improved further The
methodology IS closely supported through the use of an advanced Simulation software tool and requrres a
systematic approach from des1gn to manufacture The assembly vanal!on stud1es are currently assistmg
BAAL w1th both manufactunng and assembly process evaluatiOn leadmg to the des1gn optmusatton of both
component and toohng entities.

l. INTRODUCTION

Sources of vanal!on (d1menstonal n01se) need to be 1den1Ified and managed 1f a1rcraft manufactunng
orgamsatwns WISh to become more competitive m the global market place Vanahon can be found m
manufactured components and subassembhes, assembly toolmg and fixture, the
deslgn/manufactunng/assembly/mspectton processes, and human mterventton, (1).

Tradtttonally, destgn engmeers are pnmanly concerned wtth tssues of style, functiOn, and structural mtegnty
The commumcatton of tolerancmg, datum and locatiOn schemes of the product then becomes a secondary
concern, (2) An unportant part of the des1gn process requrres the deterrmnatton of nonunal dunens1ons and

225
the apphcatwn of tolerances. These need to be admtmstered such that each mdtvtdual part wtll meet the
performance mtent of the destgn and reqmrements of mterchangeabthty Dtmenstons therefore not only
spectfY the stze and shape of an object, but they mamtam the destgn mtegnty of the part dunng
manufactunng and assembly, (3 and 4) Many of these tssues are covered by blanket company and
mternatmnal standards and are then left to vanous mterpretahons by the down stream acttvtttes of
manufachlnng, toohng and assembly bmldmg m unnecessary cost mto some areas Thts has led to the
extstence of the 'htdden factoty' (5), where even at the late stages of productiOn operators and fitters are
"adjustmg" out vanallon problems The result of thts ts wasted lime and money, as well as comprotrusmg
product quahty

The mtroductwn of dimenstonal vanatwn from upstream manufactunng processes m addttiOn to poorly
defined locallon and datum schemes may produce maJOr problems for an assembly orgamsatwn The
dtsctphne of dtmenswnal control ts an engmeenng methodology that manages vanatton from the early
product defimllon stages through to full production Computer 3D stmulatwn tools are used to predtct how
much and where vanatton will occur m an assembly due to component vanatmn, assembly methods, and
assembly sequences, (6) A maJor objecllve of a dtmenswnal control methodology ts to expose the htdden
factory to the whole organtsatwn creatmg a network of cross functional acttvtttes that actively promotes
engmeenng concurrency

The ann of thts paper is to report a parttcular methodology adopted for a geometnc dtmenswnal analysts
study on a development wtng box structure, to htghhght why the assembly analysts was necessary, and what
was achteved The assembly analysts techntque was undertaken to mvesttgate and evaluate dtmenstonal
control ISsues surroundmg the use of new productwn techmques, technologtes and processes m the
apphcatwn of large CFC components mto future large ctvtl atrcraft structures at Bnttsh Aerospace Airbus
Ltd, (BAAL)

2. THE ASSEMBLY VARIATION MODELLING APPROACH

The appropnallon and preparatton of data necessary to facthtate the assembly vanatwn modelhng reqmres a
systemallc approach for robust results In thts case study the nature and the source of the data are outlmed
and key pomts are htghhghted The acquiSitiOn of appropnate data was lime consummg and tterattve but
became a maJOr contnbutor m helpmg focus m on factors potenttally effectmg dtmenswnal quahty pnor to
the assembly vanatwn sunulat10n and analysts Due to the amount of preparation time reqmred to undertake
an assembly analys1s, a structured approach m the mvesttgauon, preparation and presentatiOn of data wtll
save time, reduce errors and produce a clear data audtt tratl Both rnanufactunng and assembly process
knowledge (process lnmts and the expected vanatton between these levels) needed to be captured for
components and toohng at each related assembly stage

The assembly modellmg, s1mulatton and analysts were undertaken usmg the Vanation Systems Analysts
VSA-3D (7) software product embedded m the CATIA Computer Atded Engmeenng (CAE) system and

226
operated on an IBM RS/6000 AlX platform The VSA-3D product has several sections winch rely on
CATIA's FunctiOnal Dunensmnal and Tolerancmg (FD&T) module to generate feature, datum and tolerance
data

The development CFC wmg box programme at BAAL IS at an advanced stage of development Physical as
well as functiOnal detail architecture has been defined for component and fixture geometry As a
consequence, the followmg data was available for the assembly simulatmns

Detailed 3D CAD models were made available for use m CATlA for all geometry With assembly sequencmg
defmed for the development structure
The development CFC Wing box structure detailed product dimensional objeCtives were known (targets and
metncs)

In the early stages the process ts sequential but later becomes more concurrent and tterattve The assembly
modellmg process was divided mto the folloWing twelve steps.

(i) Define the engmeering challenges


The challenges were highlighted and documented by the composite Wing programme engmeermg team
resultmg mamly from work undertaken previOusly on prototype bmlds and from expenmentatmn Defimng
the engmeenng challenges allowed the formulation of aims and objectives for the analysis Tins was an
Important stage which detenruned the 'focus' and 'context' for the analysis

(ii) Define aims/objectives of the assembly analysis study


From the documented challenges a set of analysis auns and objectives were developed These defined the
analysis boundary (both physical and functiOnal) not allowmg any one model to become over complex The
challenges were grouped and several analysts studxes were plalllled, each havmg tts own set of atms and
objeCtives

(tii) Deterntine the analysis zone


Phystcal zones were created for the analysts and only the components havmg an effect on the dtmenswnal
quahty Withm that zone were mcluded The boundanes of the zones were determmed by the aims and
objectives of the analysis, but encompassed all physical components and elements of manufactunng and
assembly toohng potentially havmg an effect on the 'focus' of the mvestlgatmn

(iv) Identify all the key features for components and tooling
These were the physical features (areas of surface geometry) owned by components, manufactunng,
mspectiOn and assembly toohng that mfluence the dunenswnal quahty aspects of the zone under analysis
Component features can be spht mto the followmg categones

• Features that are associated With the manufacture of the component part
• Features assoctated With the locatiOn system

227
• Features associated with component and fixture mterfaces

• Features associated With product dllllenswnal quality control


• Features to be used as a component and fixture datum

Some features shared more than one category An example of this IS when a plane, a hole and a slot m a
component shares the status ofbemg both a locatmg and datum feature set

(v) Quantify manufacturing and assembly process variation


This was undertaken for all component parts and assembly tooling For this to be aclueved, knowledge on
all manufactunng processes used to produce the mdlVldual component and assembly toolmg key features was
reqmred This data was made available from manufactunng and assembly development process mspectwn
data

Assembly SimulatJOn mcorporatmg unreahsttc stattsttcal formulae of tolerance vanatwn can gtve mtsleadmg
results It IS Important to recognize that steps need to be taken to adequately ascertam manufactunng process
data through current SPC progranunes or other inspectiOn data systems Manufactunng and assembly
processes at BAAL were bemg measured for capability limits (Cp), how well the process IS under control
(Cpk), and how adequate samples are bemg diStnbuted Withm those limits (diStnbutwn type)

The use of the Geometnc Dunenswmng & Tolerancmg (GD&T) standard Yl4 5M-1994, ASME (8),
allowed each component to be descnbed m clear and conciSe geometriC and mathematical terms av01dmg
ambtguous tolerance and datum commumcatwn The FD&T function m CATIA automatically detemunes
the feature type and only allows the placement of datum and tolerances m accordance to a GD&T standard
Datum and component tolerances are defined m conJunctlon wtth feature defimtJons ensunng that no datum
and tolerance call outs can be created v10latmg a gtven GD&T standard

(vi) Deternune the sequence of component assembly


For each analysts study a sequence was defined representmg m whtch order components were assembled
together The components were taken from a parts hst defined at an earlier stage and were represented
diagranunatically as an assembly 'tree,' with the components and subassemblies diVergmg mto the total
assembly for that stage The assembly tree defined whether components were located directly through Jigs
and fixtures (mdependent locahon, v1a holes, pms, offsets), located VIa other component parts (!at bu1ld, VIa
component mterfaces, offsets) or located via a combmat1on of the two The assembly tree allowed
tractability at every stage and provided a tolerance map h1ghhghtmg a 'vanatwn route' progressmg through
to the final assembly via component and fixture features

(vii) Define assembly method (assembly moves) for each component part
To en~ure a dtmenstonaily ~table butld each component was con~tramed m all SIX degrees of freedom (DOF),
three m translatwn and three m rotatwn Each component was analysed for condtttons of over constramt
(more than SIX DOF) or under constramt (less than SIX DOF) If either conditiOn exiSted the assembly

228
'phtlosophy' for that component was invesltgated and revtsed where ever posstble to ensure the correct
constramt status m 3D space

VSA-3D was used to facthtate the component to fixture assembly moves usmg the followmg software
templates

Datum target based moves thts allows the placement of an unlmuted number of datum target pomts on
component and fixtures to facthtate the assembly of curved and complex features
A8'embly feature based moves allows the assembly of feature combmatwns such as hole/hole (vta pm),
hole/pme, pme/slot, plane/plane, or any combmatton of these

For any type of assembly move the nonnal convention references the component bemg moved as the
'ObJect' and the componentiftxture mto whtch tt ts to be moved as the 'Target'

(vtii) Define the analysis measurements to be undertaken


The VSA-3D software was used to measure the followmg enttltes of the development wmg box-

Gap clearance the gap condttton between nb bottom edge and skm surface to factlttate mechamcal fastener,
film and paste adhestve bondmg assembly processes
VIrtual hole condttton the vtrtual condttiOn between skm and nb holes to factlttate mechamcal fasteners

(ix) Undertake tolerance analysis simulation


When all the proceedmg steps had been completed the assembly of the wmg box sectiOn was stmulated 2,000
times for each scenano Three types of tolerance analysts were undertaken for each measurement

Monte Carlo a random number generator whtch ts related to a dtstnbutton type


Htgb-Low-Medmm thts stmulatton holds each mdtvtdual tolerance m turn at tts htgh, low and medtum level
whtle altematmg all remammg tolerances through all combmatiOns ofhtgh, low and medtum levels
Extreme case stmulates the butld settmg all vanatton at the wtdest level, i e the worst combmatwn to
produce maximum vanatwn m an assembly

(x) Analyse the simulation results


When the stmulatwn was completed a report was comptled by the VSA-3D software for each measurement
analysts Each type of analysts undertaken was presented m tts relevant report format The Monte Carlo
(MC) stmulatwn presented the results in a dtstnbutton and provtded mformatton on the magmtude of
vanatiOn assoctated wtth each measurement The Htgb-Low-Medtan (HLM) results were presented as a
Pareto analysts mdtcatmg the percentage contnbutwn to each tolerance m the analysts m relatiOn to the
vanatwn m the output measured MC and HLM analysis were undertaken for each measurement.

229
(xi) Revise parameters and repeat process for further investigation
Several analysis studies were undertaken of defined assembly zones Subsequent studies of each zone were
undertaken With the dtfferent tolerance parameters rev1o;;ed to explore the resultant effect upon the assembly
structure This process of rerunnmg the analysis models With different tolerance parameters allowed the
assessment of maJor contnbutors to geometnc vanatwn It also helps m defimng the most adequate
tolerances reqmrcd to achteve a hrgh level of confidence of assembly through the relaxatton of tolerances
that have little contributiOn and l!ghtemng of tolerances that have a maJor contribution

(xii) Document the analysis results

The final stage was to document the analysis results of the different zones m a formal report. The report
should mclude clear data trails of where the reqmred mformallon was gamed for each step of the approach,
therefore m the event of mconclus1ve results these data sources may be reviSited A typical analysis report
should contam mformat1on such as objectives, assumptiOns, modellmg methodology, draWing/CAD data,
results and conclusiOns

3. CONCLUSION

This paper outlines a systematiC approach used for the modelling of manufactunng, tooling and assembly
tssues used m support of the compostte wmg progranune at BAAL These proJects contmue to ensure the
successful development and productiOmsatiOn of future large civil aircraft and provide a base for the
development of advanced manufactunng and assembly technologies The dimensional control approach of
assembly analyM~ ts currently ass1stmg m tdenttfymg, quantlfymg and managmg sources of vanatwn
potentially mtroduced to CFC wmg box structures

Assembly vanatwn modelling reqmred both product and process knowledge most of which had been
charactenzed Such data IS formmg the basis for a better understandmg of the products control and
development resultmg m Improved robustness The credibility of any simulation result to be used for
component, subassembly and tooling design or redesign purpose relies upon model validatiOn. Tins m turn
Will be dependent upon clear audit trails m the acqUisition of data

The case study expenence also pomts to the need for a multifunctiOnal team approach for effective product
and process design and revisiOn The aim of mtroducmg the assembly vanatwn modelhng process at BAAL
IS to charactenze process and product capability and to make this mformatwn available to all engmeenng
functiOns m so far as this will enable vanatton management to be exercised at all stages of arrcraft
development and productiOn This dimensiOnal control methodology IS ensunng that tolerance verses
vanal!on related data can be fed back to the design, manufactunng, assembly and mspechon areas of the
orgamsatwn This Will provide essential mformatwn at all stages of product development necessary to
ensure total aircraft robustness

230
REFERENCES

L McCmstton P. J, 1994, "The role of dunenstonal analysts wtthm concurrent engmeenng", OhJO
Uruversity, USA (Conference paper)
2. Craig M, 1996, "Lumts oftolerance", Manufactunng Engmeer, Vol 75, No 3, 139-143
3 Henzold G , 1995, "Handbook of geometnc tolerancmg- Design, manufacture and mspectwn",
John Wiley & Sons, England, UK, ISBN 0-471-94816-0
4 Spons M F, 1983, "Dimenswrung and tolerancmg forquahtyproductwn", Prenllce-Hall, New
Jersey, ISBN 0-13-214684-3
5 Leaney P G, 1996, "Design for dimensiOnal control", m Design for X - Concurrent engmeenng
Imperatives, Ed G Q Huang, Chapman & Hall, UK, ISBN 0-412-78750-4
6 Craig M, 1992, "Controlling the vanatwn", Manuf Breakthrough, Vo11, No 6, pp. 343-348
7 Vanauon SimulatiOn Analysis Ltd, 1993, "VSA-3D Analysis Reference", St Clair Shores, Ml,
USA (Product brochure)
8 ASME, 1995, "ASME Y14 SM-1994, Dimenswnmg and Tolerancmg", The Amencan Society of
Mecharucal Engmeers, New York, USA, ISBN 0-8273-4694-8

231
Appendix 5
Dimensional control as an integral part of next
generation aircraft development

D JEFFREYS and PG LEANEY


Department ofManufactunng Engmeenng, Loughborough Umvers1ty, Le1cestersh1re, UK
Proceedmgs of the lnslltutwn of Mechamca/ Engmeers Part B Short commumcatwns m manufacturzng
and des1gn London Proc Instn Mech Engrs Vol214 Part B

ABSTRACT

The next generation of both mthtary and CIVtl a1rcraft are now bemg des1gned and developed Research
undertaken by the authors has htghhghted parttcular product and process charactenst1cs that are becommg
more cnttcal m the destgn and manufactunng of modem and future atrcraft structures A parttcular
reqUirement IS for reduced assembly vanatwn Thts reqmrement emerges, for example from design
strategtes seekmg to maxmuze the structural use of carbon fibre compostte matenals, from the mcreasmg
requirements for low observabthty m mthtary arrcraft, and from the need for fltght cnuse efficiency m ctvtl
atrcraft Large compames such as Boemg, McDonnell Douglas, BAe Systems and A1rbus have all perfonned
shld1es m dtmenswnal control One theme that emerges ts that dtmenstonal control must be exercised wtth
destgn and manufactunng workmg m close partnershtp Thts IS particularly true when constdenng the
mtroductmn of advanced matenal technologtes such as carbon fibre composttes and m the apphcat10n of
automatiOn to maJor atrcraft sub-assembhes Such challenges are found to occur across both nuhtary and
ctvil sectors Thts commumcatmn highlights some such challenges and tdentlfies a seven-pomt response for
the baste technologiCal mfrastructure to support the deployment of an effecttve dimenstonal control
methodology as an integrated part of the next generatton arrcraft product development process Further work
ts m progress on detathng and demonstratmg such a methodology

1. INTRODUCTION

Study tours of US aerospace compames by the authors has htghhghted a number of tssues relatmg to new
atrcraft structure development programs (1), (2) These tssues relate to the destgn, manufacture, mspectton
and a'\sembly processes related to next generatiOn mdttary and civil atrcraft The product and process
characteristiCS associated With these arrcraft are more radtcal and demandmg when compared to current
destgns, partly tnggered by the mtroductwn of new structure matenals such as carbon fibre composites

232
This trend towards the use of advanced matenals has also been the focus of the CIVIl sector for the last few
years for the next generatiOn of commercial aircraft Such developments present challenges, and
l
opportumttes, to engmeers across the whole spectrum of dcstgn to manufacture In today's competitive
aircraft mdustry, new design, manufactunng and assembly methods are bemg developed to lower costs and
provide a more consistent product (3) One of the biggest challenges facmg engmeers IS to ensure the
control of product dtmensmnal vanatwn Sources ofvanatwn (dtmensiOnal nOise) wtll need to be Identified,
quanllfied and managed If aerospace orgamsat1ons Wish to pursue a competitive edge m the global market
place Vanatmn extsts m all manufactunng acttvttles and IS evidenced m manufactured components and
subassemblies as well as assembly toolmg and fixture and all design, manufactunng, assembly, and
mspectton processes.

TraditiOnally, design engmeers are pnmanly concerned With Issues of style, functiOn, and structural mtegnty
The commumcatwn of product tolerance, datum and location schemes then becomes a secondary concern
(4) However, an important part of the design process is the requirement for the detenmnatwn of nommal
dimensiOns and the applicatiOn of tolerances These need to be admm1stered such that each mdividual part
will meet the performance mtent of the design Dimensions and tolerances therefore not only specifY the
allowable size and shape of an object, but they mamtam the design mtegnty of the part dunng manufactunng
and assembly (5) Many of these Issues are madequately covered by design through the use of blanket
tolerances from company and mtemattonal ~tandards and are then left to mterpretatmn by the down stream
acllvities of manufactunng, mspectwn and assembly This can lead to the existence of the 'hidden' factory
where components and subassembhes undergo 'adjustments' to facilitate assembly (6) The result IS
excesstve product and process n01se bemg mtroduced to a development program and ts a maJOr mhtbttor to
the mtroductton of more advanced automated assembly technologtes Dtmenswnal variatiOn whtch ts not
adequately dealt With dunng design and component manufacture, m additiOn to poorly defmed feature
locatiOn and datum schemes, Will mev1tably produce concerns which may only become evident late m the
assembly stage (7) The diSCipline of dimensiOnal control IS an engmeenng methodology designed to
address these Issues through the management of vanat1on from early product defimtwn stages nght through
to full senes productiOn It IS the fundamental reqmrement that must be addressed If true Design for
Manufacture and Assembly (DFMA) IS to be achieved

2. TRADITIONAL VERSUS NEXT GENERATION MILITARY AIRCRAFT

In the military sector the design and development of programs such the jOmt stnke fighter (JSF) has heralded
a departure from conventiOnal aircraft structure design A major objechve of the JSF project IS to develop
and demonstrate advanced technologies, which enable the affordable development of next generatiOn stnke
weapon systems (8) In addiiion, the JSF project auns to Iden!IfY the Impact of affordable next genera!Ion
axrframe destgns m terms of thetr suttablltty for multt-servtce conunonahty An ongomg program called the
'advanced lightweight aircraft fuselage structure' (ALAFS) has already begun to explore some of these
Issues usmg new design methodologies which mclude extensive use of carbon fibre composite (CFC)
matenals, see Ftgure 1

233
There IS a general trend to mcrease the structural efficiency of aircraft sub-assembhes Thts ts an obvtous
goal tf reduced wetght and cost targets for new structures are to be reahsed. Even after many years of
advanced research atrframe sub-structures and extenor skms are sl!ll mamly assembled usmg large numbers
of mecharucal fasteners such as nvets and bolts Sub-structure and skm components are generally dnlled
usmg 'hard' (fixed) Jig/fiXture toohng templates and the holes facthtate both assembly locatiOn and fastener
msertton access features (9) Next generatiOn arrcraft are movmg more and more away from these traditiOnal
technologtes towards hghter matenal technologtes and mtegral and bonded structures Such approaches
should not be done at the expense of dnnensmnal control but wtth careful constderatmn of dtmenswnal
control, tf the objecttves of reduced cost and wetght are to be achteved Thts move has allowed the
development of 'monocoque' type atrframe structure where a stgrnficant proportiOn of the dynamtc m fhght
loads are earned by CFC outer panels. In some mstances these advances m technologtes can be
contradtctory, for example carbon fibre composttes have the potenl!al of stgmficantly reducmg component
wetght but the matenal ttself lnstoncally has suffered greater vanatwn than metal maclnned components
Care must be apphed dunng destgn to ensure the benefits are explmted to the full whdst mtmmtsmg the
tmpact of the shortcommgs of some technologtes CFC panels must also be capable of mter vanant
mterchangeabthty, 1 e , any CFC panel of an atrcraft vanant must fit any other aucraft of that same type
These CFC panels wtll also need to be removed and replaced effictently m order to allow access for servtce
and repatr Tins has led to a destgn trade off between relaxed tolerance spectficatmns to facthtate CFC panel
to substructure mterchangeabihty, versus tighter tolerance to satisfy the transfer of stress loadmg between
mechamcal fasteners, the CFC panels, and correspondmg substructures

The mcorporal!on ofCFC matenals m both the outer flymg surface panels and mtemal substructures ts hkely
to be substanl!al Unhke conventwnal matenals such as alummmm (AI) and l!tamum (Tt), CFC matenals
wdl not readdy 'conform' vta dtstortmn of component by pressure dunng manufactunng and assembly
process fmesse. Where AI and Tt substructure subassembhes are to be used thetr part count !S hkely to be
reduced, typtcally up to 80% compared to tradtl!onal atrcraft ThiS wtll result m each structural component
becommg more geometncally complex, substanl!ally stronger, and possessmg more free state ngtdtty

3. TRADITIONAL VERSUS NEXT GENERATION LARGE CIVIL AIRCRAFT

The clV!l arrcraft manufactunng gtants such as Boemg Commerctal Alrplanes and the Atrbus Industry
consortmm are also conS!denng the next generatiOn oflarge atrcraft deSign and development concepts As
prevwusly mentwned thts aerospace sector !S alongSide the mdttary sector wtth respect to reducmg assembly
part counts, m the reductton of m process toohng and fixture, and the more extensive mtegratlon of CFC
matenals mto pnmary structural assemblies to reduce total aircraft mass

234
The successful use of CFC matenals for some arrcraft components has been evtdent for some years It has
been uhhsed for verttcal stabdtser components as well as the pnmary structure matenal for honzontal
stabthsers (Atrbus lndustnes bemg the lead of the mdustry m both these ca•e•) Atrbus are explonng the
posstbthty of producmg a wing box conshUcted predommantly from CFC matenal, Ftgure 2, a stgmficant
departure to tradttwnal wmg conshUchon whtch conststs of mostly AI and Tt matenals Tradthonally, wmg
box assembly throughout the Civil aircraft manufactunng sector mvolves the use of techmques such as
'fetthng', the removal of excess matenal from a component or subassembly feature Another techmque
known as 'shtmmmg' mvolves the deposttmg of matenal upon a feature surface to facthtate an effechve
subassembly mterface Shtmmmg techtuques are prunanly used by the mthtary sector of the mdustry wtth
the ClVll sector relymg m the mam on fetthng These techmques have evolved because certam manufacturmg
processes assoctated wtth wmg components and thetr subassembly mterface features cannot be controlled
well enough to achteve the target spectficatwn wtthout subsequent processes The need for such processes ts
reducmg as greater understandmg ts gamed as well as maturing of metal manufactunng technologtes m terms
of achievable tolerances

The mtroductton of CFC matenals wtll greatly reduce the feastbthty of fetthng techntques due to the
constrammg nahlre of the matenal Gtven thts, much greater control of component to subassembly mterface
features wdl be reqmred for a range of new and advanced assembly processes to be deployed, such as film
and paste bondmg for example

4. THE CHALLENGES FOR FUTURE AIRCRAFT PROGRAMS

The next generation of both mthtary and Civil mrcraft will be designed usmg novel concepts, new matenals,
and revtsed manufactunng and assembly strategtes It ts only through these stgmficant step changes m
technology can today's challengmg cost and wetght targets of the atrcraft customers (both mthtary and ClVll)
be fully met The challenge to new atrcraft programs can m part be charactensed by the followmg

• Integratton of dtgttal destgn and test envtromnents · VIrtual product development (VPD)
• New atrframe shUcture destgn concepts mcludmg htgh levels of atrframe comrnonahty and the
mtroductwn of 'monocoque' type CFC structures.
• Introduchon of new matenal types such as CFCs and radtcal manufactunng processes
• Development of unproved arrcraft mamtenance systems to reduce owner-operatmg costs.
• Mtgrahon from physteal to dtgttal master toohng gauge (MTG) systems - thts heralds a move from
a checkmg to a measunng philosophy Furthermore, quahty mspect10n acttvtttes should be
developed followmg a 'lean' phdosophy where atrcraft key charactensttcs remam the focus
• IntroductiOn of complex component and assembly mspectwn analysts techtuques
• Large mvestment reductwns m manufactunng and assembly toohng and fixture
• Component-to-component assembly strategy- the reliance on a component to locate the next
component wtthm an assembly process resultmg m mmtmal fixture (so called Jtg less assembly).
• Much greater emphasts on product and process costmg

235
5. ADOPTING A DIMENSIONAL CONTROL METHODOLOGY FOR NEXT GENERATION
AIRCRAFT DESIGN AND DEVELOPMENT

Some of the new product and process concepts to be mtroduced m the next generatiOn of m1htary and ciVIl
arrcraft have been h1ghhghted It IS proposed that some secl!ons of the aerospace mdustry will need to re-
evaluate a number of these engmeenng busmess processes m order to accommodate these new concepts The
need can m part be addressed through the deployment of a developmg d1menswnal control methodology to
hnk the engmeermg effort from product reqmrements, design, manufaci!!Te, through to full senes productiOn
In fact some of the pomts Identified below have already been implemented m large scale apphcatwns and the
benefits of these approaches are already startmg to be reahsed m practiCe However full ImplementatiOn of
the whole methodology still reqmres Significant effort to become reahty and only then will the full cost and
quahty benefits of the approach Will be fully reahsed A seven pomt response IS proposed as a prereqms1te
for the methodology, these bemg·

Adequately and effectzvely characterzse manufacturmg and assembly processes. the mtroductwn or
evaluatiOn of current stal!stical process capab1hty (SPC) data for all manufactunng and assembly
processes SPC has been known and Implemented for some !!me although generally post design
The ImplementatiOn of an SPC methodology early m the design process Will enables greater hnk
between design, manufacture, and assembly, ensunng that the destgn process ts complementary to
the down stream process to be employed
11 Develop new, and revzse exzstmg, manufacturmg and assembly processes, the manufacture of
components from CFC provides a number of engmeenng challenges, mcludmg the tight control of
vanat1on and the large scale mtegratwn across the manufacture and assembly of the CFC struci!!Tes
Thetr use IS relatively new for prunary structural components and maJOT sub-assemblies The
contmued development of these technologtes IS leadmg to functiOnal and cost Improvement for a
vanety of new apphcatmns
m Reduce manufacturmg and assembly )lg/future 'hard~ toolmg requzrements. a number of
manufacturers are now usmg large preciSion (5 ax1s) h1gh speed machmmg fac1hties Terms such as
hole to hole and part to part assembly are often used to descnbe th1s design concept These faciht1es
are bemg used to dnll!countersink and net profile CFC panel components m preparatiOn for
mechamcal fasteners The machmmg of accurate features for component-to-component assembly
strategies Will significantly reduce the need for location and fixture tools
tv Elzmznatzon ofphyszcal MTG and dtgllal master medza. the mtroduct1on of precisiOn and flexible
measurmg eqmpment IS already showmg stgmficant cost benefits m atdmg the mspectwn process of
jigs w1th respect to digital MTG, and hence the need for physical master toohng and statiC storage
sites can be ehmmated Th1s has reduced the requrrement for physical MTGs' These systems are
based on a number of technologies one ofwh1ch IS laser mterferometery (SMART 3 I 0 system from
Le1ca, for example) With dynamic trackmg capab1hty These systems can be used m conJunCtiOn
With the digital master models created Withm the m-house CAD/CAM/CAE envrronments
v Develop dtgztal mspectzon techmques; thts Will provide a process for component, assembly and
fixture toolmg d1menswnal quahty venficatwn The master digital data stored Withm the
CAD/CAM/CAE environment can be ul!hsed as the nommal product med1a
VI Introduce geometrzc dzmenszomng and toleranczng (GD&T) concepts, the use of a GD&T standard
provides a means of commumcatmg design mtent through the manufactunng and assembly
operahons wtth the mmtmum of ambiguity GD&T IS not about tighter control oftolerances but
more about controlhng tolerances appropnately, based on the function that feature plays m the fmal
product, 1 e where a component feature plays a key role m either the assembly or final function of
the product then particular attenl!on can be pa1d to the control ofth1s feature. Features that have
neghgible effect on performance or assembly can potenl!ally have their tolerances relaxed enabhng
reductiOn of manufactunng costs related to that feature This IS a Significant advancement over the
apphcatwn of blanket tolerances to both design and manufactunng actlvii!es
vn Efficwnt use ofvarzatzon analyszs CAD/CA£ zntegrated szmulatzon tools, e g With GD&T
specifications available m digital CAE systems 11 IS possible to undertake 30 vanatwn analysis

236
studies on cnt1cal assembly Issues Such techniques truly enable the manufactunng costs to be
evaluated Withm the product design and help to fac1htate true DFMA

Aerospace orgamsatwns are already utiliSing mtegrated product development (IPD) strategies dnven by
mtegrated product teams (JPT), (I 0) Part of the d1menswnal control process reqmres product and process
mfonnatwn from cross-functional and multldtsctplme engmeenng sectiOns making concurrent engmeenng a
fundamental requirement

Due to the reductiOn of substructure component parts m the mthtary and ctvtl aucraft sectors, a new
assembly philosophy IS emergmg This philosophy a1ms to ehmmate a Significant proportiOn of assembly
fixture toohng through the apphcatwn of new 'self toohng/locatwn' pnnc1ples Each component and
subassembly Will be produced With mherent self locatmg features and toolmg features, m addition to 1ts
nommal functional arclutecture These features wtll be used to facthtate component to component location
for assembly purposes Dimensional control techniques can agam play a maJor part m the design of
mdtvtdual components and mtmmal toohng, therr features and thetr functwnal tolerances The process
should be used to evaluate the nsk of a new design concept from component and assembly dimensional non-
conformance and the assocmted manufactunng and assembly costs that may be mcurred tf thts controlts not
mamtamed

6. CONCLUDING REMARKS

A study of a number of aerospace compames has revealed that the dimensiOnal control disc1phne needs to be
exercised early m the design phase such that the manufacture and assembly charactensllcs of senes
productiOn aircraft are effectively taken mto account Thts philosophy ts consistent With the concurrent
engmeenng Ideal and Its supportmg techniques, DFMA, for example Airbus and BAe SYSTEMS have
made stgmficant advancements over recent years and are pushmg further technology developments m these
areas. This has been precipitated by the need to develop aircraft of Improved quahty, With lower
development program costs, and m the hght of matenal and technology developments

The successful deployment of a dimensiOnal control methodology reqmres a mulllfimctwnal and mtegrallve
team approach for effecllve product and process development This Will reqmre product and process
knowledge from a dtverse number of engmeenng busmess centres and will need to mclude mfonnatiOn such
as process capability md1ces and process costs A seven-pomt response has been proposed as the bas1s of the
technological mfrastructure that rehes mcreasmgly on the extended use of digital models and digital
mspectton Further work IS m progress on detathng and demonstrating an appropnate dtmenswnal control
methodology

237
REFERENCES

P.G. Leaney, ~Assessmg best practice m dtmenstonal management and control ofvanat1on m
automobile bu1ld' !PPS Report #94/2, Department ofManufactunng Engmeenng, Loughborough
Umvers1ty, England, Nov/94
2 D. Jeffreys, 'Assessmg current best practice m dimensiOnal management techmques for atrcraft
bmld' !PPS Report #96/3, Department ofManufactunng Engmeenng, Loughborough UruvefSity,
England, Nov/96
3 Muske, S. Apphcatwn of d1menswnal management on 747 fuselage Proceedmgs of the 3'd semmar
on tolerancmg and assembly modellmg Umverslty ofM1ch1gan College ofEngmeenng. Sesswn
III. 1997
4 Craig, M. L1m1ts of tolerance Manufactunng Engmeer. 1996,75 (3), 139-143, (InstitutiOn of
Electncal Engmeers)
5 Henzold, G. Handbook ofgeometncal tolerancmg des1gn, manufacture and mspectwn, 1995
(John WJ!ey & sons Ltd, Chichester)
6 Leaney, P.G. Design for d1menswnal control In Des1gn for X- Concurrent Engmeenng
Imperatzves (Ed G Q Huang), 1996, 173-195 (Chapman & Hall, London)
7 Jeffreys, DJ., and Leaney, P.G. A dimensiOnal control methodology for the vanat10n analysis of
arrcraft structure assembly, Advances m manufactunng technology XII, Proceedmgs ofthe
Fourteenth Natwnal Conference on Manufacturmg Research, Derby, Umted Kmgdom, 7-9 Sep,
1998, 777-783, (Eds R W Bams, A Taleb-Bend~ab and Z Zhao), (Professwnal Engmeenng
Publish mg Ltd)
8 Joint Strike Fighter. Advanced bghtwe~ght azrcraft fuselage structures (ALAFS) 2000
(URL http //www 1ast miiihtml!body ala(• htm)
9 Fowler, K.J. Apphcat1on of tolerance analysis techniques to aircraft structures, Proceedmgs oflEE
I !MechE (Manufactunng Dmswn) ColloquiUm, Control of d1menswnal vanatwn, The Midlands
Engmeenng Centre, Austm Court, Bmnmgham, 4'" November, 1997, (lEE d1gest No 97\346)
I 0 Barrow, P. ImplementatiOn of new planmng and manufactunng processes In The manufacturmg
challenge m aerospace, 1997,37-55, (MechaniCal Engmeenng Pubhcatwns Limited, London)

238
Appendix 6
Stage I VSA panel to substructure tolerance
analysis results

The complete tolerance analysiS output measurements for stage I panel to substructure assembly
(see figure I) are presented below The outputs consist of process charts and HLM reports
generated form the VSA software on step, gap, and VIrtual hole conditions. An explanation of the
analysis process IS available m chapter 8 An explanatiOn the analysis results can be found m the
next section entitled 'VSA process reports'

Analysis measurements: the followmg areas were Identified for mvestigatwn These mcluded
step, gap, and virtual hole condition measurements

Figure I. Stage I panel to substructure analysis measurement pomts

239
~
-~
-·~
l'>ouu\I........ Q(flJJ

-- -
to~V-.oe OOOW

~
'~ ·~ ••
1 PHASEI P<W£1 FUP
>C>RI"~i>'IClM1 ""
PHASfi_WQSIRLJCTUFll: _,......., ~
""
··~
> Mnn01~ """0015.1POS I0 .. 002'5[MtiAIWo!ll
J'HAS[I_SU!ISIROCIUFlE ...... 8
> I~Pf WO<OIAIB!MJII;p.IJI
0~
·= I "~
~
PHASEI_fWIEUUP - -
> Mtt OlW !o'OO< OOCIO. lf'OS I0 .. 002'5[MjiAIS[Io4JI "" I "~
I'HASCI_SUBSTRLJCIIJFI[ 10<....... .,.._
> MnODI~ lo!'"'-0015. II'EA 10oo002'51AI ""
,_ "" I •m
I'HAS[I_PAN£LFUP ...... 5
>I~PI" IO{I.'OIJIIB(MJICJ.IJI ·= '"'
T PHAS£1_W9SifllJCTUFl(
>OAF IO&ip.tiQMJ "' "' ·~

.... - ·- -·= -·=


~00125
~~--

-·- -·-
·= - ...'"'
S<lOWl . .
..... ~ •w• ·~
,.,.~

.. o..ooos.c ·=
·= ·= ..... ·= •••
·l-"""":'e99no:r'-. -
2«>:l...,._l<IM-

Figure 2 Plane 4 gap measurement

-·..
Pl..oo<oV_.,.Ilw.ll
>u••- noc.m
~-
fl'..\.S/_1 P'lll.UUP
,.....,- l'l»'f;i><l
t•«~r
~

••
.. Jolor_
·-
~
~

-~
·-•.
1W>S!1 StCISIRUCTL!Rf - -
> IOnOIM M... O(J15.1POS 10oo002'5 .. JI"'~II l<sm
Plio>.>l:l SU8SIRtiC1LJR£ ...... ,
>ISPf IOWIIAIBil'llr:Jo<JI ·= I "~
f'>.IOSI:I_P>HEt_fUP ,..... ...._
> "" 0(9) lob. 0000 INS IDooOO;/'l,..JI"'Bil"ll

f'ii,>.$£1 S\.18SIRUC1llfl( - -
> ""0(11~ 10.._001~ II'ER IOooUUo.':>IAI "" ..
~

I
I
"m

'"'
~[I_P.ol<I:LFUP ...... 4
>ISPFIOO<OIAIBII'IICJolll ·~ ·= om
Pli'SEI_SU6SIRUCTUR£
> ORr IAJ![MJ:!MD "" ~

·-
-·=
... . , . _ , 000211

...............
............ ·-
·= ·= ·-
·= -
O..O.ba<n ,.,.. ""_

-·-... ·-
'IO.O..OO<s-

·- ·- ·-
·= ·~
.,.~--
·~
•••
00~

""r.m!l
lo<l;t~)<l~-
'""'
·=
Figure 3 Plane 5 gap measurement

·- ~·-
~
_,..
F'oo<.Tt<v_..,. cco:o

·-
ltLM/_,. llOO'.ll

~. f-1""' ,,,
~
1 FW.SEI_P<>N(l_fUI>
>WII"!"liJ><I£(1.11

IW.S(I_~g<;IRIJCl~ - - .. .. •=
""

·- ·- ·"~.
> -(.DI~ "•01115 lf'OS lltooO(IZ!ifi<I!AI!IIIoO)I

IW.SI:I_Sl•BSIAlJtl~ -~~
>ISI'I' IOJOOIA191"11QNII I ,~,.,.

I'HlS[I_F'IOH£l_n.JP - . - . .
-> lolnDWI><aOWJ.IPO!i IO..UII:I':iP<II.o\1 . . 11 "" "'' I ~

F'liO.S£1 ~9SIRIJCT~ - - - •• ~

..·-
-> loin ~01~ M• 01115. I?lA IO..~Gi':iiAI I ·~
IW.S£1_-l_n.P ....... )
>I >PI' IO.DIIAIEfi"JICfot)l ·~ '~
F'liO.SI:IS.U~IFIJCI~
••
_...
> DAf W!II"IQWI
·~

_,_,.,._
-·-·- ·- - -
~·~
SOO-O"'l'l

....,.,l_L_
,.c..oo-
..,., ..... ~

ow• ·-
·~
.. ~3SI!>oo- 997300'1.
~-~d<I!Ulooch
.....,
Figure 4 Plane 6 gap measurement

240
-~ -- I"~
-·~
-1-C.OCIXI

·-
IUIV~oocm

[~""""'Joi"L S..
IF>IAS!I_F'A"(lRJP
>OAf jlll,l!MIOI-Ol
PIIAS£1 SIJ9SI~LJR( .......,.,_
••
••
••
~ -m
•·=
·-
> -0Cl5M.,.00'1S.IPOSICI>o00ZS(Io<JIAISIMII

·=
0

.
"
'
PHO.>fi_SUIISTllUI;TL.IR( 11Mo11
>1>1'1 100011\iB... )IQN)I

PHA$£1-J...AS'---
> .... oO!I.l ..... Oo:a:tlfoOS IChoO(l2!.lNJIA181"11 •• ••
I
.1'BU
JS•st

Fll'tS£1_SUB';TllUC1LJRI': - . . . . -
> .. n001S "''"' 00'15.1PEA IO .. OOZSIAI
~
••
·~
F'Ho\SEI PlONEl_IUP ..... l
>I >I'! IO!WI""81WJI(Jol)l ·~ '""' ·~
FW..SEl SU9STII\JC1Ul[
>OAF Wll'"lQNI •• •• •m

... ...,Lmt
'10.C".OO!SI>o<
-··-- ·-·-·-·= ...
·- ·--·- -· -50lf0

~··
QOl~

~J--WnOO'I.
llldo:l-.....w<J(I'J';-
"'""...
Figure 5. Plane 7 gap measurement

-- -- -~

..
....... ,...,
-v...,.,. 1lOOOO

-'""' -
>l.lH~{..((..Q

" =
..·~ .,..
FW<>EIPAAEL._IlJI'JOt>n&i

·-
>llff i02WIAIBIMIIQN)I

·-·- ·-'""'
PHAS(IILJIISTIIIJ<:TUfl( - -
>lfll IO<DJI
0
FWISEU'N<H-..flF - -

·~
>lf1.l 10<001

PHO.<;{I_SLJIISTRIJCTUAI' ....,.,
>l'>l'fiOr..IOIAI!liiOJIQiolll ON

-·=..
_,,

-·=· -
.._"""",__ llro:)
.... ~0-

....
... ... ·-·= ·-...
- · - -.. su.
·-
·-
.__~.

~
"= 1107~ ·~· ·~
... o.&O<Spoe 142111:1 '""
o•~•
11!031

"""
141667

Figure 6 Plane 4 step measurement

·- -- --
-
11-..1 lo»l
"'-V-O.!JX~l

-'""' ·-
lii..Nv_... O!Wf

~ :~oo;..,.r.-

,.,..,.,,.....ElfUP-&
> l~l'r IO>illi''I"IMJIUNJI ·=
om
PHAS!:I SUBSlAUCTUFIE - -
·~
·~
>IFLii02001

~~~~~-ll.P-- "® ·~
I'HASEI SU9STR\JCTUR( ...... 5
->I~I'F IO~OIAI~)IQNII
·- '""'
I "a
·~

- ·-.. 107N1
~
-
·
"~
~
·~·
,.
•••
·~ ....
Figure 7 Plane 5 step measurement

241
·-- -~

-
~~- I OD,!
Ploce.\1--(l(.JD'j

·- ·-...
HI)!V-.:o IUW3

I ~I_P.<>Mt,JUP ....... ,
>ISf'F lOlLOIAI~IIQIOU

PHAS£1 S.U8~l~UCTOFI£ - -
,_

·-·- ...
.,llt110.2001 . .m
• f'HAS[I_f>JOH(l,_ruP - -
>lfll 101001 ·~ om
PHASfl_$111lSIROCI!.lfli:-S
.,I~Pr IOO&IIIAI~I>411QMII
·=

-·-- - -
--- "~
t:n<O
D<J59

~~--9'1""""'
121(11
G3111~

Figure 8 Plane 6 step measurement

·-- -- I"-"'
~· .. --

---
"-'V......nOIXOI

·- ...
)ljjy_..,. OOO.Q

FW,S[I f'N*:l._FUP ....,.,


>ISI'f ION:IIAIBJ.IJIQWJI ·-·
·~· ·~

·-·- ·=... ......


A-IA.SEI SU9S1AUCIURE - -
>lfll 10..001
AiOSEI _ _ l.JUP - -
>lflli010CII

I'KOSfi_SIIaSIRI!ClliRt - 7
->l!.f'f"IOOIOIIIIS~JIQI.l)l

·--·--
_,,,06
S O l - OOOS2

.... . ..__u......
... O.OO<Spoo
-..
,
,.=
10101~7

- - ·-
10ll1<7

-
ooom
o~n~
"'~
om•

Figure 9. Plane 7 step measurement.

I -.s.EI PM£l_fUP ""*2 ·~


> Wn 0(8) ,.., 00011. 11'05 ID .. OCI2SJIIIIA1~110Nlt EEl
';"r:·:~~n: I ,_
l ~!'o.i.i'~S~!Ufo\'~£1~ ro.. o0<'5"'11AI~IIl:fiJI "" I ·~
P>llt.S£1_PIII<El fU> - -
> WrtOlW- O(al H'OS ltho00251")1"'S,.Otl
AiOS£1 Wll~liii.JC1URE - -
> NnGI1l5 II•O()lS IPOS IO..U025JII)IAI8Jojtl
""
••
·=
>m
PHilSEI_SlJ!ISIRUtTUAE -~-
> lln GillS 10• 0!115 If'!: A 10.. 0~JAI "" ·~

""n-u..
.... -
'I.O.OOIS..O
..... -·-... - - -·- -
'""
""
'""
OMO
Oni>
·~lSogrooo~ . . .
·~
•••
o... ~
997:JJO'J.
lodd
~

'~<I( 100lr-'>

Figure I 0. Hole 2 virtual dmmeter measurement.

242
~

I ~~!~~
----- ..
n ~~
••' I
... 1 . ---
5R~~~
• . . " .!.
~&

1:1 1:1
"Ei "
~ ~
"""'"' ""'"'
,j
i
Ei
....
:i
i ~
"....Ei
"
;; "
;;
t Ei
! :a"
t! Ei
.a"
Cii Cii
J .......
lmi -~"' J j!!!! -~"'
I I
f M
f V)

0" 0"
m~;
i
:I: w··J
.. !;i :I:

'f
I ~
N
'f
l
M

~
m Hf 6b
.;:; l Hf 6b
.;:;
!!I
u !ili I !! ! !
ll .-·-

!!!! m~
m m !
l 1l ll
p'
Jll lll '! lll
l •3
"
•Hi
'"
l
.. '
JP
'"
-- -·
I ~~~::[~-fl.l:~""i.!os ID,.Il.02\IMIIAIBINJII)I.4)l

I'HlS£1 """'£L fU'


>Pf\F jllll(lojiQNJ 1 •a•zr
PHOSEI_$l!iiSTROC1UfiE - 6
> W.. 0015 ...._ Dlll5 IPOS ID,.DO:ZOSfN]II<IBIM)IC(N]I

PHOSEI_SlJBSTRUtlUfiE - -
'"'
> 10nll0l5- 0015 IPOS 10,.0.0211MIIA1Bfi.IH
PHOSEI f'HIO...Il.f' -,doluo
>,_,DO:O ..... DCW IPQSIO..OUZ'SIMII"UiWJI
PHOSEI_5LF>TRUI;1UIIE _.......,._
> .... 0015- Olll5. II'£R 10.. 00<!>1 ..

-··-·-
F1gure 14. Hole 6 v1rtual d1ameter measurement

244
Appendix 7
VSA process reports

The mformatton contamed m the process report ts

Sample SIZe
The number of Simulatwns performed. It IS Important to do enough sunulallons to allow good estunates of
the mean and standard devmt10n values. You should keep runmng stmulatwns until these values stabthze

Nominal
The value of the measurement when all tolerances are set at thetr nommal values

Mean
The mean for the measurement values produced dunng stmulatwns

Standard Deviation
A statistic that tells you the amount of vanatwn m the measurement values Also ca1led stgma

Upper and Lower Spec Limits (USL, LSL)


Any upper and/or lower design hm1t• from your model

Cp
A process capability mdex relatmg the allowable design vanatwn to the sample vanatiOn m the SimulatiOn
samples

An NIA Will appear If

• Both upper and lower design hm1ts have not been specified
• The number of accepted Simulatwns IS less than 5

Cpk
A process capabthty mdex relatmg allowable destgn vanatwn to sample vanatton, mcludmg a measure of
curve centralness

An NI A Will appear If

• No design lun1ts are specified It IS calculated If only one design hm1t IS specified
• The number of samples IS less than 5

Refer to the Cp and Cpk calculatiOn discussiOn at the end of this appendix

245
Distribution
Thts IS the dtstnbutton that ts fit to the data and bemg used to generate the estimated va]ues shown on the
lower part of the report The followmg alternatives are available.

Actual
When the analysis type IS actual, the data IS tested for normality. If 1t fa1ls the test, the appropnate Pearson
curve IS fit Otherwtse? a normal curve IS used for estimations

Normal
You have forced the analysis to use a Normal curve for est1matmg by settmg the analysis type to Normal

Pearson
You have forced the analysis to use an appropnate Pearson curve for estimatmg by setting the analysis type
to Pearson

Percent <Low Linut


The sample column g1ves the percent of SimulatiOn samples that fell below the lower design hm1t The
estimated column contams the percentage based on the estimated d1stnbuhon curve fit

Th1s value only appears 1f a lower design hm1t was given

Percent > High Limit


Sample and estimated percentages found to be greater than any specified upper design hm1t This value only
appears 1f an upper design hnut was given

Total % Out of Spec


These columns contam the sum of the preVIous two values It IS only shown 1f upper and/or lower design
hmJts were gtven

95. 0% Confidence Interval on the Sample% Out of Spec


This IS the 95 0 percent confidence mterval for the sample percent out of spec value Different SimulatiOn
samples from the same model WJll produce different percent out of spec values If you keep generatmg
samples, usmg a different random number seed each time, 95 0 percent of the time the value you get for the
percent out of spec Will fall m th1s range.

Low
The sample low IS the lowest value produced dunng this set of s1mulatwns The estimated low IS the low
value on the estrmated d1stnbution curve correspondmg to the range settmg (noted at the bottom of the
report) For example, 1f you have the range set to 95, the estimated low value Will be the value at the 0 025
cumulative probabilities on the estimated distnbutwn curve

246
High
The sample h1gh 1s the h1ghest value produced dunng th1s set of s1mulatwns The estimated h1gh 1s the h1gh
value on the estimated d1stnbut10n curve correspondmg to the range settmg (noted at the bottom of the
report) For example, 1fyou have the range set to 95, the estimated h1gh value wdl be the value at the 0 975
cumulative probab1hty on the estimated d1stnbut10n curve

Range
The range for each column IS the h1gh mmus the low

EstRange
The estimated range IS the value the range IS set at based on fittmg a curve to the Simulated data The
range IS used to calculate the numbers in the estnnated column

TAL LAII'I' TO hiONOSU 1 OAP

D1splays the standard deVIatiOn


value (e. g -3S, +3S) when the
estimated range corresponds to
an exact standard devtatton val-
'' ue (e g 99 7300% = 6 standard
Q
dev1at10ns) Otherw1se, URV and
u
LRV (Upper and Lower Range
'
' Value) d1splay The standard
V dev1at10n value and the URV/ LRV
destgnahons only appear when the
Analys1s field m the Report Params
tab IS set to Actual or Normal
Red areas md1cate that the
measurement has exceeded the
Nominal 2 4476 Lower Spec LJm11 1 9731 Cp 0 4251 upper or lower spectficatton lumt
Mean 2 4656 Upper Spec ltm~ 2 9731 Cpk: 0 4186
S1d OevJabcn o 3921 D1slnbllllon Tested Normal

Sample Esbmate Sample Es~mate ..


'l:> ~ LowUm~ 10~000 104577 L~ 11497 1 2893
%~ Ht~hUm1t 96500 97762 Htgh 3 9622 36419
'l.Out01Spec 2(10500 21) 2339 Range 2 8135 2 3527
··1- 3 Stgma Range

Figure I Process chart

HLM Contnbutors Report


The HLM Contnbutors report WJII tell you what tolerances are contnbutmg most s1gmficantly to
the vanatmn m a measurement Thts report gives you a hstmg m decreasmg order of all the
vanables contnbutmg more than a spec1fied cut-off percentage to the HLM vanatwn There IS also
a notation of the number ofvanables that contnbuted less than the cut-off percentage but more
than zero

If no s1gmficant Interactive Effects are present, you WJII be able to reduce any undeSirable
vanat10n m the measurement by reducmg the vanal!on m the tolerances shoWing the largest
contnbuuons Also, tolerances WJth low contnbul!ons may be cand1dates for relaxmg the!f
vanatton restncttons

You can check for the presence of stgmficant mteracttve effects by companng the HLM mam
effect vanat10n shown on the HLM Contnbutors report to the square of the standard deVIatiOn
shown on the process report or the vanatwn value ava~lable on the custom reports If they are
w1tlun I 0% of each other there are no s1gmficant mteractions affectmg the results
The mformallon on the HLM Contnbutors report 1s

HLMNominal

247
The nommal value of the measurement calculated usmg the 50th percentile values of the
tolerances.

HLM Variance
The HLM vanance IS equal to the sum of the measurement vanatwn attnbuted to each tolerance
vanatwn You can check for the presence of s1gruficant mteract1ve effects by companng the HLM
vanance to the square of the standard deviatiOn shown on the process report If they are Withm
I 0% of each other there are no Significant mteractwns affectmg the results

Tolerances
The names of all the tolerances m the model that have an effect greater than the HLM cut-off
percentage on the vanatwn m the measurement bemg analyzed They are only displayed 1fyou
have chosen to display vanable names m the display parameters ullhty

Description
The descnp!Ive comments that are m your model for each tolerance They will be 40 characters
long If you have chosen to display vanable names and 50 characters long otheTWise

%Effect
Percent effect IS the percent of the HLM vanat10n due to the vanatwn of an mdlVldual tolerance

All of the tolerances that contnbute more than the cut-off percentage shown at the bottom of the
report are displayed The length of the report Will be dependent upon the cut-off percentage and
the number of contnbutors

248
Appendix 8
Different methods of constructing tolerance zones in
CAD systems

Parametric dimenstoning and tolerancmg: Parametnc d1menswrung and tolerancing (PD&T) IS used as a
control on scalar related dunenswns such as length or angle Each hnear dimensiOn IS associated With a
fixed nommal value and a vanable component The vanable component specifies a band havmg an upper
and lower hm1t m which the nommal scalar value must he to be w1thm an allowable tolerance. Angular,
which mcludes vectors, are normally related to a radml tolerance referenced from an axts system Vector
tolerances are used to quantify the dtrectton, stze and posthon of a component relattve to an axts system A

shortfall of parametric tolerancmg IS Its mab1hty to control geometric form and Its lack of reference to datmn
systems This shortfall IS met by geometnc tolerances. An example of PD&T can be seen m the figure
below

-20+/-01

T
30+/-02

1~~
k------ --------'1
6 0 +/- 0 2

Ftgure I. Stmple example of parametric dtmenswnmg and tolerancmg.

Geometrical dtmensioning and tolerancing: GD&T IS a prectse mathematiCal language that descnbes the
Size, form, onentatwn, and locatiOn of part features. It's also a design-dimensiomng philosophy that
encourages designers to define a part based on how It functiOns m the final product

Through the use of functiOnal dimenswmng, tolerances are assigned to a part by the designer based on the
part's functiOnal reqUirements, often resultmg m a larger tolerance for manufactunng Th1s ehmmates
problems that result when a destgner asstgns arbttrary, or too tight, tolerances to a part m a drawmg because

he or she doesn't know how to determme a reasonable, functiOnal tolerance

GD&T IS used on an mdividual or a pattern of features for the control of

Form; controls the straightness, flatness, c1rculanty and cyhntnc1ty of mdividual features wtth no reference
to a datmn scheme
Profile; controls profile of a hne or a surface of a feature w1th or wtthout reference to a datum scheme or
other features

249
Orientatwn; controls angulanty, perped1culanty and parallelism of a feature With reference to a datum
scheme and related features.
Location; controls pOSitiOn, concentnctty and symmetry of a feature With reference to a datum scheme or
related features
Run out; the allowable error between two or more features wtth regard to concentnctty, perpendtcularly,
and aligument Also control roundness, straightness, flatness, angulanty and parallelism of mdiVIdual
features

A descnpt10n of these tolerances With their charactensllcs and symbols IS given m the figure below.

Symbol Characteristic Type


D Flatness

--
0
AY
(\
straghtness
Cirrulanl:y (Roundness)
CyHndrlclty
} FORM

c::::,
Profile of a line
Profile of a stl'face
} PROFILE

--L
}
Perpendi rularrl 'y'(Squareness)

L. .Angulanty ORJENT AT ION

/I Parallelism

$
}
PoSitiOn
LOCATION
© Concentncrty
S-,mmetry
I CraJ! ar runout
} RUNOUT
ll Total run DU:
None Regarcless of feature s1ze (.RFS)

@ Ma::<lmum Matenal Condrbon (MMC)

© Least Matenal Cond1t1on (LMC)

® PIOJeded tolerance :zone

CD Tangent p!a1e

e5 Dlametncal (cy1mdnca1) tolerance zone or featL¥e

® Free state

[]2] BaSic, or exact, dimenSion

~ Datum feature s-,mbol

l$10 0 1 @I A I Fealu'e coni'~ ••me


® Datum target

Figure 2. Geometnc charactenshcs and symbols.

Symbols and modifiers are commumcated through a feature control frame The control frame IS spht mto
dtfferent sections each contammg data on geometric symbol, tolerance value, any matenal modtfier and
datum references An example of a feature control frame IS gtven m figure 3

250
Tolerance value Datwn references

~/
-----1+191 o 020®Ii'JB®Ic®l
Gcomctnc ')'Jilbol \ ~---
Material modJ.fier

Tolerance zone moddier

Ftgure 3 Example GD&T feature control frame

Kinematic dimensioning and tolerancing: The mam concept of kmematic tolerance analysts can be
attnbuted to the research work of Professor Kenneth W Chase of Bngham Young Uruverslly, Utah.
Kmemattc dtmenswmng and tolerancmg (KD&T) predtcts vanatwn emergmg form small ngtd body
dtsplacements that occur m static assemblies vta a statistiCal solution Vanatwn m mechamsms can also be
modelled by descnbmg the motion of a smgle mecharusm With respect to ttme

Vanat10n modellmg can be made m 20 and 3D assemblies usmg vector loop based analysis Chase, Magleby
and Gao (1997) suggest that m th1s type of analysis all three sources of vanat10n can be accommodated, 1 e ,
parametnc, geometnc and kmemattc Kmemattc vanatlon analysts tends to be used on mechamcal based
assemblies where the mteractwns of movmg parts and the effect of vanatwn on these parts ts tmportant
Kmemattc analysts has been undertaken on ttems such as automotive dnve tram assemblies and computer
hardware mechamcal dev1ces

251
Appendix 9
Parametric Variation Assembly Analysis Tool
User manual
The PVA apphcahon tool user process IS presented in the followmg table.
Sten Process descriptiOn Instruction/comment
I Start new assembly tolerance analysis sessiOn Start MS Excel via the
Crystal Ball program Icon
The Excel environment will
automatically be started
2 The aoohcatwn Will ooen at the mam banner oaoe
3 At this pomt, If the user IS new to the application they may view The mstructwn set may be
the mstructwn set Altemal!vely, If they have used the accessed directly below the
aoohcatlon before thev can contmue to the next steo banneroage
4 The new user can read the mstructwn set to understand how to The mstructwn set IS
use the applicatiOn This may be pnnted off for reference 1f prepared m the default pnnt
necessary settmg for Excel The user
may use the FILE, PRINT,
OK options to pnnt
contents
5 Go to the application analysis area, and m the ftrst mstance to In the first mstance the used
the Wing box component gallery (WBCG), shown below m part should go to the WBCG
This can be done by
'4>9M>IC«ffjjiMN :JOHrJ pressmg F5 and selectmg
'componenet_gallery' from

~~~"
G...... , ...... --:. ~ .,...,~
the hst
....~, ,.._.i ' J • %'-&~

Q t:::::::1 ~
-~t---... """~ /'}' / .
~OQ~
/ /
J I
• 1J J ..:-...- I - I .. . . .
6 User can now If the user ha< not used the
I Create new wmg box configuration database application before, they
11 Call up a previously saved Wing box configuratiOn must complete the
database appropnate component
tables m the WBCG If the
Previously deftned databases can be recovered from the set user has used the
database These can then be transferred to the set VIewmg table apphcatwn previously and
(SVT), shown below, and finally to the WBCG via the available has saved configuration
macros data then this may be
recalled

li
11 I
7 The user can go to the D nose data table. The user can find the
appropnate table by

252
pressmg the F5 key and
selectmg
'Component_gallery_ D_No
se'.
8 The user can then defme the D nose· Define key feature
I Manufactunng process type defim!Ion
11 Nommal dimension of each cham (Nommal)
111 3 standard deVIatiOn tolerance (3 Std Dev)
IV Vanat10n dtstnbutton type
V Upper specificatiOn hm1t (USL)
VI Lower specificatiOn hm1t (LSL).
.
. - '~ .;
i
•• ... ;?-t~r..;-4""~
r \)"'~r·
.. 0).\,t:>. ~· <:;'-'"!
'!' ,~~-~....:.-::L.t~, t;--
_- _.:·!1?~- ~·,t;:.;~}l_"_

•:.r-- _,:-
I
• ~:Oli..:~:'
~ - "
~~
·\''d I\ .u.u~
:"'~' -8•·) ----. !.'::...

~
e 5°0
,. 0
• 0
0

-
' ' ' ' " ' ~~~ ~-

9 The user can go to the front spar details Part parameter defunt10n
10 Define front spar details
/As for step 8-l
11 The user can go to the A frame details Part parameter defunt10n
12 Define A frame details
/As for step8}
13 The user can go to the rear spar details Part parameter defimtmn
14 Define rear spar details.
(As for step 8}
15 The user can go to the nb details Part parameter defimtton
16 Define nb details
(A• for step 8}
17 The user can go to the top and bottom skm details Part parameter defuntwn
18 Define top and bottom skm details
{As for step 8}
19 When all the appropnate data tables m the WBCG have been
filled, thts sectiOn IS then complete
20 The user can then save the data configuratiOn to the SVT via the The user can save the data
macro command m the WBCG to the SVT
by usmg the 'Copy data
from component gallery to
SVT' macro m the WBCG
wmdow
21 The user can then save the wmg box data by transfemng the The user can save the data
data from the SVT to one of the available sets from the SVT to an
available set (set I for
example) by usmg the 'Put
data mto set 1'
22 When all the reqmred data has been mput to the WBCG (and The user can transfer the
saved If reqmred) the user Will then need to down load th1s appropnate data (data for
mformahon to the appropnate data dnver tables (DDn Thts DDTI, D nose to front

253
operatiOn w1ll need to be performed for each of the reqmred spar, for example) by usmg
analysis stage the 'Update DDTI from

23 When the data IS transferred to the DDT tables, worse case


(WC) and root sum square (RSS) levels of vanatwn are
calculated automatically

If further analys1s 1s reqmred, 1e full stal!sl!cal analys1s, they


are reqmred to undertake additional preparatiOn w1thm the
appropnate DOT Each of the nominal dunenswnal values m
the DOT table Will need to be ass1gned a statistical d1stnbutwn
The current ass1gned d1stnbutwn can be mvest1gated by
selectmg the 'Define assumptiOn' functiOn form the Crystal Ball
menu bar There are ranges of d1stnbul!on type's available (see
below) mcludmg the 'F1t' optiOn wh1ch allows the user to create
a best fit from a sample of data

When a d!stnbutwn type IS selected the user will need to ass1gn


the parameters 1fthey have not already been extracted from the
DOT The default settmg IS always a normal d1stnbutiOn (see
example below) With the mean and standard deviatiOn extracted
d1rectly from the DOT If the d1stnbul!on IS one other than
normal, then 11 should be chosen from the d1stnbutwn gallery
and the parameters ed1ted Each type of d1stnbutiOn IS set by a
number of d•fferent parameters, and the Crystal Ball manual
should be consulted for an explanatiOn on how these are
correctly defined

Assumption Name IPil -Pt Z

010
•llnhn~ty r. ~c i" Qynamic •l+l~~·~•ty
Mean I=!K2o4~ Std Devi"~P247~ _

~ ~··~•I I f.nter I Gallery j Correlate_ I ~

254
24 The output from the WC and RSS calculatiOns can be observed
m the relevant DDT
25 When all distnbut10n types and therr parameters have been Stmulahon parameters.
defined, the user Will then need to select the simulatiOn
preferences These are defined by selectmg the 'Run
preferences' optiOn from the Crystal Ball menu bar The user
can then select and mput a number of different sunulatiOn
cntenon and the sampling method (1 e Monte Carlo or Latm
Hypercube) see figure below

' ,-
'I
! - Rlmdom Number Generation------ Irials I
'r
.::;am11hng I
Use Same Sequence of Bandom Numbers, (

-m~~~~ai_Se~~~alu~- J~ ~::~-~'·.::~ Speej_


-Sampling M e t h o d - - - - - - - - -
Macros
j r. Monts:. Carlo (' Latin Hypercube

Sample Size for Correl11tJon and ' -M;;;~-,


latm Hypercube lso"~ ! ·
, II __________: ~ ~
,tan eel Help

26 Part of the 'Run preferences' set up Will require the user to mput SimulatiOn parameters
the number of stmulahons requrred The most appropnate
number Will need to be a balance between not to high- leadmg
to excesstve stmulatton time, and not to low- resultmg m poor
distnbutlon and stattsttcal defmttton
27 The user can now run the stmulatmn for the defined en tenon Frequency charts
The resultmg output IS a frequency dt~tnbutmn, an example of
which can be seen below

500 Tnals Frequency Chart 3 Outhers


036 18

:~ '"' il'
-g
"•
~
016 9

-~ 009 45 ~

28 Once a sunulatiOn has been undertaken, the user can analyze the Sensitivity charts
resultmg output The user also has the optiOn to run a sensitlVIty
anal s1s
29 The user can now run a sensitiVIty analysts Thts evaluates all SensitiVIty charts
the sources of vanance and determmes what contnbutwn each
element has made to the total measured vanance An example
of such a sensitiVIty chart IS given below

255
0 SO!flntMIJ [hAlt l!!lliJ EJ

~"~'~"-,.----------.~~--------o---o----~
Pt3 Pt4
Ptl Pt2
1'12 1'13
1'16 I'll

ttt :.5% 5ll% 75:< 100%:


M~brCorirbJJonloV~

l[l;~l'<E"I ~

30 Now the stmulatwn analysts cycle ts almost complete The user Report generatiOn
now has the option to save the smmlat10n output to a report
generated m the excel apphcatwn envrronment The user can do
thts by selectmg the 'Create report' funclton from the Crystal
Ball menu bar The user Will be offered a range of opltons
regardmg what and what not, to mclude m the report Once the
re ort has been enerated, the user ma than view the re ort.
31 When the user has viewed the report, they may then save It m an Save function
Excel file format for future reference
32 Once the sunulatiOn report ts saved 11 may be pnnted out m part Pnnt functtons
or m full at ant lime If the u&er does not save the report, and
they do not pnnt out the report m optiOn 33, they Will loose all
current stmulatwn data
33 The user may pnnt out a report by usmg the FILE, PRINT, and
ALLo t10ns
34 User now has the oplton to revtse all or some ofthe above Save analysts parameters
parameters and undertake further stmulatwn analySts Or, they
can exit out of the a It catiOn, savm all chan es tf desrred

256
Appendix 10
Publication Review

Tolerance design: A handbook for developing optimal specifications.

By C. M. Creveling (Addison Wesley), (Pp. 423), 1997.

Thts IS an excellent book for mdustnahsts and academtcs ahke lookmg for a comprehenSive mtroductton to
tolerancmg techruques for the control of component and assembly vanatwn The book follows a sound
framework contammg etghteen conciSe chapters bnngmg together the study of tolerance deSign for both
products and processes. The author does not assume the reader has a spectahsed background m thiS area but
engages the subject m enough detatls to be mformattve to the more expenenced engmeer

The book has been wntten m three sections Section one mcludec; Chapters 1 to 3 and mtroduces the concept
of tolerance destgn and apphcatton It mcludes an overvtew of product deSign and rehabthty, and goes on to
mtroduce stattsttcs and data analySis for tolerance design m the quahty engmeenng context Sectton two
mcludes Chapters 4 to I 0 and ts destgned to cover tradtttonal tolerance destgn and analysts It mtroduces the
bastes of tolerance and sensttlvtty analysts tradttlonally practised m the West, and then proceeds to revtew
some related advanced topiCS ThiS sectiOn concludes With Chapter I 0, whtch highlights the strengths and
weaknesses of the stx tradttwnal tolerancmg methodologtes, suggestmg none of winch are capable of dealmg
With multtdtsctphnary vanatwn functtons (diverse systems and sub-system elements) Sectwn three
mcorporates Chapters !I to 17, whtch comprehenstvely outhnes Gemcht Taguclu's approach to tolerance
destgn and analysts This represents the Eastern approach, whtch mcludes an mtroductwn to the quahty-loss
function approach to analyttcal and expenmental tolerance analysts Sectton four, the final sectiOn, contams
Chapter 18 The sectiOn comprehenSively outhnes five assembly tolerance analystS cases, which have been
destgned to remforce the techmques and methods previOusly outlines m sectwns one to three One of the
case studtes utilises a computer based tolerance analysts techmque to mvesttgate vanabthty levels m a dnve
module of a high volume copy machme

Thts book ts broad m scope and Will be a useful text both as a teachmg atd for engmeenng students and as a
reference handbook for mdustnal apphcattons The sub;ect area of tolerance destgn ts covered
comprehenstvely and ts supported With comprehenstvely worked examples Tins text ts therefore a good
mtroducllon to the subject of tolerance deSign and analysts

DAYID JEFFREYS
Department of Manufactunng Engmeenng
Loughborough Umversity

257
Appendix 11
Stage I - Va1isys measurement report

258
Apr 29 1999 16 56 PHASEI_PANEL1.res Page 1
Page 1
Valisys Results File Forrnatter Ver 5.3.1 1-19-1998 11:50:27

**** RESULTS FILE ****

FILE: PHASE I - PANEL1.res


FORMATTER: frntrcrnp l . vc l
DATE AND TIME OF REPORT: 1-19-1998 11:50:27
UNITS: MILLIMETERS

ANALYSIS MODE SUMMARY

TOL LOS AXIS PLANAR XTRAP


--------------------------------------
7 AVG LS BF YES
10 AVG LS BF YES
12 AVG LS BF YES
12 AVG LS BF YES
12 AVG LS BF YES
12 AVG LS BF YES
12 AVG LS BF YES
12 AVG LS BF YES
12 AVG LS BF YES
12 AVG LS BF YES
12 AVG LS BF YES
12 AVG LS BF YES
12 AVG LS BF YES
12 AVG LS BF YES
12 AVG LS BF YES
12 AVG LS BF YES
12 AVG LS BF YES
12 AVG LS BF YES
12 AVG LS BF YES
12 AVG LS BF YES
9 AVG LS BF YES
6 AVG LS BF YES
11 AVG LS BF YES
4 AVG LS BF YES
13 AVG LS BF YES
14 AVG LS BF YES
15 AVG LS BF YES
16 AVG LS BF YES
TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD
7 DIA702-7 PASSED
7 DIA6.000+/-0.09
702 DATUM B 6.09 0.09 0.09
6 00 -0 09 0.01
, Apr 29 1999 16:56 PHASEI_PANEL 1.res Page2

Page 2
Val1sys Results File Formatter Ver 5.3.1 1-19-1998 11:50:27

TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD

10 DIA703-10 PASSED
10 DIA6 000+/-0.09
703 DATUM C 6 08 0.09 0.08
6.00 -0.09 0.00

12 DIA731-12 PASSED
12 DIA6.000+/-0.09
704 HOLE 1 6.08 0.09 0.08
6.00 -0 09 0.01

12 DIA731-12 PASSED
12 DIA6 000+/-0.09
705 HOLE 2 6 08 0.09 0.08
6.00 -0.09 0 01

12 DIA731-12 PASSED
12 DIA6.000+/-0.09
706 HOLE 3 6.08 0.09 0.08
6.00 -0.09 0.01

12 DIA731-12 PASSED
12 DIA6.000+/-0 09
707 HOLE 4 6.08 0.09 0.08
6.00 -0.09 0.01

12 DIA731-12 PASSED
12 DIA6.000+/-0.09
708 HOLE 5 6.08 0 09 0.08
6 00 -0 09 0.01

12 DIA731-12 PASSED
12 DIA6.000+/-0.09
709 HOLE 6 6.08 0.09 0.08
6.00 -0.09 0.01

12 DIA731-12 PASSED
12 DIA6.000+/-0.09
710 HOLE 7 6.08 0.09 0.08
6.00 -0.09 0.01

12 DIA731-12 PASSED
12 DIA6 000+/-0.09
711 HOLE 8 6.08 0.09 0.08
6.00 -0.09 0.01
Apr 29 1999 16:56 PHASEI_PANEL 1.res Page3

Page 3
Valisys Results F1le Formatter Ver 5.3.1 1-19-1998 11:50:27

TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD

12 DIA731-12 PASSED
12 DIA6.000+/-0.09
712 HOLE 9 6.08 0.09 0.08
6.00 -0.09 0.01

12 DIA731-12 PASSED
12 DIA6.000+/-0.09
713 HOLE 10 6.08 0.09 0.08
6.00 -0.09 0.01

12 DIA731-12 PASSED
12 DIA6.000+/-0.09
714 HOLE 11 6 08 0.09 0.08
6.00 -0.09 0.01

12 DIA731-12 PASSED
12 DIA6.000+/-0.09
715 HOLE 12 6 08 0.09 0.08
6.00 -0.09 0.01

12 DIA731-12 PASSED
12 DIA6.000+/-0.09
716 HOLE 13 6.08 0.09 0.08
6.00 -0.09 0.01

12 DIA731-12 PASSED
12 DIA6.000+/-0 09
717 HOLE 14 6.08 0.09 0.08
6.00 -0 09 0.01

12 DIA731-12 PASSED
12 DIA6 000+/-0.09
718 HOLE 15 6.08 0.09 0.08
6.00 -0 09 0.01

12 DIA731-12 PASSED
12 DIA6.000+/-0.09
719 HOLE 16 6.08 0.09 0.08
6.00 -0 09 0.02

12 DIA731-12 PASSED
12 DIA6.000+/-0.09
720 HOLE 17 6.09 0.09 0.09
6.00 -0.09 0.01
Apr 29 1999 16:56 PHASEI_PANEL 1.res Page4

Page 4
Valisys Results F1le Formatter Ver 5.3.1 1-19-1998 11 50:27

TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD

12 DIA731-12 PASSED
12 DIA6.000+/-0.09
721 HOLE 18 6.09 0 09 0.09
6.00 -0.09 0.01

9 POS703-9 FAILED
9 POSIDIA0.025(M) IAIB(M) POSITION OF 1 ROUND HOLE
0.02
0.20 121.73
703 DATUM C
D: -250 00 250.00 0.00
A: -249.96 249.97 0.00 0.20 121.73

6 POS702-6 PASSED
6 POSIDIA0.025(M) lA POSITION OF 1 ROUND HOLE
0.02
0.20 0 00
702 DATUM B
D: 124.98 -125.00 0.00
A: 124 97 -124.99 -0.00 0.20 0.00
---------------------------------------------------------------------------
11 POS731-11 FAILED
11 POSIDIA0.025(M) IAIB(M) IC(M) POSITION OF 18 ROUND HOLE
0 02
0.20 121.73
704 HOLE 1
D: 35 36 35.36 0.00
A: 35.34 35.35 -0.00 0.20 121.63
705 HOLE 2
D 70.71 70.71 -6.00
A: 70.69 70.70 -6.00 0.19 121. 56
706 HOLE 3
D: 106.07 106.07 0.00
A: 106.05 106.05 -0.00 0.19 121.66
707 HOLE 4
D: 141.42 141.42 -6.00
A: 141.40 141.42 -6.00 0.20 121.61
708 HOLE 5
D: 176.78 176 78 0.00
A: 176.76 176 76 -0.00 0 20 121.65
709 HOLE 6
D: 212 13 141.42 0 00
A: 212 11 141 40 0 00 0 20 121.68
710 HOLE 7
D: 247.49 106.07 0.00
A: 247.46 106.05 0 00 0.20 121.64
711 HOLE 8
D· 282.84 70.71 0.00
Apr 29 1999 16:56 PHASEI- PANEL1.res PageS

Page 5
Val~sys Results File Formatter Ver 5.3.1 1-19-1998 11:50:27

TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD
---------------------------------------------------------------------------
A: 282.81 70.70 0.00 0.20 121 69
712 HOLE 9
D: 318.20 35.36 0.00
A: 318.16 35.35 0.00 0.20 121.69
713 HOLE 10
D: 318.20 -35.36 0.00
A: 318.16 -35.35 0.00 0.20 121.69
714 HOLE 11
D: 282.84 -70.71 0.00
A: 282.80 -70.71 0 00 0.20 121.73
715 HOLE 12
D: 247.49 -106.07 -6.00
A: 247.45 -106.07 -6.00 0.20 121.65
716 HOLE 13
D· 212.13 -141.42 0.00
A. 212.09 -141.41 0.00 0.20 121 71
717 HOLE 14
D 176.78 -176.78 0 00
A: 176 75 -176 75 0 00 0.20 121. 60
718 HOLE 15
D: 141 42 -141 42 -6.00
A: 141.38 -141.41 -6.00 0.20 121.60
719 HOLE 16
D: 106.07 -106.07 0.00
A· 106.04 -106.06 0.00 0. 20 121 65
720 HOLE 17
D. 70 71 -70 71 -6.00
A: 70.69 -70.71 -6.00 0.20 121.60
721 HOLE 18
D: 35 36 -35.36 -6.00
A: 35.33 -35.36 -6.00 0.20 121.62

4 FLT701-4 PASSED
4 FLT I 0.2
701 DATUM A 0.20 0 04

13 PER722-13 FAILED
13 PERJ0.02JA PERPENDICULARITY-PLN OF 1 PLANE
722 EDGE 1 0.02 0.03

14 PER723-14 FAILED
14 PERJO 02JA PERPENDICULARITY-PLN OF 1 PLANE
723 EDGE 2 0.02 0.09
~

Apr 29 1999 16:56 PHASEI_PANEL 1.res Page6

Page 6
Val~sys Results F~le Forrnatter Ver 5.3.1 1-19-1998 11:50:27

TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD
15 PER724-15 FAILED
15 PERI0.02IA PERPENDICULARITY-PLN OF 1 PLANE
724 EDGE 3 0.02 0.13
16 PER725-16 FAILED
16 PERI0.02IA PERPENDICULARITY-PLN OF 1 PLANE
725 EDGE 4 0.02 0.22
Apr 29 1999 16:56 PHASEI_PANEL1.res Page?

Page 7
Val1sys Results F1le Formatter Ver 5.3.1 1-19-1998 11:50:27

V:INSPECT VERSION 5.3.1_SA


PART NAME /home/catusr/PHASEI_PANEL1_INS.prt PHASEI_PANEL
PROCESS NO. 903
PROCESS NAME PHASEI_PANEL1_INS
DATE Thr Jan 15 9:28:06 1998
Apr 29 1999. 16:56 PHASEI_PANEL 1.res PageS

Page 8
Va1isys Results F11e Formatter Ver 5.3.1 1-19-1998 11:50:27

SUMMARY OF GAUGE ANALYSIS


TOL DESCRIPTION STATUS REQT FEATURES
-----------------------------------------------------------
7 SIZE DIAM PASSED 702
10 SIZE DIAM PASSED 703
12 SIZE DIAM PASSED 704
12 SIZE DIAM PASSED 705
12 SIZE DIAM PASSED 706
12 SIZE DIAM PASSED 707
12 SIZE DIAM PASSED 708
12 SIZE DIAM PASSED 709
12 SIZE DIAM PASSED 710
12 SIZE DIAM PASSED 711
12 SIZE DIAM PASSED 712
12 SIZE DIAM PASSED 713
12 SIZE DIAM PASSED 714
12 SIZE DIAM PASSED 715
12 SIZE DIAM PASSED 716
12 SIZE DIAM PASSED 717
12 SIZE DIAM PASSED 718
12 SIZE DIAM PASSED 719
12 SIZE DIAM PASSED 720
12 SIZE DIAM PASSED 721
9 POSITION FAILED <-SEP 703
6 POSITION PASSED SEP 702
11 POSITION FAILED <-SEP 704, 705, 706, 707, 708,
709, 710, 711, 712, 713
714, 715, 716, 717' 718
719, 720, 721
4 FLATNESS PASSED 701
13 PERPENDICULARITY FAILED <- 722
14 PERPENDICULARITY FAILED <- 723
15 PERPENDICULARITY FAILED <- 724
16 PERPENDICULARITY FAILED <- 725
Apr 29 1999 17:02 PHASEI_PANEL2_1NS.res Page 1
Page 1
Val~sys Results F~le Forrnatter Ver 5.3.1 1-19-1998 21:50:41

**** RESULTS FILE ****

FILE: PHASEI PANEL2_INS.res


FORMATTER: fmtrcmpl. vel
DATE AND TIME OF REPORT: 1-19-1998 21:50:41
UNITS: MILLIMETERS
ANALYSIS MODE SUMMARY

TOL LOS AXIS PLANAR XTRAP


--------------------------------------
100007 AVG LS BF YES
100010 AVG LS BF YES
100012 AVG LS BF YES
100012 AVG LS BF YES
100012 AVG LS BF YES
100012 AVG LS BF YES
100012 AVG LS BF YES
100012 AVG LS BF YES
100012 AVG LS BF YES
100012 AVG LS BF YES
100012 AVG LS BF YES
100012 AVG LS BF YES
100012 AVG LS BF YES
100012 AVG LS BF YES
100012 AVG LS BF YES
100012 AVG LS BF YES
100012 AVG LS BF YES
100012 AVG LS BF YES
100012 AVG LS BF YES
100012 AVG LS BF YES
100009 AVG LS BF YES
100006 AVG LS BF YES
100011 AVG LS BF YES
100004 AVG LS BF YES
100013 AVG LS BF YES
100014 AVG LS BF YES
100015 AVG LS BF YES
100016 AVG LS BF YES
TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD
100007 DIA702-7 PASSED
100007 DIA6.000+/-0.09
100702 DATUM B 6.09 0 09 0.09
6.00 -0.09 0.01
"
Apr 29 1999 17:02 PHASEI_PANEL2_1NS.res Page2

Page 2
Val1sys Results File Formatter Ver 5.3.1 l-19-199B 21:50:41

TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD

100010 DIA703-10 PASSED


100010 DIA6.000+/-0 09
100703 DATUM C 6.0B 0.09 0 08
6.00 -0.09 0 00

100012 DIA731-12 PASSED


100012 DIA6.000+/-0.09
100704 HOLE 1 6.0B 0.09 0.08
6.00 -0.09 0.01

100012 DIA731-12 PASSED


100012 DIA6.000+/-0 09
100705 HOLE 2 6.08 0.09 0.08
6.00 -0.09 0.01
100012 DIA731-12 PASSED
100012 DIA6 000+/-0.09
100706 HOLE 3 6.0B 0.09 O.OB
6.00 -0.09 0.01

100012 DIA731-12 PASSED


100012 DIA6.000+/-0.09
100707 HOLE 4 6 OB 0.09 0.08
6.00 -0.09 0.01

100012 DIA731-12 PASSED


100012 DIA6.000+/-0.09
100708 HOLE 5 6.08 0.09 0.08
6.00 -0 09 0.01

100012 DIA731-12 PASSED


100012 DIA6.000+/-0.09
100709 HOLE 6 6.0B 0.09 0.08
6.00 -0.09 0.01

100012 DIA731-12 PASSED


100012 DIA6.000+/-0.09
100710 HOLE 7 6.0B 0.09 0.08
6 00 -0.09 0.01

100012 DIA731-12 PASSED


100012 DIA6.000+/-0.09
100711 HOLE 8 6 OB 0.09 0.08
6 00 -0.09 0.01
Apr29199917:02 PHASEI_PANEL2_1NS.res ·Page 3

Page 3
Valisys Results File Forrnatter Ver 5.3.1 1-19-1998 21 50:41

TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD

100012 DIA731-12 PASSED


100012 DIA6.000+/-0 09
100712 HOLE 9 6.08 0.09 0.08
6.00 -0.09 0.01
100012 DIA731-12 PASSED
100012 DIA6.000+/-0.09
100713 HOLE 10 6.08 0.09 0.08
6.00 -0.09 0.01
100012 DIA731-12 PASSED
100012 DIA6.000+/-0.09
100714 HOLE 11 6.08 0.09 0.08
6.00 -0.09 0.01
100012 DIA731-12 PASSED
100012 DIA6.000+/-0.09
100715 HOLE 12 6.08 0.09 0.08
6.00 -0.09 0.01
100012 DIA731-12 PASSED
100012 DIA6 000+/-0.09
100716 HOLE 13 6.08 0 09 0.08
6.00 -0.09 0.01
100012 DIA731-12 PASSED
100012 DIA6.000+/-0.09
100717 HOLE 14 6.08 0.09 0.08
6 00 -0.09 0.01
100012 DIA731-12 PASSED
100012 DIA6.000+/-0.09
100718 HOLE 15 6.08 0.09 0.08
6.00 -0.09 0.01
100012 DIA731-12 PASSED
100012 DIA6.000+/-0.09
100719 HOLE 16 6.08 0.09 0.08
6.00 -0.09 0 02
100012 DIA731-12 PASSED
100012 DIA6.000+/-0.09
100720 HOLE 17 6 09 0.09 0 09
6 00 -0 09 0.01
~-------------------------------

Apr29199917.02 PHASEI_PANEL2_1NS.res Page4

Page 4
Val1sys Results File Formatter Ver 5.3.1 1-19-1998 21·50:41

TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD

100012 DIA731-12 PASSED


100012 DIA6.000+/-0.09
100721 HOLE 18 6.09 0.09 0.09
6.00 -0.09 0.01

100009 POS703-9 FAILED


100009 POSIDIAO 025(M) IAIB(M) POSITION OF 1 ROUND HOLE
0.02
0.20 121.73
100703 DATUM C
D: -250.00 250.00 0.00
A: -249.96 249.97 0.00 0.20 121.73

100006 POS702-6 PASSED


100006 POSIDIA0.025(M) lA POSITION OF 1 ROUND HOLE
0.02
0.20 0.00
100702 DATUM B
D. 124.98 -125.00 0 00
A: 124.97 -124.99 -0.00 0.20 0.00
---------------------------------------------------------------------------
100011 POS731-11 FAILED
100011 POSIDIA0.025(M) IAIB(M) IC(M) POSITION OF 18 ROUND HOLE
0.02
0.20 121.73
100704 HOLE 1
D: 35.36 35.36 0.00
A: 35.34 35.35 -0 00 0.20 121.63
100705 HOLE 2
D: 70.71 70.71 -6.00
A: 70.69 70.70 -6 00 0.19 121.56
100706 HOLE 3
D. 106.07 106 07 0.00
A: 106.05 106 05 -0.00 0.19 121.66
100707 HOLE 4
D: 141.42 141 42 -6.00
A: 141.40 141 42 -6.00 0.20 121.61
100708 HOLE 5
D: 176.78 176.78 0.00
A: 176.76 176.76 -0.00 0.20 121.65
100709 HOLE 6
D: 212.13 141.42 0 00
A: 212.11 141.40 0 00 0.20 121 68
100710 HOLE 7
D: 247.49 106.07 0.00
A: 247.46 106.05 0 00 0.20 121 64
100711 HOLE 8
D: 282.84 70.71 0.00
Apr 29 1999 17:02 PHASEI_PANEL2_1NS.res PageS.

Page 5
Val1sys Results F1le Formatter Ver 5.3.1 1-19-1998 21:50:41

TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD

A: 282.81 70.70 0.00 0.20 121.69


100712 HOLE 9
D: 318 20 35.36 0.00
A: 318 16 35.35 0 00 0.20 121 69
100713 HOLE 10
D: 318.20 -35.36 0.00
A: 318.16 -35.35 0.00 0.20 121.69
100714 HOLE 11
D: 282.84 -70.71 0.00
A: 282.80 -70.71 0.00 0.20 121.73
100715 HOLE 12
D: 247.49 -106.07 -6.00
A· 247.45 -106.07 -6.00 0.20 121.65
100716 HOLE 13
D. 212 .13 -141.42 0.00
A: 212.09 -141.41 0.00 0.20 121.71
100717 HOLE 14
D: 176.78 -176.78 0.00
A: 176.75 -176.75 0.00 0.20 121.60
100718 HOLE 15
D: 141.42 -141.42 -6.00
A: 141.38 -141.41 -6.00 0.20 121.60
100719 HOLE 16
D: 106.07 -106.07 0.00
A. 106.04 -106.06 0.00 0.20 121.65
100720 HOLE 17
D. 70.71 -70.71 -6.00
A: 70.69 -70.71 -6.00 0.20 121.60
100721 HOLE 18
D: 35.36 -35.36 -6.00
A: 35.33 -35.36 -6.00 0.20 121.62
100004 FLT701-4 PASSED
100004 FLTjO 2
100701 DATUM A 0.20 0.04

100013 PER722-13 FAILED


100013 PERjO 02jA PERPENDICULARITY-PLN OF 1 PLANE
100722 EDGE 1 0 02 0.03
100014 PER723-14 FAILED
100014 PERj0.02jA PERPENDICULARITY-PLN OF 1 PLANE
100723 EDGE 2 0.02 0.09
Apr29199917.02 PHASEI_PANEL2_1NS.res Page6

Page 6
Val1sys Results F1le Formatter Ver 5.3.1 1-19-1998 21:50:41

TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD
---------------------------------------------------------------------------
100015 PER724-15 FAILED
100015 PERI0.02IA PERPENDICULARITY-PLN OF 1 PLANE
100724 EDGE 3 0.02 0.13
---------------------------------------------------------------------------
100016 PER725-16 FAILED
100016 PERI0.02IA PERPENDICULARITY-PLN OF 1 PLANE
100725 EDGE 4 0. 02 0.22
' ' '
Apr 29 1999 17:02 PHASEI_PANEL2_1NS.res Page?

Page 7
Val1sys Results F1le Formatter Ver 5.3.1 1-19-1998 21:50:41

V:INSPECT VERSION 5.3.1_SA


PART NAME /home/catusr/PHASEI_PANEL1_INS.prt PHASEI_PANEL
PROCESS NO. 903
PROCESS NAME PHASEI_PANEL1_INS
DATE Thr Jan 15 9:28:06 1998
Apr 29 1999 17:02 PHASEI_PANEL2_1NS.res . Page 8

Page 8
Val1sys Results F1le Formatter Ver 5.3 1 1-19-1998 21:50:41

SUMMARY OF GAUGE ANALYSIS

TOL DESCRIPTION STATUS REQT FEATURES


-----------------------------------------------------------
100007 SIZE DIAM PASSED 100702
100010 SIZE DIAM PASSED 100703
100012 SIZE DIAM PASSED 100704
100012 SIZE DIAM PASSED 100705
100012 SIZE DIAM PASSED 100706
100012 SIZE DIAM PASSED 100707
100012 SIZE DIAM PASSED 100708
100012 SIZE DIAM PASSED 100709
100012 SIZE DIAM PASSED 100710
100012 SIZE DIAM PASSED 100711
100012 SIZE DIAM PASSED 100712
100012 SIZE DIAM PASSED 100713
100012 SIZE DIAM PASSED 100714
100012 SIZE DIAM PASSED 100715
100012 SIZE DIAM PASSED 100716
100012 SIZE DIAM PASSED 100717
100012 SIZE DIAM PASSED 100718
100012 SIZE DIAM PASSED 100719
100012 SIZE DIAM PASSED 100720
100012 SIZE DIAM PASSED 100721
100009 POSITION FAILED <-SEP 100703
100006 POSITION PASSED SEP 100702
100011 POSITION FAILED <-SEP 100704, 100705, 100706, 100707' 10070
8,
100709' 100710' 100711, 100712' 100713
100714, 100715' 100716' 100717' 100718
100719, 100720' 100721
100004 FLATNESS PASSED 100701
100013 PERPENDICULARITY FAILED <- 100722
100014 PERPENDICULARITY FAILED <- 100723
100015 PERPENDICULARITY FAILED <- 100724
100016 PERPENDICULARITY FAILED <- 100725
Apr 29 1999 17:02 PHASEI_SUBSTRUCTURE_INS.res Page 1
Page 1
Val~sys Results File Formatter Ver 5.3.1 1-19-1998 12:01:48

**** RESULTS FILE ****

FILE: PHASE! - SUBSTRUCTURE_INS.res


FORMATTER: fmtrcmpl.vcl
DATE AND TIME OF REPORT: 1-19-1998 12:01:48

UNITS MILLIMETERS

ANALYSIS MODE SUMMARY

TOL LOS AXIS PLANAR XTRAP


--------------------------------------
5 AVG LS BF YES
8 AVG LS BF YES
10 AVG LS BF YES
10 AVG LS BF YES
10 AVG LS BF YES
10 AVG LS BF YES
10 AVG LS BF YES
10 AVG LS BF YES
10 AVG LS BF YES
10 AVG LS BF YES
10 AVG LS BF YES
10 AVG LS BF YES
10 AVG LS BF YES
10 AVG LS BF YES
10 AVG LS BF YES
10 AVG LS BF YES
10 AVG LS BF YES
10 AVG LS BF YES
10 AVG LS BF YES
10 AVG LS BF YES
7 AVG LS BF YES
4 AVG LS BF YES
9 AVG LS BF YES
2 AVG LS BF YES
11 AVG LS BF YES
12 AVG LS BF YES
13 AVG LS BF YES
14 AVG LS BF YES

TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD

5 DIA702-5 FAILED
5 DIA6.000+/-0.015
702 DATUM B 6.02 0.01 0.02
6.00 -0 01 0.00
Apr29199917:02 PHASEI SUBSTRUCTURE_INS.res Page2

Page 2
Val1sys Results F1le Formatter Ver 5.3.1 1-19-1998 12:01:48

TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD

8 DIA703-8 FAILED
8 DIA6.000+/-0.015
703 DATUM C 6.02 0.01 0.02
6.00 -0.01 0.01
10 DIA722-10 PASSED
10 DIA6.000+/-0.015
704 HOLE 1 6.01 0.01 0.01
6.00 -0.01 0.00

10 DIA722-10 PASSED
10 DIA6.000+/-0.015
705 HOLE 2 6.01 0.01 0.01
6.00 -0.01 0.01
10 DIA722-10 PASSED
10 DIA6.000+/-0.015
706 HOLE 3 6.01 0.01 0.01
6.00 -0.01 0.00

10 DIA722-10 PASSED
10 DIA6.000+/-0.015
707 HOLE 4 6.01 0.01 0.01
6.00 -0.01 0.00

10 DIA722-10 FAILED
10 DIA6.000+/-0.015
708 HOLE 5 6.02 0.01 0.02
6.00 -0.01 0.00

10 DIA722-10 PASSED
10 DIA6.000+/-0.015
709 HOLE 6 6.01 0.01 0.01
6.00 -0.01 0.00

10 DIA722-10 PASSED
10 DIA6.000+/-0.015
710 HOLE 7 6.01 0.01 0.01
6.00 -0.01 0.00

10 DIA722-10 FAILED
10 DIA6.000+/-0.015
711 HOLE 8 6.02 0.01 0. 02
6.00 -0.01 0.00
. '
Apr291999 17.02.· PHASEt_SUBSTRUCTURE_INS.res Page3

Page 3
Valisys Results F1le Forrnatter Ver 5.3.1 1-19-1998 12:01·48

TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD

10 DIA722-10 PASSED
10 DIA6.000+/-0.015
712 HOLE 9 6.01 0.01 0 01
6.00 -0.01 0.00

10 DIA722-10 PASSED
10 DIA6.000+/-0.015
713 HOLE 10 6.01 0.01 0.01
6.00 -0.01 0.00

10 DIA722-10 PASSED
10 DIA6.000+/-0.015
714 HOLE 11 6.01 0.01 0.01
6.00 -0.01 0.01

10 DIA722-10 PASSED
10 DIA6.000+/-0.015
715 HOLE 12 6.01 0.01 0.01
6.00 -0.01 0.00
10 DIA722-10 PASSED
10 DIA6.000+/-0.015
716 HOLE 13 6.01 0.01 0.01
6.00 -0.01 0.00

10 DIA722-10 PASSED
10 DIA6 000+/-0.015
717 HOLE 14 6.01 0.01 0.01
6.00 -0.01 0.00

10 DIA722-10 PASSED
10 DIA6.000+/-0.015
718 HOLE 15 6.01 0.01 0.01
6.00 -0.01 0.01

10 DIA722-10 PASSED
10 DIA6.000+/-0.015
719 HOLE 16 6.01 0.01 0.01
6.00 -0.01 0.00

10 DIA722-10 PASSED
10 DIA6 000+/-0.015
720 HOLE 17 6.01 0.01 0.01
6.00 -0.01 0.01
'
' Apr 29 1999 17:02 PHASEI_SUBSTRUCTURE_INS.res Page4

Page 4
Valisys Results File Formatter Ver 5.3.1 1-19-1998 12:01:48

TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD
10 DIA722-10 PASSED
10 DIA6.000+/-0 015
721 HOLE 18 6.01 0.01 0.01
6.00 -0.01 0.00
7 POS703-7 PASSED
7 POSjDIA0.025(M) IAIB(M) POSITION OF 1 ROUND HOLE
0.02
0.05 0.00
703 DATUM C
D: -250.00 250.00 -8.00
A: -250.00 249.99 -8 00 0.05 -0.00
4 POS702-4 PASSED
4 POSjDIA0.025(M) jA POSITION OF 1 ROUND HOLE
0.02
0.05 0.00
702 DATUM B
D: 125.00 -124.99 0.00
A: 125.00 -124.99 0.00 0.05 0.00
---------------------------------------------------------------------------
9 POS722-9 PASSED
9 POSjDIA0.025(M) jAjB(M) jC(M) POSITION OF 18 ROUND HOLE
0.02
0.05 0.03
704 HOLE 1
D: 35.36 35.36 0.00
A: 35.35 35.35 0.00 0.05 0.01
705 HOLE 2
D: 70.71 70.71 0.00
A: 70.71 70.69 0.00 0.05 0.01
706 HOLE 3
D: 106.07 106.07 -8.00
A: 106.07 106.05 -8 00 0.05 0 01
707 HOLE 4
D: 141.42 141.42 0.00
A: 141.42 141.41 0.00 0.05 0.00
708 HOLE 5
D: 176.78 176.78 -8.00
A: 176.76 176.78 -8.00 0.05 0.02
709 HOLE 6
D: 212 13 141.42 -8.00
A: 212.12 141.42 -8.00 0.05 0.01
710 HOLE 7
D: 247 49 106.07 -8.00
A: 247.47 106.07 -8.00 0.05 0.02
711 HOLE 8
D: 282.84 70.71 -8.00
Apr 29 1999 17:02 PHASEI_SUBSTRUCTURE_INS.res PageS

Page 5
Val1sys Results F1le Formatter Ver 5.3.1 1-19-1998 12·01.48

TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD

A. 282.83 70.71 -8.00 0.05 0.02


712 HOLE 9
D. 318.20 35.36 -8.00
A: 318.19 35.35 -8.00 0.05 0.01
713 HOLE 10
D: 318.20 -35.36 -8.00
A· 318.21 -35 35 -8.00 0.05 0.01
714 HOLE 11
D 282.84 -70.71 0 00
A: 282.85 -70.71 0 00 0.05 0.01
715 HOLE 12
D: 247.49 -106 07 -8.00
A: 247.50 -106 08 -8.00 0.05 0.02
716 HOLE 13
D 212.13 -141.42 -8.00
A. 212.14 -141.43 -8.00 0.05 0.02
717 HOLE 14
D: 176.78 -176.78 -8 00
A: 176.78 -176.78 -8.00 0.05 0.01
718 HOLE 15
D: 141.42 -141.42 -8.00
A: 141.43 -141.43 -8.00 0.04 0.01
719 HOLE 16
D· 106.07 -106 07 -8.00
A· 106.08 -106 08 -8.00 0 05 0.02
720 HOLE 17
D. 70.71 -70.71 -8.00
A. 70.72 -70.72 -8.00 0.05 0.01
721 HOLE 18
D: 35.36 -35 36 -8.00
A: 35.37 -35 37 -8 00 0.05 0 02

2 FLT701-2 PASSED
2 FLTI 0.2
701 DATUM A 0.20 0.07

11 FLT723-11 PASSED
11 FLTJ0.04
723 EDGE 1 0.04 0.01

12 FLT724-12 PASSED
12 FLTJ0.04
724 EDGE 2 0.04 0 02
> ' '

Apr 29 1999 17:02 PHASEI_SUBSTRUCTURE_INS.res Page6

Page 6
Val1sys Results F1le Formatter Ver 5.3.1 1-19-1998 12:01:48

TOL
NUM DESCRIPTION
FEA ACTUAL/ TOL/ DEVIATION/
NUM X y z NOMINAL TOL+BONUS SPREAD
13 FLT725-13 PASSED
13 FLT!0.04
725 EDGE 3 0 04 0.04
14 FLT726-14 PASSED
14 FLT!0.04
726 EDGE 4 0,04 0.03
Apr 29 1999 17:02 PHASEI_SUBSTRUCTURE_INS.res Page?

Page 7
Valisys Results F1le Formatter Ver 5.3.1 1-19-1998 12:01:48

V:INSPECT VERSION 5.3.1_SA


PART NAME /home/catusr/PHASEI_SUBSTRUCTURE_INS.prt PHASEI_SUBSTRUCT
URE
PROCESS NO. 902
PROCESS NAME PHASEI_SUBSTRUCTURE_INS
DATE Thr Jan 15 8:58:15 1998
Apr 29 1999 17:02 PHASEI_SUBSTRUCTURE_INS.res Page 8

Page 8
Val~sys Results File Formatter Ver 5.3.1 1-19-1998 12:01:48

SUMMARY OF GAUGE ANALYSIS

TOL DESCRIPTION STATUS REQT FEATURES


-----------------------------------------------------------
5 SIZE DIAM FAILED <- 702
8 SIZE DIAM FAILED <- 703
10 SIZE DIAM PASSED 704
10 SIZE DIAM PASSED 705
10 SIZE DIAM PASSED 706
10 SIZE DIAM PASSED 707
10 SIZE DIAM FAILED <- 708
10 SIZE DIAM PASSED 709
10 SIZE DIAM PASSED 710
10 SIZE DIAM FAILED <- 711
10 SIZE DIAM PASSED 712
10 SIZE DIAM PASSED 713
10 SIZE DIAM PASSED 714
10 SIZE DIAM PASSED 715
10 SIZE DIAM PASSED 716
10 SIZE DIAM PASSED 717
10 SIZE DIAM PASSED 718
10 SIZE DIAM PASSED 719
10 SIZE DIAM PASSED 720
10 SIZE DIAM PASSED 721
7 POSITION PASSED SEP 703
4 POSITION PASSED SEP 702
9 POSITION PASSED SEP 704, 705, 706, 707, 708,
709, 710, 711, 712. 713
714, 715, 716, 717, 718
719, 720, 721
2 FLATNESS PASSED 701
11 FLATNESS PASSED 723
12 FLATNESS PASSED 724
13 FLATNESS PASSED 725
14 FLATNESS PASSED 726

You might also like