Cessna 441 Life Extension

Graham Swannell / Kim White May 2010

1
Introduction Working together

Cessna 441 is a twin turboprop, pressurised aircraft.

In 2007, as a result of a Cessna Supplemental Inspection Document (SID), all

441s in Australia with over 22,500 hours flight time were grounded by CASA.

This life is now also included in the Cessna system of maintenance.

Several aircraft were grounded immediately, and others are close to 22,500 hrs

2
CASA Instrument Working together

3
 Working together

This talk will cover:

• The basis of the life extension program
• STC holder responsibilities
• Methods used to develop the program
• Critical structural areas
• Finite element modelling
• Wing strain gauge testing
• Fuselage frame modifications
• Fatigue and damage tolerance calculation methods
• Pressure bulkhead reinforcing
• Widespread fatigue damage considerations
• Fuselage stringer considerations

4
Life extension program Working together

TAE is an organisation which, among many other activities, maintains Cessna

441s. It was previously Tenix Aviation.

TAE and AEA are cooperating to develop a life extension program for the 441

The methods of life extension include:

• Inspections
• Replacement of components with standard or redesigned new parts
• Reinforcement of existing components

Outcome of the program will be a Supplemental Type Certificate (STC ),

which will cover:
• Physical modification of aircraft
• New IFCA incorporating additional inspections

Aims of the life extension program include:

• Increase life from 22,500 hours to 40,000 hours/cycles
• Make the structure damage tolerant where practical
5
 Life extension program Working together

The process for the life extension, involving AEA, TAE and CASA, is:

What are the STC holder’s responsibilities?

6
Methods Working together

A major problem with the life extension program is that is has been done
without support from the OEM.

The general approach to the program was:

1. Establish reason for life limitation

Structure
Fatigue, Corrosion
Systems
Electrical, hydraulic etc

2. Find problem areas

3. Choose life extension methods for each area identified

4. Show that the method is effective

NOTE: It is more than just a structural evaluation.

Cover the entire aircraft.
7
Focus on maintainability.
 Working together

What are the problem areas?

There are several sources of data which indicate the locations and nature of
problem areas, both structural and other.

Cessna data
• Maintenance Manual
• Service Bulletins
• Supplemental Inspection Documents (SIDs)

Local service experience

• TAE specialises in maintenance of Cessna 441s, and has gained
wide knowledge of recurrent problems
• AEA has done a large number of repairs to 441s over many years,
and has a large database of repairs
• All systems considered

Structural analysis and test.

• Finite element analysis
• Strain gauge testing
8
 Working together

What are not problem areas?

In addition to locations which may have problems, there is evidence which
shows that an area will not have problems at the extended life.

This is more applicable to systems rather than structure.

In particular, the Cessna 404 is very similar to the 441, and has a life of 40,000
hours. Many systems on the two aircraft are the same.

The primary differences between the two aircraft are that the 404 has piston
engines rather than turboprops, and is not pressurised.

The wings have similar structure, but the 441 has increased span, higher
gross weight, greater fuel weight, lighter engines and different flight
envelope. A comparison of the wing loads on the two aircraft was done,
but the differences meant that there was no relationship between the
fatigue lives.

9
Life extension methods Working together

Inspection
• Continue existing inspections
• New inspections
• Damage tolerance principles in redesigned parts enhances inspectability

Part replacement
• Standard new parts. The new part will have a life of 22,500 hours

Modifications
• Replace sections. Some parts, such as fuselage frames, cannot be
removed and replaced in one piece. It is possible to cut out the critical
section of the part, and replace with a new section. The replacement
section can be equivalent to or the same as the original part, or may be
stronger
• Add reinforcement to reduce stress.

10
 Working together

Structure fatigue - certification history

The rules in FAR 23 relating to fatigue in Normal category aircraft have
developed progressively. The main changes were:

• From the initial issue of FAR 23 in 1964, pressurised cabins were

required to be safe life or fail safe (23.571).
• In amendment 7 in 1969, this requirement was extended to include
wing, carrythrough and attaching structure (23.572).
• Amendment 14 in 1973, required test justification for safe life structure
except for simple structures.
• Amendment 38 in 1989 added empennage.
• In amendment 45 in 1993, the option of damage tolerance was added

Although still acceptable, fail safe design by itself is generally regarded as

unsatisfactory, since it does not give a rational method of finding a failure.

11
 Working together

Structure fatigue - Certification

Dangerous

Crack Duration
Size Detectable

Time

Damage tolerance adds the requirement that damage should be found before
it becomes dangerous.

The certification basis of the Cessna 441 is FAR 23 Amendment 14.

The fuselage and wing were designed and tested to be fail safe.

In particular, the wing design was a great improvement on the Cessna 402
wing, which has a single spar, is a safe life structure, and has a short life.

12
 Working together

Structural analysis - load spectrum

An important part of the fatigue analysis is development of load spectra.

Applied loads are: Flight Length Summary

• Flight loads 40%

• Gust 35%

Proportion of Fleet Hours

• Manoeuvre 30%

25%
• Pressurisation 20%

• Ground loads 15%

• Landing 10%

5%
• Taxi 0%

• Ground - Air - Ground (GAG) 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1 2.3 2.5 2.7 2.9 3.1 3.3 3.5 3.7 3.9 4.1 4.3 4.5
Flight Time (hrs)

Spectra were derived from:

• AC23-13A Fatigue, Fail-Safe, and Damage Tolerance Evaluation
of Metallic Structure for Normal, Utility, Acrobatic, and Commuter
Category Airplanes
• Operator surveys

For most fuselage structure, pressurisation loads are dominant. 13

Fatigue spectrum Working together

AC23-13A method, which uses one stress value for whole flight, was modified
to account for different loadings.

• Max zero fuel weight

• Maximum fuel
• Half way between

Weighted mean life proportioning the 3 loadings based on operator survey

14
 Wing Loads Working together

Wing loads and stress depends on Weight W and Zero Fuel Weight ZFW

BM C441, WS 27.1, V/Vc = 0.9, BM/g BM C441, WS 120, V/Vc = 0.9, BM1g

400000 120000

350000
100000

300000

80000
250000

BM (lb in)
BM (lb in)

Max fuel Max fuel

Half 60000 Half
200000 Max ZFW
Max ZFW

150000
40000

100000

20000
50000

0 0
5000 5500 6000 6500 7000 7500 8000 8500 9000 9500 10000 5000 5500 6000 6500 7000 7500 8000 8500 9000 9500 10000

W (lb) W (lb)

Bending moment at two wing stations is plotted against aircraft weight W, for
different initial fuel weights.
Each curve corresponds to burning fuel from MTOW down to ZFW.
Maximum loads can occur at intermediate fuel loadings.
15
Operator Flight Survey Working together

From AC23-13A, flight time is assumed to be 1.1 hours, for all fuel loadings.

Shorter flights increase the fatigue damage which is caused by the GAG
cycle, so that shorter flights give lower lives.

A flight with high fuel loading may be either a single long flight, or a series of
shorter flights without refuelling.

16
 Working together

Finite Element Models

Structural analysis was done using several finite element models
These models were
• Coarse mesh model of whole airframe. For convenience, this
was divided into two parts
• Fuselage and wing
• Tailplane and fin
• Detail fine mesh models
• Fuselage frame and skin
• Forward pressure bulkhead top
• Stringer

The detail models are described later.

17
 Working together

FEM - Main Model

The main model is a coarse mesh model of whole airframe.
The model is run in two configurations:
• Standard aircraft
• Modified aircraft

18
FEM - Main Model Working together

Model follows aircraft structure

• Fuselage frames, stringers
• Wing and tail ribs, spars, stringers
• It is not possible to use a half model since fuselage is
unsymmetric due to doors

Size of model:
• 6115 nodes
• 15496 elements
• 1194 properties

A large part of the modelling time was spent calculating geometry and
properties
Data is mainly from measurements of the aircraft 19
 Working together

FEM - Main Model

For convenience, the analysis was divided into two parts
• Fuselage and wing. Tail loads were applied through a simple beam
representation of the tailplane and fin.
• Tail. The tailplane and fin models were attached to the rear fuselage.

20
 Working together
FEM - Main Model Constraints
Loads were applied as balanced load cases.
Supports were statically determinate.
Reactions were zero except for small errors.
Constraint directions
x Engines (2)
y Bottom spar cap on centreline (1)
z Main spar cap at fuselage sides and tail (3)

21
 Working together
FEM - Main Model Unit Load cases
Unit load cases were run in the model, and combined for particular loading
and flight conditions
Unit load cases are:

22
 Working together
FEM - Combined Load cases
Factors for combining load cases were derived from loading and flight
conditions

The load cases considered for fatigue and stress spectrum are:
1 g Cruise (includes pressurisaton)
Gust 1g increment
Manoeuvre 1g increment
1 g Flaps extended Vf
1 g Ground (taxi)
10 ft/s landing Type 1 (Tail down)
10 ft/s landing Type 2 (Level landing with drag)

• 1g cruise is different from 1g increment because of stress at zero g,

mainly from section pitching moment and wing twist.
• Gusts and manoeuvre loads different since tail loads are different.
• 1g taxi stress loads correspond to static equilibrium on wheels
• For landings, wing lift = 0.67 * W

23
 Working together

Wing Strain gauge testing

The FEM is validated by comparing results to
strain gauge tests.
Two sets of tests were done:
Wing
• Flight test
• Manoeuvres at various weights
and fuel load
• Speed
• Flap settings

24
 Working together

Wing centre section construction

TXXXX:
• XXXXX

25
 Working together

Outer wing construction

26
 Working together

Wing strain gauge testing

Strain gauge results are compared to FEM results in order to validate the
FEM.

In order to do this,
• Strain gauge results must be processed to get the desired
quantities, which are 1g stress and stress per g, for various
loadings and flight conditions
• Corresponding load cases must be run in the FEM
Strain gauge results must be processed
• Some gauge results were incorrect
• Gauge not working - results discarded
• Results wrong sign = multiply by -1
• Where there are two gauges on front and back flanges of T
section, the average stress from them was used
• Results from multiple test points must be combined to give the
desired quantities, which are stress per g and 1g stress. 27
 Working together
Wing strain gauge processing
Desired quantities are 1g stress and stress per g
In example below, results are reasonably linear. Best fit line is
S = 5817 * n - 499
So Spg = 5817 (lb/in2)/g
S1g = 5817 * 1 - 499 = 5318 lb/in2

Stress vs V and Flap - Flight 3 Main spar WS81.9

12000

y = 5817x - 499.35
1g Flight
10000

8000
Stress (PSI)

Push over
1g ground
Turns M 81.9
6000
Linear (M 81.9)

4000

2000
1g Ground

0
0.0 0.5 1.0 1.5 2.0 2.5
28
Load factor n (g)
Wing strain gauge correlation Working together

Strain gauge readings for forward, main

and rear spars are shown, for one Spar Stress FWD spars Case 250

flight test. 10000

9000
8000

These curves are for stress per g.

7000
6000
5000
4000

The zero value at WS120 is a result of 3000

2000
1000
the modelling technique used at the

Stress
FS
0 Test FS
-1000 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 Lines

outer wing joint. -2000

-3000
-4000
-5000
Agreement is good except for a point on -6000
-7000

the forward spar just outboard of the -8000

-9000
-10000

joint. This is due to the offset of the Y (in)

stringer at the joint.

Spar Stress REAR spars Case 250
Spar Stress MAIN spars Case 250
10000
10000 9000
9000 8000
8000 7000
7000 6000
6000 5000
5000 4000
4000 3000
3000 2000 RS
2000 Test RS
1000
Stress
1000 MS Lines
Stress

Test MS 0
0 Joint
Lines -1000 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300
-1000 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300
Joint -2000
-2000
-3000
-3000
-4000
-4000
-5000
-5000
-6000 -6000
-7000 -7000
-8000 -8000
-9000
29
-9000
-10000 -10000
Y (in)
Y (in)
 Working together
Wing strain gauge processing
It is not possible to measure 1g stress accurately

• The aircraft weight is on the undercarriage when the strain gauges

are installed, so there is an initial 1g ground stress on the wing.
• Gauges must be zeroed, also with weight on the undercarriage. This
would probably be done for each set of measurements.

The incremental stress Spg can be measured accurately, since the zero error
does not affect the slope of the curve of stress vs load factor.

If the gauges are zeroed on the ground, then the stress measurement is
relative to the 1g ground stress.

For the purpose of correlation, we can compare the increment in stress

between ground and flight, ie (S1g flight - S1g ground)

30
 Working together
Wing strain gauge correlation
Curves are for Spar Stress FWD spars Case 403

(S1g flight - S1g ground) 10000

9000
8000

These curves are for 30 degree flap, at Vf 7000

6000
5000

The effect of flap is: 4000

3000
2000

• Lift is moved inboard. The outer wing 1000

Stress
FS
0 Test FS
-1000 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 Lines

panel actually has a down load at 1g. -2000

-3000
-4000

• High torsion due to flap pitching -5000

-6000
-7000

moment is partly taken by differential -8000

-9000
-10000

bending of the main and rear spars, Y (in)

giving rapid spanwise variation of spar

stress Spar Stress MAIN spars Case 403
Spar Stress REAR spars Case 403

10000
10000 9000
9000 8000
8000 7000
7000 6000
6000 5000
5000 4000
4000 3000
3000 2000 RS
2000 Test RS
1000
Stress

1000 MS Lines
Stress

Test MS 0
0 Joint
Lines -1000 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300
-1000 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300
Joint -2000
-2000
-3000
-3000
-4000
-4000
-5000
-5000
-6000 -6000
-7000 -7000
-8000 -8000
-9000
31
-9000
-10000 -10000
Y (in)
Y (in)
 Working together
Wing fatigue
Damage tolerance analysis was done on the
wing spars and joint fittings. 2
1
5, 2
1,
2 9, 3,
1
It was found that the existing Cessna SID 6 1 2
7,
2 0 4
8 2
inspections were basically adequate to 1,
5,
ensure continued safety. 17
2
6
2
,1
1,
8
2
The majority of the wing structure has a safe 1 1
life of greater than 40,000 hours, in 5, 3,
1 1
addition to being fail safe. 6 4

1
The existing inspection of the wing attach 9,
3,
2
fittings has been extended to additional 0 4

locations.

32
 Working together
Wing attach fittings
Wing attach fittings were analysed using a simple 1 dimensional FEM, in
which fitting axial stiffness and fastener stiffness was simulated. This
enabled calculation of bearing and bypass stress at each fastener hole.

These stresses were used for the crack

growth analysis of the spar caps and
fittings.

33
 Working together

Fuselage
The main areas in the fuselage which are fatigue critical are:
• Frames
• Door and window frames
• Forward pressure bulkhead
• Bottom stringers

34
Classification of Working together

Fuselage Structure
Structure can classified according to
the load spectrum, stress level and
structural detail.

The flowchart applies to fuselage

structure, primarily loaded by
pressurisation.

Calculation of SN life showed that a

stress of less than 10 KSI for
normal pressure differential, gives
a life of greater than 40,000 cycles.

Higher stress, or a history of cracking,

required modification and further
analysis.

35
 Working together

Fuselage frames - stress

Under the effect of the cabin pressure, the fuselage, whose cross section is a
rounded rectangle , tends to become circular.
The sides and top and bottom bulge outward, and the corners straighten.

Frame deflected shape

36
 Working together

Fuselage frames - stress

Bending moments are produced in
the frame.
Stresses in the inner and outer
caps is plotted, with tension
outward from frame centreline.
Inner cap is in tension at the
corners
Outer cap is in tension on the sides,
top and bottom.
In general, the tensile stress in the
inner flange of the frames at the
corners is higher than in the
outer flange at the sides, top
and bottom.
Outer cap
stress

Inner cap
37
stress
 Working together

Fuselage frames - failsafe

There are three failsafe
mechanisms or behaviours
shown by the frames.

1. Multiple local load paths:

each frame has more than
one member or load path
at each of the critical
locations.

2. Pin joint / hinge effect:

when a flange fails, the
other flange and skin or
strap (as applicable) work
as a pin joint connection,
at the failure location. The
load redistributes along
the frame.
38
 Working together

Fuselage frames - failsafe

3. Load transfer to adjacent

frames: frames adjacent to
the failed frame will take
additional loading.

The failsafe strength of the

fuselage was simulated by
removing elements in the FEM

39
 Working together

Critical fuselage frames

40
 Working together

Fuselage frame modifications

A typical frame is at FS 223.5.
Reinforcing is added at the top and bottom
corner inner caps, and on the outer cap
on the side

External strap Straps

Angle

41
 Working together

Frame modifications - example

Section of straps at top corner
Original section had a single strap
Second strap is wider, so edge can be Straps
inspected for cracking

Straps

Frame

42
 Working together

Frame modifications - example

Straps

43
Fuselage side failsafe strap Working together

The outer flange and skin on the fuselage side is in tension.

Frames are notched to clear stringers. The skin acts as the outer flange at
these locations
A possible failure scenario is a skin crack along the rivet line to a stringer. This
crack would result in failure of the frame.
A strap has been added to the fuselage sides at the highly loaded frames to
make the frame failsafe.

POSSIBLE CRACK LOCATION

44
Damage tolerance analysis Working together

Damage tolerance analysis has been done using AFGROW.

Crack growth life was calculated at each critical location.

Residual strength was taken as limit load.

Assumptions for initial crack size are generally:

Primary crack
For existing structure: a = 0.080
This is the size which is detectable at the time of modification.

New structure: a = 0.050

Secondary crack
All structure a = 0.005

45
Damage tolerance analysis Working together

Frame 223.5 analysis locations are shown

46
Damage tolerance analysis Working together

Location 1 & 2 - Top corner inside flange strap

47
Damage tolerance analysis Working together

Location 5 & 6 - Bottom corner inner flange angle

48
Damage tolerance analysis Working together

Location 3 & 4 - Side external strap

As well as the AFGROW crack growth analysis, calculations were done for long
cracks in the skin at the strap. These analyses used a detailed model.

49
 Working together

FEM - Detail Model of Frame

A detailed, fine mesh model of a fuselage frame and
skin was made.
This model had two functions:
• Straps were added on the side of the fuselage
to make the frames damage tolerant. The
model is used in the prediction of crack growth
and residual strength near the straps.
• The coarse mesh main model does not give
accurate stress results in the outer flange of
the frames, where the skin radius is changing
rapidly. At these locations load is shearing
from hoop tension in the skin to tension in the
frame cap. The fine model captures this effect
and gave a check on stress from the main
FEM.

50
 Working together

FEM - Detail Model of Frame

Fine mesh model of fuselage frame and skin
TOP SKIN SPLICE

WINDOW ELEMENTS
NOT SHOWN FOR CLARITY
A
A
SKIN PANEL AROUND WINDOW

FRAME FS 223.50

SIDE SKIN PANEL

FRAME MODEL (UNDERNEATH)

FRAME CUTOUT FOR STRINGER (1/2 crack FS 212.10
a=
NEXT TO CRITICAL LOCATION length)
(STRINGER NOT SHOWN)

SKIN CRACK GROWTH LINE NOTE: SKINS, WINDOW SILLS, FRAMES AND THE
(STARTS IN THE MIDDLE OF THE STRAP) STRINGER NEXT TO THE CRACK LINE WERE
MODELLED WITH SHELL ELEMENTS. ALL FASTENERS
BETWEEN MEMBERS WERE MODELLED WITH SPRING 51
ELEMENTS.
 Working together
FEM - Detail Model of Frame
The stress intensity for a long skin
crack was calculated using the
crack closure technique.
This enabled calculation of crack
growth rate and residual strength. X

The critical crack length was long Y

enough to be detected from
external inspection

52
 Working together

Forward pressure bulkhead

The forward pressure bulkhead is flat, and is located level with the
forward edge of the windscreen.
A common area for cracking is the angle between the bulkhead and the
skin at the top of the bulkhead.
It is very difficult to inspect from behind, since it is covered by the
instrument panel.

53
 Working together

Forward pressure bulkhead

The angle at the top of the frame is replaced by a modified angle.

54
 Working together

Forward pressure bulkhead

A fine model of a section of angle was made.
This enabled justification of a modified angle.

55
Fuselage bottom stringers Working together

The bottom of the fuselage is flat.

Pressure loads cannot be taken by hoop stress in the skin.
Pressure on a panel is carried by membrane and bending stress to the
adjacent frames and stringers.
Stringers act as continuous beams, supported at the frames
A fine model of a stringer was made to calculate stringer bending
stress.
Nonlinear solution gave correct membrane stresses in the skin, and
hence load transfer to the stringer

56
Fuselage bottom stringers Working together

Stringer stress is maximum at the top of the stringer, where it passes

through the frames.
There are no stress concentrations at the top of the stringer, but the
stress is high enough to give a relatively short life.
The structure is failsafe, since the stringer is now simply supported at the
frame. This failure scenario is shown below

Failed stringer

57
 Working together

Fuselage strain gauge testing

The FEM is validated by comparing results to
strain gauge tests.

Fuselage
Ground test
Pressurisation

58
 Working together

Some revised inspections

• Forward pressure bulkhead
• Fuselage frame inspections
• Wing to fuselage attachment
• Tailplane inspections
• Flap debonds
• Wiring

59
 Working together

Conclusions
Wing safe life to 40,000 hours
Empennage safe life to 40,000 hours
Some additional inspections for wings, empennage, control surfaces
Fuselage modifications required for 40,000 cycles life
New IFCA required
STC in final stages of approval

Questions 60