See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/232840579

A new value stream mapping approach for complex production systems

Article  in  International Journal of Production Research · September 2006

DOI: 10.1080/00207540600690545

CITATIONS READS
175 2,417

3 authors, including:

Marcello Braglia Gionata Carmignani

Università di Pisa Università di Pisa
116 PUBLICATIONS   3,214 CITATIONS    17 PUBLICATIONS   598 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Advanced tools for lean manufacturing and Industry 4.0 View project

All content following this page was uploaded by Marcello Braglia on 22 May 2014.

The user has requested enhancement of the downloaded file.

 International
Author manuscript, Journal
published of Production
in "International Journal Research
of Production Research 44, 18-19 (2006) 3929-3952"
DOI : 10.1080/00207540600690545

Fo
rP
peer-00512904, version 1 - 1 Sep 2010

ee

18th ICPR paper: A NEW VALUE STREAM MAPPING

APPROACH FOR COMPLEX PRODUCTION SYSTEMS
rR

Journal: International Journal of Production Research

Manuscript ID: TPRS-2005-IJPR-0503

ev

Manuscript Type: Original Manuscript

Date Submitted by the

01-Dec-2005
Author:
ie

Complete List of Authors: Braglia, Marcello; Università di Pisa, Mechanical Nuclear and
Production Engineering
w

Carmignani, Gionata; Università di Pisa, Mechanical Nuclear and

Production Engineering
Zammori, Francesco; Università di Pisa, Mechanical Nuclear and
Production Engineering
On

Keywords: LEAN MANUFACTURING, ROUTING

Keywords (user): Value Stream Mapping, Temporized Bill of Material

ly

http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk

Page 1 of 45 International Journal of Production Research

1
2
3
4
A NEW VALUE STREAM MAPPING APPROACH FOR
5 COMPLEX PRODUCTION SYSTEMS
6
7
8 Marcello BRAGLIA , Gionata CARMIGNANI , Francesco
9 ZAMMORI
10
Dipartimento di Ingegneria Meccanica, Nucleare e della Produzione,
11
12 Università di Pisa,
13 Via Bonanno Pisano, 25/B - 56126 PISA (Italy)
14
15
16
Fo

17
18
19
20
rP

21
22
23
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26
27
28
rR

29
30
31
32
33
ev

34
35
36
iew

37
38
39
40
41
42
43
On

44
45
46
47
Corresponding author:
ly

48
49 Marcello BRAGLIA
50
51 Dipartimento di Ingegneria Meccanica, Nucleare e della Produzione
52 Via Bonanno, 25/B
53
54 56126 PISA (Italy)
55 phone: +39 050 913029
56
57
Fax: +39 050 913040
58 e-mail: m.braglia@ing.unipi.it
59
60

1 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 2 of 45

1
2
3
4
A NEW VALUE STREAM MAPPING APPROACH FOR
5 COMPLEX PRODUCTION SYSTEMS
6
7
8
9
10 Abstract
11
12
13 In this paper an innovative framework to apply Value Stream Mapping to products with
14
15 complex Bill of Materials is presented. Value Stream Mapping is one of the best tools
16
Fo

17
18
to map a process and to identify its main criticalities. Unfortunately it can be effectively
19
20 applied only to linear systems. When the manufacturing process is complex with flows
rP

21
22 merging together, Value Stream Mapping cannot be straightly used. Thus, the main
23
24
peer-00512904, version 1 - 1 Sep 2010

objective of this work consists in the solution of this limitation so that lean production
ee

25
26
27 can be enhanced in complex systems too. The proposed approach is based on seven
28
rR

29 iterative steps and integrates Value Stream Mapping with other tools typical of
30
31
32 industrial engineering. The basic idea is to execute a preliminary analysis to identify the
33
ev

34 “critical production path" using the Temporized Bill of Material. Then, improvements
35
36 are made considering all possible sharing with others secondary paths, as possible
iew

37
38
39 constraints. Once the critical path has been optimized, a new one may become critical.
40
41 Thus the analysis proceeds iteratively until the optimum is reached and the Work In
42
43
On

44
Process level has decreased under a desired value. A case study taken from a real
45
46 setting environment is finally presented to asses the validity of the methodology here
47
ly

48 proposed.
49
50
51 Keywords: Lean Manufacturing, Routing, Temporized Bill of Material, Value Stream
52
53 Mapping
54
55
56
57 Words Number: 6240
58
59
60

2
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 3 of 45 International Journal of Production Research

1
2
3 1 INTRODUCTION
4
5
6 Lean Production, a concept based on Toyota Production System, has emerged recently
7
8 as a global approach that integrate different tools to focus on waste elimination and to
9
10
11 manufacture products that meet customer’s needs and expectations in a better way
12
13 (Hines and Taylor, 2000). The main concept of Lean Production consists in the
14
15
specification of what create value for the end customer and in the accomplishment of
16
Fo

17
18 this specifications by aim of a production system striving for perfection and
19
20 characterized by a strained and leveled flow, pulled by customer’s demand.
rP

21
22
23 In particular, once “value” has been defined, Lean Production development requires the
24
peer-00512904, version 1 - 1 Sep 2010

ee

25 analysis of the “value stream”, being all activities – both value-added and non-value
26
27
28
added – that are currently needed to bring a specific product from raw material to the
rR

29
30 end customers. Next, wasteful steps have to be eliminated and flow can be introduced in
31
32 the remaining “value-added” processes (Womak and Jones, 1996).
33
ev

34
35 In technical literature, various authors have defined a suite of tools and techniques to
36
iew

37 implement Lean Production in a structured way. Emiliani (2000) used the primary Lean
38
39
40 Production support tools to develop a practical solution-oriented method to achieve
41
42 business goals. The final result consists in a framework that unifies technical and
43
On

44 behavioral components of management.

45
46
47 Hines and Rich (1997) proposed a set of seven tools derived from industrial engineering
ly

48
49 to support the waste removal process.
50
51
52 In this paper we focus on Value Stream Mapping (VSM) a tool more focused on the
53
54 entire value stream of a productive process. VSM is a mapping tool that is used to map a
55
56
57 productive process or an entire supply chain networks. It maps not only material flows
58
59 but also the information flows that signal and control production.
60

3 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 4 of 45

1
2
3 Tapping et al. (2002) introduced a step by step procedure to perform a VSM analysis.
4
5
6 The first step consists in the selection of a product family as the target for the
7
8 improvement and in the construction of the “Current State Map” (CSM) for the selected
9
10
product value stream. The CSM must be based on a set of data collected directly on the
11
12
13 shop floor and should be drawn using the set of standard icons shown in Figure 1.
14
15
--------------------------
16
Fo

17 Figure 1 Here
18 --------------------------
19
20
rP

21 The next step consists in the identification and analysis of the wastes encountered along
22
23 the value stream. Finally a “Future State Map” (FSM) is design to represent the ideal
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26 production process without the removed wastes.
27
28 Also the FSM should be drawn using a set of standard icons shown in Figure 2 and it
rR

29
30
31 can be obtained answering eight questions reported in Table 1 (Lian and Van
32
33 Landeghem, 2002).
ev

34
35
36 --------------------------
Figure 2 Here
iew

37
38
39
--------------------------
40
41 --------------------------
42
43 Table 1 Here
On

44 --------------------------
45
46
47 In respect to other mapping techniques, VSM offers several advantages:
ly

48
49
50
• it forms the basis for Lean Production implementation;
51
52 • it relates the manufacturing process internal to the facility to the whole Supply
53
54
55
Chain;
56
57 • it displays both the product flow and the information flow;
58
59
60 • it links “Products Planning” and “Demand Forecast” to “Production Scheduling” and

to “Flow Shop Control”;

4
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 5 of 45 International Journal of Production Research

1
2
3 • it includes information related to production time as well as information related to
4
5
6 inventory levels.
7
8 Unfortunately VSM has also two main drawbacks:
9
10
11 • basically it is a paper and pencil based technique, thus the accuracy level is limited
12
13 and the number of versions that can be handled is low;
14
15
16 • in real settings, many companies are of a “high variety-low volume type”, meaning
Fo

17
18 that many value streams are composed of hundreds of industrial parts and products.
19
20
Thus this complication cannot be addressed with the standard method.
rP

21
22
23 In particular VSM can be effectively used only for productive systems characterized by
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26 linear product routings. If the production process is complex VSM application breaks
27
28 down as it fails to map value streams characterized by multiple flows that merge. This
rR

29
30 typically happens for products described by a complex Bill of Material (BOM),
31
32
33 manufactured in a job shop facility. In such cases Rother and Shook (1999) suggested to
ev

34
35 map only the key elements of the flow and to “draw one flow over another” if
36
iew

37
38
necessary. Still no decisional process is proposed to choose the key elements of the
39
40 value stream. Furthermore, identifying which flows are identical or even similar is not a
41
42 trivial activity for any but the easiest of the real production process.
43
On

44
45 In technical literature these problems have been firstly addressed in three different
46
47 works.
ly

48
49
50 McDonald et al (2002) applied VSM to a “three parallel lines assembly process” using
51
52 discrete event simulation to define the basic parameters for the FSM. They
53
54 demonstrated that, in case of production complexity, simulation can provide important
55
56
57 information for the FSM implementation.
58
59 Lyan and Van Landeghem (2002), followed a similar approach to map a “two parallel
60
line” push system. Two simulation models were built for the push and pull systems

5 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 6 of 45

1
2
3 respectively and the key measurements such as lead times, throughput rates, value-
4
5
6 added ratios were compared as well as evaluated.
7
8 Khaswala and Irani (2004) improved a welding job shop facility by aim of a new
9
10
mapping approach called “Value Network Mapping”. This technique was derived from
11
12
13 an integration of PFAST (Production Flow Analysis and Simplification Toolkit) and
14
15 VSM (Irani and Zhou, 2003). It was proved that this approach supports facility
16
Fo

17
18
improvements, such as the creation of manufacturing cells and improvements in the
19
20 current material handling methods.
rP

21
22
23
In this work, an alternative and innovative framework for a structured application of
24
peer-00512904, version 1 - 1 Sep 2010

VSM to products requiring non-linear value streams, is proposed. The framework is

ee

25
26
27 based on a recursive procedure and integrates the classic VSM technique with different
28
rR

29
30 tools derived from the manufacturing engineering area. The underlying idea among the
31
32 method is to execute a preliminary analysis to identify the longer critical production
33
ev

34 path by aim of the Temporized Bill of Material (TBOM). Clearly the improvement
35
36
process will start from the critical path that is responsible for the whole Lead Time of
iew

37
38
39 the productive process. Once the critical path has been identified, possible
40
41
improvements are searched, considering all sharing with secondary paths as further
42
43
On

44 constraints. Finally, when the main value stream has been improved, a new path may
45
46 become the critical one. Thus the analysis proceeds iteratively until the optimum is
47
ly

48
49
reached or the Work in Process (WIP) level has decreased under the desired level. In
50
51 that way the framework makes it possible to explore the overall production process
52
53 determining the correct order of the path to be improved.
54
55
56
57
58
59
60

6
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 7 of 45 International Journal of Production Research

1
2
3 2 IMPROVED VALUE STREAM MAPPING PROCEDURE (IVSM)
4
5
6 The framework follows an iterative procedure based on the set of seven steps listed
7
8 below:
9
10
11 1. select a product family;
12
13
14 2. identify machine sharing;
15
16 3. identify the main value stream;
Fo

17
18
19 4. map the critical path;
20
rP

21
5. identify and analyze wastes;
22
23
24 6. map the future state for the critical/sub-critical path;
peer-00512904, version 1 - 1 Sep 2010

ee

25
26
27 7. identify the new critical path and iterate the process;
28
rR

29 The procedure is shown as a flow diagram in Figure 3.

30
31
32 --------------------------
33 Figure 3 Here
ev

34
35 --------------------------
36
iew

37
2.1 Step 1: “Select a Product Family”
38
39
40 The first step of the framework consists in the identification of the product families and
41
42
in the selection of one as the initial target for the improvement.
43
On

44
45 Performing a Product Quantity Analysis (PQ Analysis) is a classical approach for
46
47
product mix segmentation. This method is used to display the product mix in the form
ly

48
49
50 of a Pareto chart, sorting the products in order of decreasing production quantity. In that
51
52 way it is possible to understand how the total quantity is distributed among different
53
54
55
product types. The underlying logic of PQ analysis is that high volume products are
56
57 responsible for the largest part of non-value added costs such as: material handling,
58
59 WIP, queuing and other operational costs. Therefore, focusing on high volume value
60
stream should boost the overall performances of the factory. When a more

7 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 8 of 45

1
2
3 representative product set is needed, it is possible to include the sell revenue as an
4
5
6 additional parameter for the ABC analysis. To do that, two distinct Pareto Analysis”
7
8 (PQ$ analysis) must be performed. The first one should sort the products in order of
9
10
decreasing production quantity, while the second one should sort the products in order
11
12
13 of decreasing selling revenue. Therefore, the combination of the two Pareto analysis
14
15 leads to the definition of nine distinct products classes ranging from AA (upper grade)
16
Fo

17
18
to CC (lower grade). Such interrelation can then be expressed using a classification
19
20 matrix, which show how products are distributed among each of the nine classes
rP

21
22 previously defined. An example of a PQ$ Pareto Analysis is shown in Figure 4.
23
24
peer-00512904, version 1 - 1 Sep 2010

--------------------------
ee

25
26 Figure 4 Here
27
28 --------------------------
rR

29
30
31 Although PQ$ analysis is the easiest way for product mix segmentation, when the
32
33 Pareto chart displays a 40:60 or even less PQ ratio (as typically happens in job shop
ev

34
35 facility), a different approach is needed to choose the target value stream.
36
iew

37
38 Since a product family is defined as “a group of products that pass through similar
39
40 processing steps and over common equipment in the downstream processes”, it is
41
42
43 possible to group product into families through the analysis of the Multi Product
On

44
45 Process Chart (MPPC) or, alternatively, of the Machine-Part Matrix (MPM). An
46
47
example of a Multi product Matrix is shown in Table 2.
ly

48
49
50
51
52
53
54
55 --------------------------
56
57 Table 2 Here
58 --------------------------
59
60

8
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 9 of 45 International Journal of Production Research

1
2
3 It is important to note that the use of a MPM is preferable if dealing with a great number
4
5
6 of products, for it can be directly used as the input for a “cluster analysis” procedure,
7
8 making easier the identification process. To do that, part families are typically obtained
9
10
grouping items by aim of hierarchical methods on the basis of the Pair-Wise Similarity
11
12
13 Coefficients (PWSC) between parts, obtained from the MPM matrix.
14
15
Generally the PWSC are obtained through the “Jaccard” similarity function that is
16
Fo

17
18 reported below:
19
20
Xij + Xij Yij
rP

21
Sij = (1)
22
23
Xij + Xi + Xj + Xij Yij
24 where:
peer-00512904, version 1 - 1 Sep 2010

ee

25
26
27 • 0<= Sij <=1;
28
rR

29 • Xij = Number of machines used by both part “i” and part “j” (number of matches);
30
31
32 • Xi = Number of machines used by part “i” only;
33
ev

34
35
• Xj = Number of machines used by part “j” only;
36
iew

37 • Yij =Number of machines that are used neither by part “i” nor by part “j” (number of
38
39
40
misses);
41
42 For further explanation on the most used similarity function please see also Veramani
43
On

44
45
and Mani (1996) and Offodile and Grznar (1997).
46
47 An example of cluster analysis application concerning the formation of part families is
ly

48
49
50
briefly reported in Table 3 and Figure 5. In particular, Table 3 shows the PWSC
51
52 obtained applying the Jaccard similarity function to the MPM presented in Table 2,
53
54 while shows the resulting dendrogram of the part families , obtained using the “single
55
56
57 linkage clustering algorithm (SLCA). Figure 5 shows the resulting dendrogram for the
58
59 obtained part families.
60

9 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 10 of 45

1
2
3 --------------------------
4
5 Table 3 Here
6 --------------------------
7
8
9 --------------------------
10 Figure 5 Here
11
12 --------------------------
13
14 Finally, once items have been grouped into clusters, the target value stream can be
15
16
Fo

17 identified simply applying a PQ$ analysis to the part families themselves, as previously
18
19 explained.
20
rP

21
22
23
24 2.2 Step 2: “Identify machine sharing”
peer-00512904, version 1 - 1 Sep 2010

ee

25
26
27 The second step of the framework consists in the identification of the machines that,
28
rR

29 being shared by more than a product family, act as a possible constraint for lean
30
31
32 production implementation. In particular, to achieve a continuous and leveled flow, a
33
ev

34 facility layout must possess the following characteristics:

35
36
1) flow between consecutive pairs of operations takes place over short travel distances;
iew

37
38
39 2) flow is unidirectional with minimum backtracking or cross flows between machines;
40
41
42 3) flow paths in the cell must consist of smooth contours.
43
On

44
45
Machines that are dedicated to a single product family represent an opportunity to
46
47 improve the layout structure through the implementation of autonomous “Cellular
ly

48
49 Manufacturing System” (CMS). On the contrary, machines that are shared by products
50
51
52 belonging to different families, act as constrains to the fulfillment of the layout
53
54 requirements. This is called “external sharing” due to the fact that, unless machine
55
56 duplication, it will be impossible to develop a set of completely unrelated
57
58
59 manufacturing cell. Thus some products will be forced to move from one cell to another
60
one, to complete its own processing sequence. Furthermore, external sharing represents

10
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 11 of 45 International Journal of Production Research

1
2
3 an additional constrain to flow synchronization between cells. Thus, before proceeding
4
5
6 with VSM, it is necessary to identify all the machines that are used by more than a
7
8 family.
9
10
11 External sharing can be easily derived from the analysis of the MPM, after the order of
12
13 both its columns and its rows have been rearranged in the following way:
14
15
16 • the columns (part labels) should be interchanged in order to reproduce the part
Fo

17
18 families previously identified;
19
20
• the rows (machine labels) should be interchanged in order to gather along the main
rP

21
22
23 diagonal the maximum amount of positive elements (elements equals to 1).
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26 In this way the machine tends to cluster into autonomous CMS, each one corresponding
27
28 to a single part family. On the contrary the positive elements distributed far from the
rR

29
30 diagonal represent the parts that need to be moved from one cell to another, or machines
31
32
33 that need to be replicate in another cell. The MPM can be reallocated through an
ev

34
35 automatic procedure essentially in two different way: “array based method” or
36
iew

37
38
“hierarchical clustering method”.
39
40 The array-based methods can be stated as follows: “given a Boolean incidence matrix
41
42
43
(MPM), partition it into a maximal number (C) of clusters, C > 2, such that the number
On

44
45 of exceptional elements is minimized” (Selim et al, 2003). Many array based methods
46
47 are reported in literature (Veramani and Mani, 1996; Offodile and Grznar 1997; Selim
ly

48
49
50 et al, 2003). All array based methods proceed by rearranging the rows with the leftmost
51
52 positive cells to the top and the columns with the top-most positive cells to the left of
53
54
the MPM, so that, after several iterations, all the positive cells will form diagonal blocks
55
56
57 from the top left corner to the bottom-right corner of the matrix. On the other hand,
58
59 hierarchical clustering methods make a partition into clusters for both machines and
60

11 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 12 of 45

1
2
3 parts and then rearrange the rows and the columns of the MPM, accordingly with the
4
5
6 defined families, to obtain the same “diagonal block form” of the matrix.
7
8 In Table 4 is shown a possible solution to the stated problem for the MPM matrix
9
10
11 reported in Table 2.
12
13 --------------------------
14
15 Table 4 Here
16 --------------------------
Fo

17
18
19 As displayed by the matrix, there are six (out of fifteen) machines affected by “external
20
rP

21 sharing”. Considering Family Three (F#3) as an example, one can notice that M#15 and
22
23
24
M#12 are used also by products belonging to F#1 and F#2 respectively. For this reason
peer-00512904, version 1 - 1 Sep 2010

ee

25
26 it would not be possible to analyze F#3 value stream as if it was independent from the
27
28 other families. In particular, due to machine sharing, two additional constrain will have
rR

29
30
31 to be added.
32
33 In agreement with the “Lean principle” that states to substitute single high volume
ev

34
35
36 equipment with a set of similar, yet smaller and more flexible machines, it seems
iew

37
38 plausible to replace M#15 with two independent machines. In fact in this case both the
39
40 new machine would experiment a high level of usage being used by two and three
41
42
43 products respectively. The same cannot be said for M#12 which is used by P#3 only. In
On

44
45 this case, if an additional machine was placed in the third cell it will experiment a low
46
47
usage rate. A possible solution to synchronize the flow and to develop a pull system
ly

48
49
50 without an additional machine is displayed in Figure 6.
51
52
--------------------------
53
54 Figure 6 Here
55 --------------------------
56
57
58 Due to the fact that M#12 is shared, it cannot be leveled on the mix micro and must
59
60 produce in batch, multiple of the pitch. Still M#12 needs to be linked to the other cell in

12
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 13 of 45 International Journal of Production Research

1
2
3 order to form a pull system. For that reason, when the stock quantity in front of M#13
4
5
6 decreases under a certain level, a Signal Kanban (SK) is delivered to M#12. Please note
7
8 that a SK is a printed card indicating the cumulative number of products that need to be
9
10
produced to replenish what have been pulled from the downstream process. The same
11
12
13 process connects M#12 to M#11. It is important to underline that even though M#11
14
15 receives order for P#3 in the form of a SK, it must produce maintaining a leveled
16
Fo

17
18
production sequence (mix micro), without manufacturing the whole batch reported in
19
20 the SK at the same time. The first reason to do that is to maintain the “Takt Image” (that
rP

21
22 is the capability to produce each part every fixed time), while the second is to prevent
23
24
peer-00512904, version 1 - 1 Sep 2010

M#13 from starving.

ee

25
26
27
28
rR

29
30 2.3 Step 3: “Identify the main value stream”
31
32
33
As previously stated, when dealing with non-linear production processes characterized
ev

34
35 by multiple flows that merge, mapping and improving the whole process at the same
36
iew

37 time, is not easy and usually not even feasible. Thus the mapping process should begin
38
39
40 from the main (critical) value stream. Afterwards, following an iterative process, the
41
42 analysis should embrace the other branches of the process in a second moment and only
43
On

44 if needed to improve the overall performances of the system.

45
46
47 One of the main objective of VSM is to reduce the time it takes a piece to move all the
ly

48
49 way through the process (i.e. Lead Time) and consequently to cut down the WIP as well
50
51
52 as the finished good inventories. Thus the critical path can be defined as the processing
53
54 sequence which is responsible for the total production time that determines the
55
56 minimum time frame needed to schedule production in advance. In that sense, the
57
58
59 underlying logic beyond the proposed approach states that if the critical path is targeted
60
as the starting point for the analysis, then any improvements obtained through VSM is

13 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 14 of 45

1
2
3 directly converted in an overall LT reduction. The BOM is the element that links the
4
5
6 productive sequence with the LT needed to perform each manufacturing operation, so it
7
8 is natural to derive the critical path from the analysis of the BOM itself. For that
9
10
purpose in particular, the TBOM is preferable. This is because the TBOM is a graphical
11
12
13 representation of the operations as well as the time taken to manufacture a product. In
14
15 particular, in a TBOM the vertical segments linking each level of the productive
16
Fo

17
18
process, are proportional in length to the LT requested to perform the corresponding
19
20 operation. Therefore, as shown in the Figure 7, the critical path can be simply obtained
rP

21
22 selecting the production sequence characterized by the maximum vertical length.
23
24
peer-00512904, version 1 - 1 Sep 2010

--------------------------
ee

25
26 Figure 7 Here
27
28 --------------------------
rR

29
30 The TBOM shows both the supply and the production LT with a hatch and straight line
31
32
33 respectively. Therefore, before choosing the critical path, it is necessary to asses if the
ev

34
35 VSM is intended to cover the global value stream, or if it is just intended to cover the
36
iew

37
38
manufacturing value stream inside the facility. In the second case, the first level of the
39
40 TBOM referred to suppliers should not be considered in the critical path selection.
41
42
43
As exemplified in Figure 7, in order to simplify the mapping process, it is advisable to
On

44
45 identify on the critical path the manufacturing operations that are performed with a
46
47 shared resource. It is important to make a clear distinction between internal sharing (i.e.
ly

48
49
50 a machine that is used by different components of the same product) and external
51
52 sharing (i.e. a machine that is used by different product families).
53
54
55
56
57 2.4 Step 4: “Critical path mapping”
58
59
60 The fourth step consist in the construction of the “Current-State-Map” for the critical

value stream.

14
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 15 of 45 International Journal of Production Research

1
2
3 Data collection should follow the approach recommended by Rother and Shook (1999).
4
5
6 It should begin at shipping, working backwards in the production process to row
7
8 materials or suppliers and collecting snapshot data on inventory level at each stage of
9
10
the value stream. Since VSM has the unique feature to record the information flows
11
12
13 associated with the material flows in the same map, the following data should be
14
15 collected:
16
Fo

17
18 • Information Flow:
19
20
- type of orders released by customers (ordinary, planned, opens etc.);
rP

21
22
23 - type of orders released to first tier suppliers (ordinary, planned, opens etc.);
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26 - ordering frequency;
27
28 - forecast frequency;
rR

29
30
31 - system used to plan production;
32
33
ev

34
- time frame to plan production;
35
36 • Machines:
iew

37
38
39 - set up time;
40
41 - up time;
42
43
On

44 - cycle time;
45
46
47 - n° of operators;
ly

48
49 - n° of shifts/hours per day;
50
51
52 - cycle time;
53
54
55
• Production Flow:
56
57 - average customers demand;
58
59
60 - shipping frequency;

15 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 16 of 45

1
2
3 - pallets dimension;
4
5
6 - production batches;
7
8
9 - inventory levels;
10
11 - type of flow between machine (push-pull).
12
13
14 When all the data have been collected, the CSM can be constructed using the VSM
15
16 standard icons. Due to the fact that the obtained representation is limited to the critical
Fo

17
18
19 path, there is the additional need to identify on the map all the machines that are shared,
20
rP

21 as well as the insertion points where secondary branches merge with the main value
22
23 stream. Moreover, when a machine is shared, the percentage of time that is used by each
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26 family must be inserted as an additional data in the corresponding “Machine Data Box”.
27
28
rR

29
30
31 2.5 Step 5: “Waste identification”
32
33
ev

34
The fifth step of the framework consist in the analysis of non value added activities that
35
36 are displayed on the actual state map. To do that it can be helpful to use a suite of tools
iew

37
38 expressly designed to identify and to remove waste origin in individual value stream
39
40
41 (Hines and Rich, 1997).
42
43 Once the wastes and the relative causes have been detected, it is necessary to evaluate if
On

44
45
46 such problems are linked to inefficiency spread among the whole critical stream or if
47
they are concentrated in correspondence of the insertion points.
ly

48
49
50 This should be the case of an assembly station fed by two items “A” and “B” coming
51
52
53 from the main and a secondary value stream respectively. If the average LT for that
54
55 station is high due to frequent starving caused by lacking in product B replenishment,
56
57
then the cause of the waste (which correspond in an abnormal increase of A stock level
58
59
60 in front of the station) must be searched in the secondary branch of the process. To

highlight such situation it could be useful to split the downtime recorded for each

16
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 17 of 45 International Journal of Production Research

1
2
3 production station into its main components that are: “breakage”, “idle time” and “lack
4
5
6 of operator”. If no major wastes linked with inefficiencies on the secondary streams
7
8 have been detected, the framework proceeds with Step 6.1. Vice-versa the frameworks
9
10
proceeds with Step 6.2.
11
12
13
14
15
16 2.6 Step 6.1: “Main Stream Future state mapping”
Fo

17
18 The sixth step consists in the construction of the FSM for the critical value stream. The
19
20
map should be defined using the VSM standard icons and trying to answer the “eight
rP

21
22
23 future state question”.
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26 To do that the following approach can be used:
27
28 a) Using the TBOM, divide the critical value stream into a set of adjacent segments.
rR

29
30
31 Each segment must be confined between two consecutive insertion point, with the
32
33 exception of the first one. This one starts with the first operation of the process and
ev

34
35 end at the first insertion point.
36
iew

37
38 b) Select the last segment as the future-state implementation starting point:
39
40
41 b.1.) If machines are not shared, join them into a cell (continuous flow processing)
42
43 whenever possible. “One by one manufacturing is in fact the best way to
On

44
45 level flow and to reduce inventory level. Alternatively use a Kanban pull
46
47
system to link machines and to synchronize flow between them.
ly

48
49
50 b.2.) If a machine is affected by internal sharing, as shown in Figure 8 (where both
51
52
53 RM1 and B1 are components of the same product), two alternative solutions
54
55 are possible:
56
57
58
59
60 --------------------------
Figure 8 Here

17 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 18 of 45

1
2
3 --------------------------
4
5
6 b.2.1.) Due to the fact that the Takt Time is the same for both operation on M2
7
8 (internal sharing), if the Cycle Time of the three machines considered is
9
10
11 equal or less a half Takt Time (CT1&CT2&CT3 <= 0.5*Takt Time), then
12
13 M1, M2, M3 can be joined into a manufacturing cell and a continuous flow
14
15
process can be established. In this hypothetical situation two operators are
16
Fo

17
18 needed. The first one is fixed on M2 while the second skips alternatively
19
20 between M1 and M2. The main constrain to this solution is constituted by
rP

21
22
23
machines Cycle Time and by Set-Up time that must be practically reduced
24
peer-00512904, version 1 - 1 Sep 2010

to zero.
ee

25
26
27
28
b.2.2.) Alternatively it is possible to use a “FIFO Lane” served by three operators,
rR

29
30 to connect each machine. The only constrain to this solution is represented
31
32 by M2 Cycle Time that ,as previously, must be equal or less than a half
33
ev

34
35 Takt Time. Using this solution is also advisable to define “a priori” the
36
production sequence on M2, in order to maintain the correct Takt Image.
iew

37
38
39
40 b.3.) If a machine is affected by external sharing, generally the only available
41
42 solution is represented by the use of a “Signal Kanban”. In that way, the
43
On

44 shared resource is allowed to produce in batches bigger then the optimum

45
46
47 pitch, still pulled by the downstream resources.
ly

48
49 c) Typically, in correspondence of an insertion point a continuous flow is not feasible.
50
51
52 This is due to the fact that production information must be simultaneously
53
54 communicated to two or even more value streams that merge at that point of the line.
55
56 Thus, on the main value stream, the station at the insertion point can be included in a
57
58
59 continuous flow only if a visual control based on a reorder point is sufficient to pull
60
production of the other value stream and a leveled pace is not essential. In all other

18
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 19 of 45 International Journal of Production Research

1
2
3 cases, an alternative solution is to establish flow using different lean techniques such
4
5
6 as the single – double Kanban pull system.
7
8 d) Choose the next segment of the original value stream and apply iteratively the
9
10
11 method starting from step “b” to step “d” until all the adjacent segments have been
12
13 considered.
14
15
16
Fo

17
18 2.7 Step 6.2: “Secondary Stream Future state mapping”
19
20
rP

21 When the main problems are caused by inefficiency located on a secondary branch of
22
23 the process, this branch must be analyzed before performing any attempt to improve the
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26 main value stream. To do that, an CSM of the secondary value stream must be construct
27
28 following the approach described in Step 4 and 5. When the CSM has been made, the
rR

29
30 cause of waste must be identified and a FSM derived to. eliminate the inefficiencies for
31
32
33 the critical path. The FSM of the secondary branch must be developed following the
ev

34
35 approach described in Step 6.1.
36
iew

37
38 Finally, when the secondary branch of the process has been optimized, the analysis skip
39
40 back to the main value stream. If its external problems have been solved it can be finally
41
42
43
improved following the approach described in “Step 6.1”.
On

44
45
46
47
2.8 Step 7: “Process Iteration”
ly

48
49
50 The seventh step of the framework consists in the application of the method to other
51
52
53 branches of the productive process. To choose the next value stream to be taken as the
54
55 targeted for the improvement, the TBOM need to be modified in relationship to the LT
56
57
reduction obtained through the modifications introduced in the process. In that way it is
58
59
60 possible to asses if the new total LT is still determined by the improved main value

stream or if it is now determined by another value stream.

19 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 20 of 45

1
2
3 In the first case, the procedure can be arrested. In fact, from a further application of the
4
5
6 method, a reduction in inventory and handling cost should be obtainable, yet it would be
7
8 impossible to reduce the total production time any more.
9
10
11 On the contrary, if the total production time is now determined by another branch, this
12
13 becomes the new critical path and the procedure restart from Step 4. Proceeding in this
14
15
way it is possible to analyze the various branches of the productive process, following a
16
Fo

17
18 structured iterative procedure that comes to an end when the global process has been
19
20 optimized or, alternatively, when the total production time can not be shortened any
rP

21
22
23
more.
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26
27
28 3 DESCRIPTION OF CASE APPLICATION
rR

29
30 To asses the validity of the method here proposed, a real productive process was
31
32
33 analyzed and a possible lean solution developed. This application was implemented
ev

34
35 within an electro domestic manufacturing firm considering in particular the plant
36
iew

37
38
devoted to refrigerators production. The organization’s identity is protected and
39
40 productive information cannot be released, thus some data were purposely altered. In
41
42 the next part of the paper, we will simply refer to the plant as Refrigerator Manufacture
43
On

44
45 (RM).
46
47 To execute the IVSM approach, a design team was assembled with people having
ly

48
49
50 different skills and competences, with the aim to define the actual state map and the
51
52 desired future state map for the 150 liters refrigerator line. This kind of domestic
53
54 refrigerator was chosen for it accounts for 60% of the overall production. In particular,
55
56
57 RM manufacturers three families of 150 liters refrigerators, type A, type B and type C
58
59 in same quantities. The productive process is similar for all three types and, across each
60
families, there are just little differences both in processing time and in set up times.

20
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 21 of 45 International Journal of Production Research

1
2
3 As shown in Figure 9, each 150 liters refrigerator is manufactured as four parallel
4
5
6 subassemblies (body, door, condenser pipe, electronic devices), each one converging in
7
8 a different stages of the main assembly process.
9
10
11 --------------------------
12 Figure 9 Here
13
14 --------------------------
15
16 A weekly MRP schedule, based on demand forecast, is generated for each stream of the
Fo

17
18
19 process to push orders through production. In addition, a supervisor is in charge to
20
rP

21 control production and to make daily adjustments to solve possible problems and to
22
23 fulfill urgent orders.
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26
27
28
3.1 Actual State Mapping
rR

29
30
31 To built the CSM, the TBOM, was firstly analyzed to identify the critical value stream
32
33
ev

34
determining the total production LT. As clearly shown in Figure 10, the critical value
35
36 stream begins with sheets cutting, proceeds with the refrigerator body assembly and
iew

37
38 ends at the final check and packaging. The estimated Lead Time is equal to seven days,
39
40
41 that is the period used to schedule production in advance.
42
43 --------------------------
On

44
45 Figure 10 Here
46 --------------------------
47
ly

48
49 Once the critical path was identified, the next step was to collect all data necessary to
50
51 create the CSM of the process. This was done walking backwards from the “packaging”
52
53
54
to the “sheets cutting”, taking snapshot of the encountered WIP and inventories level.
55
56 The obtained CSM for the critical value stream is shown in Figure 11.
57
58
--------------------------
59
60 Figure 11 Here
--------------------------

21 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 22 of 45

1
2
3 Please note that only the operations belonging to the critical path are represented, while
4
5
6 the secondary branches are drawn as “empty boxes”, just to show their insertion points
7
8 in the main value stream. Another difference with traditional CSM consists in the fact
9
10
that some process boxes are represented with a dotted interior. This icon is used to
11
12
13 identify the shared resources. In particular RM manufactures also commercial
14
15 refrigerators with an interior capacity of 300 liters. These are manufactured on a
16
Fo

17
18
dedicated assembly line, with the exception of the sheets cutting operations, that are
19
20 made with the same equipments used for 150 liters refrigerators. Sharing between 150
rP

21
22 and 300 liters refrigerators is present also in the secondary branches. In particular “tubes
23
24
peer-00512904, version 1 - 1 Sep 2010

production” and “electronic devices” assembly line are completely shared, while only
ee

25
26
27 the last operation of the “door production” line is shared.
28
rR

29
30 For what concerns the data reported inside the process boxes, both Cycle Time (CT) and
31
32 Change Over Time (CO) are determined based on a weighted average of the three
33
ev

34 families of 150 liters refrigerators. This seemed to be a meaningful approach because

35
36
across each families, there are just little differences in times. Furthermore Up Time
iew

37
38
39 (UT) denotes the percentage of time that a resource is available for production and it
40
41
takes into account time lost for breakages, set up and, in the case of shared resource, the
42
43
On

44 percentage of productive capacity dedicated to a product family.

45
46
The time line at the bottom represents manufacturing processing time and inventory
47
ly

48
49 time. The estimated LT is equal to 41 hours while the value added time is only 39
50
51 minutes, so there is a great margin of improvement.
52
53
54 3.2 Criticalities analysis
55
56 The next step was the definition of a new system, to pull orders through the line. Before
57
58
59 doing this it was necessary to analyze the inefficiencies encountered on the process and
60
evaluate if such problems were due to inefficiency spread among the whole critical

22
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 23 of 45 International Journal of Production Research

1
2
3 value stream or due to inefficiencies concentrated in correspondence of the insertion
4
5
6 points. As a results it was observed that the insertion points corresponding to the
7
8 conjunction of the main value stream with the “tube assembly” lines and with the
9
10
“electronic devices preassembly” line were not critical ones. Although both processes
11
12
13 are shared (a typical cause of criticality) operations are manual with short cycle time
14
15 and qualified worker are not requested. Thus it is easy to reallocate workers in case of
16
Fo

17
18
necessity, so that the main value stream can always be feeded. In addition, the output of
19
20 both these processes is a low cost generic WIP, so maintaining a buffer to separate
rP

21
22 processes is acceptable, for it does not increase costs and facilitates production control.
23
24
peer-00512904, version 1 - 1 Sep 2010

On the contrary, the insertion point corresponding to the conjunction of the main stream
ee

25
26
27 with the “door assembly” line was judged to be critical. In fact the “door assembly”
28
rR

29
30 station, that is the process bottle neck, was usually starving for shortage in doors
31
32 replenishment. For that reason, the throughput was reduced and the LT increased.
33
ev

34
35 Thus, an additional map was developed to analyze the causes of this inefficiency. The
36
obtained CSM for the sub- critical value stream is shown in Figure 12.
iew

37
38
39
40 --------------------------
41 Figure 12 Here
42
43
--------------------------
On

44
45 It was found that the problem was linked with the presence of sharing at the “seal
46
47
insertion” machine. To minimize set up when production skipped from 150 liters to 300
ly

48
49
50 liters refrigerators, orders were scheduled in large batch, increasing the probability of
51
52 starving at the “door assembly station”.
53
54
55 3.3 Future State Mapping
56
57
Following the proposed approach, the future state map was developed splitting the value
58
59
60 stream into three consecutive parts, that were improved in the following order:

23 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 24 of 45

1
2
3 1. the process upstream the critical insertion point;
4
5
6 2. the secondary but critical value stream (door assemble line);
7
8
9 3. the process downstream the critical insertion point.
10
11 Figure 13 shows the obtained future state map for the process.
12
13
14 --------------------------
15 Figure 13 Here
16
Fo

17 --------------------------
18
19 It can be noted that the “final assembly” station was chosen as the pacemaker process to
20
rP

21
22 pull orders through the line. Order are released from the production control and inserted
23
24 in the backlog list at the pacemaker process. Then order are released backward in the
peer-00512904, version 1 - 1 Sep 2010

ee

25
26
line by aim of a kanban system.
27
28
rR

29 Whenever possible production station have been re-organized into “U shaped cell” in
30
31
order to obtained a “one by one” production flow. Alternatively productions were
32
33
ev

34 connected using FIFO lines. In that way the process was divided into four “micro
35
36 factories” synchronized through a kanban system.
iew

37
38
39 To level the production mix at the pacemaker process, an EPE (every part every ) of 4
40
41 hours was established. So all three families will be manufactured twice a day. This was
42
43
On

44
due to the fact that, even reducing change over time of 30-40%, the maximum number
45
46 of set up at the foaming machine is equal to eight, as shown in the following relation:
47
ly

48
49 Tot. AvailableTime Tot.ThroughputTime 8 60 (1 300 + 15)
N °SetUP = = = 8 (2)
50 AverageSetUpTime 20
51
52
53
54
55 In particular choosing 6 set up per day gives an addition margin of security, to face
56
57 potential problems or variation in the demand level.
58
59
60 As shown in Figure 13, a Signal Kanban corresponding to a batch of 150 items, is

needed to link the “sheets cutting” machines with the assemble line. This was due to the

24
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 25 of 45 International Journal of Production Research

1
2
3 fact that, even reducing change over time of 50%, the maximum number of set up is
4
5
6 equal to 4, thus assuring the production of two families per day.
7
8 This is shown in the following relation:
9
10
11 8 60 0.7 (25 600) / 60
12 N °SetUP = =4 (3)
13 20
14
15 where:
16
Fo

17
18 70% is the productive capacity assured to the 150 liters refrigerators line;
19
20
600 is the number of items processed (300 refrigerator bodies + 300 refrigerator
rP

21
22
23 doors).
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26 Please note that this solution requires the presence of a buffer with an average WIP
27
28 level equal to 225 items corresponding to 6 hours. This is acceptable, for the WIP at this
rR

29
30 stage is not expensive and it does not increase manufacturing costs but facilitates
31
32
33 production control.
ev

34
35 By aim of these improvements, the total productive LT has been reduced to 13 hours,
36
iew

37
38 corresponding to a 68% reduction in respect to the initial 41.5 hours.
39
40
41
42
43 4 CONCLUSIONS
On

44
45
46
The paper focused on VSM that has been indicated as one of the best tool for Lean
47
Production implementation in a facility. This method is not easy to use in the case of
ly

48
49
50 complex production processes characterized by multiple flows that merge. To address
51
52
53 this problem a new framework “Improved Value Stream Mapping” has been developed.
54
55 This was done through an integration of the standard VSM approach with a set of
56
57 additional tools derived from the manufacturing engineering area. In particular the
58
59
60 framework is structured in order to explore the various branches of the whole value

stream only if necessary and following an optimized order.

25 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 26 of 45

1
2
3 To be proved, the framework was finally applied to a refrigerators production process.
4
5
6 As a results it was demonstrated that IVSM has the following main advantages:
7
8 • it offers a structured way to choose the key elements of a complex production
9
10
11 process and to perform an optimized analysis of the whole value stream;
12
13 • it is successful in handling multiple products having non identical routings;
14
15
16 • it is successful in handling complex BOM with multiple level as well as multiple
Fo

17
18 flows that merge.
19
20
rP

21 Future works are intended to integrate the approach with economic measures to express
22
23 both value added and non value added costs sustained through the process. An
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26 additional field for future research can be that to evaluate also the variances of the
27
28 analyzed manufacturing process, by aim of a statistical and/or fuzzy approach.
rR

29
30
31
32
33
ev

34
35
36 5 REFERENCES
iew

37
38
39
Emiliani M., 2000, Cracking the code of business, Management Decision, 38, 60-79.
40
41 Hines P., Rich N., 1997, The seven value stream mapping tool, International Journal of
42
43
On

44
Operations & Production Management, 17, 46-64.
45
46 Hines P., Taylor D., 2000, Going Lean, Cardiff Lean Enterprise Research Center,
47
ly

48
49
Cardiff Business School.
50
51 Irani A, Khaswala N.,., 2000, Value Network Mapping (VNM): visualization and
52
53
54
analysis of multiple flows in Value Stream Maps, www-iwse.eng.ohio-state.edu/
55
56 (Department of Industrial, Welding and System Engineering, The Ohio State University
57
58 Columbus Ohio 43210).
59
60

26
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 27 of 45 International Journal of Production Research

1
2
3 Lian Y., Van Landeghem H., 2002, An application of simulation and Value Stream
4
5
6 Mapping in Lean Manufacturing, 14th European Simulation Symposium, SCS Europe
7
8 BVBA.
9
10
11 McDonald T., Van Aken E.M., Rentes A.F., 2002, Utilising Simulation to Enhance
12
13 Value Stream Mapping: A Manufacturing Case Application, International Journal of
14
15
Logistic: Research and Application, 5, 213-232.
16
Fo

17
18 Offodile O.F., Grznar J., Part family formation for variety reduction in flexible
19
20
manufacturing systems, 1997, International Journal of Operations & Production
rP

21
22
23 Management, 3, 291-304.
24
peer-00512904, version 1 - 1 Sep 2010

ee

25 Tapping D., Luyster T., Shuker T., 2002 Value Stream Management, New York,
26
27
28 Productivity Press
rR

29
30 Rother M., Shook J., 1999, Learning to see: Value Stream Mapping to Add Value and
31
32
33 Eliminate Muda, Brookline, MA, The Lean Enterprise Institute.
ev

34
35 Selim H., Reda M.S., Mahdi A., 2003, Formation of machine groups and part families:
36
iew

37
38 a modified SLC method and comparative study, Integrated Manufacturing Systems, 14,
39
40 123-137
41
42
43 Veramani D., Mani K., 1996, `Optimal clustering in vertex-tree graphic matrices,
On

44
45 International Journal of Production Research, 34, 2587-2611.
46
47
Womak J.P., Jones D., 1996, Lean Thinking: Banish Waste and Create Wealth in Your
ly

48
49
50 Corporation, New York: Simon & Schuster.
51
52
53 Zhou J., Irani S.A., 2003, A New Flow Diagramming Scheme For Mapping And
54
55 Analysis Of Multi-Product Flows In A Facility, Process Journal of Integrated Design
56
57
and Process Science, 7, 26-58.
58
59
60

27 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 28 of 45

1
2
3
4
5
6 Future State Questions
7
8 1) What is the takt time?
9
10 2) Will production produce to finished goods supermarket or directly to shipping?
11 Basic 3) Where can continuous flow processing be utilized?
12
13 4) Is there a need for a supermarket pull system within the value stream (kanban – conwip) ?
14 5) What single point in the production chain will be used to schedule production ?
15
16
Fo

17 6) How will the production be leveled at the pacemaker process ?

Heijunka
18 7) What increment of work will be consistently released from the pacemaker process(pitch) ?
19
20
Kaizen 8) What process improvements will be needed?
rP

21
22
23
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26
27
28
rR

29
30
31
32
33
ev

34
35
36
iew

37
38
39
40
41
42
43
On

44
45
46
47
ly

48
49
50
51
52
53
54
55
56 Table 1: Design Question for Future State Map
57
58
59
60

28
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 29 of 45 International Journal of Production Research

1
2
3
4
5
M#1 M#2 M#3 M#4 M#5 M#6 M#7 M#8 M#9 M#10 M#11 M#12 M#13 M#14 M#15
6
7 P#1 1 1 1 1 1 1
8 P#2 1 1 1 1 1 1
9 P#3 1 1 1 1
10 P#4 1 1 1 1 1
11 P#5 1 1 1 1 1 1 1
12
P#6 1 1 1 1 1 1
13
14 P#7 1 1 1 1 1 1
15 P#8 1 1 1 1
16 P#9 1 1 1 1 1
Fo

17 P#10 1 1 1 1 1
18
19
20
rP

21
22
23
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26
27
28
rR

29
30
31
32
33
ev

34
35
36
iew

37
38
39
40
41
42
43
On

44
45
46
47
ly

48
49
50
51
52
53
54
55
56 Table 2: Machine Part Matrix
57
58
59
60

29 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 30 of 45

1
2
3
4
5
6
P#1 P#2 P#3 P#4 P#5 P#6 P#7 P#8 P#9 P#10
7
8 P#1 1 0,231 0,301 0,763 0,199 0,231 0,85 0,301 0,921 0,264
9 P#2 1 0 0 0,928 0,85 0 0 0,264 0,264
10 P#3 1 0,342 0 0,301 0,301 0,809 0 0,732
11 P#4 1 0,231 0 0,921 0 0,672 0
12 P#5 1 0,784 0,199 0 0,231 0,231
13
P#6 1 0 0,301 0,264 0,438
14
15 P#7 1 0,301 0,763 0,264
16 P#8 1 0,342 0,912
Fo

17 P#9 1 0,301
18 P#10 1
19
20
rP

21
22
23
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26
27
28
rR

29
30
31
32
33
ev

34
35
36
iew

37
38
39
40
41
42
43
On

44
45
46
47
ly

48
49
50
51
52
53
54
55
56 Table 3: Similarity Matrix
57
58
59
60

30
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 31 of 45 International Journal of Production Research

1
2
3
4
5
6 Family #1 Family #2 Family #3
7 P#5 P#2 P#6 P#4 P#7 P#1 P#9 P#8 P#10 P#3
8 M#1 1 1 1
9
M#14 1 1 1
10
11 Cell #1 M#9 1 1 1
12 M#7 1 1 1 1 1
13 M#15 1 1 1 1 1
14 M#3 1 1
15
M#5 1 1 1 1
16
Fo

17 M#6 1 1 1 1
Cell #2

18 M#10 1 1 1 1
19 M#12 1 1 1 1
20
rP

M#8 1 1 1
21
22 M#13 1 1 1 1
Cell #3

23 M#4 1 1 1 1
24
peer-00512904, version 1 - 1 Sep 2010

M#2 1 1 1
ee

25 M#11 1 1 1
26
27
28
rR

29
30
31
32
33
ev

34
35
36
iew

37
38
39
40
41
42
43
On

44
45
46
47
ly

48
49
50
51
52
53
54
55
56
57 Table 4: Rearranged Machine Part Matrix
58
59
60

31 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 32 of 45

1
2
3
4
Figures Captions
5
6 Figure 1: Actual State Map Icons
7
8
9 Figure 2: Future State Map Icons
10
11
12 Figure 3: Improved Value Stream Mapping Procedure
13
14 Figure 4: PQ$ Analysis Chart
15
16
Fo

17 Figure 5: Part Families Dendrogram

18
19
20
Figure 6: Possible Solution to Synchronize Flow
rP

21
22 Figure 7: Example of Internal Machine Sharing
23
24
peer-00512904, version 1 - 1 Sep 2010

Figure 8: Example of TBOM

ee

25
26
27 Figure 9: Refrigerators Production Process
28
rR

29
30 Figure 10: Refrigerator Simplified TBOM
31
32
33 Figure 11: Critical Path CSM
ev

34
35 Figure 12: Sub - Critical Path CSM
36
iew

37
38 Figure 13: Refrigerators FSM
39
40
41
42
43
On

44
45
46
47
ly

48
49
50
51
52
53
54
55
56
57
58
59
60

32
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 33 of 45 International Journal of Production Research

1
2
3
4
5 Assembly CT = 10 sec
6 CO = 15 min
7 Vendor I 2 x Weeks
8 UT = 85%
9 EPE = 3 days 5 days
10
Manufacturing
11 Outside Source Data Box Inventory Truck Shipment
12 Process Box
13
14
15
16 Push Production Finished Goods Manual Electronic
Fo
Operator
17 Arrow to Custumer Information Flow Inforamtion Flow
18
19
20
rP

21
22
23
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26
27
28
rR

29
30
31
32
33
ev

34
35
36
iew

37
38
39
40
41
42
43
On

44
45
46
47
ly

48
49
50
51
52
53
54
55
56
57
58 Figure 1: Actual State Map Icons
59
60

33 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 34 of 45

1
2
3
4
5
6
7
8
9
10
11
12 Supermarket Pull Withdrawal Kanban Signal Kanban
13
14
15
Reduce Set Up
16
Fo

17
18
Load Leveling
19 FIFO Line Kanban Post Kaizen Burst
20 Box
rP

21
22
23
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26
27
28
rR

29
30
31
32
33
ev

34
35
36
iew

37
38
39
40
41
42
43
On

44
45
46
47
ly

48
49
50
51
52
53
54
55
56
57
58 Figure 2: Future State Map Icons
59
60

34
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 35 of 45 International Journal of Production Research

1
2
3 Choose a Family
4
5
6
7 Identify Machine Sharing
8
9
10
Identify Critical Path
11
(Maximum LT)
12
13 The secondary path
14 becomes the critical one
15 Map the Current State
(for the critical path only)
16
Fo

17
18
19 Analyze Inefficiencies
20
rP

21
22
23
24
peer-00512904, version 1 - 1 Sep 2010

No Yes
ee

25 Are inefficiencies due to a

26 secondary path?
27
28 Map The Future State of Map the Actual State of the
rR

29 the critical path secondary path

30
31
32
33 Evaluate LT reduction Analyze Inefficiencies
ev

34
35
36
Map The Future State
iew

37
38 Yes Is the new LT due to a No
39 secondary path?
40 Analyze residual
41 inefficiencies on the critical
42 path
43
On

44
45
46 No Yes
Are inefficiencies due to
47 another secondary path?
ly

48
49
50
51
52
53
54
55 End Procedure

56
57
58 Figure 3: Improved Value Stream Mapping Procedure
59
60

35 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 36 of 45

1
2
3
4
5
6 P-Q Chart
7
8 100%
2500
9
10 80%
Quantity
2000

11 A B C
12 1500
60%

13 P1
1000 40%
14
A P9
15 20%
16
500
P2
Fo

17 0 0%

18 P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15

19 P4

Revenue
20
rP

B P3 P5 P10
21 P-R Chart

22 P6
100%
23 200000
24 P11
peer-00512904, version 1 - 1 Sep 2010

80%
ee

25 150000 P7 P12
26 60%
C P13
27 100000

28 40% P8 P14
rR

29 50000
P15
20%
30
31 0 0%
P1 P2 P9 P3 P5 P6 P10 P4 P11 P15 P7 P8 P14 P12 P13
32
33
ev

34
35
36
iew

37
38
39
40
41
42
43
On

44
45
46
47
ly

48
49
50
51
52
53
54
55
56
57
58 Figure 4: PQ$ Analysis Chart
59
60

36
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 37 of 45 International Journal of Production Research

1
2
3
4
5
6
7

Similarity
8
9
10
11 0.15
12
13 0.23
14
15
16
Fo

17
18
19
20
rP

21
22 0.762
0.771
23
24
peer-00512904, version 1 - 1 Sep 2010

0.817
ee

25
26
27
0.912
28
rR

0.921
29
0.928
30
31
32
Clusters
33
ev

34
P#2 P#5 P#6 P#4 P#7 P#1 P#9 P#8 P#10 P#3
35
36
iew

37
38
39
40
41
42
43
On

44
45
46
47
ly

48
49
50
51
52
53
54
55
56
57
58
Figure 5: Part Families Dendrogram
59
60

37 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 38 of 45

1
2
3
4
5
P#3
6
P#10 P#10
7
P#8 P#8
8 P#3
9 P#10 P#8
10 P#8
11 P#8
12 M#13 P#10
13 P#8 P#10
14
15 P#3
16 P#10 M#11 I
Fo

17
18 P#3 P#3
P#3
19 P#3
20
rP

21

Family #2
22 M#12
23 I
24
peer-00512904, version 1 - 1 Sep 2010

ee

25 P#3
26
27
28 Family #2
rR

29
30
31
32
33
ev

34
35
36
iew

37
38
39
40
41
42
43
On

44
45
46
47
ly

48
49
50
51
52
53
54
55
56
57
58
59
60 Figure 6: Possible Solution to Synchronize Flow

38
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 39 of 45 International Journal of Production Research

1
2
3
4
5
6
7
8 RM1 RM2 C1 B3
9
10
11
12 M1 M2 B2 M3
13
14
15
16 B1
Fo

17
18
19
20
rP

21
22
23
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26
27
28
rR

29
30
31
32
33
ev

34
35
36
iew

37
38
39
40
41
42
43
On

44
45
46
47
ly

48
49
50
51
52
53
54
55
56
57 Figure 7: An Example of Internal Machine Sharing
58
59
60

39 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 40 of 45

1
2
3
4
5
6
7
8
9
10
11 Finished item
12
Time

13
14
15 A1 A2 A3
16
Fo

17
18 B1 B2 RM1 B3 B4
19
20
rP

21
22
23 RM1 RM2 RM3
24 RM1
peer-00512904, version 1 - 1 Sep 2010

RM3 C4
ee

25
26
27 RM1
28 RM2
rR

29 RM5
30
31 RM3 RM3
32
33
ev

34
35 Internally manufactured part Production LT
RM5
36
Supplier LT
iew

37 Externally bought part

38 Critical Path
39 Internal Sharing External Sharing
40
41
42
43
On

44
45
46
47
ly

48
49
50
51
52
53
54
55
56 Figure 8: Example of TBOM
57
58
59
60

40
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
Page 41 of 45 International Journal of Production Research

1
2
3
4
5
6
7
8
9
10
11 Condenser pipe production line Electronic devices preassembly
12
13
14 Output
15
16 Refrigerator body production line
Fo

17
18
19 Doors production line
20
rP

21
22
23
24
peer-00512904, version 1 - 1 Sep 2010

ee

25
26
27
28
rR

29
30
31
32
33
ev

34
35
36
iew

37
38
39
40
41
42
43
On

44
45
46
47
ly

48
49
50
51
52
53
54
55
56
57
58 Figure 9: Refrigerators Production Process
59
60

41 / 45
http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
 International Journal of Production Research Page 42 of 45

1
2
3 LT F in is h e d R e frig e ra to r
4
5
5 h
6 A s s e m b le d R e frig e ra to r

Fo
8 E le c tro n ic D e v ic e s
9 10 h

10 P re a s s e m b le d R e frig e ra to r
11
12
13
15 h
rP
C o m p le a te d D o o r C ondenser M o to r
C a b le s Top C om m ands
peer-00512904, version 1 - 1 Sep 2010

14
15
16 20 h ee
17
18
19 25 h
Foam ed door

P re a s s e m b le d D o o r
S e a ls

rR Foam ed Body

20
21 H in g e s P la s tic
ev
iew
22 R e frig e ra to r B o d y
30 h
23
24
D o o r In te rn a l S tr.
25
B ase F eet R e frig e ra to r C e ll E x te rn a l
26 M o to r B a s e C o n d e n s e r P ip e

On
35 h S h e e ts
27 E x te rn a l F ra m e S h e e ts
28 B a s e A s s e m b ly W e ld e d P ip e s A l P ip e s
29

ly
40 h
30 C a p illa r
E v a p o ra to r T h e rm o s ta t C u P ip e s
31 P ip e s

32
33 In te rn a l S tru c tu re

34 50 h

35 E x te rn a l F ra m e S h e e ts

36
37
38 Figure 10: Refrigerator Simplified TBOM
39
40
41
42
43
44 42
45 http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
46
47
Page 43 of 45 International Journal of Production Research

1
2
3
4
5
6
7

Fo
8
9
10
11
12
13 rP
peer-00512904, version 1 - 1 Sep 2010

14
15
16 ee
17
18
19
rR
20
21
ev
iew
22
23
24
25
26
27
28
29
On
30
31
32
33
ly
34
35
36
37
38
39 Figure 11: Critical Path CSM
40
41
42
43
44 43 / 45
45 http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
46
47
 International Journal of Production Research Page 44 of 45

1
2
3
4
5
6
7

Fo
8
9
10
11
12
13 rP
peer-00512904, version 1 - 1 Sep 2010

14
15
16 ee
17
18
19
rR
20
21
ev
iew
22
23
24
25
26
27
28
29
On
30
31
32
33
ly
34
35
36
37
38
39 Figure 12: Sub - Critical Path CSM
40
41
42
43
44 44
45 http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
46
47
Page 45 of 45 International Journal of Production Research

1
2
3
4
5
6
7

Fo
8
9
10
11
12
13 rP
peer-00512904, version 1 - 1 Sep 2010

14
15
16 ee
17
18
19
rR
20
21
ev
iew
22
23
24
25
26
27
28
29
On
30
31
32
33
ly
34
35
36
37
38
39 Figure 13: Refrigerators FSM
40
41
42
43
44 45 / 45
45 http://mc.manuscriptcentral.com/tprs Email: ijpr@lboro.ac.uk
46
47 View publication stats