Int J Adv Manuf Technol (2018) 94:433–450

DOI 10.1007/s00170-017-0827-7

ORIGINAL ARTICLE

Changeover process improvement based on modified SMED

method and other process improvement tools application:
an improvement project of 5-axis CNC machine operation
in advanced composite manufacturing industry
Rosmaini Ahmad 1 & Mohd Syazwan Faiz Soberi 1

Received: 21 March 2017 / Accepted: 17 July 2017 / Published online: 11 August 2017
# Springer-Verlag London Ltd. 2017

Abstract This paper shares a real industry experience for an 1 Introduction

improvement project of the changeover process through a
conventional Single-Minute Exchange Die (SMED) method The effective and efficient changeover process is an important
and other process improvement application tools. This project element that supports the production control process for most
focuses on the changeover activities for one of the advanced of the manufacturing shop floor. It is even more important if
composite manufacturing processes, called trimming process the demand is highly complex (e.g., fluctuate, varieties of
that involves 5-axis CNC machine operation. The cause and products specification). Thus, it is able to shorten the produc-
effect and five whys analysis techniques and some steps of tion lead times and help in achieving higher quality standards
conventional SMED are applied in this improvement project. [1]. The most significant positive impact of having a small
Four standard strategies and priorities sequence are introduced changeover time is the batch size reduction in a production
for an extension step of conventional SMED. The conceptual system [2, 3]. In other words, it continuously supports the one
decision model is then proposed to present the systematic piece flow concept that is critically needed for a high product
improvement process carried out in this project. Four specific variety demand type of production. The other related benefits
solutions from the selected strategies are then recommended of the short changeover time include the expenses reduction,
to be implemented. The current implementation results show faster production speed, increased output, lead times reduc-
that the total changeover time is reduced to 44% and internal tion, smoother flows, a broader range of lot sizes, lower in-
type activities time is reduced to 48%. The paper discusses ventory, and increased customer satisfaction [1].
some key success factors and future recommendations of the Literature study shows that the terms “changeover” and
current project. “setup” has been used interchangeably by many researchers.
Generally, setup time is the time required for preparing the
necessary resources to perform a task. It also refers to the time
Keywords Single-Minute Exchange of Dies (SMED) . Lean elapsed between producing the last good part of the first lot
manufacturing . Cause and effect analysis . Five whys analysis and the first good part of the next lot [4–6]. According to
method . 5S . Visual workplace system . Trimming process . McIntosh et al. [7], setup is one of the three elements in a
5-axis machine changeover process. The first element is “run-down” which
is running the last product of the last batch production in a
manufacturing line. It may include the initial preparation ac-
tivities (e.g., looking the right tools/jigs) for the next element
of the changeover process. The second element is “setup”
which is the activities to remove current tools and other related
* Rosmaini Ahmad equipment and replacing with new tools including the rough-
rosmainiahmad@unimap.edu.my cut setting and adjustments. The last element is “run-up”
which involves the initial running or testing of the machine
1
School of Manufacturing Engineering, Alam Pauh Putra Campus, operation until it reaches an acceptable level of output quality
Universiti Malaysia Perlis, 02600 Arau, Perlis, Malaysia and the machine speed. Shingo [8] often used the term
434 Int J Adv Manuf Technol (2018) 94:433–450

“changeover” because in their cases under consideration are 2 Literature review

associated with the three mentioned elements. Therefore, from
the authors’ point of view, the term “setup” and “changeover” The conventional SMED consists of four phases as depicted in
must be used carefully to avoid misunderstanding the scope of Fig. 1. Phase 1 is the current changeover process mapping.
the case study under consideration. The main idea of this phase is to reveal an overall picture of
The Lean manufacturing approach introduced a well- the changeover activities. Gathering related information and
known methodology to solve the abovementioned issue, data about the current changeover procedure is the first step of
called Single-Minute Exchange of Dies (SMED). The this phase. It is followed by decomposing the changeover
SMED was developed in the 1950s by Shigeo Shingo, operations into a series of actions. The next step of this phase
an Industrial Engineer for Toyota Company where the is to conduct time and motion study to determine the standard
main focus was to practically minimize the overall time for each operation. Phase 2 is classifying the internal and
changeover time. The popular theory of SMED is to externals activities. It begins with listing all the changeover
make the changeover process completed in below activities in a sequence followed by classifying and analyzing
10 min (in other words, “single digit minute”); thus, it the internal and external activities. This phase ended with
gives the name of this method. One of the analogies separating the external and internal activities.
that can be presented through the SMED application is The next phase (phase 3) focuses on transferring the inter-
the Formula One (F1) pit stop scenario where those pit nal to external activities. As been highlighted in the previous
crews systematically refueling, changing new tyres, section, this phase is the key improvement strategy to mini-
repairing, and for mechanical adjustments. mize the machine downtime and thus reduce the overall
Under the SMED method, all of the changeover ac- changeover time that is currently offered by the conventional
tivities are classified into the internal and external type’s SMED. The straightforward way of this phase is to identify
activities. All of the activities that can be completed any possible internal activities that can be done before the
only when the machine is shutting down are defined machine stops or after the machine starts to operate.
as internal type activities. Meanwhile, the activities that Almonani et al. [11] stated that this phase also can be carried
are performed while the machine is in its operating state out through equipment modification and discarding some set-
are classified as external type activities. The main goal ting. The final phase (phase 4) focuses on streamlining all
of any changeover process improvement is to minimize internal and external activities. This phase can be accom-
the machine downtime due to the internal activities. plished by eliminating any unnecessary operations (e.g., re-
Therefore, the key strategy of SMED is looking for dundant activities) and improving the original procedures of
conversion of the internal activities to the external ac- the changeover process (standardize time of activities).
tivities; thus, it directly reduces the downtime of the Another action that can add value in this phase is executed
machine. through the parallel operation wherever possible.
Although SMED is promoted as one of the practical Various types of case studies involving the conventional
methods for the changeover time reduction, it is not a generic SMED application are presented. The literature reveals that the
method to be applied for all cases. It is supported by case studies from the automotive industry are dominated by the
Goubergen [9] who stated that the conventional SMED is only conventional SMED success application. For instance, Deros
suitable in the systems consisting of only one machine and et al. [12] demonstrated the application of SMED to reduce the
one worker. As been mentioned by Karasu et al. [10], this setup time in an automotive battery assembly line. Desai [13]
method still can be improved and open for future improve- presented productivity improvement for the automobile industry
ment. Therefore, this study presents the improved methodol- in India by applying SMED method. Kumar and Abuthakeer
ogy version to improve the changeover process for 5-axis [14] also reported the setup time reduction in the automobile
CNC machine operation. The proposed methodology inte- industry by applying the SMED method. The case study was
grates few steps from the conventional SMED method with done in fagor press, involved in the machining of the evaporator.
other process improvement tools. Another application of SMED in the automobile industry was
The remaining part of this paper is organized as follows: done by Souza et al. [15] in the cell of shock absorbers structure.
Section 2 presents the literature studies on conventional In the manufacturing industry, Kušar et al. [16] reported the
SMED and its applications. Section 3 introduces the improve- positive effect of the SMED application in reducing the machine
ment project background and the current setup application setup time. Another case study from the manufacturing industry
process. Section 4 presents the step-by-step applied method- is presented by Moreira and Pais [17], where the application of
ology in the case study under consideration. Section 5 dis- SMED methodology has reduced the setup time in a manufactur-
cusses critically the results before and after the improvement. ing firm located in North Portugal. The example of the SMED
Section 6 ends with a conclusion and future recommendations application in the electronic industry has been discussed by
for the next project improvement. Dhake and Rajebhosale [18] in a solder paste printing machine.
Int J Adv Manuf Technol (2018) 94:433–450 435

Fig. 1 Conventional SMED Phase 1: Phase 2:

Current setup process map Separate internal from external setups
approach framework [11]

Observe current methodology List all setup activities

and decompose setup
operations into a series of
actions
Analyze all activities
Conduct times and motion
study to determine standard
time for each step Separate external from
internal
Create standard operation
setup check sheet

Phase 3: Phase 4:
Transferring internal to external Streamlining all setup operations

Eliminating any
Prepare operating conditions
unnecessary operations
in advance

Improve original setup

Standardizing setups tasks procedures (external and
(equipment, tool, parts, internal)
process, supplies and inventory
locations) Executions of parallel
operations wherever
possible

The initial step was gathering all the information about the cur- setup process. Karasu et al. [26] integrated the conventional
rent. Benjamin et al. [19] has discussed the practical use of SMED with Taguchi methods in an injection molding produc-
SMED in eliminating small stops in a manufacturing firm which tion. Almomani et al. [11] considered Multiple Criteria
manufactures metal barrel in Malaysia. Other types of industries Decision-Making (MCDM) in a PVC industry. The purpose
reported that the success of the SMED application is given by of the MCDM technique is to select the best setup technique
Sayem et al. [20] who stated that the application of SMED is one among the available alternatives which was adopted in the
of the best ways to enhance productivity through the changeover third step of the conventional SMED. Braglia et al. [27] intro-
time reduction in the furniture industry. duced the Changeover Out of the Machine Evaluation
On the other hand, the improved version of the SMED Technique (COMET) to support the third step of the conven-
methodology is also found in the literature review. The im- tional SMED. The proposed approach helps decision-makers
proved version of the SMED methodology can be defined to evaluate all the interventions on the system to externalize as
when the conventional SMED steps (as presented in Fig. 1) many tasks as possible and at the same time evaluating the
is integrated with other tool, approach, concept, or even mod- cost-benefit ratio and any operational issue of the intervention.
ify the steps of the conventional SMED. For example, Patel These literature studies conclude that there is an important
et al. [21] applied the mistake proofing (also known as poka- need to improve the conventional SMED methodology based
yoke) to the concept as a strategy to reduce the setup time. The on the specific case under consideration. Therefore, this paper
Mistake proofing device is introduced to prevent the compo- presents another improved version of SMED methodology so
nent from being positioned and machined incorrectly. Also, a that the changeover process improvement for specific case
similar integrated concept has been proposed by Singh and highlighted in this paper can be carried out as effective as
Khanduja [22] in a nonferrous foundry in India, where it re- possible.
sulted in 10 to 15% changeover time reduction. Meanwhile,
Kumar and Bajaj [23] highlighted the 5S concept to be inte-
grated into the conventional SMED steps to reduce the setup
time for the mechanical press machine in the Gill Agro 3 Improvement project background
Industry located in Punjab. The application of the 5S concept
in the conventional SMED is also given by Ibrahim et al. [24] The improvement project reported in this paper is carried out
to improve the time-consuming changeover process at the in a manufacturing company classified as advanced composite
stamping machine of a textile manufacturing company. manufacturing industry. Generally, the company producing
Stadnicka [25] combined the conventional SMED with other various types of products for the commercial airline compo-
tools including FMEA, survey method, five whys analysis, nents. The current focus of this company is to manufacture the
Pareto analysis, and statistical analysis. The seven steps are flat and contoured structure composite bond. Their production
proposed for the identification and elimination of waste in the processes are divided into several production departments and
436 Int J Adv Manuf Technol (2018) 94:433–450

one of them is called trimming process that is the focus of the Six Sigma methodology that is “define.” As has been
presents project case. highlighted in the conventional Six Sigma methodology,
Figure 2 shows the current procedure of the changeover define the problem is a very important step to identify,
process for 5-axis trimming operation. The first process is to clarify, and summarize the real issue to be solved. The
load the Drilling Router Jig (DRJ). The DRJ is taken from the detail discussion of this step is given in [28]. Other
rack by using a forklift. There are two workers assigned to related information and implementation discussion of
complete the trimming process which is called the pilot and this step are given by [29–31].
buddy. Usually, the buddy will do the loading and unload the In this project, this step is carried out based on the
DRJ process assisted by the pilot. The pilot will give instruc- teamwork basis, where the team was comprised of six
tion to the buddy for placing the DRJ into the bed of the 5-axis people which is lead by an improvement (CI) manager,
machine. three engineers from the trimming process department
After that, both workers will place the DRJ with “T” bolts and two researchers from a local technical-based insti-
into the bed slot and tighten the holding nut. Next, the workers tution. Each team member has their own functions and
will do the process of cleaning the DRJ and ensuring the DRJ roles; for example, the CI manager monitors the overall
holes and routing path is cleared from the entrapped dust. progress of the project and report to the top manage-
Then, workers will perform the best fit process. The best fit ment in terms of current costs and benefits implication.
process is done to make sure that the DRJ is installed in a flat Meanwhile, the engineers are the key persons to collect
position on the machine bed. Therefore, the panel (machined the data and identify the possible solutions from a prac-
part) can be trimmed according to the dimension of the panel. tical perspective. Finally, researchers that have industrial
After all the setup activities are completed, the automatic trim- engineering background are used to guide the team from
ming process is then followed. the theoretical project point of view and help the engi-
neers in analyzing the collected data as well as propos-
ing the solutions to the team.
The problem definition process of this project starts with
4 Methodology of improvement project
the basic questions such as “what we want?”, “why?,” and etc.
Specifically, in this project, the aims are to identify and clarify
This section presents the methodology applied in the present
the following elements: the actual issue of the project, scope,
improvement project. It is divided into five steps. The meth-
boundary, goals, benefits, and time frame of the project. The
odology steps discuss in this section applied two general pro-
summary of the problem definition statement is then presented
cess improvement approaches from Lean and Six Sigma. It is
in a project charter form [32], as shown in Table 1. From the
followed by applying various process improvement concepts
project discussion team, all the group members agreed that the
and tools such as production “wastes,” value and non-value
actual issue of this project is about “time,” where the produc-
added, and SMED.
tion records showed that for all the types of products, the
average time taken to perform the changeover process that
4.1 Define the problem leads to machine downtime has contributed between 40 to
60% of the overall trimming process time. For example, if
Define the problem is the first step applied in this presented
improvement project. This step is inspired by the first step of
Table 1 Project charter

Project Improvement of changeover process

title

Issue Long changeover process time

Scope Product type X, trimming process
Boundary Focus in minimizing the downtime of trimming machine due
to changeover process
Goals Reduce at least 30% of changeover time of trimming process
for product type X
Benefits Reduce (work in progress) WIP of the product, therefore
company be able to increase the production, thus increase
the profit
Time 4 months
frame
Fig. 2 Current changeover procedure of 5-axis trimming operation
Int J Adv Manuf Technol (2018) 94:433–450 437

the overall trimming process time for a product is 120 min for a variety of cases are given by [33–37]. The CE analysis
(2 h), the changeover process that causes machine downtime can be carried out by dividing four major groups (man, ma-
is within 48 to 72 min. chine, method, environment, and people) that are possible to
From “Lean” perspective, the percentage of time taken to contribute to the “causes” of the problem.
perform the changeover activities that lead to machine down- The overall result of the CE analysis for the present im-
time is classified as non-value added scenario, which should provement project is shown in Fig. 3. In the analysis group of
be eliminated or minimized. Since this pioneer improvement “method,” where the focused is given on the steps that the
project focus on the changeover process, the project team then workers performed the changeover activities is analyzed.
focus on the critical product, namely product X (also named as The results of the CE analysis under this focus group found
machined part/panel). This product has been chosen due to that the workers are currently performing many non-value
the highest changeover process time recorded. Regarding added activities such as repetitively walking from the trim-
the element of the boundary, this project was focused on ming machine to the PC table. From this discussion, all the
minimizing the downtime duration of the trimming ma- team members agreed to conclude that this non-value added
chine due to the changeover process. In other words, activity has not contributed any value during the best fit setup
any proposed solutions for this project tries to minimize activities. Other than that, an observation also found that the
the trimming machine downtime due to the changeover worker currently needs to move the others (unused) Drilling
process. The goal of this project has been decided to Router Jig (DRJ) to find the desired DRJ. This is another non-
reduce at least 30% of the downtime due to the change- value added activity that is believed to contribute to the prob-
over process. The determination of this goal is based on lem defined.
“rule-of-thumb” approach of the project team, where based In the “man” group, the focused is given on any activity that
on their experience and observation; this is the realistic relates to the performance, consistency, attitude, and the skills
target for the pioneer level of the improvement project. of the workers while performing the changeover activities. The
The implication of the goal of this project directly reflects observation and video record study found that the workers
on the production benefit, where work in progress (WIP) looked tired especially after completing the best fit setup activ-
of the product X will be reduced. Therefore, this is the ities. Thus, this scenario affected the time of the following
final key performance index that the top management in- changeover activities by delaying the following tasks.
tent to look. Finally, the project team is confident that this Another point that falls under the “man” group analysis is the
project can be completed in 4 months. quality of the activities that have been done by the workers. All
the team members of this project agreed that this is the leak of
4.2 Problem root cause identification multi-skills element among the workers, where it is supported
by the fact that workers that have been assigned to the trimming
In this project, the second step after the problem has process department is always changing from time to time.
been defined is problem root cause identification. Two In the “environment” group, the focus of the analysis is on
common root cause analysis methods have been applied; the current workplace layout. Observation found two impor-
cause and effect and five whys analyses. The aim of tant causes that contribute to the long changeover time, where
this step is to identify any possible root cause of the both of these causes are related to the distance of two loca-
defined problem and it patterns from two different per- tions. The first cause is the distance between machine control-
spectives based on the two methods that have been ler (MC) and PC table locations, where all project team mem-
mentioned. bers concluded that there is currently unnecessary distance
been located. Another cause goes to the distance issue be-
4.2.1 Cause and effect (CE) analysis tween MC and DRJ rack locations. Currently, it was found
that there is quite a long distance between these locations,
The other name of CE analysis is the Ishikawa or fishbone thus, the worker is required to drive the forklift for a long
diagram-based analysis. The CE analysis is a graphic tool that distance to take the DRJ from the rack to the trimming ma-
is used to explore and depict the information and possible chine area.
causes on a stated problem statement. The CE analysis plays In the final analysis of the “machine” group, where the
an important role because it shows the direct results of various focus of the CE analysis is given on the current procedure that
actions and processes and allows us to look at them objective- relates to the equipment or tools operations to support the
ly before those actions are even put into place. The informa- changeover process. Two causes have been finalized, the first
tion of the CE analysis can be gathered through observation, is regarding the procedure to find the right computer program
survey, interviews, and video record study. Thus, the best for the product that is being used for the machine (trimming
linkage of important causes can be identified through the team process). Currently, it is performed manually, where workers
discussions session. Examples of the CE analysis applications find the computer program in the database folder before
438 Int J Adv Manuf Technol (2018) 94:433–450

Fig. 3 Cause and effect diagram

Method (Me) Man (Ma)

(Ma 3)
(Me 1) (Ma 1) Worker delay
Repetitive best fit Inconsistent the works
activities speed of works

(Me 2) (Ma 2)
Selection the desired Low level of multi-
DRJ at the DRJ rack skills workers
Long setup
(E 1) (M/E 1) time
Distance from trimming Need to find the
machine area to DRJ computer
rack is far program manually
(E 2)
Distance from trimming
(M/E 2)
machine area to PC DRJ defect
table is far

Environment (E) Machine/Equipment (M/E)

running the program. This current procedure directly takes to the defined problem. An example of the five whys analysis
time and cause the overall changeover time longer. Another application for industry case is reported by [39].
cause is related to the improper condition of DRJ, where the In this improvement project, the simple procedure was ap-
observation found that some DRJ are bent and not flat. Thus, plied to perform the five whys analysis, where all the project
the workers need a long time to set up the DRJ to the trimming group members have to answer individually based on the fol-
machine bed. lowing initial question; “Why is the changeover process re-
The result of the CE analysis given in Fig. 3 presents the ported too long?” and then they have to ask the next sequence
initial views of the root cause of the problem. The project team questions of “why...” based on the previous answers given as
members believed that CE analysis is a good start towards the many as possible. Figure 4 presents four (A, B, C, and D)
changeover process improvement. However, it is highly pos- answers scripts that have been summarized and finalized.
sible that the causes summarized in the CE analysis related In script A, the first answer gives the reason why workers
from one to another are high. Therefore, the root cause iden- take a long time, to identify the desired DRJ. The second
tification by using the five whys analysis method is then car- “why” question is based on the first answer, given two differ-
ried out. This method helps to peel away the layers of symp- ent answers directions; the first sub answer (2a) goes to the
toms which can lead to the root cause of a problem. The reason of the worker that they need more time to move un-
method is also to determine the sequence relationship between wanted DRJs before the desired DRJ can be identified. The
one root causes and another. second sub-answer (2b) directly focuses on the storage envi-
ronment of DRJs issue, where currently are not well-arranged
at the rack. The next answer of “why” based on the first sub-
4.2.2 Five whys analysis answer (3(2a)) is that work needs time to move the other DRJs
to the other rack. Then, the last answer (4(2a)) of “why” is
The five whys analysis is another root cause analysis method based on the previous answer that the worker wants to take the
applied in this improvement project. This analysis is used to desired DRJ that is blocked by other DRJs. Meanwhile, the
support the CE analysis presented in the previous section. next answer of “why” is based on the second sub-answer
Andersen and Fagerhaug [38] described that this analysis is (3(2b)) which is because the worker just places the other
a sequential process of asking “why” five times in a belief that DRJ randomly. The following reason of the “why” question
the actual root cause of the defined problem will appear after (4(2b)) highlighted there is no specific place to store the DRJ
asking “why” five times. However, from the author’s point of at the rack. The final answer (5(2b)) is based on the answer to
view, this description regarding five whys analysis procedure the previous question which is due to the improper arrange-
is not always necessary where it depends on how the analysis ment system to store the DRJs. The five whys analysis pre-
has been arranged and managed. The authors believed that it is sented in script A found an important scenario that contributed
not a significant issue of how many “why” is been asked, to the long time setup process. The scenario is that the workers
instead of the pattern of answers from a sequence of asking have to spend some significant amount of time to identify the
“why” towards identifying possible root causes that contribute desired DRJ. While identification of desired DRJ is carried
Int J Adv Manuf Technol (2018) 94:433–450 439

Fig. 4 Five (5) whys analysis 5 Whys (Sciprt A)

script

1. Why the setup time is reported too long?

It is because the worker takes long time to take the desired DRJ from the rack.

2. Why the worker takes long time to take the desired DRJ from the racks?
a. It is because worker needs more time to move the other DRJs.
b. It is because all the DRJs are not well-arranged at the rack.

3(2a). Why the worker needs time to move the other DRJs?
It is because the worker needs time to move the other DRJs to the other rack.

4(2a). Why the worker needs time to move the other DRJs to the other rack?
It is because the worker wants to take the desired DRJ that blocked by other DRJs.

*********************************************************************

3(2b). Why all DRJs are not well-arranged at the rack?

It is because the worker just places the other DRJ randomly.

4(2b). Why the worker just place the DRJ randomly?

It is because there is no specific place to store the DRJ at the rack.

5(2b). Why there is no specific place to store the DRJ at the rack?
It is because there is improper arrangement system to store the DRJs.

5 Whys (Sciprt B)

1. Why the setup time is reported too long?

It is because the worker needs long time to set the computer program for the DRJ before
starting the automatic trimming process.

2. Why the worker needs long time to set the computer program for the DRJ before starting the
automatic trimming process?
It is because the worker needs to travel between the machine controller (MC) and the PC table
many times.

3(a). Why the worker needs to travel from the machine controller (MC) and the PC table many
times?
It is because the worker needs to walk for a long distance from the machine controller to the PC
table.

4(3a).Why the worker needs to walk for a long distance from the machine controller to the PC
table?
It is because the PC table is not strategically located.

*********************************************************************

3(b). Why the worker needs to travel from the machine controller (MC) and the PC table many
times?
It is because there have improper communication system between two staffs in performing jobs.

5 Whys (Sciprt C)
1. Why the setup time is reported too long?
It is because the worker takes long time to find right computer program for the part number
that need to be processed.

2. Why the worker takes long times to find right computer program for the part?
It is because the workers do it manually.

3. Why the worker do it manually?

It is because there is no quick and effective method/system to find the computer program for
the part number that need to be processed.

5 Whys (Sciprt D)
1. Why the setup time is reported too long?
It is because the worker delays the works.

2. Why the worker delay the works?

It is because the worker gets tired

3. Why the worker get tired?

It is because there are too many manual works and repetitive movement from one location to
another.
440 Int J Adv Manuf Technol (2018) 94:433–450

out, the workers have to manage and arrange others DRJs in the aim of this mapping process is to present systematically
order to get the right DRJ. the overall changeover activities patterns. In this step, the time
In script B, the first answer highlighted the reason that the for each changeover activity in a sequence form has been
workers need a long time to set the computer program for the taken and recorded. The clustering of the changeover activi-
DRJ before starting the automatic trimming process. The next ties is then performed and the trend of times taken for each
answer of “why” based on the first answer is that the workers activity is then presented.
need to travel between the machine controller (MC) to PC table The biggest trick to carry out this step is to define the scope
many times to align the DRJ with the specific trimming process of the changeover activities. It means that “from what activity
program. In the following question of “why” based on the pre- to what activity” should be included in this mapping process.
vious answer, it found that there are two different answers record- The justification of this issue is that if the scope of the change-
ed: the first reason is because the worker needs to walk for a long over activities is wrongly defined, the actual benefit from the
distance from the machine controller to the PC table, while the changeover time reduction project cannot be achieved proper-
second reason is that they have improper communication system ly. In order to solve this tricky issue, the project team decided
between the two staffs in performing jobs. The last question to consider all the activities that are carried out for the purpose
(4(3a)) that come from the answer to the question 3a give the of the changeover process for product X. It includes the activ-
reason that the PC table is not strategically located. The analysis ities that contribute to the downtime or uptime of the trimming
result presented in script B highlighted the two important points machine due to the changeover process.
that contribute to the long setup process. The first is the distance Table 2 presents the sequence of 64 activities of the change-
issue between the PC table and MC. The second is the improper over process for product X. The recorded changeover activi-
and ineffective communication system between two operators ties were then clustered into seven groups: the cleanup pro-
that work in the PC table and MC, where currently operator at cess, DRJ identification, DRJ movement, DRJ setup, compos-
MC has to go to the PC table many times to confirm related ite panel setup, computer program setup, and the previous
coordination changes to be adjusted at the MC. storage DRJ. The time taken for each activity is also stated
In script C, only three sequences of “why” questions have in the table and the overall time distribution for each activity is
been finalized. The first answer for the general initial question presented in Fig. 5.
is because the workers take a longer time to find the right The time pattern distribution of the changeover activities giv-
computer program for the part number that needs to be proc- en in Fig. 5 shows that 25% or 16 activities currently took more
essed. The following answer of the “why” question based on than 100 s to be completed, where the highest time recorded goes
the previous answer is because the workers do it (find right to activity no. 61 for 533 s (choose and find the correct program
computer program) manually. The last answer of the third for the composite panel). The second and third highest times are
“why” question is because there is no quick and effective given by activity no. 25 for 276 s (move DRJ to another rack)
method/system to find the computer program for the part num- and no. 54 for 249 s (check the best fit flatness), respectively. Up
ber that needs to be processed. to this point, the top three of the high times changeover activities
In script D, three sequences of “why” questions have been that have been recorded are matching with the results of root
recorded. The answer to the first “why” question is because cause analysis presented in Section 4.2.
workers delay their works. Then, the answer for the next Figure 6 shows more details interpretation of the time dis-
“why” question is due to tiredness of workers and the final an- tribution for the changeover activities. The highest total time is
swer of the third “why” question give the reason that there are too given by activities in the group D for 1614 s (DRJ setup).
many manual works and repetitive movement from one location Meanwhile, the second and third top high total times are given
to another. The result given in script D clearly shows that the by group B for 1444 s (DRJ identification) and group A for
workers significantly reduce the speed of the setup activities 597 s (cleanup process), respectively. This clustering pattern
especially for the activities involving manual action. gives useful information to the project team towards looking
The result presented from the five whys analysis gives for the best and systematic solutions that are presented in the
better views for the root cause of the identification process. following section.
The relationship between the causes presented in the CE anal-
ysis has been revealed in the five whys analysis. This infor-
mation helps significantly the project team members for better 4.4 Decision-making process
ideas in looking for the solutions to the problem.
As presented in Sections 4.2.1 and 4.2.2, it reveals that the
4.3 Current changeover process mapping long time of the current changeover process comes from var-
ious factors. It includes the factors of the working system,
This step is the first step from the conventional SMED appli- workflow, equipment arrangement, and location, etc. This sec-
cation (as presented in Fig. 1). In this improvement project, tion presents a conceptual decision model development that
Int J Adv Manuf Technol (2018) 94:433–450 441

Table 2 Overall changeover process activities

No. Activity Time taken (s) Description Task type

1 Switch off the related operation buttons of trimming process 6 A. Cleanup process Internal
2 Open the machine door 6 Cleanup previous trimming
3 Remove the machined parts 185 process. Remove the
machined (trimming) parts
4 Open up the DRJ clamps 19
and its DRJ from the
5 Adjust the overhead crane 14 trimming process area
6 Hook four strings of the overhead crane to DRJ 77 (Total time taken, 597 s)
7 Adjust the hooks 95
8 Lift up the DRJ 13
9 Place the DRJ outside the trimming process area 20
10 Release the hooks 112
11 Move the overhead crane 50
12 Drive forklift to the DRJ rack 34 B. DRJ identification Internal
13 Lift a long DRJ 146 Identify the desired DRJ
14 Move the DRJ to another rack 25 (Total time taken, 1444 s)
15 Lift a new long DRJ 73
16 Move the DRJ within rack 30
17 Lift another long DRJ 102
18 Adjust forklift to reverse 105
19 Move the DRJ to another rack 19
20 Drive forklift back to origin 23
21 Check DRJ part number 68
22 Drive forklift back to rack 75
23 Take a new long DRJ 16
24 Adjust forklift to reverse 33
25 Move DRJ to other rack 276
26 Reverse forklift 75
27 Lift another DRJ 112
28 Move DRJ to another rack 31
29 Take the desired DRJ 59
30 Adjust forklift to reverse 25
31 Bring the desired DRJ to the trimming machine 117
32 Adjust the overhead crane 14 C. DRJ installation Internal
33 Hook four strings of the overhead crane to DRJ 77 Take the desired DRJ to
34 Adjust the hooks 95 trimming machine
(Total time taken, 356 s)
35 Lift the DRJ 7
36 Place the DRJ onto the bed of the machine 12
37 Release the hooks 101
38 Move the overhead crane 50
39 Tighten the clamps for DRJ 19 D. DRJ setup Internal
40 Adjust the tool bit using machine controller 93 Setting, adjusting, and
41 Touch offset the surface of the DRJ 102 balancing the DRJ to the
machine bed
42 Set the machine program 9
(Total time taken, 1614 s)
43 Adjust the tool for best fit process 73
44 Best fit process of one side DRJ 38
45 Best fit another side of DRJ 111
46 Check DRJ part number 62
47 Checking the flat point on PC 210
48 Enter flatness points data into PC 12
49 Adjust flatness points in the PC 186
442 Int J Adv Manuf Technol (2018) 94:433–450

Table 2 (continued)

No. Activity Time taken (s) Description Task type

50 Best fit again for all points 7

51 Enter data into PC 88
52 Best fit process for another part number 137
53 Setting the program for the part number 65
54 Check the best fit flatness 249
55 Setting the tool bit 153
56 Takes the composite panel 9 E. Composite panel setup Internal
57 Lift the composite panel and place it onto the DRJ 28 Install the composite panel to
58 Fix the vacuum hose 92 the DRJ
(Total time taken, 252 s)
59 Place another panel 123
60 Close the machine door 6 F. Computer program setup Internal
61 Choose and find the correct program for the composite panel 533 Final setting before start the
62 Push start button for trimming process 4 trimming process
(Total time taken, 543 s)
63 Drive forklift to the previous DRJ located as stated in activity (9) 40 G. Previous DRJ storage External
64 Move the previous DRJ to the rack 120 Store the previous DRJ
(Total time taken, 160 s)
Total time taken for overall setup process = 4966 s / 82.8 min / 1.38 h

has been proposed and applied in the present improvement where the values of a process can be increased with waste
project. elimination. This lean concept highlighted that elimination is
a supreme strategy, which directly increases the values, if the
elimination strategy cannot be done then the strategy for min-
4.4.1 Improvement strategies and priorities imizing the waste will then become the next alternative.
Therefore, in this project, the “eliminate-activities” strategy
This section presents four improvement strategies concept that is considered as a top decision priority towards the changeover
has been considered in this project towards minimizing the time process improvement. The benefit of this strategy is that the
of the changeover process. These strategies are then been con- more the setup activities can be eliminated the more overall
sidered as series of decision flow priorities for the decision- changeover time can be reduced.
making process that is presented in the following section. The second decision priority towards changeover time re-
The first decision priority applied in this project is based on duction that is considered in this project is based on the “con-
the “eliminate-activities” strategy. The concept of this strategy vert-activities” strategy. In the conventional SMED model
is inspired by the general concept of lean manufacturing, (refers to Fig. 1, phase 2), the concept of this strategy is to

Fig. 5 Overall changeover

Changeover activities distribution
activities distribution
600

500

400
Time in second

300

200

100

0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63

Number of changeover activities

Int J Adv Manuf Technol (2018) 94:433–450 443

Fig. 6 Clustering of changeover

Changeover activities distribution
activities
600
F

500

400

Time in second
B
300
D
A
200
C E G
100

0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63

Number of changeover activities

convert the internal setup to external setup activities. As has section is given in Fig. 7. Generally, the flow of the proposed
been described in Section 2, internal activities refer to the decision model is a continuous loop-based process, where it
setup activities that are carried out in the downtime state of begins with the node of “start” and end with the node “con-
machine operation (in this project it refers to trimming process sider other group activities/activity.”
of 5-axis CNC machine). Meanwhile, the external activities The application of the proposed decision model begins
refer to the setup activities that are performed during the up- with the group level of the changeover activities (e.g., cleanup
time state of machine operation. The benefit of this strategy is process, DRJ identification process) to choose the best im-
that the more the internal activities can be converted to exter- provement strategy. If the group level activities are not
nal activities the more downtime of the machine operation can allowed to be done (it means in group-based consideration),
be reduced. Although it may not reduce the overall time of the then the consideration of “activity-by-activity” will be carried
changeover process through this strategy, the production out- out. As been described in the previous section, the first strat-
put can be increased due to less downtime of the machine and egy to be considered is the “eliminated” strategy and then
more production time available. followed by the “converted” strategy. For these two decisions
The third decision priority used in this project is based on strategies, there are three directions of options; “yes,” “possi-
the “combine-activities” strategy. This strategy refers to two or ble,” and “no.” The direction of node “yes” is used if one of
more activities that are carried out in parallel. This strategy is these strategies is appropriate with clear and no other serious
motivated by the F1 race pit stop setup activities, where mul- issue to be done. Then it is followed by the process of imple-
tiple setup activities are performed in parallel, systematic, and mentation strategy (refers to the node of “implement”).
quickly. The benefit of this strategy is that it combines two or Meanwhile, the direction of the node “possible” is used if
more activities at a one time; thus, it directly reduces the over- one of these strategies is basically carried out but have signif-
all changeover time. However, this strategy will deal with icant issues to be clarified and considered before or along the
other issues such as multi-worker skills ability, identifying implementation of the strategy. For example, activity A is
the right activities that are appropriate to be combined and “possible” to be “eliminated” but specific advanced material
also to determine the right synchronization of the setup activ- handling equipment is required. Therefore, for this example,
ities that have been combined. the improvement team should consider related worthiness fac-
The last decision priority applied in this project is based on tors (e.g., cost, profit, safety, time-saving) or specific solutions
the “simplify-activities” strategy. The idea behind this strategy towards achieving the maximum benefits of the strategy.
is to simplify the steps of a changeover activity; thus, it direct- On the other hand, if one of these strategies (“eliminated”
ly reduces the time to accomplish the activity. This strategy or “converted”) is totally inappropriate then the following
may involve workflow revision and physical modification of strategy of the “combined” and “simplified” will be consid-
machine parts or material of equipment handling. The appli- ered. Unlike the first two strategies, these strategies have two
cation of the new technology towards implementing this strat- directions of options; “possible” and “no.” The justifications
egy is also possible to be applied. of these options are similar as before. The reason why there is
no “yes” option is that the nature of these strategies (“com-
4.4.2 Conceptual decision model bined” and “simplified”) requires further clarifications and
investigation. For example, if the strategy of “combined” is
The overall conceptual decision model based on decision “possible” to be done, the improvement project team have to
strategies and priority elements presented in the previous analyze with what activity/activities is the best to be
444 Int J Adv Manuf Technol (2018) 94:433–450

Fig. 7 Conceptual decision

Start
model

Does it
able to be…

Eliminated? Implement
Yes

No Possible

Does it
able to be…

Changeover Implement

Further investigation and consideration needed

time Converted? Yes
reduction
achieved
& No Possible
Consider Does it
other group able to be…
activities /
activity Combined?
Possible Implement

No

Does it
able to be…

Simplified?
Possible Implement

No

Consider other
group activities /
activity

combined. The results of conceptual decision model valida- The overall decision mapping for the activity level is present-
tion for group activities and activity levels are presented in ed in Table 4. The application of the conceptual decision mod-
Tables 3 and 4, respectively. el for the activity level focuses on the activity in the classified
In other words, the results validation given in Tables 3 and groups that currently recorded the significant time taken. For
4 presented the decision mapping process that guides the im- example, in group A, the activities 3, 6, 7, and 10 become the
provement project team to focus on the best solutions to be focus to identify the appropriate strategy. Based on the
planned and implemented. Table 3 shows that for the group decision-making flow given in the conceptual decision model,
activities level, group activities B is possible (P) to be elimi- the appropriate strategy for activity 3 is “combined” and
nated or converted. The justification of this decision for group
B can be described based on the value and non-value added
Table 3 Validation of conceptual decision model for group activities
concept in lean thinking approach. After comprehensive dis- level
cussion between the academician and practitioner (e.g., engi-
neers) members, the team agreed that majority of the activities Decision options Eliminate Convert Combine Simplify
Group activities
in group B (except activities no. 12 and 31) be classified under
the non-value added activities. Therefore, the suggestion of A: Cleanup process N N N P
group activities B is “possible” to be eliminated is preferred, B: DRJ Identification P P – –
where the used of the decision direction “possible” means the C: DRJ Installation N N N P
team has to focus on the best solution so that the group activ- D: DRJ Setup N N P P
ities B can be eliminated. E: Composite panel setup N N P P
On the other hand, there are variety of strategies given for F: Computer program setup N N P P
the group activities A, C, D, E, F, and G. Therefore, the deci- G: Previous DRJ storage N N N P
sions mapping for the activity level is recommended for these
groups to determine the right activity for the right strategy. Y yes, P possible, N no, − no longer considered
Int J Adv Manuf Technol (2018) 94:433–450 445

Table 4 Validation of conceptual decision model for activity-by-activity level

No. Decision options Eliminate Convert Combine Simplify

Activity Time taken (s) Group activities

1 Switch off the related operation buttons of trimming process 6 A (Internal) – – – –

2 Open the machine door 6 – – – –
3 Remove the machined parts 185 – – P P
4 Open up the DRJ clamps 19 – – – –
5 Adjust the overhead crane 14 – – – –
6 Hook four strings of the overhead crane to DRJ 77 – – – P
7 Adjust the hooks 95 – – – P
8 Lift up the DRJ 13 – – – –
9 Place the DRJ outside the trimming process area 20 – – – –
10 Release the hooks 112 – – – P
11 Move the overhead crane 50 – – – –
12 Drive forklift to the DRJ rack 34 B (Internal) – Y – –
13 Lift a long DRJ 146 P – – –
14 Move the DRJ to another rack 25 P – – –
15 Lift a new long DRJ 73 P – – –
16 Move the DRJ within rack 30 P – – –
17 Lift another long DRJ 102 P – – –
18 Adjust forklift to reverse 105 P – – –
19 Move the DRJ to another rack 19 P – – –
20 Drive forklift back to origin 23 P – – –
21 Check DRJ part number 68 P – – –
22 Drive forklift back to rack 75 P – – –
23 Take a new long DRJ 16 P – – –
24 Adjust forklift to reverse 33 P – – –
25 Move DRJ to other rack 276 P – – –
26 Reverse forklift 75 P – –– –
27 Lift another DRJ 112 P – – –
28 Move DRJ to another rack 31 P – – –
29 Take the desired DRJ 59 – Y – –
30 Adjust forklift to reverse 25 – Y – –
31 Bring the desired DRJ to the H&H Trimming Machine 117 – Y – –
32 Adjust the overhead crane 14 C (Internal) – – – –
33 Hook four strings of the overhead crane to DRJ 77 – – – P
34 Adjust the hooks 95 – – – P
35 Lift the DRJ 7 – – – –
36 Place the DRJ onto the bed of the machine 12 – – – –
37 Release the hooks 101 – – – P
38 Move the overhead crane 50 – – – –
39 Tighten the clamps for DRJ 19 – – – –
40 Adjust the tool bit using machine controller 93 – – P P
41 Touch offset the surface of the DRJ 102 – – P P
42 Set the machine program 9 – – – –
43 Adjust the tool for best fit process 73 P P
44 Best fit process of one side DRJ 38 – – P P
45 Best fit another side of DRJ 111 D (Internal) – – P P
46 Check DRJ part number 62 – – P P
47 Checking the flat point on PC 210 – – P P
446 Int J Adv Manuf Technol (2018) 94:433–450

Table 4 (continued)

No. Decision options Eliminate Convert Combine Simplify

Activity Time taken (s) Group activities

48 Enter flatness points data into PC 12 – – P P

49 Adjust flatness points in the PC 186 – – P P
50 Best fit again for all points 7 P – P P
51 Enter data into PC 88 – – P P
52 Best fit process for another part number 137 – – P P
53 Setting the program for the part number 65 – – P P
54 Check the best fit flatness 249 – – P P
55 Setting the tool bit 153 – – P P
56 Takes the composite panel 9 E (Internal) – – – –
57 Lift the composite panel and place it onto the DRJ 28 – – P P
58 Fix the vacuum hose 92 – – P –
59 Place another panel 123 – – P P
60 Close the machine door 6 F (Internal) – – – –
61 Choose and find the correct program for the composite panel 533 – – – P
62 Push start button for trimming process 4 – – – –
63 Drive forklift to the previous DRJ located as stated in activity (9) 40 G (External) – – – –
64 Move the previous DRJ to the rack 120 – – – –

activities 6, 7, and 10 are “simplified,” respectively. These two process involved, and storage area. The implementation result
strategies are dominated by activities in other groups (D, E, F, found that the application of this tagging system also elimi-
and G).The following section presents the specific solutions nated the activity no. 46 in the group activities D. It is because
towards achieving the strategies that have been mapped. in current practice, activity no. 46 is carried out to confirm that
they (workers) are installed in the right type of DRJ.
4.5 Solutions towards improvement strategies
4.5.2 Scanning system application
This section presents three current solutions that have been
implemented and one solution that is planned to be done in The application of scanning system is another solution that
the near future. All the solutions are based on the decisions was recommended by the improvement project team based
mapping process presented in Section 4.4.2. on the strategy of “simplified” as given in Table 4. This solu-
tion directly reduce the time of activity no. 61 (choose and find
4.5.1 Tagging system design the correct program for the composite panel) in group F (com-
puter program setup), wherein the current practice this activity
The tagging system is designed in order to implement the contributes 98% of the overall internal type of activities in
strategy of “eliminated” for the group activities B (except group F. Figure 9 shows the physical picture of scanning
activities no. 12 and 31) as suggested in the conceptual deci-
sion model result given in Table 3. The development of tag-
ging system is focused on two areas: tagging for DRJs and
tagging for the DRJs racks. The idea behind this tagging sys-
tem is that once the worker received the order of the next type
of product to be trimmed, then the forklift is directly driven to
the right rack and take the right DRJ based on the information
presented in the tagging. Figure 8 shows the tagging that has
been designed. The tagging cards containing all of the infor-
mation of the DRJ needed by workers in order to take and
place the DRJ. The tagging card contains the name of the
product, specific model, type, part number, part name, the Fig. 8 Tagging card
Int J Adv Manuf Technol (2018) 94:433–450 447

system that is applied and implemented. After the scanning

system is introduced, the steps of activity no. 61 are reduced
and hence it reduced the overall time of activities in group F.

4.5.3 Relocate and reposition the machine control and PC

table

This solution is related to the improvement strategy for activ-

ities no. 44, 45, 47, 49, 50, 52, and 54 of group D. As sug-
gested by the conceptual decision model, all these activities
are recommended either to implement the “combined” or
“simplified” strategies.
The justification behind this solution is as follows. During the
Fig. 9 Scanning system
setup activities of group D (DRJ setup), the workers need to walk
over and over again between the machine controller and the PC
table to calculate the aerodynamic, to scan the right computer improvement (changeover process time reduction including
program and to see the flatness of the DRJ on the machine bed external and internal activities) is 44%, where 88% time re-
during the best fit process. Therefore, the improvement solution duction is contributed by the improvement of group activity B
suggested by the project team is to relocate and reposition the PC (DRJ identification process), followed by group activity F
table and machine controller. It means that the PC table be moved (computer program setup process) that achieved 68% im-
closer to the machine controller and the screen of PC table provement and, finally, group activity D (DRJ setup process)
repositioned so that the movement of the workers during DRJ which improved to 32% of the time reduction. According to
setup activities is minimized and the communication between the project charter presented in Table 1, the overall recorded
workers is more effective. improvement exceeded the target of the project goal where at
least 30% of the overall improvement is targeted.
4.5.4 DRJ calibration and handling The overall time of group activity B is reduced from 1444
to 176 s, where it was achieved due to the application of
During the DRJ setup process (group D), the times to accom- tagging card system on DRJs and DRJ racks. Currently, the
plish activities no. 39, 44, 45, 47, 49, 50, 52, and 54 are also tagging cards system is an effective solution to improve the
contributed by the physical issue of DRJ that is currently used. DRJ identification process. This solution results in the elimi-
Observations and video record found that the DRJ was not nation of activities no. 13 to 29, thus contributing to the time
handled properly and stored in the right position, thus some of reduction in this group activities to 88%. Another time
parts of the DRJ were bent. Therefore, the current setup pro- reduction in this group is contributed with the solution of
cess of DRJ required physical adjustment to reach the right conversion from the internal type activities to external type
flatness rate. For example, one of the repetitive actions to be activities for activities no 12, 30, and 31. Thus, it directly
carried out in activities no. 44, 45, 47, 49, 50, 52, and 54 is that reduces the downtimes of the machine operations.
the workers need to adjust and re-adjust the clamps of DRJ on The second higher of time reduction comes from group
the machine bed in order to reach the right flatness point. activity F which is reduced from 543 to 173 s by applying
Therefore, two solutions strategies are proposed by the im- the scanning system. Although the initial cost of the scanning
provement project team. The first is to develop comprehensive system is quite significant, the positive effects from this solu-
planning and scheduling of the DRJ calibration. The second is tion are still worthwhile for the company in the perspective of
to revise the current procedure of DRJ handling including productivity. The implementation of this solution specifically
current ways of DRJ to be stored and also the two-way process reduced the time of activity no. 61 from 533 to 163 s.
of loading-moving-unloading the DRJ from the rack to the Meanwhile, the following time reduction of this project comes
trimming machine area. The goal of these proposed solutions from group activity D, which it reduced from 1614 to 1099 s
strategy is to avoid the DRJ from bending thus will minimize (32% of time reduction) through the suggested solution to
the overall DRJ setup process. revise the working layout. The detail of this solution is to
relocate and reposition the machine control and PC table; thus,
the activities no 45, 47, 49, 52, and 54 are reduced from 111 to
5 Results and discussion 80, 210 to 87, 186 to 137 to 95, and 249 to 89 s, respectively.
Moreover, the time reduction in group activity D also is con-
The comparison results before and after improvement based tributed to the implementation solution to eliminate the activ-
on the proposed solutions are presented in Table 5. The overall ity no 46, where it is a “chain effect” benefit from the
448 Int J Adv Manuf Technol (2018) 94:433–450

Table 5 Results of before and after improvement

Group activity Improvement strategy/solution description Time taken (s)

Before Improvement After Improvement

A. Cleanup process 597 (internal) 597 (internal)

B. DRJ identification i. Activities no. 13 to 29 applied eliminated strategy. Solution 1444 (internal) 176 (external)
for this strategy is implementation of tagging card system Time reduction percentage is 88%
on DRJ and DRJ rack.
ii. Activities no. 12, 30, and 31 applied converted strategy
from internal to external type activity.
C. DRJ installation 356 (internal) 356 (internal)
D. DRJ setup i. Applied the strategy of simplified, where the solution of this 1614 (internal) 1099 (internal)
strategy is relocate and reposition the machine control and Time reduction percentage is 32%
PC table. Therefore, times taken to accomplish the
activities no. 45, 47, 49, 52, and 54 are reduced from 111 to
80, 210 to 87, 186 to 89, 137 to 95 and 249 to 89 s,
respectively.
ii. Activity no. 46 is eliminated, where it is positive effect
from tagging card system implementation.
E. Composite panel setup 252 (internal) 252 (internal)
F. Computer program setup Applied the simplified strategy, where the implementation of 543 (internal) 173 (internal)
scanning system reduced the time of activity no. 61 from Time reduction percentage is 68%
533 to 163 s.
G. Previous DRJ storage 160 (external) 160 (external)
Total overall changeover time before and after improvement 4966 (82.8 min) 2792 (46.5 min)
Overall changeover process time reduction (%) 44% improvement

application of the tagging card system as has been discussed in The next example is the DRJ installation (group activity C)
the previous paragraph. process, which currently takes 356 s to complete the process.
On the other hand, another achievement in this current Suggested strategy from the decision model is to “simplify”
project is the significant reduction of overall internal type the activities no. 33, 34, and 37 that can be looked forward to
activities to 48%. Before the improvement is carried out, the in further improvement projects by designing a single lock
total time of internal activities is recorded as 4806 s and after and unlock equipment to make the installation process faster
the improvement, the total time of the internal activities is and easier. Also, further improvement in the DRJ setup pro-
reduced to 2477 s. As been discussed in Section 4.4.1, the cess (group activity D) that includes activities no. 39, 44, 45,
more the times of internal activities are reduced the more 47, 49, 50, 52, and 54 can be improved by applying more than
production time are available. Thus, in long-term effects, it one strategies (e.g., combine and simplify strategies) and re-
will increase the productivity. lated solutions such jig and clamps design based on poka-
Although the present project shows a significant improve- yoke concept.
ment in terms of the overall changeover process time reduc- Another example of the internal type group activity that can
tion, the project team believe that some activities especially be considered for further improvement project is the compos-
the internal type are possible to be improved further in future ite panel setup (group activity E) process. Currently, this pro-
improvement projects. For example, the cleanup process cess takes 252 s to be completed. In this process, activities no.
(group activity A) which currently takes 579 s to be complet- 57, 58, and 59 are possible to apply “combine” and/or “sim-
ed, where 32% (185 s) of the time comes from activity no. 3 plify” strategies as suggested by the decision model. Like
(remove the machined parts). The analysis of the decision discussion given for groups A, C and D previously, the pos-
model suggested that the activities no 3, 6, 7 and 10 are highly sible solution is to re-design the material handling equipment
recommended for “simplify” strategy (activity no 3 is also (e.g., clamps) to make the movement faster.
possible to “combine”). The possible solutions that can be From the scientific perspective, although the application of
considered in the further project are to simplify the workflow the conceptual decision model in this project helps the im-
and re-design the support tools (e.g., clamps to remove and provement project team towards identifying and focusing the
setup machined parts) towards single “lock in - lockout” right strategies, the scientific element of this process can be
design. enhanced in future projects. For example, the application of
Int J Adv Manuf Technol (2018) 94:433–450 449

fuzzy-based analytical hierarchy process (FAHP) is one of the Acknowledgements The authors wish to acknowledge the support
from the Universiti Malaysia Perlis (UniMAP) through the Research
possible methods to be considered to systematically analyze
Acculturation Grant Scheme (RAGS 9018-00076). The authors also
and evaluate the related criteria for selecting the best solutions would like to thank the Case Study Company and their team as well the
of the strategy. For a more complex case such as multi- anonymous referees for their helpful suggestions and comments, which
strategies and solutions consideration will make the applica- have improved the contents and style of this paper.
tion of FAHP more challenging and interesting to be applied
in future improvement projects.
The final discussion point is related to the team set- References
ting and management approach used in the current pro-
ject. The project team believes that one of the key fac- 1. Allahverdi A, Soroush HM (2008) The significance of reducing
setup times/setup costs. Eur J Oper Res 187:978–984
tors that contributed to the success of the current project
2. Mackelprang AW, Nair A (2010) Relationship between just-in-time
is the combination of two groups of team members that manufacturing practices and performance. A meta-analytic investi-
from practitioners (engineers/manager) and academician. gation. J Oper Manag 28:283–302
The combination of ideas, knowledge, and experiences 3. Gross R, Maxim C, Adhikari CD, Rothe J (2010) Leveraging new
of these groups directly drove the project towards opti- SEMI standard to reduce waste and improve flow for semiconduc-
tor manufacturing. Robot Comput Integr Manuf 26:658–664
mizing the theory and practice elements. For example,
4. Liu CY, Chang SC (2000) Scheduling flexible flow shops with
academic-related team members guided the project to- sequence-dependent setup effects. IEEE Trans Robot Autom
wards scientific perspective, while the engineers/ 16(4):408–419
manager verify and validate the solutions process to- 5. Eren T, Guner E (2006) A bicriteria flow shops scheduling problem
wards practical and business perspective. with setup times. Appl Math Comput 183:1292–1300
6. Chen WJ (2009) Scheduling with dependent setups and mainte-
nance in a textile company. Comput Ind Eng 57:867–873
7. McIntosh R, Culley S, Gest G, Mileham T, Owen G (1996) An
assessment of the role of design in the improvement of changeover
6 Conclusions performance. Int J Oper Prod Manag 16(9):5–22
8. Shingo S (1985) A revolution in manufacturing: the SMED system.
Productivity Press, Cambridge
This paper reported an improvement project to reduce the
9. Goubergen D (2008) Set-up reduction for lean cells and multi-
changeover time for the trimming process in the advanced machine situations. Int Fed Inf Proc 257:295–303
composite manufacturing industry. The following conclusions 10. Karasu MK, Cakmakci M, Cakiroglu MB, Ayva E, Demirel-
from this study are as follows. Ortabas N (2013) Improvement of changeover times via Taguchi
empowered SMED—case study on injection moulding production.
Measurement 47:741–748
& The presented project applied various continuous im-
11. Almonani MA, Aladeemy M, Abdelhadi A, Mumani A (2013) A
provement tools such as the cause and effect analysis, five proposed approach for setup time reduction through integrating
whys analysis, and some steps of conventional SMED conventional SMED method with multiple criteria decision-
method. making techniques. Comput Ind Eng 66:461–469
& The proposed four standard improvement strategies and 12. Deros BM, Mohammad D, Idris MHM, Rahman MNA, Ghani JA,
Ismail AR (2011) Setup time reduction in an automobile battery
its priorities are introduced, as well as the conceptual de- assembly line. Int J Syst App Eng Dev 5(5):618–625
cision model for best strategies selection. 13. Desai MS (2012) Productivity enhancement by reducing setup time
& The proposed conceptual decision model guided the im- - SMED: case study in the automobile factory. Glob J Res Eng
provement project team to identify and focus the right Mech Mechn Eng 12:5A
strategies and the best solutions. 14. Kumar BS, Abuthakeer SS (2012) Productivity enhancement by
implementing lean tools and techniques in an automotive industry.
& Four specific solutions have been recommended based on Int J Eng 10:167–172
the mapping strategies given in the conceptual decision 15. Souza M, Brandão LC, Rascalha A, Fernandes W, Souza MG
model and three of them have been implemented (2012) Implantation of the SMED technology into an auto parts
successfully. industry: a case study. Proceeding of the International Conference
on Industrial Engineering and Operations Management (pp. 1–10),
& Current implementation results found that the overall
Portugal
changeover time is reduced to 44% and the total internal 16. Kušar J, Berlec T, Žefran F, Starbek M (2010) Reduction of ma-
type activities time is reduced to 48%. chine setup time. J Mech Eng 56(12):833–845
& One of the keys that contributed to the success of the 17. Moreira AC, Pais GCS (2011) Single minute exchange of die. A
project is the combination of team members that form case study implementation. J Technol Manag Innov 6(1):129–146
academicians and practitioners. 18. Dhake R, Rajebhosale V (2013) Setup time reduction on solder
paste printing machine—a case study. Int J Lean Thin 4(1):90–97
& The method presented in this paper is also possible to be 19. Benjamin SJ, Murugaiah U, Marathamuthu SM (2013) The use of
applied for others cases related to the changeover im- SMED to eliminate small stops in a manufacturing firm. J Manuf
provement process. Technol Manag 24(5):792–807
450 Int J Adv Manuf Technol (2018) 94:433–450

20. Sayem A, Islam MA, Khan MMA (2014) Productivity enhance- rubber gloves manufacturing process. Int J Lean Six Sigma 5(1):
ment through reduction of changeover time by implementing 2–21
SMED technique—in furniture industry. Int J Ind Syst Eng 17(1). 30. Gijo EV, Scaria J, Antony J (2011) Application of Six Sigma meth-
doi:10.1504/IJISE.2014.060820 odology to reduce defects of a grinding process. Qual Reliab Eng
21. Patel S, Shaw P, Dale B (2001) Set-up time reduction and mistake Int 27(8):1221–1234
proofing methods. A study of application in a small. Bus Process 31. Nonthaleerak P, Henry L (2008) Exploring the Six Sigma phenom-
Manag J 7(1):65–75 enon using multiple case study evidence. Int J Oper Prod Manag
22. Singh BJ, Khanduja D (2010) SMED: for quick changeovers in 28(3):279–303
foundry SMEs. Int J Product Perform Manag 59(1):98–116 32. Pande PS, Neuman RP, Cavanagh RR (2000) The Six Sigma way:
23. Kumar V, Bajaj A (2015) The implementation of single minute how GE, Motorola, and other top companies are honing their per-
exchange of die with 5’s in machining processes for reduction of formance. McGraw-Hill, New York
setup time. Int J Rec Tech Mech Elect Eng 2(2):32–39 33. Doggett AM (2005) Root cause analysis: a framework for tool
24. Ibrahim MA, Mohamad E, Arzmi MH, Rahman MAA, Saptari A, selection. Qual Manag J 12(4):34–45
Shibghatullah AS, Sulaiman MA, Ali MAM (2015) Enhancing
34. Ahmad R, Kamaruddin S, Khan ZA, Mokthar M, Almanar IP
efficiency of die exchange process through single minute exchang-
(2006) Implementation of dust control system using management
ing die at a textile manufacturing company in Malaysia. J Appl Sci
and planning tools (MPT): a case study. Manag Env Qual Int J
15(3):456–464
17(4):390–408
25. Stadnicka D (2015) Setup analysis: combining smed with other
tools. Manag Prod Eng Rev 6(1):36–50 35. Tan CM, Raghavan N (2007) Root cause analysis based mainte-
26. Karasu MK, Cakmakci M, Cakiroglu MB, Ayva E (2014) nance policy. Int J Qual Reliab Manag 24(2):203–226
Improvement of changeover times via Taguchi empowered 36. Mahto D, Kumar A (2008) Application of root cause analysis in
SMED/case study on injection molding production. J Measure improvement of product quality and productivity. J Ind Eng Manag
47:741–748 1(2):16–53
27. Braglia M, Frosolini M, Gallo M (2016) Enhancing SMED: 37. Jabrouni H, Kamsu-Foguem B, Geneste L, Vaysse C (2011)
changeover out of machine evaluation technique to implement the Continuous improvement through knowledge-guided analysis in
duplication strategy. Prod Plan Control 27:328–342 experience feedback. Eng App Artif Intell 24(8):1419–1431
28. Pyzdek T (2003) The six sigma handbook: a complete guide for 38. Andersen B, Fagerhaug T (2000) Root cause analysis: simplified
green belts, black belts, and managers at all levels. The McGraw- tools and techniques. ASQ Quality Press, Milwaukee
HIll Companies, Inc. doi:10.1036/0071415963 39. Murugaiah U, Benjamin SJ, Marathamuthu MS, Muthaiyah S
29. Jirasukprasert P, Garza-Reyes JA, Kumar V, Lim MK (2014) A Six (2010) Scrap loss reduction using the 5-whys analysis. Int J Qual
Sigma and DMAIC application for the reduction of defects in a Reliab Manag 27(5):527–540