Sie sind auf Seite 1von 48

International Journal of Quality & Reliability Management

Integration of Taguchi and Shainin DOE for Six Sigma improvement: An Indian case
Anupama Prashar,
Article information:
To cite this document:
Anupama Prashar, "Integration of Taguchi and Shainin DOE for Six Sigma improvement: An Indian case", International
Journal of Quality & Reliability Management, https://doi.org/10.1108/IJQRM-08-2015-0116
Permanent link to this document:
https://doi.org/10.1108/IJQRM-08-2015-0116
Downloaded on: 07 July 2017, At: 00:15 (PT)
References: this document contains references to 0 other documents.
To copy this document: permissions@emeraldinsight.com
The fulltext of this document has been downloaded 1 times since 2017*
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

Access to this document was granted through an Emerald subscription provided by emerald-srm:149425 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service
information about how to choose which publication to write for and submission guidelines are available for all. Please
visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company manages a portfolio of
more than 290 journals and over 2,350 books and book series volumes, as well as providing an extensive range of online
products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee on Publication
Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive preservation.

*Related content and download information correct at time of download.


Title: Integration of Taguchi and Shainin DOE for Six Sigma improvement: An Indian case

Abstract

Purpose:

The goal of the study was to demonstrate the application of Six Sigma/DOE Hybrid framework
for improving damping force generation process in a shock absorber assembly unit.

Methodology/Design
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

The study adopted case study research method with single case (holistic) design. This research
design was found to be appropriate for testing the projected framework for integrating DOE
approaches within Six Sigma DMAIC (Define-Measure-Analyse-Improve-Control) cycle. In the
proposed framework, Shainin’s Component Search Technique (CST) was deployed at the
‘analysis’ phase of DMAIC for first stage filtering of process parameters, followed by the use of
Taguchi Orthogonal Arrays (OA) at the ‘improve’ phase for identifying the optimal setting of the
parameters

Findings

The application of Shanin CST facilitated in ascertaining that assembly component (piston with
rebound stopper) and not the assembly process was causing the variation. Further, the use of
Taguchi OA at the improve phase, allowed the collection of necessary data to determine the
significant piston parameters with minimum experimentation (eight experimental runs in this
case as opposed to the expected 64) and ANOVA on the collected data facilitated the selection of
parameter settings to optimize the CTQ i.e. rebound damping force.

Originality

This study provided a stimulus for wider application of integrated DOE approaches by the
engineering community in the problem-solving and the identification of parameters responsible
for poor performance of the process.

1
Keywords: Shainin Component Search Technique (CST), Taguchi Orthogonal Arrays; Shock
absorber assembly; damping force

1. Introduction

It has been more than nine decades since Sir Ronald Fisher put Design of Experiments(DOE) in
the hands of statisticians thereby making the study of complex systems more efficient and
effective (Fisher, 1935). During the 1920s and early 1930s, Fisher introduced principles of
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

statistical thinking into designing experimental investigations mainly in agriculture and other
related life sciences. Most methods of experimentation such as factorial design and analysis of
variance (ANOVA), developed during this era, are classified as Classical DOE (Fisher 1958,
1966). The application of statistically designed experiments from agriculture to more complex
industrial experimentation was driven by the development of Response Surface Methodology
(RSM) in 1950s (Box & Wilson, 1951). Over the next thirty years, RSM and other Classical
DOE techniques spread throughout the chemical and process industries, mainly for the purpose
of research and development. However, due to lack of computing resources and adequate
training in basic statistical concepts, the application of statistically designed experiments was not
widespread at manufacturing process level (Johnson et al, 2011).

By late 1970s, statisticians switched their attention to more standardized approaches of designed
experiments, popularly known as Taguchi methods (Kackar 1985; Taguchi 1987, 1991; Taguchi
& Wu 1980). Developed by Dr. G. Taguchi, this experimentation approach (also known as
robust design) suggested highly fractionated factorial designs and orthogonal arrays(OA) along
with some novel statistical methods for discrete part industries including automotive, aerospace,
electronics and semiconductors.

Furthermore, the need for simpler alternatives for designed experiments led to the development
of Shainin System (SS), named after its creator - Dorian Shainin, a well-known quality
consultant in US and Europe (Bhote, 1990; 1991; Bhote & Bhote, 2000). The Shainin System
(SS) emphasized on using observational investigations prior to experimental investigations; and
searching for a dominant cause using the process of elimination and leveraging (Shainin, 1993b;

2
De Mast, 2004; Steiner et al, 2008). Whilst the work of Fisher and Taguchi is well acclaimed in
contemporary manufacturing industry, the work of Dorian Shainin is criticized for being
overstated and unsubstantiated (Zeigel, 2001; Moore, 1993; Nelson, 1991; Hockman, 1994.

All these statistical experimental design approaches (Classical, Taguchi & Shainin) have their
proponents and opponents; and there is sufficient literature that debates the advantage of one
over the other for industrial experimentation (Tanco et al, 2008; Thompson & Antony, 2005;
Steiner & MacKay, 2005; De Mast et al, 2004, 2000; Bhote, 2000; Verma, et al, 2004; Ledolter
& Swersey, 1997). Nevertheless, instead of approving one experimental design approach over
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

the other, these powerful tools needs be integrated to achieve cost effective experimentation
(Tanco et al, 2009; Thomas & Braton, 2006; Schmidt & Lausnby, 2005; De Mast et al, 2000). So
a meaningful strategy is to integrate the best DOE tools from an engineering point of view to
ensure their ease of use and implementation.

For instance, the SS offers clue generation tools to systematically reduce the number of
parameters involved in the problem to a manageable number (Thompson & Antony, 2005).
Taguchi’s OA together with ANOVA allows the running of experiments to identify significant
parameters and their optimal setting for product/process quality with minimum amount of
experimentation. Thus, these tools help in saving cost and time substantially. Although the merits
and demerits of the different approaches of designing experiments are extensively researched
(Tanco et al, 2008; Thompson & Antony, 2005; Steiner & MacKay, 2005; Verma, et al, 2004;
Bhote, 2000; Ledolter & Swersey, 1997), there are hardly any studies on disciplined integration
of the different DOE tools into a functional framework (Thomas & Braton, 2006; De Mast et al,
2000).

In that direction, Six Sigma has turned out to be a process improvement strategy, which
competently integrates the established quality tools such as Design of Experiments (DOE),
Quality Function Deployment (QFD), Failure Mode and Effect Analysis (FMEA), and
Measurement System Analysis (MSA) into a structured methodology (Goh, 2002). In such
implementations, DOE features prominently in the ‘analysis’ and ‘improve’ phases of Sig Sigma
DMAIC (Define-Measure-Analyze-Improve-Control) cycle.

3
Addressing the paucity of literature on the topic, the current study proposed a Six Sigma/DOE
Hybrid framework for integrating Taguchi and Shainin DOE tools into Six Sigma DMAIC cycle
and illustrated the application of such a framework in reducing rework in an automotive
assembly. The rest of the paper is organized in three sections; beginning with the review of
literature on Six Sigma Hybrid improvement methodologies.The next section described an
integrated experimental approach employed in a Six Sigma project aiming to reduce the rework
in automobile shock absorber assembly. This is followed by conclusion and direction for future
research.
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

2. Literature Review

The literature has been reviewed from the following perspectives:

• Application of Six Sigma Hybrid frameworks


• Integration of design of experimentation within Six Sigma process improvement

2.1 Application of Six Sigma Hybrid frameworks

Illustrations of Six Sigma Hybrid methodologies for process improvement are not new in
literature. These studies draw attention to some of the limitations of long-established Six Sigma
tools & techniques (Aboelmaged, 2010; George et al. 2005; Senapti, 2004; De Mast, 2004;
Pande et al., 2000). Some of these limitations are ambiguity and vagueness resulting from the
process of Six Sigma project selection (Bilgen & Sen, 2012; Breyfogle et al. 2001), subjectivity
in the generation of hypothesis proposing relationships between variables involved (De mast,
2004), and lack of integration of improvement initiatives with organizational strategy (Antony et
al 2006; Asif et al 2009; Coronado & Antony, 2002).

Lean Six Sigma (LSS) approach is one of the most popular hybrid methodologies combining the
synergies of Six Sigma and Lean management to improve the business opportunities in customer
satisfaction, cost and process-pace (Salah et al, 2010; Pepper & Spedding, 2010; George et al,

4
2005; Arnheiter & Maleyeff, 2005; Sheridan, 2000). On similar lines of LSS, Banawi & Bilec
(2014) integrated Lean, Six Sigma and Green (assessment of environmental aspects) in a
systematic approach with the goal of improving the quality and environmental impacts of the
construction processes.

Studies demonstrated that integration of Six Sigma philosophy into the quality standard (ISO
9000:2000) implementation process facilitates the achievement of optimal results in quality
progress, thereby ensuring customer satisfaction (Lupan et al, 2005; Pfeifer et al, 2004; Dey,
2002; Munro, 2000). There is ample evidence that combination of Six Sigma with Total Quality
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

Management (TQM) program facilitates process improvement through a detailed data analysis.
Six Sigma’s proven drill-down approach to root-cause analysis greatly supplements the
conventional TQM approach and enhances its effectiveness (Green, 2006; Lee & Ken, 2003).

Kumar et al. (2009) proposed & tested a hybrid methodology combining Analytic Hierarchy
Process (AHP) with project desirability matrix (PDM) for Six Sigma project selection in
automotive industry. Tka´c & Lyo´csa (2009) employed mathematical optimization modeling
techniques and real options theory in designing a Six Sigma project evaluation model. Kumar et
al. (2007) combined data envelopment analysis (DEA) with Six Sigma methodology for the
identification of important inputs and outputs for project selection. Tang et al (2007) explored
the possibilities of integrating Operations Research/Management Science (OR/MS) techniques to
enhance Six Sigma deployment in operational and transactional environments.

In a recent study, Krueger et al (2014) adopted qualitative case study methodology to explore the
process of implementing Six Sigma in a manufacturing plant. The study used grounded theory to
gain insights from the behavior of employees involved in the implementation process. Heavey &
Murphy (2012) demonstrated that the combination of Balanced Score Card (BSC) and Six Sigma
provides organizations the capability to influence & control their high-level performance metrics
in a systematic and structured manner. Ghosh & Maiti (2012) proposed a Six Sigma DMAIC
framework, driven by data mining techniques for defect diagnosis and quality improvement
using online process data. The study employed two decision tree algorithms viz. Classification

5
and Regression Tree and Chi-squared Automatic Interaction Detection in developing the
proposed framework.

Kumar et al (2011) used a three-tiered approach exploiting the synergies between Six Sigma,
discrete event simulation, and DOE to develop a flexible distribution network. Park et al (2009)
proposed and illustrated an integrated management strategy for digital society namely
Knowledge-based Digital Six Sigma (KDSS). It combined Six Sigma, Data technology (DT) and
Knowledge Management (KM). The authors suggested an alternative process flow viz. DMAMP
(Define-Measure-Analysis-Model-Predict) cycle for implementation of KDSS strategy.
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

2.2 Integration of DOE within Six Sigma framework

Literature includes several comprehensive texts on development of common industrial


experimental design and analysis methodologies (Allen, 2003; Wilcox, 2003; Bennett et al,
2003; Stamatis, 2002). Although a number of versions of design of experiments used by quality
practitioners since the “pre-Six Sigma” years, the ‘improvement’ had always been the impetus
for experimental design applications, regardless of the specific format of DEO used (Goh, 2002;
2001).

The review noted several Six Sigma case illustrations employing full and factorial experimental
design for process optimization (Vinodh et al, 2014; Gijo et al, 2014; Zhang et al, 2014; Johnson
et al, 2012; Gijo & Scaria 2010; Chen & Lyu, 2009; Lee et al, 2009; Johnson et al, 2006a;
2006b; Leitnaker & Cooper, 2005; Gijo & Rao 2005). Many of these industrial experimentation
studies used Taguchi orthogonal arrays (OA) and S/N ratio (signal to noise ratio) for reducing
variation in the processes (Vinodh et al, 2014; Saravanan et al, 2012; Gijo & Scaria 2010; Anawa
et al, 2008; Singh et al, 2007; Gijo 2005).

Although, the applications of the Classical and Taguchi methods within Six Sigma projects have
been extensively researched, but only a handful of contributions on Shainin DOE applications
have been reviewed and reported (Sharma & Chetiya, 2009). In one of the significant studies,

6
Bhote (1990) compared Shainin DOE techniques with Taguchi methods, in the context of the
electronics industry and concluded that the Shainin tools are simpler, less costly, and statistically
more powerful than Taguchi methods. Steiner et al (2008) provided a comprehensive overview
of Shainin System (SS) and critically assessed Six Sigma (SS) tools with other industrial
problem solving systems. The study suggested that SS is valuable for many types of industrial
problems and appropriate for integration with other process improvement methodologies.
However, the study emphasized that many of the SS tools are not novel and not necessarily the
best. Logothetis (1990) also evaluated the Shainin techniques in relation to the Taguchi methods
and statistical process control methods. Verma, et al (2004) used three cases of Taguchi
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

experiments from the available literature and reapplied Shainin method to find out whether it had
an edge over the other DOE techniques. Sharma & Chetiya (2009) examined two industrial
projects using Shainin tools within Six Sigma DMAIC improvement cycle.The study found that
these tools are simple and easy to implement. Goodman & Wyld (2001) demonstrated analysis of
industrial honing operation using Shainin experimental design. This study suggested that the
Shainin methodology is very practical and easily executable in a variety of industrial settings.

Review also showed studies which described and critically reviewed the controversial tools
associated with Shainin DOE namely: Isoplot (Traver, 1995, Shainin, 1992), Multivari chart (De
Mast et al., 2001; Seder 1990), Component Search (Shainin and Shainin, 1988), Variable Search
(Thompson & Antony, 2004; Anthony, 1999; Ledolter & Swersey, 1997b), Paired or group
comparison (Bhote & Bhote, 2000; Shainin, 1993b;), B vs. C (Bhote, 1991) and Pre-control
(Steiner, 1997-1998; Ledolter & Swersey,1997a; Mackertich, N. A. 1990; Ermer & Roepke;
1990; Traver, 1985).

The literature so far demonstrated that the comparison and assessment of established
experimental design approaches viz. Taguchi & Shainin, for their value in industrial
experimentation has been successfully done (Bhote, 1990; Logothetis, 1990; Goodman & Wyld,
2001; Verma, et al, 2004). However, there are hardly any studies which illustrate the integration
of these DOE approaches in a functional framework (Thomas & Braton, 2006; De Mast et al,
2000).

7
3. Research Methodology

The study adopted case study research method with single case (holistic) design (Yin, 2009;
Eisenhardt, 1989). This research design was found to be appropriate for testing the projected
framework for integrating DOE approaches within Six Sigma DMAIC (Define-Measure-
Analyse-Improve-Control) cycle.

3.1 Background of case


Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

The case study was conducted at an automotive component supplier unit in North India in the
year 2013.The supplier was facing challenge of high rework cost of front shock absorber
assembly of Model X of their 4-wheelers.This was also adversely affecting their productivity
since every suspected part needed to be tested twice before rework. The problem was responsible
for the damping rework warranty claims of 3300ppm (parts per million) (defective rate of 10% in
Nov. 2013) and consequently, an annual cost of external failures of $5,000. Owning to the
chronic nature of the problem, the company initiated a Six Sigma improvement project. A cross-
functional team involving representatives from product & process design, engineering
&production, quality management, marketing, sales and distribution was formed to work on this
project. The facilities, materials, and time required to carry out the project was approved by
management, thus enabling the project team to begin its work. For a better understanding of the
problem, the damping force generation process is explained in the next subsection.

3.2 Damping force generation process in Shock Absorber assembly

In an automobile, the suspension system serves as a forced damping system, where force is
exerted from the road surface. It plays the dual role:
• contributing to the vehicle's road holding or handling and braking for safety and driving
pleasure, and
• keeping vehicle occupants comfortable by isolating them from road noise, bumps, and
vibrations.
The suspension system, basically, has two main components:
8
• springs to absorb impacts, and
• shock absorber or shock absorber to control unwanted spring oscillations and dampen the
variations in automobile wheels & chassis.

A hydraulic Twin-tube shock absorber is presented in Figure 1. It comprises a piston with built
in valves and orifices, which moves within a liquid filled inner (cylindrical) tube. Valves are
built into the piston itself and the piston is forced to move within the tube as the car’s wheel
suddenly hits a bump on the road. In this process a pressure differential is built across the piston
and the liquid is forced to flow through valves located in the piston. This stroke of the piston is
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

called ‘Compression Stroke’. Further, since this flow of liquid from chamber above the piston to
the chamber below it is restricted because of smaller size of the piston orifice and valve, the
sudden shock because of the bump of the road gets countered or absorbed. Additionally, in a
twin-tube shock absorber also has base valves to further absorb the extra pressure. These base
valves which open up and push the liquid to the outer tube. The complete assembly does have a
coil spring whose tension increased as it is compressed by the compression stroke. This
compressed spring tried to bring the piston back into the original position. This stroke is called
‘Rebound Stroke’. Once again, piston and base valves allow the liquid to shift between the
chambers in such a way that adequate rebound forces are generated as a result.

Thus, if the rebound forces are too low or too high respectively leading to the car losing its
contact from the ground or passenger feeling a jerk. The figure below (Figure 1) showed the
difference between the recovery of piston’s original position if adequate or low rebound force is
generated in the shock absorber.

9
Figure 1: Adequate & low rebound force
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

3.3 Six Sigma – DOE hybrid framework

In order to integrate Six Sigma (SS) and Taguchi DOE tools within Six Sigma DMAIC cycle, the
team proposed a Six Sigma – DOE hybrid framework. The Six Sigma DMAIC methodology was
used as the core of the framework, since overall it is a strong methodology that has already
achieved a greater degree of acceptability across industries (De Mast, 2004). SS & Taguchi DOE
were incorporated within the DMAIC cycle to bridge the ‘analyse’ and ‘improve’ phases of the
cycle by narrowing down the source of variation (Thomas & Braton, 2006).
.
In the proposed framework, Shainin Component Search tool (CST) was employed at the first
stage of experimentation to identify whether the primary cause of failure of the front shock
absorber lay within the assembly process itself or within the assembly components (Bhote,
2000). After the filtration in the first stage, Taguchi Orthogonal arrays (OA) were used to

10
identify the key factors and interactions that existed between them (Antony & Kaye, 1995;
Montgomery, 1992; Taguchi, 1987; Phadke, 1989) (Figure 2).

Figure 2: Six Sigma/ DOE Hybrid framework


Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

Source: Adopted from Thomas & Braton, 2006

11
The key tools & techniques used in each phase of the proposed framework are listed in Table
1(Table 1 about here).

A detailed description of each of these phases is given in the following sections:

3.3.1 The define phase

The define phase of the Six Sigma methodology aims to define the scope and goals of the
improvement project in terms of customer requirements (‘critical to quality’ or CTQs) and
identify the underlying process that needs improvement. In order to achieve this aim, the team
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

used the following tools:

3.3.3.1 Pareto analysis for CTQ identification: The team plotted a Pareto chart using the shock
absorber rejection & rework data for three months (July-September, 2013) (Figure 3).

Figure 3: Pareto chart for shock absorber assembly rejections

12
The analysis revealed that “damping force (DF)” was the key contributor (61.9%) towards
rejection of the shock absorbers. Further analysis of possible types of DF in a shock absorber (or
shock absorber) and related defect data showed that low rebound (67.4%) was the major source
of DF defect (Figure 4).

Figure 4: Pareto chart for DF defect


Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

3.3.1.2 Project Charter: After CTQ identification, the team developed a project charter detailing
the problem definition, project goals, process boundaries, team composition and project schedule
(Annexure I). Considering the aspects of time-period, specificity and measurability (Snee, 2001),
the team defined the problem & mission statement as follows:

Problem statement: High damping force warranty claims (defective rate of 10% in Nov.
2013) ensuing a cost of external failures of $5,000 per annum.

Mission statement: Based on a rule of thumb for setting Six Sigma project target of solvingat
least 50% of the problem within four to five months of time (Eckes, 2001), the team set the
following target:

13
Reduction in the field rejection rate of shock absorber assembly from 10% to 3% by the end
of Feburary, 2014.

3.3.2 Measure

In this phase, the team focused on validation of measurement system and capability analysis of
CTQs defined in the previous stage (Breyfogle, 1999). The tools employed are described below:

3.3.2.1 Statistical Process Control (SPC) for assessing measurement system stability: SPC charts
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

help ensure that the measurements taken for the process are stable and consistent over time
(Doty, 1996). The team used p-chart to assess the stability of ‘damping force testing machine’
(Figure 5).

Figure 5: Damping force testing machine

The team monitored the stability of the testing machine over time by plotting the p-chart (Doty,
1996). In order to plot the p-chart, the team selected 23 shock absorber subgroups of variable

14
size and recorded the number of defective shock absorbers in each subgroup after testing (Table
2 about here).

Using the observations (Table 2), the team plotted the p-chart (Figure 6) and computed the
control limits. It was found that proportion defectives ( ) in DF testing was 11% which
indicated that 89% defective units were correctly tested. So, it was concluded that the
measurement system was stable and acceptable for further data collection.

Figure 6: p-chart for DF testing


Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

P-Chart for DF testing


0.25

0.20

UCL=0.1527
Proportion

0.15

_
P=0.1124
0.10

LCL=0.0721

0.05
1
1

0.00
1 3 5 7 9 11 13 15 17 19 21 23
Sample
Tests performed with unequal sample sizes

3.3.2.2 Baseline performance & Capability analysis of CTQ: The Minitab output of process
capability analysis indicated a Total observed performance PPM (parts per million) of 101,986
(or 2.85 sigma) and Cpk (process capability index) of 0.46. Based on these observations, the
team concluded that the process of damping force (DF) generation was incapable.

15
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

Figure 7: Capability plot for DF generation process

16
3.3.2 Analysis

The aim of the analyze phase in a Six Sigma project is to identify the potential causes for the
process problem being studied and then select the root causes with the help of data and their
analysis. Six Sigma literature outlined many tools & techniques that are used at each stage for
data driven improvement (Pande et al, 2000). However when it comes to the ‘analyse’ phase,
subjective approach based on brainstorming and cause-and-effect diagrams is employed to form
casual hypothesis of relationships between factors involved in this problem (Aboelmaged, 2010,
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

Senapti, 2004, De Mast et al, 2000). Designing complex experiments in such situation may be
cumbersome due to large number of factors involved. This limitation is even more critical when
the cost of sampling is high or a low volume of product is available to test (De Mast et al, 2000).

SS offers clue generation techniques which involves offline experimentation using a small
sample of good and bad parts (called ‘best-of-best’ and ‘worst-of-worst’ parts) to eliminate
variables in a process which do not have an effect on the overall variation without disrupting
process settings (Steiner et al, 2008; De Mast, 2004). Shainin Component Search technique
(CST) is a clue generation tool, which is used when two or more assemblies are available that
can be disassembled and reassembled without effecting the CTQ (Steiner et al, 2009; Bhote,
2000; Shainin & Shainin, 1988).

Details of the tools employed at this phase are given below:

3.3.2.1 Brainstorming and Cause & Effect analysis: To start with, a brainstorming session was
planned and conducted by the team with the involvement of engineers in the company and a list
of potential causes for variation in CTQ was generated. A cause-and-effect diagram was drawn
based on these causes, which is presented in Figure 8. However, the plethora of possible
contributing educated guesses demanded heavily on time and hence, cost.

17
Figure 8: Cause and effect diagram
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

3.3.2.2 Shainin CST: In order to avoid making educated guesses for factors and estimating their
levels for DOE, the team wanted a simple experimentation approach to narrow down the factors
to a manageable number. Thus, it was decided to employ Shainin CST for the first stage filter of
factors responsible for DF defect.

The Shainin CST, as described in Shainin & Shainin (1988) and Bhote (2000), consists of four
stages. Stage one, which is known as the ballpark stage, tries to determine if the root causes of
defects are among the components being considered in the assembly or the assembly process.
Stage two, which is called the elimination stage, seeks to eliminate the unimportant components
from the search. Stage three, termed as ‘the capping run’, ensures and verifies whether the
important components have been identified or not. Finally, in the stage four, the factorial
analysis is done to identify the optimum levels of important components and check, if there are
any interaction effects between them. The team used the first three stages of Shainin CST
procedure in order to determine a list of important factors for further experimentation.

Stage 1: Ball Park stage

18
The team selected two sample assembly units from production – a Best of the Best (BOB) unit
and a Worst of Worst (WOW) unit, based on the CTQ specifications i.e. DF (55kgf – 75kgf). An
initial DF value for both the units was recorded before experimentation.

Then, the sample assemblies were disassembled and reassembled twice and their DF was
measured after each reassembly. After this initial experimentation, a discriminate ratio, D/d
(where D is the difference of median of BOB & WOW; d is the average range of BOB & WOW)
was calculated (Table 3). A value of D/d < 1.25 would mean that problem is with the assembly
process; else the team must drill down to establish that the problem is with components (Bhote &
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

Bhote, 2000; Antony, 1999).

Since the D/d ratio > 1.25, the team concluded that assembly was not contributing to the DF
issue.

Using the data from the Ball Park stage, the decision limits were calculated for the elimination
stage. For the components search involving two assemblies, this can be calculated using the
following formula (which uses the median of the disassembly/reassembly results and value of d)
(Bhote & Bhote, 2000; Antony, 1999).


 
    =  ± 2.776 ×
1.81

Decision limits for BOB: Max.= 20.815; Min = 20.585


Decision limits for WOW: Max.= 45.115; Min = 44.886

Stage 2& 3: Elimination & Capping run

This stage involved ranking of assembly components in descending order of significance through
engineering judgment. Although ranking components in this way brings about an element of
subjectivity, but it can only affect the time taken to reach the root cause and not the root cause
itself (Bhote & Bhote, 2000; Antony, 1999).

19
The team ranked the components of shock absorber assembly, based on their knowledge and
experience, for experimentation (Table 4 about here).

After the ranking is done, the top-ranked component from the BOB unit was swapped with the
WOW unit and vice versa and the CTQ was recorded. The decision for swapping of the most
critical component between BOB and WOW was based on the decision limits, which
weredetermined in the previous stage (Bhote & Bhote, 2000; Antony, 1999).
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

There are three possible outcomes of first swap and subsequent running of the test:
• No Change in reading observed: If no change is observed in the CTQ, the tested
component is swapped back to its original assembly and test moves on to the next
component in the ranked list.

• Small change is observed: If a small change is observed in the CTQ in one or


more assemblies, it infers that the component swapped is important but not the
only component responsible for the variation seen. Therefore, the tested
component is swapped back and experiment moves on to the next component.

• Complete reversal is observed: If there is a complete reversal measured in the


CTQ – that is, a BOB assembly becomes a WOW assembly and the WOW
assembly becomes a BOB assembly, it becomes obvious that the component
under testing is responsible for the variation seen. The search is stopped there and
then.

In the case under study, the complete reversal of DF was observed in the very first trial, when the
PV assembly (top ranked component) was swapped between BOB and WOW assembly units
(Table 5). So, the team concluded that PV assembly was causing the problem.

Further, to confirm that all significant components have been identified, a capping run is
performed (Bhote & Bhote, 2000; Antony, 1999). In this experimental run, all significant

20
components are swapped between the BOB and WOW assembly. This is done to check if the
respective CTQs get completely swapped.

In the case under study, the capping run re-affirmed that PV assembly was the significant
component causing DF problem (Table 5 about here).

3.3.2.3 Shainin CST for PV assembly: For further drilling down to the root cause, the team
continued the Shainin CST procedure with the PV assembly.
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

Thus, the team selected a BOB and WOW assembly unit based on the response. The D/d ratio
was found greater than 1.25, indicating that assembly was not the problem. The assembly
components were ranked (Table 6) and swapped one at a time between BOB & WOW. It was
found that the complete reversal in the response took place when piston was swapped (Table 7).

Based on the observations (Table 7), the team concluded that Piston (with rebound stopper) was
causing the main problem. Thus, using Shainin CST, the team was able to filter the factors for
further investigation.

3.3.3 Improve

During the improve phase of the project, solutions for the selected root causes are to be identified
and implemented to observe the results.The team planned and conducted Taguchi DOE
experiments for identifying optimum setting of parameters for ‘piston and its stopper’ which
were identified as the key contributor of DF defect by means of Shainin CST (Figure 9).

21
Figure 9: Piston and Rebound Stopper

Piston Rebound Stopper


Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

Taguchi DOE approach is useful to determine a feasible combination of factors, which when
controlled can help reduce the variability in product responses (Antony & Kaye, 1995;
Montgomery, 1992; Phadke, 1989; Taguchi, 1987). It allows extracting maximum information
with minimum number of experiments using standard Orthogonal Arrays (OA) (Wu & Hamada,
2011).

Based on a detailed brainstorming session involving the engineers in the company, the team
selected five piston parameters for experimentation. During the brainstorming session, the team
felt that the interaction of these piston parameters was not significant for DF generation. So, the
interactions were not considered for further study.The response of the experiment was decided as
‘Rebound DF’ measured on the shock absorber with the dynamometer. The control factors along
with the levels are listed in Table 8 (Table 8 about here).

3.3.3.1 Experimental Design Layout: For conducting a full factorial experiment with six factors,
each at two levels, it required conducting a total of 64 (26) experiments. In view of the fact that
conducting 64 experiments will be very costly and time consuming, the team decided to adopt

22
orthogonal arrays (OA) for conducting the fractional factorial experiments (Wu & Hamada,
2011).

Estimation of main effects of 6 factors each at two levels requires 6 degrees of freedom (1df per
factor) and requires minimum of 7 experimental runs. So, the team decided to use standard L8
OA for designing this experiment (Wu and Hamada, 2011).The design layout (Table 8) for the
experiment was prepared by allocating the factors and level to the L8 (26) OA. The experimental
sequence given in the design layout was randomized and response was recorded after
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

experimentation. A sample of two tests was conducted at each experimental point. Table 9 shows
the OA design matrix and response values for the experiment.

3.3.3.2 Statistical Analysis: The mean DF response obtained using L8 OA was analyzed through
Analysis of Variance (ANOVA) (Table 10). ANOVA is a suitable technique to determine and
interpret the variation in average performance of tested groups (Fisher, 1960). At this point,
further analysis into process improvement could be carried out through Signal to Noise ratio
calculation and pooled ANOVA, with the aim of reducing variation in the response
(Montgomery & Runger, 2007). However, the team preferred not to delay the implementation of
solutions by getting further into deeper analysis which might threaten the involvement of
stakeholders of the project.

From the ANOVA table, it was found that piston parameters ‘piston orifice diameter (A)’,
‘rebound stopper hardness(C)’ and ‘piston valve depth (D)’ were significantly contributing to
damping force at 10 percent level of significance. Further the percentage contribution values
affirmed ‘stopper hardness’ (48.27%) as the most significant contributor followed by ‘piston
valve depth’ (20.42%) and ‘piston orifice diameter’ (9.38%).

The team selected the setting of significant piston parameter by analyzing the mean response
table with the objective of optimizing the rebound DF. They kept the technical and economic
feasibility into consideration while setting the specifications. The predicted DF was optimal at
factors levels:

23
Piston orifice diameter: 11.12 to 1.08 mm
Rebound stopper hardness: 90 to 95 HBR
Piston valve depth: 5.3 to 5.5 mm

3.3.3.3 Validation of solutions: Based on the optimum settings obtained from experimentation
results, a confirmatory run was carried out using a sample of 30 assemblies and the process
capability was observed (Figure 10).
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

24
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

Figure 10: Process capability inconfirmatory run

25
3.3.3.4 Improvement actions: Once the confirmatory runs showed a positive response, the team
implemented the following improvement actions:

Action 1: The tolerance of piston orifice diameter was reduced from 50 microns to 20 microns.

Action 2: The hardness of RD Stopper increased from 75-80 HRB to 90~95 HRB.

Action 3: Apart from these design modifications, the team also established preventive measures
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

for preventing contamination of internal components of shock absorber assembly.


• Dip cleaning replaced with Ultrasonic washing machine for effective cleaning of
assembly components (Figure 11).

Figure 11: Ultrasonic washing machine

Before After

• Washing media contamination in the cleaning machine (Ultrasonic washing machine)


was monitored on daily basis and the media was restored every month to prevent
contamination of assembly components)

• In order to prevent impurities on the outer tube, a six stage cleaning & ph checking
mechanism was installed (Figure 12).

26
Figure 12: Cleaning & Ph monitoring machine

Before After
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

• The piston packaging lot size of reduced from 1000 to 500 in order to control
contamination.

Action 5: 100% inspection of piston rod riveting diameter using snap gauge (previously checked
with Vernier calipers during process audits) was implemented (Figure 13).

Figure 13: Riveting diameter inspection


Before After

Action 5: Daily recording of process parameters was implemented (Figure 14).

27
Figure 14: Daily report
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

3.3.3.5 Gains: The implementation of remedial actions reduced the warranty claims due to
damping force from 3300ppm (defective rate 10% in Nov. 2013) to 2700ppm (defective rate 2%
in Feb. 2013), resulting in a cost of external failure saving of $5,000 (Figure 15). Additionally,
these corrective actions may well be horizontally deployed across other models ensuing large
savings.

Figure 15: Improvement

28
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

3.3.4 Control

The primary aim of this phase is to ensure that the gains obtained during the improve phase are
maintained long after the project has ended. To that end, it is necessary to standardize and
document procedures, to make sure that all employees are trained, and in the end, communicate
the project’s results.

In addition, the project team must create a plan for regular monitoring of the process (Gijo &
Rao, 2005). In order to retain the gains achieved, the team prepared a control plan standardizing
the improvements that were rolled out along with details of roles, tasks, frequency and methods
(Table 11).Training about the improved methods,was provided to the operators working with the
process to enhance their confidence levels in working with the new process.

29
4. Conclusion and future research directions

The goal of the study was to demonstrate the application of Six Sigma/DOE Hybrid framework
for improving damping force generation process in a shock absorber assembly unit. As a result of
this adoption, the field rejection of shock absorber assembly of Model X due to inadequate
damping force reduced from 10% to 3%. This recorded a financial saving of $5,000 per year due
to the reduction in rework cost. These results motivated the management to deploy the
improvement actions horizontally across other models as well, ensuing large savings.
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

The study demonstrated that the integration of Shainin & Taguchi DOE tools reinforced the
established Six Sigma methodology in two ways; firstly, by narrowing down the potential
sources of variation and secondly by optimizing the response. The application of Shainin clue
generation techniques at the analysis phase of DMAIC improvement cycle allowed a more
efficient root cause analysis of the variation. In the proposed framework, the analysis phase
started with reducing the numbers of factors under consideration by eliminating unimportant
factors objectively through Shanin CST. The application of Shanin CST facilitated in
ascertaining that assembly component (piston with rebound stopper) and not the assembly
process was causing the variation. Thus, the case illustrated that Shainin clue generation
technique worked well in accomplishing first stage filter of factors using small samples (two
assembly units, one good & one bad in this case) from ongoing production through offline
experiments without interrupting production (Steiner, et al, 2008). Further, the use of Taguchi
OA at the improve phase, allowed the collection of necessary data to determine the significant
piston parameters with minimum experimentation (eight experimental runs in this case as
opposed to the expected 64) and ANOVA on the collected data facilitated the selection of
parameter settings to optimize the CTQ i.e. rebound damping force (Taguchi et al, 2005).

The successful implementation of this framework led to the conclusion that rather than treating
Taguchi and Shainin as competing DOE approaches, aspects of these approaches can, in fact, be
blended with Six Sigma to produce a more powerful, efficient and systematic improvement
approach (termed as Six Sigma/DOE Hybrid in the case) (De Mast et al, 2000).

30
The proposed framework employed Shainin CST for identification of potential sources of
variation. The scope of application and improvement can be extended by integration of other
Shainin tools at various phases of DMAIC cycle. For instance, Shainin Isoplot facilitates in
validating the measurement system (Shainin, 1992), Shainin B & C allows to confirm whether
the actions taken have actually improved the process (Steiner et al, 2009), and Shainin pre-
control signals the need of process adjustment at the control phase (Traver, 1985; Steiner, S. H.
1997-1998).
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

References

• Aboelmaged, M. G. 2010. “Six Sigma Quality: A Structured Review and Implications for
Future Research.” International Journal of Quality & Reliability Management 27(3):
268-317.

31
• Allen, T. T. 2003. Design of Experiments and Other Six Sigma Related Methods
Explained and Derived. Florence, KY: Thomson Learning.
• Andrew, T., and Richard, B. 2006. “Developing an SME based six sigma strategy.”
Journal of Manufacturing Technology Management 17(4): 417-433.
• Antony, J., Banuelas, R., and Kumar, A. 2006. World Class Applications of Six Sigma.
Oxford: Butterworth-Heinemann Ltd.
• Anthony, J., and Cheng, A. H. Y. 2003. “Training for Shainin’s Approach to
Experimental Design Using a Catapult.” Journal of European Industrial Training 7(8):
405-412.
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

• Antony, J. 1999. “Spotting the Key Variables using Shainin’s Variables Search Design,”
Logistics Information Management 12(4): 325-331.
• Antony, J., and Kaye, M. 1995. “Experimental quality.” Journal of Manufacturing
Engineer 74(4): 178-81.
• Arnheiter, E. D., and Maleyeff, J. 2005. "The integration of lean management and Six
Sigma", The TQM Magazine 17(1): 5 - 18
• Asif, M., Bruijn, E. J., Douglas, A. and Fisscher, O. A. M. 2009. “Why quality
management programs fail: a strategic and operations management
perspective.” International Journal of Quality and Reliability Management 26(8):778-94.
• Banawi, A., and Bilec, M. M. 2014. “A framework to improve construction processes:
Integrating Lean, Green and Six Sigma.” International Journal of Construction
Management 14(1): 45-55.
• Bennett, J. O., Briggs, W. L., and Triola, M. F. 2003. Statistical Reasoning for
EverydayLife 2nd ed. Boston: Addison-Wesley.
• Bhote, K. R. and Bhote, A. K. 2000. World Class Quality: Using Design of Experiments
to Make It Happen. 2nd ed. AMACOM.
• Bhote, K. R. 1991. World-Class Quality: Using design of experiments to make it happen.
American Management Association (AMACOM): New York.
• Bhote, K. R. 1990. “A More Cost-Effective Approach to DOE than Taguchi.” Annual
Quality Congress San Francisco. 44: 857-862.

32
• Bilgen, B., and Şen, M. 2012. “Project selection through fuzzy analytic hierarchy process
and a case study on Six Sigma implementation in an automotive industry.” Production
Planning& Control: The Management of Operations 23(1): 2-25,
• Box, G. E. P., and K. B. Wilson 1951. “On the Experimental Attainment of Optimum
Conditions.” Journal of the Royal Statistical Society Series B (13): p. 1-45.
• Breyfogle, F. W., 1999. Implementing Six Sigma: smarter solutions using statistical
methods. New York, NY: John Wiley.
• Chen, M., and Lyu, J. J. 2009. “A Lean Six-Sigma approach to touch panel quality
improvement.” Production Planning & Control: The Management of Operations, 20(5):
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

445-454.
• Coronado, R. B., and Antony, J. 2002. “Critical success factors for the successful
implementation of Six Sigma projects in organizations.” Total Quality Management
Magazine 14(2): 92-9.
• De Mast, J. 2004. “A methodological comparison of three strategies for quality
improvement.” International Journal of Quality & Reliability Management 21(2): 198-
213.
• De Mast, J., Schippers, W., Does, R., and Van den Heuvel, E. 2000. “Steps and strategies
in process improvement”, Quality and Reliability Engineering International, Vol. 16 No.
4, pp. 301-11.
• Dey, P. (2002), “How to complement ISO 9001:2000 with six sigma”, iSixSigma LLC,
available at: http://isixsigma.com/ library/content/c020211a.asp
• Doty, L. A. (1996). Statistical Process Control, 2nd ed. NY: Industrial Press Inc.
• Eckes, G. 2001. The Six Sigma Revolution, How General Electric and Others Turned
Process Into Profits. USA: John Wiley & Sons Inc.
• Eisenhardt, K. M. 1989. “Building theories from case study research.” The Academy of
Management Review 14(4): 532-50.
• Fisher, R. A. 1935. The Design of Experiments. New York: John Wiley.
• Fisher, R. A. 1958. Statistical Methods for Research Workers, 13th ed. Edinburgh,
Scotland: Oliver & Boyd.
• Fisher, R. A. 1966. The Design of Experiments 8th ed. New York: Hafner.

33
• George, M. L., Rowlands, D., Price, M., and Maxey, J. 2005. The Lean Six Sigma Pocket
Tool book: A Quick Reference Guide to 70 Tools for Improving Quality and Speed.
McGraw-Hill Professional, p. 225.
• Ghosh, S., and Maiti, J. 2014. “Data mining driven DMAIC framework for improving
foundry quality – a case study.” Production Planning & Control: The Management of
Operations 25(6): 478-493, DOI: 10.1080/09537287.2012.709642
• Green, F. B. 2006. “Six-Sigma and the Revival of TQM.” Total Quality Management &
Business Excellence 17(10): 1281-1286.
• Gijo, E. V., Bhat, S., and Jnanesh, N. A. 2014. “Application of Six Sigma methodology
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

in a small-scale foundry industry.” International Journal of Lean Six Sigma 5(2): 193-
211.
• Gijo, E. V., and Scaria, J. 2010. “Reducing rejection and rework by application of Six
Sigma methodology in manufacturing process.” International Journal Six Sigma and
Competitive Advantage 6(1/2): 77–90.
• Gijo, E. V., and Rao, T.S., 2005. “Six Sigma implementation – hurdles and more
hurdles.” Total Quality Management and Business Excellence, 16 (6), 721–725.
• Gijo. E. V., 2005. “Improving process capability of manufacturing process by application
of statistical techniques.” Quality Engineering 17, 309–315.
• Goodman, J., and Wyld, D. C. 2001. “The Hunt for Red X: A Case Study in the Use of
Shainin Design of Experiments (DOE) in an Industrial Honing Operation.” Management
Research News 4(8/9): 1-17.
• Goh, T. N. 2002. “The Role of Statistical Design of Experiments in Six Sigma:
Perspectives of a Practitioner.” Quality Engineering 14(4): 659-671. DOI: 10.1081/QEN-
120003565
• Goh, T. N. 2001. “A pragmatic approach to experimental design in industry.” Journal of
Applied Statistics 28(3&4): 391–398.
• Hahn, G. J., Doganaksoy, N., and Hoerl, R. W. 2000. “The evolution of Six Sigma.”
Quality Engineering. 12(3): 317–326.
• Heavey, C., and Murphy, E. 2012. "Integrating the Balanced Scorecard with Six Sigma."
The TQM Journal 24(2): 108 – 122.

34
• Hockman, K. 1994. “Review of World Class Quality by K. Bhote.” Technometrics 36:
425-426.
• Johnson. R. T., Hutto, G. T., Simpson, J. R. & Montgomery, D. C. 2012. “Designed
Experiments for the Defense Community.” Quality Engineering 24(1): 60-79.
• Johnson, J. A., Widener, S., Gitlow, H., and Popovich, E. 2006a. “A Six Sigma Case
Study: G.E.P. Box's Paper Helicopter Experiment—Part B.” Quality Engineering
18(4):413-430.
• Johnson, J. A., Widener, S., Gitlow, H., and Popovich, E. 2006b. “A Six Sigma Case
Study: G.E.P. Box's Paper Helicopter Experiment—Part B.” Quality Engineering 18(4):
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

431-442.
• Kackar, R. N. 1985. “Off-line quality control, parameter design, and the Taguchi
method.” Journal of Quality Technology 17:176–188.
• Krueger, D. C., Parast, M. M., and Adams, S. 2014. Six Sigma implementation: a
qualitative case study using grounded theory. Production Planning & Control: The
Management of Operations 25:10, 873-889.
• Kumar, S., McCreary, M. L., and Nottestad, D. A. 2011. “Quantifying Supply Chain
Trade-Offs Using Six Sigma, Simulation, and Designed Experiments to Develop a
Flexible Distribution Network.” Quality Engineering 23(2): 180-203.
• Kumar, M., Antony, J., and Cho, B. R., 2009. “Project selection and its impact on the
successful deployment of Six Sigma.” Business Process Management Journal 15 (5):
669–686.
• Kumar, U.D., Saranga, H., Ramírez-Márquez, J. E., and Nowicki, D. 2007. "Six sigma
project selection using data envelopment analysis." The TQM Magazine 19(5): 419 – 441.
• Ledolter, J., and Swersey, A. 1997a. “An Evaluation of Pre-control.” Journal of Quality
Technology 29: 163-171.
• Ledolter, J. and Swersey, A. 1997b. “Dorian Shainin’s Variable Search Procedure: A
Critical Assessment.” Journal of Quality Technology 29: 237-247.
• Logothetis, N. 1990. “A Perspective on Shainin’s Approach to Experimental Design for
Quality Improvement.” 6(3): 195-202.

35
• Lee, Kuo-Liang, Wei, Chun-Chin, and Lee, Hsu-Hua, 2009. “Reducing exposed copper
on annular rings in a PCB factory through implementation of a Six Sigma project.” Total
Quality Management & Business Excellence 20(8): 863-876.
• Lee, R., and Ken, B. 2003. “Integrating Six Sigma with total Quality Management: A
Case Example for Measuring Medication Errors.” Journal of Healthcare Management
48(6): 377-91.
• Leitnaker, M. G., and Cooper, A. 2005. “Using Statistical Thinking and Designed
Experiments to Understand Process Operation.” Quality Engineering 17(2): 279-289
• Linderman, K., Schroeder, R.G., Zaheer, S. and Choo, A.S., 2003. “Six Sigma: a goal-
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

theoretic perspective.” Journal of Operations Management 21: 193–203.


• Lupan, R., Bacivarof, L. C., and Kobi, A., and Robledo, C. 2005. “A Relationship
between Six Sigma and ISO 9000:2000.” Quality Engineering 17(4): 719-725.
• Mackertich, N. A. 1990. “Pre-control vs. Control Charting: A Critical Comparison.”
Quality
Engineering 2: 253-260.
• Montgomery, D. C. 1992. “The use of statistical process control and design of
experiments in product and process improvement.” IIE Transactions 24(5): 4-17.
Phadke, M. S. 1989. Quality Engineering using Robust Design. Englewood Cliffs, NJ:
Prentice-Hall.
• Montgomery, D. C., Woodall, W. H. 2008. “An overview of Six Sigma. International
Statistical Review.” 76(3): 329–346.
• Montgomery, D., and Runger, G. 2007. Applied statistics and probability for engineers
(4th ed.). Hoboken, Netherlands: Wiley & Sons
• Moore, D. 1993. “Review of World Class Quality by K Bhote.” Journal of Quality
Technology 21: 76-79.
• Munro, R.A. (2000), “Linking six sigma with QS 9000: Auto industry adds new tool for
quality improvement.” Quality Progress 5: 47-53.
• Nelson, L. S. 1991. “Review of World Class Quality by K Bhote.” Journal of Quality
Technology. 25: 152-153.
• Pande, P. S., Neuman, R. P., and Cavenagh, R. R. 2000. The Six Sigma Way: How GE,
Motorola, and Other Top Companies are Honing Their Performance. McGraw-Hill

36
Professional, 2000, p. 448.
• Park, S. H., Ntuen, C. A., and Park, E. H. 2009. “A new paradigm of Six Sigma:
Knowledge-based Digital Six Sigma.” Total Quality Management & Business Excellence
20(9): 945-952.
• Pepper, M. P. J. and Spedding, T. A. 2010. “The evolution of lean Six Sigma.”
International Journal of Quality & Reliability Management 27(2): 138-155.
• Pfeifer, T., Reissiger, W., and Canales, C. 2004."Integrating six sigma with quality
management systems.” The TQM Magazine 16(4): 241 – 249.
• Salah, S., Rahim, A., and Carretero, J. A. 2010. “The integration of Six Sigma and lean
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

management." International Journal of Lean Six Sigma 1(3): .249 – 274.


• Saravanan, S., Mahadevan, M., Suratkar, P., and Gijo, E. V. 2012. “Efficiency
improvement on the multicrystalline silicon wafer through six sigma methodology.”
International Journal of Sustainable Energy 31(3): 143-153.
• Schmidt, S. R. and Lausnby, R. G. 2005. Understanding Industrial Designed
Experiments. 4th ed. Colorado, USA: Air Academy Press.
• Senapati, N. R. (2004). “Six Sigma: Myths and Realities.” International Journal of
Quality
and Reliability Management 21(6): 683.
• Sheridan, J.H. (2000), “Lean Sigma synergy”, Industry Week, Vol. 249 No. 17, pp. 81-2.
• Shainin, R. D. 1993b. “Strategies for Technical Problem Solving.” Quality Engineering
5: 433-438.
• Shainin P.D. 1992 “Managing SPC, A Critical Quality System Element,” 46th Annual
Quality Congress Proceedings, ASQC, pp. 251-257
• Shainin, D. and Shainin, P. 1988. “Better than Taguchi Orthogonal tables.” Quality and
Reliability Engineering International 4: 143-149.
• Singh, S; Maheswari, S., and Pandey, P. C. 2007. “Optimization of Multi performance
Characteristics in Electric Discharge Machining of Aluminium Matrix Composites
(AMCs) using Taguchi DOE Methodology.” International Journal of Manufacturing
Research 2(2): 138-161.
• Snee, R. D. 2001. “Dealing with the Achilles heel of Six Sigma initiatives.” Quality
Progress 34(3): 66-72.

37
• Stamatis, D. H. 2002. Six Sigma and Beyond: Design of Experiments. Boca Raton, FL:
CRC Press.
• Steiner, S. H., MacKay, R. J., and Ramberg, J. S. 2008. “An Overview of the Shainin
System for Quality Improvement.” Quality Engineering 20(1): 6-99.
• Steiner S.H. and MacKay, R. J. 2005. Statistical Engineering: An Algorithm for Reducing
Variation in Manufacturing Processes. Milwaukee: Quality Press.
• Steiner, S. H. 1997-1998. “Pre-control and Some Simple Alternatives.” Quality
Engineering 10: 65-74.
Taco, M., Viles, E. and Pozueta, M. 2008. “Are All Designs of Experiments Approaches
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)


Suitable for Your Company.” Proceedings of the World Congress on Engineering Vol II,
WCE 2008, July 2 - 4, 2008, London, U.K.
• Taguchi, G., Chowdhury, S., and Wu, Y. 2005. TAGUCHI’S Quality Engineering
Handbook. New Jersey: John Wiley & Sons.
• Taguchi, G. 1987. System of Experimental Design: Engineering Methods to Optimize
Quality and Minimize Cost. White Plains, NY: UniPub.
• Taguchi, G. (1991). Introduction to Quality Engineering. White Plains, NY: Asian
Productivity Organization.
• Taguchi, G., and Wu, Y. 1980. Introduction to Off-Line Quality Control. Nagoya: Central
Japan Quality Control Association.
• Tanco, M., Viles, E., llzarbe, L, and A’lvarez, M. J. 2008. “How is experimentation
carried out by companies? A survey of three European regions.” Quality and Reliability
Engineering International 24(8): 973–981.
• Tanco, M., Viles, E., llzarbe, L, and A’lvarez, M. J. 2009. “Implementation of Design of
Experiments projects in industry.” Applied Stochastic Models Business and Industry
25(4): 478–505.

38
• Tang, L. C., Goh T. N., Lam, S. W., and Zhang, C. W. 2007. “Fortification of Six Sigma:
expanding the DMAIC toolset.” Quality and Reliability Engineering International
23(1): 3–18.

• Thomas, A. J., and Antony, J. 2005. “A comparative analysis of the Taguchi and Shainin
DOE techniques in an aerospace environment.” International Journal of Productivity and
Performance Management 54(8): 658-678
• Thompson, A., and Barton, R. 2006. “Developing an SME based six sigma strategy.”
Journal of Manufacturing Technology Management. 17(4): 147-433
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

• Thomas, A., and Antony, J. 2004. “Applying Shainin’s variables search methodology in
aerospace applications.” Assembly Automation 24(2): 184–191.
• Tka´ c, M. and Lyo´ csa, S., 2009. “On the evaluation of Six Sigma projects. Quality and
Reliability Engineering International.” 26: 115–124.
• Traver, R. W. 1985. “Pre-control: A Good Alternative to X -R Charts.” Quality Progress
September: 11-14.
• Verma, A K; Srividya, A; Mannikar, A V; Pankhawala, V A and Rathanraj, K J (2004).
‘Shainin Method: Edge over other DOE techniques’, Engineering Management Conf.
2004 Proceedings, IEEE International, Vol. 3, No. 18-21, pp. 1110- 1113.
• Vinodh, S. Kumar, S. V., and Vimal, K. E. K. 2014. “Implementing lean sigma in an
Indian rotary switches manufacturing organization.” Production Planning & Control:
The Management of Operations 25(4): 288-302.
• Wilcox, R. R. 2003. Applying Contemporary Statistical Techniques. San Diego, CA:
Academic Press.
• Wu, C. F. J., and Hamada, M. 2011. Experiments-Planning, Analysis, and Parameter
Design Optimization. New York: John Wiley.
• Yin, R. K. 2009. Case Study Research: Design and Methods, 4th ed. United States of
America: SAGE Publications Inc.
• Zeigel, E. R. 2001. A Review of World Class Quality by K. Bhote and A. Bhote. 2nd
Edition Technometrics 43: 112-113.

39
• Zhang, M., Wang, W., Goh, T. N., and He, Z. 2015. “Comprehensive Six Sigma
application: a case study.” Production Planning & Control: The Management of
Operations 26(3): 219-234.
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

Tables

Table 1: DMAIC phases

Phases Purpose Tools used


To spot customer requirement (CTQ) and • Project charter
determine the process problem (CTP), • Project plan
Define
establish project scope, goals, and financial • Pareto analysis
& performance targets
• Pareto analysis
To Identify and classify factors, validate the • Brainstorming
Measure
measurement system • Cause & Effect analysis
• SPC(p chart)
• Shainin Component
Search technique(CST)
for first stage filter of
factors
To determine active primary factors, • Taguchi Orthogonal
Analysis
interpret results and determine solutions Arrays(OA) using factors
identified from Shainin
CST
• Analysis of Variance
(ANOVA) to study

40
statistical significance of
factors.
• Confirmatory runs for
verification of results

Improve To validate solutions • Process Capability Studies


To implement controls and evaluate • Control and
Control
experimentation iteration standardization
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

Table 2: Observations of non-conforming shock absorbers


Date Defectives Subgroup size
18.11.13 26 276
19.11.13 14 171
19.11.13 18 190
20.11.13 31 367
20.11.13 28 304
21.11.13 35 304
21.11.13 18 190
22.11.13 9 285
24.11.13 20 209
24.11.13 31 209
25.11.13 36 400
25.11.13 22 197
26.11.13 23 171
26.11.13 8 76
27.11.13 72 437
27.11.13 51 320
28.11.13 45 272

41
28.11.13 21 154
29.11.13 40 247
29.11.13 51 320
30.11.13 16 171
30.11.13 55 553
Proportion defective, =0.1124
Standard Deviation, σ = 0.0426
Upper Control Limit, UCL = + 3 σ = 0.1527
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

Lower Control Limit, LCL = - 3 σ = 0.0721

Table 3: Observations for Ball Park stage


BOB(DF in kgf) WOW(DF in kgf)

Initial assembly 65.200 45.000

After 1st re-assembly 65.700 44.800

After 2nd re-assembly 66.200 45.300

Median 65.700 45.000

Range 1.000 0.500

D= difference in median= 65.700-45.000=20.700

d= average range=(1+0.5)/2 = 0.075

D/d test= 20.700/0.075= 27.60 > 1.25; assembly is NOT the dominant cause.

Table 4: Ranking of Shock absorber assembly components


Rank Component
1 Piston valve(PV) assembly
2 Compression valve assembly
3 Cylinder
4 Rod guide

42
Table 5: Observations for Elimination & Capping run stage

BOB (DF in kgf) WOW(DF in kgf)

65 60 55 50 45 40

After 1st re-assembly 65.700 44.800

After 2nd re-assembly 66.200 45.300

PV assembly swapped 45.200 64.300


Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

Complete reversal on first swap↑

Capping run 45.600 Complete reversal on capping run 63.450

Table 6: Ranking of PV assembly components

Rank Component Figure

1 Moon Valve

2 Piston (with rebound stopper)

3 Deflection valve discs

4 Back up disc

5 Piston rod nut

Table7: Shainin CST on PV assembly

43
BOB(DF in kgf) WOW(DF in kgf)

65 60 55 50 45 40

After 1st re-assembly 65.700 44.800

After 2nd re-assembly 66.200 45.300

Moon Valve swapped 65.900 45.400

Bring back to original 66.000 45.200


Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

Piston (with stopper) swapped 45.400 65.300

Complete reversal on second


swap ↑

Bring back to original 64.200 45.300

Complete reversal on
Capping run 45.600 64.450
capping run

Table 8: Factors & levels used in the experiment

Factors Units Level 1 Level 2


A Piston orifice diameter mm 1.08 – 1.12 1.05 – 1.15
B Piston disc depth mm 0.9 1.1
HRB(Rockwell
C Rebound stopper hardness 90 – 95 75 – 80
Hardness)

D Piston valve depth mm Existing production 5.3 – 5.5

E Piston groove diameter mm 18.95 19.05


F Priming cycles Number 3 6

Table 9: OA design matrix & response values

44
Factors Response(Rebound DF)
Experimental
A B C D E G R1 R2 Mean
run
1 1 1 1 1 1 1 71.0 77.7 74.35
2 1 1 1 2 2 2 71.8 72.2 72.00
3 1 2 2 1 2 2 69.1 67.4 68.25
4 1 2 2 2 1 1 66.4 64.2 65.30
5 2 1 2 2 1 2 63.2 63.8 63.50
6 2 1 2 1 2 1 68.5 68.7 68.60
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

7 2 2 1 2 2 1 67.3 69 68.15
8 2 2 1 1 1 2 69.9 72.2 71.05

Table 10: ANOVA for mean response of OA

Fcritical (10%
Percentage contribution
Source DF SS MS F significance
[SS/SS(Total)*100]
level)

A 1 19.58 19.58 5.05 3.46* =19.58/208.47 = 9.38%

B 1 8.85 8.85 2.28 3.46

C 1 101.50 101.50 26.19 3.46* =101.50/208.47=48.72%

D 1 42.57 42.57 10.99 3.46* =42.57/208.47=20.42%

E 1 2.32 2.32 1.67 59.4

F 1 0.45 0.45 8.51 59.4

Error 8 31.00

Total 15 208.47
*Factor significant at 10% significance level.

Table 11: Control plan


S.No What When Who Where How
45
1 Revalidation of Bi-monthly Quality QA standard Inspection
piston Assurance(QA) room check sheet
specifications Engineer
2 Revalidation of Bi-monthly QA Engineer QA standard Inspection
Rebound stopper room check sheet
3 Process Daily Production Shock Checksheet
parameter Supervisor absorber (DIR01)
monitoring Assembly
4 Contamination Fortnightly QA Engineer QA standard Checksheet
checking room (DIR01) &
standard
5 Reviting tool Monthly Production Shock Checksheet
inspection for Supervisor absorber
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

wear & tear Assembly


6 Contamination Daily QA Engineer QA standard Checksheet
checking(internal room (DIR01) &
components) standard

Annexure I: Project charter

Project Charter Worksheet Date:04/11/13

Project Title: Integration of Taguchi & Shainin DOE for Six Sigma process improvement: An Indian
case.

Problem Statement: High damping force warranty claims (defective rate of 10% in Nov. 2013) ensuing
a cost of external failures of $5,000 per annum.

Mission Statement: “Reduction in the field rejection rate of shock absorber assembly from 10% to
3% by the end of January, 2014.

Process Boundaries

Start point:

46
Stop Point:

Impacted Functions: product and process design, engineering and production, quality management,
marketing, sales and distribution

Mentor Name: Team Leader:

Team Members Functional Areas of Team Members

Project Milestones
Downloaded by Chinese University of Hong Kong At 00:15 07 July 2017 (PT)

Target Completion
Name of Phase Actual Completion Date
Date

Define Phase 4.11.13 4.11.13

Measure& Analysis Phase 20.12.13 22.12.13

Improve Phase 10.1.14 12.1.14

Control Phase 28.2.14 28.2.14

47

Das könnte Ihnen auch gefallen