You are on page 1of 150

The Pennsylvania State University

The Graduate School

Department of Industrial and Manufacturing Engineering

A SUSTAINABILE MODULAR PRODUCT DESIGN APPROACH WITH

KEY COMPONENTS AND UNCERTAIN END-OF-LIFE OPTIONS CON-

SIDERATION

A Dissertation in

Industrial Engineering and Operations Research

by

Junfeng Ma

© 2016 Junfeng Ma

Submitted in Partial Fulfillment


of the Requirements
for the Degree of

Doctor of Philosophy

May 2016
The dissertation of Junfeng Ma was reviewed and approved* by the following:

Gül Okudan Kremer


Professor of Industrial and Manufacturing Engineering
Dissertation Advisor
Chair of Committee

M. Jeya Chandra
Professor of Industrial and Manufacturing Engineering

Charles David Ray


Associate Professor of Ecosystem Science and Management

Conrad S. Tucker
Assistant Professor of Industrial and Manufacturing Engineering

Janis Terpenny
Peter and Angela Dal Pezzo Department Head of Industrial and Manufacturing
Engineering

*Signatures are on file in the Graduate School.

ii
ABSTRACT

An expanding awareness of manufacturing’s short-term and long-term environmental im-

pacts continues to drive manufacturers to reconsider the product design process with a view toward

sustainability. An emphasis on sustainability is imperative for the competitive success of modern

design and manufacturing enterprises, and incorporating sustainability into the product design pro-

cess continues to warrant the increasing attention. Sustainability includes three dimensions: eco-

nomic, environmental, and social. To the best of our knowledge, no existing product design

method yet involves these three dimensions of sustainability, partly because social sustainability

is difficult to evaluate and measure. This research gap motivates this study.

Product life cycle management, especially the end-of-life (EOL) stage management, with its

benefits of cost savings and increased environmental awareness performance, has emerged as a

critical component of the product design process. However, determining EOL options during the

product design stage can present a complex challenge because of vagueness in the design environ-

ment—for example, an imprecise factor such as product maintenance. Therefore, developing a

comprehensive framework to help understand uncertain EOL options is also a foundational moti-

vation of this study.

In today’s challenging marketplace, key components play a significantly important role in

product competitiveness. These components may represent a core technology and affect the prod-

uct’s functional performance, such as the turbocharger in an automobile engine; or they may be

expensive or complicated to assemble and consequently influence the product’s market perfor-

mance, such as the cabinet of a refrigerator. However, based on existing literature, scant research

iii
discusses key components in the product design process. Therefore, involving a focus on key com-

ponents in the design process also stimulates this study.

Modular product design (MPD)—splitting a product into several components and then an-

alyzing each separately—has proven to be a widely accepted design tool due to its advantages

across the entire product life cycle, evincing benefits to manufacturing, to assembly, and to the

product supply chain itself. In this study, a new MPD approach is presented that considers uncer-

tain EOL options and key component specifications to improve three dimensions of sustainability.

Three indicators—cost, environmental impact, and labor time—are employed to measure

economic, environmental and social sustainability as individual elements. The environmental im-

pact assessment is adopted using an open source software, Eco-99, which measures environmental

sustainability in a comprehensive manner. Fuzzy logic is used to estimate the probability of each

EOL option and to determine the expected EOL sustainability values (cost, environmental impact

and labor time) for each component. Three clustering algorithms, with respect to the three sustain-

ability aspects, are developed for specific key components; component modules are then grouped

based on the identified key components and their sustainability optimization at all three stages of

the total life cycle (manufacturing, maintenance, and EOL). The tri-criteria model builds toward

optimizing sustainability in order to derive the compromised module structure.

Based on this investigation into three-dimension sustainability, the economic, environmen-

tal, and social aspects are first quantified and measured simultaneously from the view of the prod-

uct life cycle, and then improved through a product module generation process. Key components

are also emphasized in the design stage, and key-component-based clustering algorithms are in-

troduced to generate product modules. Moreover, uncertain EOL options are taken into account in

iv
the module forming process. This new MPD approach has proven to generate better performance

than previously existing methods (e.g., decomposition approach) in terms of sustainability.

Two case studies are used to demonstrate the implementation of the proposed methodology,

based on the perspective of product scale. A coffee maker is used to illustrate the analysis of a

small and simple product, and a refrigerator is used to illustrate the analysis of a large and complex

product. Given pre-determined sustainability indicator weights, both the coffee maker and the re-

frigerator are decomposed into several modules, as such modular architectures can benefit product

life cycle full-dimension sustainability performance assessment.

v
Table of Contents

List of Figures............................................................................................................................................. ix
List of Tables ............................................................................................................................................... x
Acknowledgements .................................................................................................................................... xi

Chapter 1 Introduction............................................................................................................................... 1
1.1 Modular Product Design ..................................................................................................................... 2
1.2 Sustainability....................................................................................................................................... 4
1.3 Research Problem ............................................................................................................................... 7
1.4 Dissertation Organization ................................................................................................................... 9

Chapter 2 Literature Review ................................................................................................................... 11


2.1 Product End-of-Life Options ........................................................................................................... 11
2.1.1 Product End-of-Life Options Definition ..................................................................................... 11
2.1.2 Product EOL Options Selection .................................................................................................. 17
2.2 Traditional Modular Product Design Methodology ......................................................................... 19
2.2.1 Matrix-Function Classification ................................................................................................... 19
2.2.2 Cluster-Graph-Math-Artificial-Genetic (CGMAG) Classification ............................................. 22
2.3 Sustainability Related Modular Product Design .............................................................................. 23
2.3.1 Literature Search Method............................................................................................................ 23
2.3.2 Sustainability Focused Modular Product Design ........................................................................ 24
2.3.3 Product Lifecycle Management Related Modular Product Design ............................................. 32
2.4 Literature Review Summary ............................................................................................................ 45

Chapter 3 Methodology ............................................................................................................................ 47


3.1 EOL Options Fuzzy Representation ................................................................................................ 48
3.1.1 EOL Options Definition .............................................................................................................. 50
3.1.2 EOL Options’ Fuzzy Characteristics ......................................................................................... 51
3.1.3 EOL Options Fuzzy Evaluation .................................................................................................. 55
3.1.4 EOL Options Transfer Equations ................................................................................................ 57
3.1.5 Left-Right Defuzzification Method ............................................................................................ 59
3.2 Failure Rate Representation ............................................................................................................. 59
3.3 Life Cycle Sustainability Model ...................................................................................................... 61

vi
3.3.1 Economic Sustainability Model .................................................................................................. 62
3.3.1.1 Manufacturing Stage Cost ..................................................................................................... 63
3.3.1.2 Maintenance Stage Cost ........................................................................................................ 63
3.3.1.3 End-of-Life Stage Cost ......................................................................................................... 64
3.3.2 Environment Sustainability Model ............................................................................................ 65
3.3.2.1 Manufacturing Stage Environmental Impact ...................................................................... 65
3.3.2.2 Maintenance Stage Environmental Impact ......................................................................... 65
3.3.2.3 EOL Stage Environmental Impact ...................................................................................... 66
3.3.3 Social Sustainability Model ....................................................................................................... 66
3.3.3.1 Manufacturing Stage Labor Hour ........................................................................................ 67
3.3.3.2 Maintenance Stage Labor Hour ........................................................................................... 67
3.3.3.3 EOL Stage Labor Hour ........................................................................................................ 68
3.4. Key Component based Hierarchy Clustering Algorithms ............................................................. 68
3.4.1 Key Component Determination ................................................................................................ 68
3.4.2 Economic Sustainability Clustering Algorithm ........................................................................ 69
3.4.3 Environmental Sustainability Clustering Algorithm ................................................................ 70
3.4.4 Social Sustainability Clustering Algorithm ............................................................................. 72
3.5. Compromised Optimal Module Structure Determination .............................................................. 73

Chapter 4 Results ...................................................................................................................................... 74


4.1 Coffee Maker Case Study ............................................................................................................. 75
4.1.1 Case Study Introduction ........................................................................................................... 76
4.1.2 Coffee Maker EOL Options Analysis ...................................................................................... 79
4.1.3 Coffee Maker Life Cycle Sustainability Model ....................................................................... 83
4.1.4 Coffee Maker Key Component Determination ........................................................................ 83
4.1.5 Coffee Maker Key Component Based Cluster Algorithm Results .......................................... 84
4.1.6 Coffee Maker Compromised Optimal Module Structure ......................................................... 85
4.2 Refrigerator Case Study ................................................................................................................ 86
4.2.1 Case Study Introduction ........................................................................................................... 87
4.2.2 Refrigerator EOL Options Analysis ......................................................................................... 93
4.2.3 Refrigerator Life Cycle Sustainability Model .......................................................................... 96
4.2.4 Refrigerator Key Component Determination ........................................................................... 96
4.2.5 Refrigerator Component Based Cluster Algorithm Results ..................................................... 96
4.2.6 Refrigerator Compromised Optimal Module Structure ........................................................... 98

vii
Chapter 5 Sensitivity Analysis and Comparison .................................................................................... 99
5.1 Sensitivity Analysis ....................................................................................................................... 99
5.2 Comparison with DA Approach and ASCEM Approach Results................................................. 103
5.3 Response to Research Questions................................................................................................... 106

Chapter 6 Conclusion and Future Work .............................................................................................. 107


6.1 Conclusion .................................................................................................................................... 107
6.2 Future Work .................................................................................................................................. 108

Reference ................................................................................................................................................. 110

Appendix A Coffee Maker EOL Strategy Sustainability Value ......................................................... 119

Appendix B Coffee Maker EOL Strategy Characteristics Fuzzy Evaluation ................................... 121

Appendix C Coffee Maker Fuzzy Calculation ..................................................................................... 122

Appendix D Refrigerator EOL Strategy Sustainability Value ........................................................... 125

Appendix E Refrigerator EOL Strategy Characteristics Fuzzy Evaluation ..................................... 127

Appendix F Refrigerator Fuzzy Calculation ........................................................................................ 128

Appendix G Decomposition Approach ................................................................................................. 134

Appendix H ASCEM Approach ............................................................................................................ 136

viii
List of Figures

Figure 1-1 Three Spheres of Sustainability-Version 1 (Adopted from Rodriguez et al., 2002) ................... 5
Figure 1-2 Three Spheres of Sustainability-Version 2 (Adopted from Geniecafe, 2011) ............................ 6
Figure 2-1 Sustainability & Modular Product Design Publications............................................................ 24
Figure 2-2 Framework of Supply Chain Structure. M: Manufacturing; A: Assembly; P: Packing
(Adopted from Ernst and Kamrad, 2000) ................................................................................. 28
Figure 2-3 Life Cycle Objective Structure for Modular Design (Adopted from Li et al., 2008) ................ 34
Figure 3-1 Methodology Flow Chart .......................................................................................................... 48
Figure 3-2 EOL Option Expected Value Calculation ................................................................................. 50
Figure 3-3 Characteristic Relationship ....................................................................................................... 51
Figure 3-4 EOL Evaluation Characteristics ................................................................................................ 52
Figure 3-5 Triangular Membership Function.............................................................................................. 56
Figure 3-6 DSM Example (Adopted from Yan et al., 2012) ...................................................................... 60
Figure 3-7 Failure Rate DSM w.r.t. Two Components ............................................................................... 61
Figure 3-8 Product Life Cycle .................................................................................................................... 62
Figure 3-9 Manufacturing Cost Organization ............................................................................................. 63
Figure 3-10 Economic Sustainability Cluster Algorithm Flow Chart ........................................................ 70
Figure 3-11 Environmental Sustainability Clustering Algorithm Flowchart .............................................. 71
Figure 3-12 Social Sustainability Clustering Algorithm Flowchart ........................................................... 72
Figure 4-1 Mr.Coffee PR15 ........................................................................................................................ 75
Figure 4-2 Connection Relation DSM (Adopted from Chung, 2012) ........................................................ 77
Figure 4-3 Assembly Cost DSM (Adopted from Chung, 2012) ................................................................. 77
Figure 4-4 Assembly Labor Time DSM (Adopted from Chung, 2012)...................................................... 78
Figure 4-5 Disassembly Cost DSM (Adopted from Chung, 2012) ............................................................. 78
Figure 4-6 Disassembly Labor Time DSM (Adopted from Chung, 2012) ................................................. 78
Figure 4-7 Assembly Environmental Impact DSM .................................................................................... 79
Figure 4-8 Disassembly Environmental Impact DSM ................................................................................ 79
Figure 4-9 Refrigerator Dissection Sketch (Adopted from Chung, 2012) .................................................. 87
Figure 4-10 Connectivity Graph of Refrigerator (Adopted from Chung, 2012) ......................................... 89
Figure 4-11 Refrigerator Connection DSM (Adopted from Chung, 2012)................................................. 90
Figure 4-12 Refrigerator Assembly Time DSM (Adopted from Chung, 2012).......................................... 90
Figure 4-13 Refrigerator Assembly Cost DSM (Adopted from Chung, 2012) ........................................... 91
Figure 4-14 Refrigerator Disassembly Time DSM (Adopted from Chung, 2012) ..................................... 91
Figure 4-15 Refrigerator Disassembly Cost DSM (Adopted from Chung, 2012) ...................................... 92
Figure 4-16 Refrigerator Assembly Environmental Impact DSM .............................................................. 92
Figure 4-17 Refrigerator Disassembly Environmental Impact DSM ......................................................... 93
Figure 5-1 4 Dimensional Plot .................................................................................................................. 101
Figure 5-2 MSSI Index v.s. Pre-defined Key Component Numbers ........................................................ 102
Figure 5-3 Coffee Maker DA Module Structure ....................................................................................... 103

ix
List of Tables
Table 1-1 Sustainability Indicators & Major Themes ................................................................................... 7
Table 2-1 EOL Option Definitions ............................................................................................................. 12
Table 2-2 EOL Suitable Product/Component Types (Summarized from CRR, 2013) ............................... 14
Table 2-3 EOL Options Advantages & Disadvantages (Summarized from CRR, 2013) ........................... 15
Table 2-4 EOL Options Challenges & Optimization Strategies (Summarized from CRR, 2013) .............. 16
Table 2-5 ELDA EOL Evaluation Characteristics (Adopted from Ross, 2001) ......................................... 18
Table 2-6 Traditional MPD based on Matrix-Function Classification ....................................................... 21
Table 2-7 Sustainability Issue & Modular Product Life Cycle Stage
(Adopted from Yan et al., 2012) ............................................................................................... 37
Table 2-8 Attribute list for all life cycle (Adopted from Ji et al., 2012) ..................................................... 38
Table 2-9 Product Life Cycle & Product EOL Researches......................................................................... 44
Table 3-1 EOL Options Definition (Adopted from Marco et al., 1994) ..................................................... 50
Table 3-2 EOL Evaluation Characteristics and Description ....................................................................... 52
Table 3-3 EOL Characteristics Linguistic Variables and Level Used ........................................................ 55
Table 3-4 Linguistic Variable Level and corresponding Fuzzy Set ............................................................ 56
Table 3-5 Transfer Equations...................................................................................................................... 58
Table 4-1 Product Category (Adopted from Wolf, 2001) ........................................................................... 74
Table 4-2 The Components’ Attributes of Coffee Maker
(Adopted from Chung, 2012 with MTBF Extension) ............................................................... 76
Table 4-3 Material Market Price, Scrap Price and Landfill Price ............................................................... 80
Table 4-4 Filter Basket Fuzzy Set After Transferring ................................................................................ 81
Table 4-5 Filter Basket Defuzzification Numbers ...................................................................................... 81
Table 4-6 Filter Basket Sustainability Value Weight ................................................................................. 81
Table 4-7 Coffee Maker Excepted EOL Sustainability Value .................................................................... 82
Table 4-8 Non-Key Component Order ....................................................................................................... 84
Table 4-9 Optimal Module Structure and Corresponding Disassembly Sustainability Value .................... 85
Table 4-10 The Sub-Functions of the Components in the Refrigerator (Adopted from Chung, 2012) ...... 87
Table 4-11 Refrigerator Attributes (Adopted from Chung, 2012) .............................................................. 88
Table 4-12 Cabinet Frame Fuzzy Set After Transferring ........................................................................... 93
Table 4-13 Cabinet Frame Defuzzification Numbers ................................................................................. 94
Table 4-14 Cabinet Frame Sustainability Value Weight ............................................................................ 94
Table 4-15 Refrigerator Expected EOL Sustainability Value..................................................................... 94
Table 4-16 Non-Key Component Order ..................................................................................................... 97
Table 4-17 Optimal Module Structure and Corresponding Disassembly Sustainability Value .................. 97
Table 5-1 Sustainability Weights v.s. MSSI Index v.s. Optimal Module Structure ................................... 99
Table 5-2 Key Component Number v.s. Minimum MSSI Index .............................................................. 101
Table 5-3 ASCEM Result (Adopted from Chung, 2012) ......................................................................... 104
Table 5-4 Comparison between DA, ASCEM and Proposed Method ...................................................... 105

x
ACKNOWLEDGEMENTS
Pursuing the Ph.D. degree is a long and painful journey. I would like to thank all people who

helped me during this journey.

First and foremost, I would like to express my sincere gratitude to my advisor, Dr. Gül Oku-

dan Kremer, who was always there for me during the entire dissertation time. She has dedicated

an enormous amount of time, numerous efforts and great patience to guide me on my dissertation

research. Without her strong support and continuous help, I would not be able to reach this mile-

stone. She is a great mentor not only in academics but also in life, and she is my academic mother

in U.S. It is my great honor to have her as my advisor and spend several years working with and

learn from her.

I would also like to thank Dr. M. Jeya Chandra, Dr. Charles David Ray and Dr. Conrad S.

Tucker for their serving on my dissertation committee, offering insightful comments and providing

continuous encouragement. I am also grateful to Dr. Kathy L. Jackson and Dr. Sven Bilen for their

encouragement and generous help during my job search.

I would like to take this opportunity to thank all my friends at Penn State University for their

help. Special thanks to Kijung Park and other members in Applied Decision Analysis for Improved

Products & Systems (ADAPS) group for their friendship and support during this journey.

I would like to thank my parents, my parents-in-law, my sister and other family members

for their unreserved love and unconditional support they gave me. I really appreciate all the un-

derstanding and encouragement from them during dissertation journey.

xi
Deepest thanks to my wife, Yu Wang. She always believes in me and gives me vigorous

support. Her continuous encouragement and patience were extremely important when I experi-

enced setbacks. She deserves as much credit as I do in this dissertation.

xii
DEDICATION

To my wife and my family …

xiii
Chapter 1

Introduction

Product design is a creative process which combines the needs of customers, the strategic

requirements of companies, and the environmental constraints of government regulating agencies.

Four stages comprise product design: problem definition, conceptual design, preliminary design

and detail design. All design requirements should be fulfilled through these four stages. Because

approximately 70% of product cost (Appelqvist et al., 2004) and 80% of product quality (Dow-

latshahi, 1992) are determined during the design stage of product planning, product design is a

critical concern for developers.

Modularity is a methodology used to reduce complexity by decomposing a complicated

whole into several relatively simple parts (Steward, 1965. Modular product design (MPD) involves

applying the module concept to product design in order to increase the efficiency and effectiveness

of the design process as well as to facilitate production and expand marketing appeal. Many aspects

of MPD methods benefit product design, such as easy assembly, low supply chain cost and mass

customization. Therefore, applying an MPD method during the design stage has become a trend.

Paralleling the movement of increasing environmental and social concerns, sustainability

management and product life cycle management continually attract a great deal of attention. Sus-

tainability includes three dimensions: economic, environmental, and social aspects (Rodriguez et

al., 2002; Zamagni et al., 2009). Product life cycle management involves a product’s entire life

and end-of-life (EOL) options. EOL options can create a significant amount of energy waste, en-

vironmental pollution, and cost within a product’s life cycle (Li et al., 2008). Research efforts have

increasingly fixated on the analysis of sustainability performance throughout the product life cycle,

1
and have lately focused on taking product EOL options into account during the design stage. The

aggregation of MPD and life cycle management—as well as sustainability improvement in the

design stage—is currently popular, and many researchers have developed new product design

methodologies. However, there still exist some research gaps. First, both the entire-life-cycle-

based and the EOL-based methods hold a critical assumption: that EOL options have typically

been determined before the methodologies are developed, meaning that EOL is fixed and treated

as a constant input (e.g., Hata et al., 2001; Bryant et al, 2004; Kreng and Lee, 2004a, 2004b; Li et

al., 2008; Lai and Gershenson, 2009; Yan et al., 2012; Ji et al., 2012; Means and Guggemos, 2015;

Ameli et al., 2016). Another gap is that three dimensions of sustainability has not been analyzed

yet in relation to the product design field. Much existing research has been done to improve a

single dimension (e.g., Krikke et al., 2004; Fernandez and Kekale, 2005; Lau et al., 2007; Tseng

et al., 2008; Umeda et al., 2009; Shin et al., 2011) or in some cases two dimensions (e.g., Lee et

al., 2001; Li et al., 2008; Chung et al., 2011; Yan et al., 2012). The third gap is that key compo-

nents have not yet attracted sufficient attention. Key components represent product core competi-

tiveness and strongly influence product marketing performance. At present, there is almost no

research simultaneously incorporating uncertain EOL options, full-dimension sustainability im-

provement and key component specification within the framework of MPD. Therefore, the purpose

of this research is to take into account uncertain EOL options, full-dimension sustainability and

key components at the same time and to develop a novel modular design methodology.

1.1 Modular Product Design

As technology becomes more advanced, design becomes more important and complex. One

common way to reduce complexity is through decomposition, which splits a harder, larger system

2
into easier, smaller subsystems. By applying this philosophy inversely to design engineering, mod-

ular product design has evolved. Initially, this approach was proposed by Steward (1965), who

introduced the philosophy of system partition and testing. MPD involves clustering simple and

small product components into more complex subassemblies, and then combining these subassem-

blies to create a complete product. In modular product architecture, each functional product com-

ponent is implemented in exactly one subassembly, with few interactions between subassemblies

(Ulrich and Eppinger, 2000).

Many practical advantages of modularity have been examined in recent research. MPD has

been shown to increase manufacturing efficiency (Okudan et al., 2012); it can benefit the supply

chain by reducing inventory cost and saving distribution time (Feitzinger and Lee, 1997; Ernst and

Kamrad, 2000; Kamrani and Salhieh, 2008; Chiu and Okudan, 2014; Lavigne et al., 2014; Lewis,

et al., 2015). It can also satisfy the demand for mass customization (Gershenson et al., 2003;

Gershenson et al., 2004; Lau et al., 2007; Smith et al., 2013; Kuo, 2013; Wang and Tseng, 2013).

Given the advantages of MPD, much related research has been conducted as evidenced in

the literature. There are various ways to categorize the findings. Zhang and Gershenson (2003)

classified modular design methods into two groups: (1) matrix based, and (2) function based. Ma-

trix-based methods cluster product components using matrix or matrix-related techniques (Kusiak

and Chow, 1987; Kusiak and Wang, 1993; Pimmler and Eppinger, 1994; Newcomb et al., 1996;

Gu et al., 1997; Huang and Kusiak, 1998; Ji et al., 2013; Halstenberg, et al., 2015). Function-based

methods group product components according to their independent functions (Ishii et al., 1995;

Marshall et al., 1998; Stone et al., 1998; Jeong et al., 2015). Jose and Tollenaere (2005) distin-

guished modularization techniques into five categories: (1) clustering methods (Kusiak and Chow,

1987; Baldwin and Clark, 2000; Okudan et al., 2012;), (2) graph and matrix partitioning methods

3
(Kumar and Chandrasekharan, 1990; Huang and Kusiak, 1998; Li et al., 2008; Fan et al., 2013),

(3) mathematical programming methods (Kusiak and Wang, 1993; Shin et al., 2011), (4) artificial

intelligence methods (Zhang et al., 2006), and (5) genetic algorithms and heuristics (Kreng and

Lee, 2004a, 2004b; Tseng et al., 2008; Smith and Yen, 2010; Yan et al., 2012; Fujita et al., 2014).

MPD has a variety of merits. This research adapts the idea of MPD and combines it with an

uncertain EOL options concern and key components specification factor to improve the product

life cycle three-dimension sustainability performance while development is at the design stage.

1.2 Sustainability

Sustainability-related research has commanded global attention, due largely to the fact that

the environment and environment-related issues have increasingly become a matter of concern.

Sustainable development, as the World Commission on Environment and Development (WCED)

defined it more than a quarter century ago, is “meeting present needs without compromising the

ability of future generations to meet their needs” (1987). More recently, the US Environmental

Protection Agency (EPA) described sustainability as “the satisfaction of basic economic, social,

and security needs now and in the future without undermining the natural resource base and envi-

ronmental quality on which life depends” (1994). Sandborn and Myers (2008) modified the defi-

nition of sustainability using a more technical vantage point, stating that sustainability means keep-

ing an existing system operational and maintaining field versions of the system such that the orig-

inal requirements are satisfied. Based on this definition, they classified sustainability into three

broad groups: environmental, business or corporate, and technology. Additional popular and ac-

ceptable categorization includes economic sustainability, environmental sustainability and social

4
sustainability (Rodriguez et al., 2002; Zamagni et al., 2009), and its proponents suggest sustaina-

bility should consider these three factors simultaneously. Figure 1-1 and Figure 1-2 show the

details for each of these factors and their overlapping regions.

Environmental
Social-Environmental Social-Economic
Environmental Justice Natural Resource Use Energy Efficiency
Natural Resources Stewardship Environmental Management Subsidies/Incentives for
Locally & Globally Pollution Prevention use of Natural Resources
(air, water, land, waste)

Social Economic
Standard of Living Profit
Education Cost Savings
Community Economic Growth
Equal Opportunity R&D

Economic-Social
Business Ethics
Fair Trade
Worker’s Rights

Figure 1-1 Three Spheres of Sustainability—Version 1 (Adopted from Rodriguez et al., 2002)

5
Innovation
Capital Efficiency
Risk Management
Margin Improvement
Growth Enhancement
Total Shareholder Return

Job Creation Economic Growth Resource Efficiency


Skills Enhancement
Product Stewardship
Local Economic Impacts
Life Cycle Management
Social Investments
Margin Improvement
Business Ethics
Products to Services
Security Socio-Eco- Eco-Effi-
nomic ciency

Environmental Stew-
Social Progress Social-Environ-
ardship
mental
Diversity
Clear Air, Water & Land
Human Rights
Emissions Reductions
Community Outreach
Zero Waste, Releases & Spills
Indigenous Communities
Biodiversity
Labor Relations
Safety & Health
Environmental Regulations
Global Climate Change
Crisis Management
Environmental Justice

Figure 1-2 Three Spheres of Sustainability-Version 2

(Adopted from Geniecafe, 2011)

By combining and eliminating the overlapping elements from Figure 1-1 and Figure 1-2,

sustainability indicators and their corresponding major themes are summarized in Table 1-1. In

each table cell, key items associated with a major sustainability factor or their intersecting factors

are listed.

6
Table 1-1 Sustainability Indicators & Major Themes
Sustainabil- Eco-Environ- Social-Environ- Social-Eco-
Economic Environmental Social
ity Indictors mental mental nomic
Emissions Re-
duction;
Risk Manage- Human Di-
Environment Safety and
ment; Energy Effi- versity;
Management; Health;
Major Profit; ciency; Human Business
Environmental Global Climate
Themes Cost Saving; Life-Cycle Rights; Ethics;
Assessment; Change;
Product Inno- Management Labor Rela- Security
Natural Re-
vation; tions
source Manage-
ment;

1.3 Research Problem

In today’s challenging market, product competitiveness depends strongly on a few key com-

ponents. These may represent a core technology and affects the product function performance,

such as the turbocharger in automobile engine for driving technology. They may be expensive or

complicated to assemble or pollute the environment and consequently influence product market

performance, such as the cabinet of a refrigerator. Therefore, design products with a concentrated

consideration of their key components has become necessary and important. However, to the best

of knowledge, scant research available today addresses key components, with only one paper par-

tially alluding to the idea of focusing on them during product design. In it, Huang and Kusiak

(1998) provided a decomposition approach to clustering product components. In addition to con-

sidering the interactions among the components, they took people’s preferences into account dur-

ing analytic module formation. Thus, the key components were identified and handled as prefer-

ences. But their method is really subjective and not sustainability-oriented (Ma and Okudan, 2013).

Therefore, developing a sustainability-oriented product design method which incorporates key

components specification is the primary motivation for this research.

7
Today, companies must consider sustainability issues in relation to their own internal re-

quirements (e.g., cost saving and profit increasing) and to outside constraints (e.g., environmental

protection and social responsibility). Life cycle management (specifically EOL options) has

emerged as a critical factor in sustainability. However, determining EOL options at the product

design stage is a complex challenge. Take, for example, design-stage EOL options for retired bi-

cycle components in perspective with geography. In Asia, the environmental policy is relatively

loose and disposal is the most inexpensive option; thus, disposal is always adopted as the EOL

option there. In Europe, however, the environmental policy is stricter and recycling is the preferred

EOL option. As such, fixed EOL options in the design stage may not always be appropriate. Lim-

ited research exists discussing uncertainty in EOL options at the design stage. This research gap

is related to practice in both industry and academia; therefore, it motivates this investigation, too.

As mentioned above, MPD can be an effective and powerful product design method resulting

in significant benefits. Therefore, the aggregation of MPD, key components specification, and un-

certain EOL options during the design stage is consequently important. Subsequently, the research

questions in this dissertation may be derived as:

Can the new MPD method specify and concentrate on the key components in the design

process?

Can the new MPD method take uncertain EOL options into account in life cycle sustaina-

bility?

Can the new MPD method improve three-dimension sustainability performance with key

components specification and uncertain EOL options consideration?

8
In this research, fuzzy logic is adopted to represent uncertainty in EOL options. Selected

fuzzy characteristics will be summarized to evaluate uncertainty (e.g., component quality, mate-

rial). EOL options’ excepted sustainability performance value (cost, environmental impact and

labor time) will be obtained using this comprehensive fuzzy decision-making framework. A sus-

tainable model will be developed to quantify environmental, economic and social impacts. Three

clustering algorithms will be developed to specify key components and to group non-key compo-

nents into key component modules with regards to sustainability dimensions. These algorithms

will provide optimal module structures with respect to each sustainability dimension. A tri-criteria

sustainable decision-making process will be developed to balance sustainability performance and

to generate compromised module structure. It is posited that by using the proposed methodology,

companies will be able to improve their economic performance, without sacrificing environmental

performance (e.g., increasing emission) and social performance (e.g., firing labors), and therefore

reach an equilibrium among three sustainability dimensions. Moreover, companies can not only

avoid making several EOL options assumptions at the design stage that might be harmful to entire

life cycle, but can also specify key product components in order to emphasize on product core

competitiveness. It is additionally posited that the proposed methodology may apply to companies

which operate under different product strategies. For example, product maintenance may involve

the adaptation of disassembly and assembly methods in response to local, regional, or national

environmental regulations.

1.4 Dissertation Organization

This dissertation entails six parts: introduction, literature review, methodology, result, sen-

sitive analysis and comparison, and conclusion and future work. The introduction presents the

9
importance of MPD and sustainability, and introduce the research questions accordingly. The lit-

erature review section examines related research topics, such as EOL definition and selection, as

well as MPD methods and sustainability related MPD methods, and will illustrate the research

gaps. The methodology section discusses our proposed methodology in detail. The result section

applies the proposed methodology to two case studies. The sensitive analysis and comparison sec-

tion varies the parameters and/or coefficients in the proposed methodology in order to validate it,

and compares it to other methods. The conclusion summarizes this dissertation and points the way

toward future research.

10
Chapter 2

Literature Review

This literature review addresses product end of life (EOL) options, modular product design

(MPD) methods and sustainability-related MPD methods. Section 1 covers product EOL options

definitions and selection method. Section 2 covers traditional MPD methodologies: matrix-func-

tion classification and Cluster-Graph-Math-Artificial-Genetic (CGMAG) classification. Section 3

covers works related to sustainability issues especially focused on product lifecycle management.

2.1 Product End of Life Options

Demographic growth, combined with increasing product requirements, generates an extreme

amount of retired products each year. Due to the concern of environmental pollution, EU has for-

mulated new regulations based on the principle of extended producer responsibility (Walls, 2006),

which requires the manufacturer to be more responsible for the life cycle of their products, espe-

cially the end of life (EOL) phase. In response to this requirement, many researchers have con-

ducted investigations on EOL phase. In this section, several product EOL options definition and

selection methods will be reviewed.

2.1.1 Product End-of-Life Options Definitions

Due to the importance of EOL options, several EOL option definition and categorizing meth-

ods were developed by former researchers. Marco et al. (1994) defined seven types of EOL options:

reuse, remanufacturing, primary recycle, secondary recycle, incineration, landfill and special han-

dling. Ishii et al. (1995) summarized Marco et al.’s work into four types of EOL options: remanu-

facturing/reuse, primary recycling, secondary recycling, and disposal. Stevels (1997) gave also

four types of EOL options: reuse, remanufacturing/refurbishing, recycling, and incineration.

11
Ijomah et al. (1999) provided five types of EOL options: reuse, repair, reconditioning, remanufac-

ture and recycling. Rose (2001) classified EOL options into six types: reuse, service, remanufac-

turing, recycling with disassembly, recycling without disassembly and disposal based on analysis

of hundreds of products. Li et al. (2008) considered three EOL options: reuse, recycle and disposal.

Chung et al. (2011) adopted this classification and applied it to product life cycle improvement.

Remery et al. (2012) categorized six EOL options: reuse, remanufacturing, recycling with disas-

sembly, recycling without disassembly, incineration with energy recovery and disposal. The latest

report of Center of Remanufacturing and Reuse (CRR) (2013) defines nine EOL options: reman-

ufacturing, reconditioning/refurbishing, reuse, repurposing, repair, recycling, composting, incin-

eration and landfill based on the classification and summarization of former work in both industries

and academia. All definitions are summarized in Table 2-1.

Table 2-1 EOL Options Definitions


Author/ Description /Notice
EOL Option
Year
Reuse Using in the same (direct reuse) or another (indirect reuse) application.
Retaining serviceable parts, refurbishing usable parts, or replacing identical or reworked com-
Remanufacture
ponents from obsolete products.
Reprocessing a material into a form that can be used in the same or another “high” value prod-
Primary Recycle
Marco et al. uct.
(1994) Reprocessing a material into a “low” value product, e.g., industrial grade rubber being repro-
Secondary Recycle
cessed into a general grade rubber.
Incineration Incinerating a material to produce heat and electricity.
Landfills Landfilling waste products with no intrinsic value.
Special Handling Mandatory for all toxic or hazardous materials.
Remanufacturing/ Using parts “as-is” or after minor repairs or overhauls.
Reuse
Ishii et al.
Primary Recycling Processing materials into another “high-value” application.
(1995)
Secondary Recycling Processing materials into a “low-value” application.
Disposal Landfilling or incineration.
Reuse Using manufacturer’s name in its second life too, not the name of repairing firm.
Remanufacturing/ Quality and volume guarantee required.
Stevels
Refurbishing
(1997)
Recycling Technology available and sufficient economy of scale concerned.
Incineration Restrictions for acceptance required.
Reuse Using a functional component from a retired assembly.
Repair Bringing damaged components back to a functional condition.
Ijomah et
Restoring components to a functional and/or satisfactory state but not above original specifica-
al. Reconditioning
tion, common methods included resurfacing, repainting, sleeving, etc.
(1999)
Bringing an assembly to like-new condition through replacing and rebuilding component parts
Remanufacture
at least to the current specification.

12
Taking component materials and processing them to make the same materials or useful de-
Recycling
graded materials.
Reuse Second-hand trading of product using as originally designed.
Service Repairing or rebuilding the product using service parts at product used location.
Remanufacturing Bringing large quantities of similar products into a central facility and disassembled.
Rose Recycling with Disas- Disassembly into material fractions by removing material contaminants, hazardous materials,
(2001) sembly or high-value components.
Recycling without Reducing material size for facilitate sorting. The shredded material methods are based on mag-
Disassembly netic, density or other properties of the materials.
Disposal Landfilling or incineration the product with or without energy recovery.
Li et al. Reuse Using in the same (direct reuse) or another (indirect reuse) application.
(2008) & . Taking component materials and processing them to make the same materials or useful de-
Recycle
Chung et al. graded materials.
(2011) Disposal Landfilling or incineration.

Used for product modules without performing any repair or renovation operation other than
Reuse
cleaning.
Using refurbished parts and new ones, new modules are built and sold as remanufactured prod-
Remanufacturing
ucts.
Recycling with Disas- Reusing of a material for its original application required.
sembly
Remery et
Used for product modules which are shredded or compacted without any previous disassembly
al. (2012)
Recycling without operation, and the different kinds of materials are then recovered: materials containing iron are
Disassembly separated magnetically and others are separated with methods like wind sifting, vibration meth-
ods,
Incineration with En- Used for product modules containing materials with a high calorific value, generally plastics.
ergy Recovery
Disposal Landfill as solid waste.
Process of returning “a used product to at least its original performance with a warranty that is
Remanufacturing
equivalent or better than that of the newly manufactured product”.
Process of returning “a used product to a satisfactory working condition by rebuilding or re-
Reconditioning
pairing major components that are close to failure, even where there are no reported or appar-
/Refurbishing
ent faults in those components”.
Process by which “a product or its components are put back into use for the same purpose at
Reuse
EOL”
Process of “utilizing a product or its components in a role that it was not originally designed to
CRR Repurposing
perform”.
(2013)
Repair Process of returning “a faulty or broken product or component back to a usable state”.
Process of returning “waste materials for their original purpose or for other purposes, excluding
Recycling
energy recovery”.
Process of converting organic matter (via controlled aerobic decomposition) to create a soil ad-
Composting
ditive which improves soil structure and provides nutrients for plants.
Process of combustion of organic waste materials to generate electric power (or combined heat
Incineration
and power).
Landfill Process of disposing of waste by burial.

The report of Center of Remanufacturing and Reuse (CRR) (2013) not only gave the defini-

tion of nine EOL options, but also provided individual suitable product examples for each EOL

options, advantage, disadvantage, challenge and optimization method. All of these are presented

in Tables 2-2, 2-3, and 2-4.

13
Table 2-2 EOL Suitable Product/Component Types (Summarized from CRR, 2013)
EOL Options Suitable Product/Component Type
 Have low technological obsolescence
 Have a high inherent material value and production cost
 Be sold in a market tolerant of as-new products
Remanufacturing
 Be feasible to collect at EOL
 Contain a durable core which can be reused many times
 Have the potential for disassembly down to component level (page 4 & 5)
 Have low technological obsolescence
 Have a high inherent material value
Reconditioning
 Be sold in a market tolerant of second-life products
/refurbishing
 Be feasible to collect at EOL
 Have the potential for disassembly down to sub-assembly level (page 6)
 Have low technological and aesthetic obsolescence
Reuse  Be sold in a market tolerant of second-life products
 Be feasible to collect at EOL for resale (page 8)
 Have modular functionality (so that functions unnecessary in the new
purpose can be easily disabled)
 Have multiple functions
Repurposing
 Support standard interfaces (e.g., USB for computing)
 Be controlled by a software based operating system (e.g., Symbian
the operating system for mobile phones) (page 9)
 Have a high inherent value (either financial or sentimental)
 Have high initial cost
Repair
 Be difficult or impossible to replace
 Contain components that can be replaced (via non-destructive means) (page 11)
 Contain components not suitable for more profitable reuse strategies
 Contain components made from materials that are recyclable and that can be col-
lected at EOL for recycling
Recycling  Be constructed from a single or few separable materials
 Have the potential for disassembly down to component level
 Contain high-value materials (and in large enough quantity to enable cost effective
extraction) (page 12)
 Organic or certified compostable materials
 Materials which will compost in 3 years or less (for domestic composting)
Composting
 No synthetic adhesives and material coatings that could potentially contaminate
compost (page 14)
 Be products for which no other higher value EOL option is currently feasible and
for which landfill is undesirable
Incineration
 Have primarily organic (containing carbon) components with little heavy metal
content (page 15)
 Contain inert, non-toxic materials
Landfill  Not contain electronics (WEEE) or other regulated components/materials
 Be products for which no other EOL treatment is currently feasible (page 17)

14
Table 2-3 EOL Options Advantages & Disadvantages (Summarized from CRR, 2013)
EOL Options Advantages Disadvantages
Ability to charge an as-new price, whilst benefiting from reduced production Only suited to products with specific characteristics;
Remanufac- costs and reduced environmental impacts; Product must be disassembled down to component level;
turing New business models based on Product Service Systems (PSS); Significant time and financial investment to set up the system.
Increased skilled employment. (page 5) (page 5)
Reduced production costs and environmental impacts as only parts which re- Limited warranty and lack of standardized terminology lead to concerns over
Recondition-
quire attention are remediated; quality;
ing
New business models based on Product Service Systems; Market size is small compared to the market for new products;
/refurbishing
Increased skilled employment. (page 7) Aesthetics of product may not be as-new. (page 7)
Little or no remediation required; Only appropriate for products with very low technological and aesthetic obso-
Minimal production costs and reduced environmental impacts; lescence;
Reuse
New business models based on PSS. Concerns over quality and fitness for purpose;
(page 8) Some faults may remain undetected during testing. (page 8)
Can provide a novel market for outdated technologies and aesthetics; Difficulty of identifying appropriate Markets;
Repurposing May require little or no remediation; Concerns over quality and fitness for purpose;
Reduced environmental impacts of production. (page 10) Legislation may be different in new product sector. (page 10)
Strong but small market for repair of products that cannot be easily replaced; No large scale systems in place for repair, therefore, costs may be high;
Repair Increased local skilled jobs; Concerns over quality and fitness for purpose;
Reduced environmental impacts of production. (page 11) Product likely to command a low resale value. (page 11)
OEMs‟ recovery obligations relatively easy to meet through recycling; May not meet increasing producer responsibility obligations;
Positive branding opportunities due to widespread knowledge of recycling; Quantity of recycling is often higher than demand;
Recycling
Diverts material from landfill and reduces production of virgin materials. Mass recycling of mixed waste can lead to contamination and quality varia-
(page 13) tions. (page 13)
Domestic composting avoids financial cost of waste disposal & transport Compostable materials require specific environmental conditions to decom-
impacts at EOL; pose (air, water, bacteria);
Composting Sale of compost can offset a portion of collection costs (municipal compost- Difficult to prevent contamination with non-compostable materials;
ing); (page 14) Relative environmental benefits compared to non-compostable materials are
Use of plant-based compostable biopolymers reduces reliance on crude oil. not yet clear. (page 14)
Heat and power generation can subsidize the treatment of waste; Toxic ash produced is hazardous to human health;
Relatively low land use; Removes the major driver for the development of clean technologies;
Incineration
Reduces the volume of the original waste by up to 95%. (page 16) Expert and local community concern about the environmental impact of
Incineration. (page 16)
Existing infrastructure; Convenience; Valuable resources (such as copper and aluminum) are not reused/recycled;
Landfill Possibility of future gas harvesting and landfill mining. (page 17) Potential water pollution and soil Contamination;
Increasing costs of landfill taxes. (page 17)

15
Table 2-4 EOL Options Challenges & Optimization Strategies (Summarized from CRR, 2013)

EOL Options Main Challenges Optimization Strategy


Customer perceptions of quality; Offer appropriate guarantees;
Remanufactur
Levels of Core return; Build Core return into the business model (via lease or financial incentives);
-ing
Unknown Core quality. (page 6) Incorporate wear markers into the product. (page 6)
Perceptions of quality; Give appropriate guarantees;
Reconditioning
Levels of Core return; Build core return into the business model (via lease or financial incentives);
/refurbishing
Some faults may remain undetected after testing. (page 7) Increase durability and reduce the complexity of disassembly and testing. (page 7)
Technological and aesthetic obsolescence; Design product for easy upgrade and repair;
Reuse Limited markets; Market to less technologically advanced global regions;
Some faults may remain undetected after testing. (page 9) Increase durability and reduce the complexity of disassembly and testing. (page 9)
Likely need for cross-sector partnerships in establishing new product; Create partnerships through industry bodies and networking hubs;
Repurposing Predicting possible second life uses during original design; Design product for easy upgrade and repair;
Some faults may remain undetected after testing. (page 10) Increase durability and reduce the complexity of disassembly and testing (page 10)
Relatively small market for repaired products; Design business model with incentives for repair. Rebrand 'repair';
Repair High costs of one-off repair; Design product for easy upgrade, disassembly, and repair by non-specialists;
Some faults may remain undetected after testing. (page 12) Increase durability and reduce the complexity of disassembly and testing. (page 12)
Correct collection and separation of recyclable materials; Design for easy disassembly and/or sorting by householders and apply appropriate
Contamination of materials; labeling;
Recycling
Low market value for recycling. (page 13) Specify single material types or compatible materials (when recycled together);
Specify recycled materials to create demand for recycles. (page 13)
Preventing non-compostable materials from entering composting Correctly label product and packaging. Design for easy separation of non-composta-
stream; ble material from product;
Composting Lack of composting facilities accredited to PAS 100; Design products for both composting and recycling;
Preventing compostable materials being sent to landfill. (page 15) Consider the appropriateness of specifying a compostable component, given the
likely EoL treatment of the product. (page 15)
Emissions and hazardous waste materials; Specify organic materials with little or no heavy metal content;
Quantity of waste material; Create markets for recycling and lightweight your products;
Incineration
Opposition to incineration. (page 16) Specify recycled, compostable and recyclable materials. (page 16)

Lack of space; Design to encourage other EOL Strategies;


Landfill Quantity of waste material; (page 18) Create markets for recycling and lightweight your products; (page 18)
Valuable resources (such as copper and aluminum) are not recycled. Correctly label product and packaging. Design for disassembly, reuse and recycling.

16
From the literature review in this section, there are eight EOL options definitions, and they

cover all possible handling methods for retired products. The definition in CRR (2013) is complete

and detailed, not only because it is by an authoritative organization, but also it summarized the

former works and provided advantages and disadvantages for each EOL option. However, CRR

(2013)’s definition is based on product characteristics summarization and lacks numerical analysis;

therefore, it is good for qualitative analysis but not good for quantitative evaluation. Although

Marco et al. (1994)’s work is two decades ago, their work has been developed as calculation for-

mulas by Lee et al. (2001), and therefore it could be used as quantitative analysis. Hence, in order

to assess EOL options quantitatively, the methodology adopt Marco et al. (1994)’s seven EOL

options definition.

2.1.2 Product EOL Options Selection

When the definition of EOL options is done, it is reasonable to select EOL options for each

of product components in the next step. In this section, we will review former works of EOL op-

tions selection. Several researchers provided evaluation criteria for EOL option selection. Keeney

and Raiffa (1976) provided five rules for criteria selection:

 Completeness: all important points of view are covered;

 Non-redundancy: two or more criteria should not measure the same thing;

 Minimality: the dimension of the problem should be kept to a minimum;

 Operationality: the set of criteria can be measured and meaningfully used in the analysis;

 Discriminality: the criteria should discriminate between EOL alternatives: if all EOL alter-

natives have the same value on a certain criterion then this criterion will not play any role

in the comparison of the EOL alternatives.

17
Lee et al. (2001) assigned seven EOL options to components based on materials, such as

metal without other alloy is recommended to primary recycling, alloy is recommended to second-

ary recycling, ceramic is recommended to secondary recycling or landfill, etc.

Rose (2001) summarized five groups of evaluation characteristics. The details are shown in

Table 2-5.

Table 2-5 ELDA EOL Evaluation Characteristics (Adopted from Rose, 2001)
Factors Characteristics
Cycle design; wear-out life; technology cycle; tech-
External nology research focus; company design focus; rea-
son for obsolescence; functional complexity
Penalties associated with recycling; cleanliness of
Material the product; number of materials; recycling value
drivers; EOL path categorization
Disassembly time; number of connectors; number
Disassembly
of modules; number of parts; disassembly time/step
Disassembly Disassembly motivation; overall access to compo-
Continued nents
Responsibility for recycling costs; responsibility
Inverse Supply for collection and transportation costs; recycling
Chain drop off centers; trade in possibilities; beneficiaries
of recycling

Bufardi et al. (2004) described three criteria: the direction of preference, the scale of meas-

urement and unit of measurement. Li et al. (2008) provided five criteria for EOL evaluation: com-

ponent classification, life cycle spanning, recycling methods, material compatibility, and special

handling and material classification.

From the review in this section, it is evident that EOL options determination needs to con-

sider several criteria. The criteria should consider all possible impacts of components completely,

which should cover the influence of component inside, such as material, the component itself,

among components but inside of the product, and outside of the product. However, no studies in

the published literature involve these four aspects. Therefore, this study will re-consider EOL de-

termination criteria in newly developed modular design method.

18
2.2 Traditional Modular Product Design Methodology

MPD is a widely applied methodology in design engineering, and many MPD methods have

evolved over the past several decades. Here some of these traditional methods are selected and

discussed based on the categories of matrix-function classification and Cluster-Graph-Math-Arti-

ficial-Genetic (CGMAG) classifications.

2.2.1 Matrix-Function Classification

As noted in Chapter 1, Zhang and Gershenson (2003) categorized MPD methods into two

groups: (1) matrix based, and (2) function based. Matrix-based methods sort product components

into one module using a clustering method based on matrix characteristics. Function-based meth-

ods require intrinsic knowledge of the complex product or system to identify functions and de-

compose products. This study identifies this method as the matrix-function classification.

Among classic matrix-based MPD methods, Kusiak and Chow (1987) developed a cluster

identification algorithm and a cost analysis algorithm to group components. In order to increase

design efficiency, Kusiak and Wang (1993) developed the triangularization algorithm based on

depth-first search and applied this method along with a decoupling algorithm to group and opti-

mize modules. Newcomb et al. (1996) defined two indexes: CR (Correspondence Ratio) and CI

(Cluster Independence) to measure modularity and then applied cluster identification algorithm to

re-design products. Gu et al. (1997) took several design criteria into account and used integrated

modular design methodology for life cycle engineering. Huang and Kusiak (1998) modified the

triangularization algorithm by considering interaction and suitability matrices. Pimmler and Ep-

pinger (1994) applied a heuristic swapping algorithm to measure interaction among components

according to five different integers (-2, -1, 0, 1, 2) and divided the interaction into four types, and

put them into one entry with four numbers.

19
Among function-based MPD methods, Ishii et al. (1995) used a fishbone diagram to repre-

sent the relationships among modules. Marshall et al. (1998) checked whether the corporate goals

and product requirements were matched or not in the modular design. Stone et al. (1998) used a

function structure diagram to identify dominant flows, branching flows and conversion transmis-

sion flows, where each flow is a potential module or module type.

The matrix-based MPD methods focus on the similarities and differences among compo-

nents, but these methods direct limited attention to the function-level relationships. For example,

Kusiak and Chow (1987) developed two matrix-based algorithms from only the perspective of the

components’ physical relationships. In contrast, function-based modular design methods concen-

trate on functional or group relations while mostly ignoring component-level properties. For ex-

ample, Marshall et al. (1998) considered only functional criteria to check the implications of mod-

ular design. Until recently, no methods have taken these two factors into account simultaneously.

Table 2-6 summarizes published examples of matrix-based and function-based modular design

methods.

As may be seen in Table 2-6, the traditional MPD methods discuss basic module forming

procedures and do not concentrate on sustainability at the design stage. Thus, there is a need to

develop a new MPD method that targets on sustainability improvement.

20
Table 2-6 Traditional MPD based on Matrix-Function Classification
Authors
Type Method Merits Limitations
/Date
Cluster Identi-
Kusiak and fication Algo- No guarantee for more
Matrix- Develop simpler group
Chow rithm + Cost complex modular design
Based methods
(1987) Analysis Al- problems
gorithm
Triangulariza-
Kusiak and
Matrix- tion Algorithm Two methods increase No guarantee for all
Wang
Based + Decoupling design efficiency cases
(1993)
Algorithm
Five levels interactions
(-2~2); four types in-
Pimmler and Heuristic Ignores similarity; no
Matrix- teractions (spatial, en-
Eppinger Swapping Al- guarantee for the final
Based ergy, information, ma-
(1994) gorithm form
terials); one entry with
four numbers
Gu, Hashe- No similarity notes
Simulated An- Integrated modular de-
mian and Matrix- among modules and
nealing Algo- sign methodology for
Sosale Based component incompatibil-
rithm life cycle engineering
(1997) ity
The suitability may not
Triangulariza-
Huang and Two matrices: interac- contribute to design; the
Matrix- tion Algorithm
Kusiak tion and suitability ma- deletion + duplication
Based + deletion+du-
(1998) trix cannot satisfy all condi-
plication
tions
The algorithm may not
Newcomb, Two indexes CR (cor-
Cluster Identi- be feasible with no inter-
Bras and Matrix- respondence ratio) and
fication Algo- mediate steps to allow a
Rosen Based CI (Cluster Independ-
rithm designer to move to a
(1998) ence)
better design
Ishii, Juen- Func- Fishbone diagram rep- May not apply in other
Fishbone Dia-
gel and Eu- tion- resents module rela- life cycle applications
gram
banks (1995) Based tionships except product recycling
Func- Holonic Prod- Check the matches of
Marshall No specific modulariza-
tion- uct Design corporate goals and
(1998) tion method
Based Method product requirements
Heuristic
Module identification
Stone, Wood Func- Method+
is unique; each func- The heuristic approach is
and Craw- tion- Function
tion is a potential mod- not easy to quantify
ford (1999) Based Structure Dia-
ule or module
gram

21
2.2.2 Cluster-Graph-Math-Artificial-Genetic (CGMAG) Classification

As noted in Chapter 1, Jose and Tollenaere (2005) proposed categorizing MPD methods.

Their analysis divides MPD methods into five groups: (1) clustering methods, (2) graph and matrix

partitioning methods, (3) mathematical programming methods, (4) artificial intelligence methods,

and (5) genetic algorithms and heuristics. Herein this study refers to this as CGMAG, based on the

first letter of each group.

Clustering methods group components into clusters according to similarities and differences

based on different design criteria (Kusiak and Chow, 1987; Baldwin and Clark, 2000; Okudan et

al., 2012). Graph and matrix partitioning methods apply graph or matrix-based methods to sort

components. The investigations of Kumar and Chandrasekharan (1990) and Huang and Kusiak

(1998) can be incorporated into graph and matrix partitioning methods. Fan et al. (2013) employed

network graph methodology to solve the structure-oriented MPD planning. Most matrix-based

methods could be categorized into this group (Pimmler and Eppinger, 1994; Gu et al., 1997; New-

comb et al., 1998). Mathematical programming methods are used to form component groups.

Kusiak and Wang (1993) presented a mathematical programming method which searches for mod-

ules through the use of linear programming. Chiu et al. (2016) provided a product modularity

redesign approach to improve both assembly and sustainability performance based on particle

swarm optimization (PSO). Artificial intelligence, as a branch of computer science, is also a useful

tool for clustering components. Zhang et al. (2006) discussed an evolving knowledge-based arti-

ficial intelligence technique for the modularization of components. Genetic algorithms and heuris-

tics are a set of widely used methods to solve the optimizing problems. Kreng and Lee (2004b)

proposed an MPD method that uses nonlinear programming to construct an objective function that

is subject to certain constraints, and then applies a grouping genetic algorithm heuristic to search

22
for an optimal or near-optimal modular design. Fujita et al. (2014) combined genetic algorithm

and a simplex method to solve the simultaneous design problem of module communalization strat-

egies under the given product architecture and supply chain configuration.

In this section, traditional MPD methods based on two categories: matrix-function and

CGMAG were reviewed, and they almost covered all current MPD methods. Key component may

represent a core technology and affect the product function performance, such as the turbocharger

in automobile engine for driving technology (Micheletti, 1985); or they may be expensive or com-

plicated to assemble or pollute the environment and consequently influence product market per-

formance, such as the cabinet of a refrigerator (Umeda, et al., 2000). However, few papers availa-

ble today address these key components, with only one paper partially alluding to the idea of em-

phasizing them during product design. Huang and Kusiak (1998) provided a decomposition ap-

proach to cluster product components. Besides interaction among components, they took designers’

preference into account in module forming. Key components could be identified and handled as

preference. But their method is not sustainability oriented. Therefore, a key components specified

and sustainability focused MPD is needed to fill this gap.

2.3 Sustainability Related Modular Product Design

2.3.1 Literature Search Method

The popularity of combining sustainability and MPD is evident. Based on investigation us-

ing the database Compendex, the number of papers in this field has steadily increased during last

two decades. Figure 2-1 shows this trend, based on a search using the keywords “sustainability”

and “modular product design”.

23
Sustainability MPD Publications
171
180
160
140
# of Publications
120
100
69
80
60
40 20
20 4 4
0
before 1990 1990~1995 1995~2000 2000~2005 2005~now
Year

Figure 2-1 Sustainability & Modular Product Design Publications


To generate Figure 2-1, the literature related to both indicator subjects in the Compendex database

was searched, combining the auto-stemming function and limiting the date range from 01/01/1980

to the present. For example, to search for product innovation, it used the keywords modular prod-

uct design and product innovation, and located 116 papers. The majority of papers addressed other

domains and did not relate to main concern. To narrow the field, it refined the search by selecting

only the classification codes product design and product development, which yielded 67 papers,

many of which were also irrelevant. Therefore, it continued to pare down the field by incorporating

the controlling vocabulary to include industrial engineering and management, product engineer-

ing and industrial economics, and 65 papers remained. Based on the 65 sustainability papers, this

study selected some of the classical methods to discuss in sub-section 2 and detail the product

lifecycle management in sub-section 3.

2.3.2 Sustainability Focused Modular Product Design

As mentioned in Chapter 1, sustainability indicators include 17 issues, however, only some

of them are strongly related to MPD, such as those involving product innovation, profit, cost saving,

24
environment management and lifecycle management. Other indicators such as risk management,

security and labor relationship are weakly associated with MPD, and very few papers were found

that addressed these areas. Therefore, product innovation, profit and cost saving MPD are focused

in this sub-section and emphasis will be moved to lifecycle management in sub-section 3.

Product Innovation is a sustainability indicator which focuses on improving product perfor-

mance by taking innovation into account. One of the most commonly used definitions of product

innovation is product newness (Danneels and Kleinschmidt, 2001) which can be operationalized

as newness to the customer, to the firm or to the industry. Therefore, when identifying product

innovation, both marketing and technological perspectives as well as macro level and micro level

perspectives should be taken into account (Garcia and Calantone, 2002).

Lau (2009) analyzed MPD from the perspective of the managerial side and addressed the

whole supply chain. He summarized seven critical factors used to explore the management of MPD,

and then addressed product innovation by comparing and analyzing six cases of MPD with differ-

ent fields, sizes, competitive strategies and market segments. Lau’s seven factors are: 1) pre-de-

fined product advantage; 2) selectively used design rules; 3) module definition; 4) system integra-

tion; 5) technological newness; 6) internal communication; and 7) supplier and customer involve-

ment. Lau derived a managerial guide for MPD that elaborates on these seven critical factors. This

guide includes six criteria: 1) product modularity decision; 2) MPD on the product level; 3) MPD

on the organizational level; 4) internal coordination on MPD; 5) supplier coordination on MPD;

and 6) customer coordination on MPD. This guide provides management direction of MPD, and

could have benefits for both industrial and academic fields. However, Lau does not provide a spe-

cific management method for MPD, and therefore his work can only be regarded as a guideline.

Moreover, Lau’s work is based on one critical assumption: MPD is positive relative to product

25
innovation. While this assumption will be later proven to be partially correct, it is discussed in his

later publications.

The theory of inventive problem solving (TRIZ) is another good method for analyzing product

innovation with MPD consideration. The underlying idea of TRIZ is that invention has logical

rules and principles that lead from problem to solution. To achieve its basic goal, TRIZ uses a

system involving abstract principles and laws together with a huge collection of facts and examples

in a readable application manner (Rantanen and Domb, 2002). There are several TRIZ-related

MPD methods. Regazzoni and Rizzi (2008) proposed a roadmap for complex MPD based on

combination of modular function deployment (MFD), TRIZ, and design structure matrix (DSM).

Their design paradigm is built on the scheme of MFD based on Ericsson and Erixon (1999)’s work,

which consists of five steps:

1) Collect and formalize customer requirements (using Quality Function Deployment (QFD))

2) Analyze and select a technical solution (using a Pugh matrix and function decomposition)

3) Define modules (using Module Interface Matrix (MIM))

4) Evaluate modules

5) Optimize modules

Regazzoni and Rizzi (2008) noted that there are some drawbacks in Ericsson and Erixon’s

original MFD. In step 1, using QFD makes it hard to find all the real needs of eventual customers

because simply asking the most important customers to define their “wants” and “wishes” rarely

yields complete satisfaction. Therefore, Regazzoni and Rizzi applied Human Centered Design

(HCD) in step 1. In step 2, the Pugh matrix presents two significant disadvantages: the results

obtained rely strongly on expert experience and understanding of the problem, and psychological

inertia limits engineers to the set of known solutions and excluding new ideas.

26
To overcome these drawbacks, Regazzoni and Rizzi (2008) adopted TRIZ-based analysis

which incorporates a Function Tree Diagram, a TRIZ functional model, RCA+, and a RelEvent

Diagram. In step 3, the original MFD uses Module Drivers and MIM, although module drivers do

not consider basic interactions among components. Therefore, Regazzoni and Rizzi applied DSM

at this step. Step 4 and Step 5 remained the same as they appeared in the original MFD. Regazzoni

and Rizzi then proposed a new roadmap of MPD involving the following 5 revised steps:

1) Collect and formalize customer requirements (using HCD)

2) Choose a technical solution (using TRIZ based analysis)

3) Define modules (using DSM)

4) Evaluate modules

5) Optimize modules

Regazzoni and Rizzi’s TRIZ-based MPD method is creative and could improve product innovation.

However, they did not provide a specific method for module forming.

Cost saving is another important sustainability indicator. MPD can generate savings in sup-

ply chain costs and product development costs. It may lower the supply chain cost as it can reduce

inventory cost and save distribution time (Feitzinger and Lee, 1997; Ernst and Kamrad, 2000;

Ulrich and Eppinger, 2000; Kamrani and Salhieh, 2008). However, the way in which MPD lowers

the cost of product development and production is a bit obscure. It needs to make the connection

through product variety. The increasing demand for product variety from customers forces com-

panies to concentrate on product families rather than single products. However, high product va-

riety can greatly increase administrative and manufacturing costs due to the need for more special-

ized materials, process and quality control ways (Lee and Billington, 1994). MPD is one of the

27
approaches that can be used to save cost and still offer a variety of products (Kim and Chhajed,

1999).

Ernst and Kamrad (2000) proposed a conceptual framework for evaluating product supply

chain structures through product modularization and postponement. They defined modularization

as inbound logistics, which combines different components or modules to assemble the final prod-

uct. Postponement is associated with outbound logistics because it is the distribution function that

specific customer demand is satisfied. This definition of modularization necessitates MPD. There-

fore, based on the classification of modularization (inbound logistic) and postponement (outbound

logistic), they introduced a framework of supply chain as shown in Fig. 2-2.

Figure 2-2 Framework of Supply Chain Structure. M: Manufacturing; A: Assembly; P: Packing


(Adopted from Ernst and Kamrad, 2000)

In Figure 2-2, RIGID represents a traditional vertical supply chain structure where the ob-

jective is maximizing economy by keeping a high level of finished products inventory. FLEXIBLE

represents an opposite supply chain structure, which produces and stores different components and

then assembles different entire products in response to specific demands. These two supply chain
28
structures show two opposite and extreme cases. POSTPONED and MODULARIZED supply

chain structures represent the intermediate cases. Modularized structure means multiple sources

for components, and assembly output is the finished product; postponed structure involves a single

source for components, and assembly outputs are developed according to different demands.

Ernst and Kamrad (2000) used total cost to evaluate these four supply chain structures by

building a mathematical model. Their results show that different supply chain structures are ef-

fective in different cases. Even though they proposed a framework of supply chain structure and

evaluated using a total cost model, their evaluation model is limited to comparing only vertical or

horizontal supply chain structures. In addition, their framework is summarized based on empirical

studies.

Kim and Chhajed (2000) developed a model to determine when a modular product should

be introduced and how much modularity should be offered. They used MPD as an approach for

product platform design to improve commonality in product variants. They posited that higher

commonality will lead to product cannibalization as well as lower cost, because commonality

might decrease production costs but will also make the products more indistinguishable from one

another. A model was built to trade off cost savings and product cannibalization through two types

of cannibalization problems associated with MPD for two product classes: (1) low market segment,

which has low market share, and (2) high market segment, which has high market share. High

commonality products correspond to low market segment and low commonality products relate to

high market segment. Using the model, Kim and Chhajed provided a framework to find the com-

monality condition that helps the firm to increase revenue the most while showing product simi-

larity would cause product cannibalization. Even though Kim and Chhajed’s (1999) work provides

29
a tradeoff between modularity level and market performance, their work is based on two assump-

tions: first, a product can be represented by a single dimension “quality”; and second, the total

market is divided into high segments and low segments, and there is no middle segment between

them. Therefore, a more general case which represents a segment between low segment and high

segment should be taken into account in order to analyze completely.

Profit is another critical theme of sustainability associated with MPD. Profit is related to cost

in that it is equal to total revenue minus total cost; therefore, cost saving is generally a benefit to

profit increasing. However, profit increasing is not only due to cost saving. Thus, profit and cost

saving are separately discussed.

Fixson and Park (2008) showed the relationship between product architecture, innovation

and industry structure using the case of development of a bicycle drivetrain system. They added

integrating as a new design operator in product architecture, as well as splitting, substitution, aug-

menting, excluding and inverting (Baldwin and Clark, 2000). From common knowledge, product

architecture migrates to more modular structures in the long run; however, Fixson and Park’s work

shows a reverse process: the product architecture tends to become integral in the long run, and the

corresponding industry tends to be a monopoly. Integration over the long run improves industrial

competition and profit. Fixson and Park’s (2008) work shows a distinct conclusion about modu-

larity and industry competition over the long run, which brings an additional concern when dis-

cussing product architecture and competition. However, their work is based only on one industry

and is particular to one case study.

Das and Chowdhury (2012) proposed a reverse supply chain planning process associated

with MPD. Their work is based on two MPD advantages: reduced lead time (Kamrani and Salhieh,

2008) and ease of manufacturing (Okudan et al., 2012). They considered return products collection

30
and the recovery process in the total supply chain. They identified the return products as products

after use (EOL or before EOL), products returned under warranty, defective products, obsolete

products returned by retailers, and products returned by customers. The collections of returned

products are derived from third party logistics, wherein collection centers are opened by remanu-

facturer/manufacturer and retailers. The recovery process is determined by original manufacturers

or third party logistics. Since returned products can enter the market after recovery, the reverse

supply chain organizes products into three categories: (1) products that use all new compo-

nents/modules; (2) products that use a mixture of new and recovered components/modules; and (3)

products that use only recovered components/modules. Das and Chowdhury then maximized the

total profit through mixed-integer programming.

Das and Chowdhury’s (2012) model provides three different products to markets with re-

covered components/modules consideration. This contribution fits very well in the real market.

Moreover, they considered the reverse logistics from the perspective of the total supply chain.

However, they considered the retailer as the most efficient collection option and third party logis-

tics as the most effective recovery option and there may be a negative correlation between these

two options. In addition, they assumed that product cost has a positive relationship with module

numbers, so the correction of assumption may depend on the number of modules.

Mukhopadhyay and Setoputro (2005) treated MPD as a competitive tool for build-to-order

(BTO) products because MPD can bring benefits to a customer return policy as well as reduce lead

time and production cost. They considered three factors in their BTO model to maximize profit—

return policy, modularity level, and product price—and focused mainly on the first two factors.

They believed that customer demand is positively related to profit; therefore, they used customer

demand as an intermediary to build the relationship between the three BTO factors and profit.

31
From their model, they suggested several cases when both return policy and modularity level

should increase: 1) the market is more sensitive to return policy or modularity, 2) the sellers want

to decrease product development cost, 3) the sellers want to salvage more, and 4) the constant

market price is increasing.

Mukhopadhyay and Setoputro’s (2005) work provides guidelines to help increase profit in

the BTO business. However, their model is empirically derived and the demand function is a linear

function of return policy, modularity, and price. The corresponding conclusions are based on the

assumptions that return policy, modularity level and price are all positive in relation to customer

demand. The assumptions need proof and verification. In addition, they set the market price as a

constant.

2.3.3 Product Lifecycle Management Related Modular Product Design

Life cycle analysis is regarded as one of the most important and efficient sustainability man-

agement tools. Life cycle management-related MPD considers how MPD affects the life cycle,

including the product design stage, product updating, and product EOL options. Many MPD-based

tools have been adopted for use in this area such as green design, module updatability, design for

End of Life Options (DfEOL) and reverse supply chain design. Life cycle management involves

two parts: the whole life cycle and the EOL options. The whole life cycle refers to all phases that

a product goes through, including design, manufacturing, assembly, service, maintenance, and re-

cycling, etc. (Umeda et al., 2008; Tseng et al., 2008; Go et al., 2015). EOL options address only

the last phase of the life cycle, generally including reuse, recycle and disposal (Li et al., 2008). As

reviewed in this section, certain research discusses how MPD affects the entire life cycle (e.g.,

Bryant et al, 2004; Kreng and Lee, 2004b; Ji et al., 2012); other research focuses on how to deal

32
with retired products (e.g., Li et al., 2008; Lai and Gershenson, 2009; Yan et al., 2012; Ji et al.,

2013), the details will be shown in the left part of this section.

Gonzalez and Adenso (2005) developed a bill of materials based approach to determine

product / component end-of-life strategy. They considered reuse, recycling and remanufacturing

as three good end-of-life options for reducing environmental problems from landfills. The meth-

odology was built based on product structure (obtained from the bill of materials), and the joining

and geometrical relationship among components (obtained from 3D CAD representation). The

proposed approach addressed not only the product / component end-of-life option, but also the

most profit disassembly sequence. The scatter search meta-heuristic was employed to determine

the disassembly cost at each level of bill of materials. Overall, Gonzalez and Adenso’s (2005)

work presented a new end-of-life option determination approach with the bill of materials consid-

eration, which is a new angle of design for environment. They also applied scatter search meta-

heuristic algorithm to estimate disassembly cost. However, the scatter search requires more calcu-

lation work, and some other alternative algorithm might be better in terms of cost evaluation.

Li et al. (2008) proposed a fuzzy graph based MPD method involving product life cycle

consideration. They summarized the relationship between life-cycle performance and the modu-

larity of a product into three objective levels in Figure 2-3. The primary objective relates to the

aggregate value, which represents the total life-cycle performance of the modular product. The

second level objectives are divided into four sub-items: disassembly, reuse, material selection, and

serviceability. Further, second-tier objectives are decomposed into lower level criteria, such as

energy and standard time in disassembly, human factors and facility factors in serviceability at the

third level. Based on these, Li et al. defined an index for each third level sub-objective and applied

Analytic Hierarchy Process (AHP) to aggregate the index of second level sub-objectives and to

33
generate a life cycle performance index. Then they set up the objective functions, where one min-

imizes inter-cluster distance and the other maximizes extra-cluster distance by taking the index

information and fuzzy graph product representation into account, and then applied an adopted K

ordered greedy clustering algorithm to form modules.

Li et al.’s (2008) work provides a framework of design for environment (DfE) based on

fuzzy graph application. However, when calculating the index of life cycle performance, they did

not rank the four second-level sub-factors, which means that all four have the same importance. In

addition, they did not provide the calculation of the serviceability index.

Figure 2-3 Life Cycle Objective Structure for Modular Design (Adopted from Li et al., 2008)

Lai and Gershenson (2009) proposed an MPD method based on the perspective of the prod-

uct retirement process, represented by two aspects: similarity and dependency. They considered

post-life intent (e.g., recycle, reuse, dispose), material compatibility and components connection

34
type/disassembly direction as factors to quantify similarity sequentially; and they considered ac-

cessibility, disassembly force, positioning, tool requirements, material handling and fastening to

quantify dependency. Based on the quantification of similarity and dependency, Lai and Gershen-

son set the objectives as maximizing component similarity in each module and minimizing com-

ponent dependency out of the modules, and applied a DSM-based clustering method to form the

modules.

Lai and Gershenson’s (2009) work quantified similarity and dependency by forming a design

matrix which shows a metric view for evaluating the product retirement process. However, for

quantification of similarity, they did not give a specific method to rank the importance order of

three factors (post life intent, material compatibility, and components connection type/disassembly

direction). In their analysis, they assumed post-life intent to be more important than material com-

patibility, which in turn is more important than component connection type/disassembly direction.

Therefore, based on their method, the evaluation of similarity may yield several different results.

Bryant et al. (2004) presented an MPD-based redesign tool which both reduces part count

and improves the life cycle impact of a product. They combined functional-based MPD with an

Elimination Preference Index (EPI) metric to measure the life cycle impacts of a design. Their

work considered six life-cycle factors (assembly time, part necessity, ease of component handling

and manipulation, ease of component insertion for assembly, recyclability, and dismantleability)

to quantify and assess the life cycle. They calculated EPI values to generate an EPI metric. Based

on their work, components with high EPI values were deemed to be candidates for elimination,

while low EPI value components were deemed to present the potential to reduce any environmental

impact.

35
Bryant et al.’s (2004) work applies EPI to measure and assess life-cycle impacts and to quan-

tify the disassembly of a product. However, their six life-cycle factors are summarized based on

an empirical study which may change for different products. In addition, while the method works

well for products with small numbers of component products, more details are required for prod-

ucts with larger numbers of components.

Kreng and Lee (2004b) proposed a QFD and linear integer programming (LIP)-based MPD

method. They asserted that all MPD methods should consider some of the following modular de-

sign drivers as objectives: carryover, technology evolution, planned product changes, standardiza-

tion of common modules, product variety, customization, flexibility in use, product development

management, styling, purchasing modularity components, manufacturability refinement and qual-

ity assurance, quick services and maintenance, product upgrading, recycling, reuse and disposal.

Using QFD, Kreng and Lee divided MPD into two phases: modular driver selection and modular

design. In the first phase, customer requirements, company requirements, and design requirements

are transferred and summarized as modular design drivers. In the second phase, Kreng and Lee

derived linear integer programming (LIP) based on two relationships—modular driver and com-

ponents, and component and component—and the idea that a modular product should maximize

similarity within each module and minimize interaction among modules. Then from the calcula-

tion of LIP, modules are formed.

Kreng and Lee’s (2004b) work aggregates market requirements and competitive strategies

into the modular design which originally includes only physical and functional relationships in

terms of modularity drivers. However, QFD is not a perfect method for use to collect final cus-

tomer needs data because the simple responses of important customers to survey questions seldom

yield completed satisfaction input.

36
Yan et al. (2012) proposed a new sustainability-oriented MPD method. The authors took

several sustainability issues into account such as society, economic, environment, material, man-

ufacturing and end-of-life options. A quantified sustainability assessment method was provided to

constrain design criteria. Table 2-7 shows the assessment attributes. A kernel-based fuzzy c-means

algorithm was used to integrate components of a product into different modules based on their

correlation distance. A genetic algorithm was employed to determine the optimal clustering num-

ber based on its efficiency in coming up with the global solutions.

Modular Product’s Mainly Life Cycle Stages


Material Manufacturing End-of-life
Renewable or Technology Reuse
Environment

Nonrenewable
Material Process Process Recycle
Recyclability Energy Used Remanufacturing
Redesign
Sustainability Criteria

Disposal
Raw Material Production Cost Reuse Cost
Cost
Economic

Labor Cost Energy Cost Recycle Cost


Packaging Cost Remanufacture
Cost
Transportation Redesign Cost
Cost
Worker Health Worker Health Reuse
Society

Safety Safety Remanufacture


Recycle
Redesign

Table 2-7 Sustainability Issue & Modular Product Life Cycle Stage
(Adopted from Yan et al., 2012)

Yan et al.’s (2012) work combines several sustainability factors into MPD. However, their

corresponding scores are based on an empirical study.

Ji et al. (2012) proposed an effectiveness-driven modular design method in order to solve

the problem posed by the requirement of different module forms due to diverse design objectives

of different phases. The effectiveness of a module is regarded as the ability to which a module

37
fulfills the expected objective at a certain phase of the entire life cycle. The life cycle used in this

research includes design, manufacturing, and assembly, service, maintenance, recycling. Ji et al.

took all possible effectiveness scenarios for all life cycle phases into account and balanced the

granularity and compositions of each module during the clustering process. They used a product

descriptive model composed of components, interaction attributes, and liaison graph as the basis

for an effectiveness-driven modular design. The interaction attribute shows the relationship be-

tween two components. Table 2-8 summarizes all interaction attributes in the life cycle. A liaison

graph represents a product structure that includes all interaction attributes. They applied a quanti-

tative split method for liaison structure to cluster components into modules. Since all effectiveness

scenarios were considered, they used three aggregation rules to maximize and finalize the effec-

tiveness of the modules.

Table 2-8 Attribute list for all life cycle (Adopted from Ji et al., 2012)

Phase Interaction Attributes


Transverse seriating level; longitudinal seriating level;
Design
upgrade level; customer participation level
Manufacturing & Assembly Manufacturing location; assembly location; process
Service Value lifetime; upgrade level
Value lifetime; maintenance frequency; maintenance
Maintenance
technology level
Material compatibility; value lifetime; material value;
Recycling
eco-indicator; processing mode

Ji et al.’s (2012) work combines all life cycle phases and tries to derive an appropriate mod-

ule forming method. However, they did not provide any ranking method or guideline for weights

determination in each phase, and the attributes were randomly chosen. In addition, for the last

phase of the product life cycle, they only considered recycling; reuse and disposal should also be

considered.

38
Tseng et al. (2008) proposed a green life cycle-driven MPD method. They pointed out that

product use and recycle or disposal will largely determine green life cycle. To conduct the research,

they applied a liaison graph model with four types of engineering attributes involved to represent

product structure: contact type, combination type, tool type and accessed direction. Based on the

liaison graph, they derived an index of liaison intensity (LI) to show component relationships.

Then, Tseng et al. adapted Falkenaur’s (1998) grouping genetic algorithm (GGA) to cluster com-

ponents into modules based on liaison graph information. The GGA can overcome several limita-

tions of the traditional genetic algorithm (GA); for instance, module numbers and module size can

be set previously. Tseng et al. also balanced the green design and cost issue by comparing pollu-

tion value and total cost. The pollution value was calculated using a Simapro-based eco-indicator.

The goal of green product design was shown to be achievable by replacing materials and updating

liaison intensity continuously from the initial module results in GGA.

Tseng et al.’s (2008) work provides the framework of green life cycle based MPD, and de-

signers can apply this framework to reduce assembly time and to control life cycle cost. However,

their approach is based on replacing and updating components’ materials and structure iteratively,

which is not effective. In addition, like other methods which use GGA (e.g., Kreng and Lee, 2004a;

Meehan et al., 2007), when the number of components in product increases, computational time

increases dramatically.

Umeda et al. (2008) proposed an MPD methodology with aggregated product life cycle-

related attributes and component/module geometric feasibility. They employed self-organizing

maps (SOM) to integrate product life cycle attributes, such as life cycle options (LCOPs), materials,

physical lifetime, etc. The SOM will cluster components into groups based on similarity of life

39
cycle attributes. Umeda et al. took geometric feasibility into account by introducing several in-

dexes such as combination, density, and connection. These indexes show the relationship between

components in the same group which is formed by SOM.

Umeda et al.’s (2008) work takes both product life cycle and component/module geometric

feasibility into account in the module forming. However, their method still requires designers to

set module numbers in advance, which may not optimize the module design.

Based on Umeda et al.’s (2008) MPD methodology, Umeda et al. (2009) provided a life

cycle evaluation method for MPD from the perspective of resource efficiency in order to clarify

the effectiveness of modularity on environmental consciousness. They introduced the Resource

Efficiency (RE) index, which represents the resource time length of utilization per amount of re-

sources consumed during the whole product life cycle, to evaluate the modular structure. Since life

cycle options have a great impact on modular structure forming, and since some external factors

such as labor cost in a different region will affect the life cycle options, there is an equilibrium

between life cycle cost and environmental load. Umeda et al. (2009) introduced probability for the

life cycle options of modules to represent this equilibrium. They employed the probability function

for each module and derived the RE value for the product. The higher the RE value is, the better

life cycle a product has. Therefore, designers can improve the RE value by removing or inserting

components to modules.

Umeda et al.’s (2009) work provides a useful evaluation method for the modular design with

life cycle consideration. However, their work is based on the assumption that a product’s lifetime

is equal to the lifetime of the component with the shortest life in the product. In addition, they

emphasize only on components in modules and ignore single components.

40
Smith and Yen (2010) employed the atomic theory-based MPD method from their earlier

work (Yen and Smith, 2009), and added green design constraints such as material compatibility,

part recyclability and part disassemblability to form a new green design method. In the atomic

theory-based MPD, a module represents a subassembly of a product and it is formed by consider-

ing components’ spatial locations, structures, and lifecycle options. Smith and Yen built a touch

matrix based on atomic theory to show the geometric relationship between two components in

which 0 means no relationship and 1 means they connect with each other, and then defining green

objectives as design constraints, they formed modules.

Smith and Yen’s (2010) work provides a new atomic theory-based green MPD method

which overcomes several limitations of GGA and DSM MPD, such as dramatic computational

time increases that result from increasing the number of components. However, the touch matrix

simply represents components connection relationship by binary numbers, which cannot show how

tight two components connected. Therefore, one of the possible improvements for this method is

replacing binary numbers by numerical numbers.

Ji et al. (2013) developed an MPD related methodology to facilitate life cycle material effi-

ciency by considering component material reuse and minimizing resource commitment throughout

product realization process. They emphasized on leader-follower joint optimization and leveraged

technical system modularity (TSM) and material reuse modularity (MRM), and therefore proposed

a comprehensive framework. They employed modularity metrics taxonomy to measure component

interaction and grouped into modules accordingly. Multi-attribute utilities of different dimensions

of component similarity were used to quantified and aggregated modularity measures. A bi-level

constrained genetic algorithm was put forward for the joint decisions of TSM and MRM.

41
Ji et al.’s (2013) work is motivated by material efficiency improvement from the view of the

life cycle, which is different from most traditional MPD methods’ low cost concentration. The

only possible drawback is that genetic algorithm might require huge computation efforts when the

component number increases.

Koga and Aoyama (2008) investigated how to balance the long life cycle of product and

market changes using modules. They did not adapt traditional methods such as modular updating

to conduct the research. Instead, they tried to predict the market change in the life span of a product

family and based on an estimation of the market, they developed a modular product family design

method. They assumed that market change is modeled as the quantity change of sold products, and

they set life cycle and product sale scenarios as input for product family design in order to draw a

product family graph.

Koga and Aoyam’s (2008) work aggregates product life cycle and market change issues in

the product family modular design. However, they focus on product family graph drawing and

ignore the challenge of how to predict market change and how to form modules specifically.

Cebon et al. (2008) conducted research on the impact of product modularity on product life

cycle. They noted that many product life cycle methods, such as technical innovation related, mar-

keting related, strategy related and product development related, are based on the assumption that

products are integrated wholes. However, modularization of a product undermines specific syner-

gies and aggregation. Cebon et al.’s argument suggests that the product life cycle will be dramati-

cally altered and attenuated if a product has full non-specific synergy between modules and sub-

systems, even though other effects, such as product brand, market etc., still work for the product

life cycle.

42
Cebon et al.’s (2008) work investigates the relationship between product modularity and

product life cycle. However, their work is based only on an empirical study and former literature,

and therefore, the qualitative conclusions need to be approved by case studies and improved by

quantitative analysis.

Chung et al. (2011) proposed a robust modular architecture methodology based on life cycle

assessment from the perspective of the supply chain. They evaluated the life cycle from the view

of life cycle cost (LCC) and life cycle energy consumption (LCEC) in a closed-loop supply chain.

Chung et al. initialized modular architecture based on the physical connectivity graph, which co-

vers component attributes (Vertices), and component interactions (Edges). A supply chain optimi-

zation model was set up and could be used to evaluate all possible modular architectures in terms

of LCC and LCEC. Starting from an initial modular architecture, a robust modular structure could

be derived by evaluating the supply chain model heuristically. Chung et al.’s work provides an

optimal modular forming method based on the evaluation of supply chain and sustainability. How-

ever, the heuristic method may require considerable computational efforts.

Shin et al. (2011) proposed a product concept selection methodology based on extended

QFD and mixed integer non-linear programming (MINLP) with consideration of life cycle and

resource allocation. They extracted data on customer requirements and product life cycle require-

ments from an extended QFD and then used it to represent MINLP constraints. The objective of

MINLP is to maximize product design satisfaction, which depends on the degree of design quali-

ties for engineering characteristics including customer and product life cycle requirements from

the extended QFD. Investment budget and resource allocation involving a number of human re-

sources or software licenses are also constraints. A greedy algorithm and a net search algorithm

are used to find the optimal solution for MINLP, which provides a robust product concept.

43
Shin et al.’s (2011) work combines an extended QFD and MINLP to search for the optimal

product concept by considering both life cycle and resources. However, since there are several

uncertainties at the conceptual stage, certain optimal values derived from this method may not

make sense, such as investment cost. In addition, the proposed approach may be regarded as a one-

shot approach for generating an optimal product concept.

The product life cycle related papers are compared and summarized in Table 2-9.

Table 2-9 Product Life Cycle & Product EOL Researches


EOL Re-
Author/
Life Cycle Stages/Attributes Life Cycle Research Method search
Year
Method
Assembly time + Part necessity + Ease Elimination Preference Index Not Specify
of component handling and manipula- (EPI)
Bryant et al.
tion + Ease of component insertion for
(2004)
assembly + Recyclability + Dismantle-
ability
Kreng and Not Specify Module Drivers + QFD + LIP Not Specify
Lee(2004b)
Li et al. Disassembly + Recycle/reuse/disposal Fuzzy Graph+Fuzzy AHP+K Fixed EOL
(2008) + Material Selection + Serviceablility Ordered Greedy Algorithm Options
Product Use + Disposal/Recycle Liaison Graph Model + Liaison Fixed EOL
Tseng et al.
Intensity (LI) + Group Genetic Options
(2008)
Algorithm
Umeda et al. Life Cycle Options (LCOPs)+Materi- Self-Organizing Maps (SOM) Not Specify
(2008) als+Physical Lifetime
Koga and Not Specify Product Life Cycle + Market Not Specify
Aoyama Change
(2008)
Cebon et al. Not Specify Product Life Cycle + Product Not Specify
(2008) Modularity
Lai and Post life intent + Material compatibility Max Component Similarity + Fixed EOL
Gershenson + Components connection type/disas- Min Component Dependency Options
(2009) sembly direction
Umeda et al. Life Cycle Options (LCOPs)+Materi- Resource Efficiency (RE) Not Specify
(2009) als+Physical Lifetime
Smith and Material Compatibility+ Part Recycla- Atomic Theory Not Specify
Yen (2010) bility+Part Disassemblability
Manufacturing + Maintenance + End of Life Cycle Cost (LCC) + Life Fixed EOL
Chung et al.
Life Cycle Energy Consumption Options
(2011)
(LCEC)
Shin et al. Not Specify QFD + MINLP + Greedy Algo- Not Specify
(2011) rithm + Net Search Algorithm

44
Yan et al. Material + Manufacturing + End of Kernel based fuzzy c-mean al- Fixed EOL
(2012) Life gorithm Options
Ji et al. Design+Manufacturing&Assembly+ Component Liaison Structure + Not Specify
(2012) Service + Maintenance + Recycling Quantitive Split Method

From Table 2-9, the papers take the life cycle and/or the EOL into account, and combine

analysis with mathematical programming or other related methods to assist designers during the

design stage. However, there is a premier assumption among them: the life cycle or EOL options

are fixed. This assumption does not fit global practice very well due to geographic issues. There-

fore, this study will consider uncertainty in EOL during the design stage and aggregate it with

MPD to form a new modular design method.

2.4 Literature Review Summary

In this chapter, EOL options definitions and determination methods, traditional MPD meth-

ods based on two different categorizations, and sustainability related MPD were reviewed. The

traditional MPD methods are reviewed from the perspective of pure theory, and sustainability re-

lated MPD methods are the application of MPD focus on sustainability. From the literature review,

two main research gaps in this area are able to be summarized:

 Most of the MPD papers do not specify key components in design stage and none of them

optimize MPD based on key components;

 None of the MPD papers consider uncertainty in EOL options.

These two gaps have bad influences to both academia and industries. No concentration of

key components in the module forming will cause two key components in the same module, which

may increase the assembly difficulty and cost when doing the maintenance. For example, com-

pressor and control unit in the refrigerator are expensive for both manufacturing and assembling.

45
If two components are in the same module, when we maintain one of them, we need to disassembly

and assembly another one, which costs more compared with they are in separate modules. The

proposed clustering algorithm could identify key components in the algorithm initialization step.

Components interactions in the same module and sustainability values optimization have been

considered in algorithm development. The algorithm could be not only used in module structure

forming, but also extended to supply chain network optimization, such as transportation optimiza-

tion and warehouse allocation problem. Ignoring the uncertainty in EOL options will cause wrong

product strategy. For example, the bicycle frame EOL options are different when putting into dif-

ferent markets. To fill these two gaps, we will develop a new MPD method. This new method

could specify key components from the beginning of design, and provide the optimal product mod-

ules structure. In the meantime, the new method takes the uncertain EOL options into account in

the product design stage.

46
Chapter 3

Methodology

In this chapter, a new sustainable MPD approach with uncertain EOL options consideration

and key components specification will be proposed. By addressing the research questions pre-

sented in Chapter 1 (Can the new modular product design method specify and concentrate on the

key components in the design process? Can the new modular product design method take the un-

certain EOL into account in life cycle sustainability? Can the new modular product design method

improve full-dimension sustainability performance with key components specification and uncer-

tain EOL options consideration?), the methodology is introduced.

For considering uncertainty in EOL options, two analytic methods are potential candidates:

probability theory and fuzzy logic. In our methodology, we opt to apply fuzzy logic for two reasons:

(1) vagueness of design environment is an issue, and (2) a probability model requires specificity

and it is almost impossible to determine the correct probability in the early design stage. For sus-

tainable design, according to Rodriguez et al. (2002) and Zamagni et al. (2009), sustainability

covers three dimensions, each of which appears in Table1-1: the environment, the economy, and

society. The new MPD approach involves all of them. The complete life cycle involves the man-

ufacturing process, the maintenance process, and the EOL process. The methodology employs

measures of life cycle cost (LCC), life cycle environmental impact (LCEI) and life cycle labor

time (LCLT) to evaluate modular design impact in terms of environmental, economic, and societal

sustainability. Also, since maintenance depends largely on component failure frequency, the

failure rate has been taken into account in the maintenance process, using fuzzy logic to accom-

modate imprecise and inexact environments. For module forming, three key components based

clustering algorithm have been developed to group product components. The tri-criteria involve

47
minimizing LCC, minimizing LCEI and maximizing LCLT. Figure 3-1 shows the flow chart of

methodology.

The structure of this chapter divides into three sections: 1) fuzzy representation; 2) sustain-

ability model setup based on life cycle analysis; 3) tri-criteria integer programming model.

Figure 3-1 Methodology Flow Chart

3.1 EOL Options Fuzzy Representation

As a result of the increasing population growth and increasing product requirements, a huge

quantity of retired products has been generated. These have been found to cause sustainability

problems, such as environmental pollution, economic waste and, adverse effects on human health.

Due to the significance of these retired products, researchers have investigated how to determine

and select EOL options to address and meet practical requirements. One widely used method

involves applying multi-criteria decision making in EOL selection (e.g., Lee et al., 2001; Bufardi

48
et al., 2004; Remery et al., 2012). This stream of research adopts MCDM methods to balance

several design criteria and reaches a compromising EOL alternative; the advantage of this method

is that it can take many design objectives into account. Another set of methods applies pure math-

ematical algorithms to select optimal EOL alternatives (e.g., Erdos et al., 2001). However, many

qualitative and quantitative methods for determining an appropriate EOL option suffer from

vagueness due to either the use of incomplete data sets or the unavailability of data expressed as

exact numbers (Yang and Li, 2002). Therefore, linguistic assessment is recommended instead of

a methodology requiring exact data (Beach et al., 2000). Fuzzy numbers and membership functions

are widely used in the linguistic expression. In order to overcome the ambiguity in linguistic as-

sessment, triangular and trapezoidal membership functions can be developed (Delgado et al., 1993).

These membership functions are used to transform the linguistic variables into fuzzy numbers

(Singh et al., 2006). This study uses Fuzzy logic to represent EOL options.

As discussed at the end of Chapter 2, it is difficult to specify EOL option for each component

in the design stage. However, it is easy to find all possible EOL options information for each

component easily, such as cost, environmental impact for each EOL options regarding each com-

ponent. Therefore, one of the possible ways to deal with uncertainty in EOL options is using ex-

pected value. Expected value is calculated by multiplication of each EOL options information and

corresponding weights or possibility. Figure 3-2 presents EOL expected value calculation. Hence,

the difficulty is transferred from uncertainty in EOL options to uncertainty in EOL options’

weights.

49
Figure 3-2 EOL Option Expected Value Calculation

3.1.1 EOL Options Definition

Before representing EOL options via fuzzy logic, EOL options should be identified. In Chap-

ter 2’s literature review, there are several EOL options definition methods. This study opts to adopt

Marco et al. (1994)’s definition because EOL options are able to be quantified based on the devel-

oped calculation formula by Lee et al. (2001). Seven options and their descriptions are shown in

Table 3-1.

Table 3-1 EOL Options Definition (Adopted from Marco et al., 1994)

EOL Options Definition


Reuse Using in the same (direct reuse) or another (indirect reuse) application.
Retaining serviceable parts, refurbishing usable parts, or replacing identical
Remanufacture
or reworked components from obsolete products.
Reprocessing a material into a form that can be used in the same or another
Primary Recycle
“high” value product.
Secondary Recy- Reprocessing a material into a “low” value product, e.g., industrial grade
cle rubber being reprocessed into a general grade rubber.
Incinerated Incinerating a material to produce heat and electricity.
Landfills Landfilling waste products with no intrinsic value.
Special Handling Mandatory for all toxic or hazardous materials.

50
3.1.2 EOL Options’ Fuzzy Characteristics

En route to defining EOL options in fuzzy representations, one needs to determine fuzzy

characteristics. There are several fuzzy characteristics available (Rose, 2001; Bufardi et al., 2004;

Li et al., 2008). Although these may relate to corresponding EOL definitions, they may not work

in this study’s definition. Therefore, new fuzzy characteristics for this study’s EOL options should

be re-defined. It is evident that four streams of characteristics are needed to evaluate EOL options.

One stream must demonstrate the characteristics of the inside of the component, which is material

(e.g., how does a hazardous material affect EOL option selection?) One stream must describe the

characteristics of the component itself (e.g., how does component’s remaining life cycle affect

EOL option selection?) One stream should discuss the relationship between components in the

product (e.g., how does component disassembly affect EOL option selection?) Finally, there

should be another stream which takes external impacts into account (e.g., how does customer pref-

erence affect EOL option selection?). Figure3-3 shows the relationship among these four streams.

Figure 3-3 Characteristic Relationship


Based on analysis of Tables 2-2, 2-3, 2-4 and Remery et al. (2012), new fuzzy characteristics

are shown in Table 3-2 and EOL evaluation characteristics are shown in Figure 3-4. EOL options’

weights are calculated based on EOL evaluation characteristics fuzzy assessment.

51
Table 3-2 EOL Evaluation Characteristics and Description

Characteristics Description
Component Quality Component durability, component EOL condition
Functional/Module Complexity Difficulty level of assembly, disassembly, etc.
Material Material properties and related characteristics
External Market acceptance and customer preference

Figure 3-4 EOL Evaluation Characteristics


Each characteristic has been explained following by modifying from Remery et al. (2012) and
making necessary extension:

Component Durability: component’s residual value after it is first used, and defined as the ratio

between the component’s wear-out life and the product’s useful life, which has to be as high as

possible, so that component can still be functional after the product’s EOL and be reused.

52
Component EOL Condition: The state of the component at the EOL of the product. The EOL

condition is associated with the component’s reliability and the circumstances in which it was used

during its lifetime, and indicates the possibilities for reuse and the potential remanufacturing cost,

environmental impact and labor time required.

Difficulty with the Component’s Disassembly: Takes into account the overall access to the com-

ponent, as well as the number and type of attachments that have to be dismantled in order to dis-

sociate them from the product. This is directly related to component’s disassembly cost, environ-

mental impact and labor time used.

Level of Integration: Linked to the number of functions realized by component and represents its

complexity. The higher the number of functions performed by component, the greater the level of

integration will be. The characteristic is used when determining the remanufacturing sustainability

values.

Quantity of Parts: Defined as the in-dissociable elements of the component that perform one or

more functions, except the connection function. Generally, the quantity of parts is an indication of

the component’s disassembly cost, environmental impact and labor time used, and therefore is

related to sustainability values of remanufacturing and primary recycle.

Difficulty of Dismantling Part Attachments: Refers to the sustainability values of dismantling the

connectors that link the various parts of the component. This parameter is involved in component’s

part disassembly sustainability values, which has to be taken into account in remanufacturing and

primary recycle in terms of social sustainability.

Component Weight: Influences landfill and incineration.

53
Quantity of High-Value Materials: These can be resold at a high price after recovery. Materials

with a very high value, like gold, palladium, and silver, are considered as precious materials. Com-

ponent made with these materials are usually primary recycle. Other materials that can easily be

resold are special metallic alloys (e.g., copper, aeronautic aluminum, iron), some plastics (e.g.,

PEE, PC, PM, ABS), and glass.

Material Calorific Capacity: When the value is high, it is preferable to incinerate the component,

rather than dispose of it in a landfill, because this will permit the recovery of a substantial amount

of energy. Generally, when calorific capacity is higher than 8 MJ/kg, the option of incineration is

preferred over landfill disposal.

Amount of Hazardous Materials: Hazards are governed by laws and regulations which differ from

one country to another. Generally, regulations stipulate that hazardous component must be disas-

sembled and treated separately.

Amount of Different Materials: A component with few materials can be easily recycled since its

treatment will require less separation effort. Another important benefit, with an appropriate sepa-

ration method, a significant amount of particular materials will be recovered and potentially resold.

Regulation Support: Some component with special materials will adapt primary or secondary

recycle since some specific regulation has been issued with environment protection consideration.

Customer Preference: This is positively related to component economic performance, which

means such component will always be reused due to the preference of the customer.

Succedaneum Price: This characteristic is negatively related to component economic performance,

which means the higher the succedaneum price is, the lower reuse rate component has.

54
3.1.3 EOL Options Fuzzy Evaluation

To capture the uncertainties associated with EOL options’ weights, this study gets through

evaluation of EOL characteristics by fuzzy logic. The fuzzy set theory provides a technical basis

to evaluate and derive an approximate conclusion (Bellman and Zadeh, 1995). These fuzzy varia-

bles are also known as linguistic variables in fuzzy set theory (Zadeh, 1965). This study measures

the EOL characteristics via a rating scale, and the ratings are measured at five levels: “very low”,

“low”, “moderate”, “high” and “very high”. Table 3-3 shows the linguistic variables and their

related fuzzy levels.

The uncertainties of fuzzy numbers are evaluated using membership functions. Widely used

membership functions include triangular and trapezoidal membership functions (Delgado et al.,

1993). The advantages of triangular membership functions are their simplicity and the fact that

they are commonly used in product development analysis (Shehab and Abdalla, 2001; Yadav et

al., 2002). Figure 3-5 shows a triangular membership function.

This study evaluates every fuzzy characteristic in order to get each EOL option’s score.

Therefore, the fuzzy input is fuzzy characteristic evaluation and fuzzy output is a score of EOL

option. All of them come up with fuzzy set. By normalizing the EOL option scores, EOL option

weights will be obtained. The triangular membership function is used for both inputs and output.

The linguistic variable levels and corresponding fuzzy sets are shown in Table 3-4.

Table 3-3 EOL Characteristics Linguistic Variables and Level Used

EOL Characteristics EOL Sub and Sub-Sub Char-


Level Used
Linguistic Variables acteristics Linguistic Variables
Component durability; compo- Very low; low; moderate;
Component Quality
nent EOL condition high; very high

55
Difficulty with component’s
disassembly; level of integra-
Function/Module Very low; low; moderate;
tion; quantity of parts; diffi-
Complexity high; very high
culty of dismantling part at-
tachment; component weight
Quantity of high-value materi-
als; material calorific capacity;
Very low; low; moderate;
Material amount of hazardous materi-
high; very high
als; amount of different mate-
rials
Regulation support; customer Very low; low; moderate;
External
preference; succedaneum price high; very high

Figure 3-5 Triangular Membership Function


Table 3-4 Linguistic Variable Level and corresponding Fuzzy Set

EOL Characteristic Fuzzy Score (0~10)


Linguistic Variable Level Fuzzy Set
Very Low (Very Poor) (0, 0, 2.5)
Low (Poor) (0, 2.5, 5)
Moderate (Fair) (2.5, 5, 7.5)
High (Good) (5, 7.5, 10)
Very High (Very Good) (7.5, 10, 10)

56
3.1.4 EOL Options Transfer Equations

The relationships between EOL characteristics and sustainability are determined using the

quantitative equations shown in Table 3-5. The transfer equations are developed by adopting

Remery et al. (2012)’s work with the necessary extension. The results of the equations are repre-

sented as fuzzy set combinations; therefore, fuzzy arithmetic should be applied to simplify them.

The fuzzy arithmetic is developed by 𝛼-cut method for which definitions are provided as follows:

𝑎 = (𝑎1 , 𝑎2 ) 𝑏 = (𝑏1 , 𝑏2 ) are two fuzzy sets,

𝑎 + 𝑏 = (𝑎1 + 𝑏1 , 𝑎2 + 𝑏2 ) (3-1)

𝑎 − 𝑏 = (𝑎1 − 𝑏2 , 𝑎2 − 𝑏1 ) (3-2)

𝑎 × 𝑏 = (min{𝑎1 × 𝑏1 , 𝑎2 × 𝑏2 , 𝑎1 × 𝑏2 , 𝑎2 × 𝑏1 }, (3-3)

max{𝑎1 × 𝑏1 , 𝑎2 × 𝑏2 , 𝑎1 × 𝑏2 , 𝑎2 × 𝑏1 })

𝑎 𝑎 𝑎 𝑎 𝑎 𝑎 𝑎 𝑎 𝑎
= (min {𝑏1 , 𝑏2 , 𝑏1 , 𝑏2 } , max {𝑏1 , 𝑏2 , 𝑏1 , 𝑏2 }), (3-4)
𝑏 2 1 1 2 2 1 1 2

if𝑏1 , 𝑏2 are not 0.

57
Table 3-5 Transfer Equations
Sustainability Secondary Re-
Reuse (A1) Remanufacturing (A2) Primary Recycle (A3) Incinerate (A5) Landfill (A6)
Dimension cycle (A4)

Economic Per- N13 × VG × √(N2 − VP)(N1 − VP) F


G N3 N3 × N4 VP
formance N14 × (VG − VP) VG
Environmental P VP
VG G N12×F N12×P −1
N10 −1
× N11 × −1
N10 −1
× N11 ×
Performance VG 2 VG 2
Social Perfor- N5 + N6 + N7 + N8 N5 + N6 + N7 + N9
N7 N9 VP VP
mance 4 × N22 4
N6: Difficulty of N7: Difficulty with
N1: Component N3: Quantity of High- N4: Component N5: Level of
N2: Component EOL Condition Dismantling Part the Component’s
Durability Value Materials Calorific Capacity Integration
Attachments Disassembly
N8: Quantity of N10: Amount of N11: Component N12: Regulation N13: Customer N14: Succedaneum
N9: Amount of Different Materials
Parts Hazardous Materials Weight Support Preference Price

The notation Nn−1 means that the higher the assessment of the parameter, the lower its performance. If Nn=VG, then Nn−1 =VP; if Nn=G, then Nn−1 =P; if Nn=F, then Nn−1 =F

58
3.1.5 Left-Right Defuzzification Method

Each fuzzy set combination is converted to a single fuzzy set by employing fuzzy operations

(through equations 3-1 to 3-4); however, comparisons are still complex. The left-right fuzzy

method was developed by Chen and Hwang (1992) to rank the fuzzy numbers. To defuzzify a

fuzzy number, the fuzzy maximizing and minimizing sets are defined, respectively, as:

𝑥, 0 ≤ 𝑥 ≤ 10,
𝑓𝑚𝑎𝑥 (𝑥) = { (3-5)
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

10 − 𝑥, 0 ≤ 𝑥 ≤ 10,
𝑓𝑚𝑖𝑛 (𝑥) = { (3-6)
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

When given a triangular fuzzy number FPII is defined as 𝑓𝐹𝑃𝐼𝐼 : 𝑅 → [0,10], with a triangu-

lar membership function, the right and left scores of FPII can be obtained, respectively, as

𝑈𝑅 (𝐹𝑃𝐼𝐼) = 𝑠𝑢𝑝𝑥 [𝑓𝐹𝑃𝐼𝐼 (𝑥) ∧ 𝑓𝑚𝑎𝑥 (𝑥)] (3-7)

𝑈𝐿 (𝐹𝑃𝐼𝐼) = 𝑠𝑢𝑝𝑥 [𝑓𝐹𝑃𝐼𝐼 (𝑥) ∧ 𝑓𝑚𝑖𝑛 (𝑥)] (3-8)

The total score of FPII can be obtained by combining the right and left scores. The total

score of FPII is used to determine the fuzzy number ranking, which is defined as:

𝑈𝑇 (𝐹𝑃𝐼𝐼) = [𝑈𝑅 (𝐹𝑃𝐼𝐼) + 10 − 𝑈𝐿 (𝐹𝑃𝐼𝐼)]/2 (3-9)

For example, fuzzy set (0.2, 0.23, 0.5) is converted to a single number 0.35 by using the

left-right fuzzy method and this number can be used to represent the fuzzy set and rank. The sin-

gle numbers will be normalized in order to transfer to weights of EOL strategies’ weights.

3.2 Failure Rate Representation

Failure rate influences the maintenance frequency, and maintenance cost, energy consump-

tion and labor hour requirements depend on the maintenance frequency; thus, failure rate plays a

significant role in the maintenance stage. Accordingly, failure rate must be taken into account in

maintenance related calculations. Many researches are conducted to investigate how to estimate

59
failure rate. Nicholls (2007) introduced five failure rate related factors: operational stress, environ-

mental stress, cycling stress, induced stress and solder joints stress. Based on these five factors

concern, Nicholls developed RIAC 217PlusPM Component Failure Rate Model. Some other

researches focused on estimating failure rate by empirical mathematical methods (Hutchison et al.,

2008; Han, 2012). This study will apply mean time between failures (MTBF) to estimate failure

rate because MTBF data is easy to get.

From the reliability theory, the failure rate is the inverse of MTBF as shown in Eq.3-10.

𝐹𝑎𝑖𝑙𝑢𝑟𝑒 𝑅𝑎𝑡𝑒 = 1⁄𝑀𝑇𝐵𝐹 (3-10)

Since failure rate depends on components’ connection, we establish a failure rate correlation

matrix between components. A DSM is a widely accepted method for analyzing the design of

products and systems which is not only used in modularization but also widely employed for man-

aging the complexity of components (Avnet and Weigel, 2009). Figure 3-5 shows a DSM example

with both a connectivity graph and a DSM table.

Figure 3-6 DSM Example (Adopted from Yan et al., 2012)


However, the elements in failure rate DSM is not static, they will change as product lifetime

goes on. For example, MTBFs for component A and B are 10 months and 8 months separately,

60
component B goes to failure first at 8th month and component A goes to failure first at 10th month.

Therefore, the failure rate DSM at 8th month and 10th month are different as shown in Figure 3-

7.

Figure 3-7 Failure Rate DSM w.r.t Two Components

3.3 Life Cycle Sustainability Model

This section will introduce sustainability model. Global environment challenges make it nec-

essary for all product manufacturers to reduce the environmental impacts of their products, whilst

maintaining the profit margins (CRR, 2013). In addition, all companies are required to shoulder

responsibility and contribute to society in ways such as improving the hiring rate and taking care

of workers’ health. Responding to this trend, many researchers have investigated how to achieve

sustainability goals for products. Kiritsls et al. (2003) considered three criteria based on environ-

mental, economic and social elements: CO2 emissions, disassembly cost, and number of employ-

ees, with minimizing, minimizing and maximizing as objectives. Chung et al. (2011) took eco-

nomic and environmental impacts into account in life cycle analysis. They used LCC calculated

by summing up manufacturing cost, maintenance cost and EOL cost, and LCEC calculated by

summing manufacturing, maintenance and EOL stages consumed energy, to measure these two

impacts. Both of calculations have minimize as objectives. This study will consider three criteria

involving the environment, the economy, and the society, with total environmental impact in the

61
life cycle, total cost in the life cycle and total employees hired in the life cycle as measurements.

For the third measure, the total labor time used in the life cycle is positively related to the total

employees hired in the life cycle; therefore, total labor time used is able to represent total employ-

ees hired. Product life cycle involves three stages: manufacturing, maintenance and EOL stages.

Figure 3-8 shows the product life cycle.

Manufacturing Maintenance
Stagte
EOL Stage Life Cycle
Stage

Figure 3-8 Product Life Cycle

3.3.1 Economic Sustainability Model

In considering the economic diversion of sustainability, total product life cycle cost is used

as a measurement. In Figure 3-8, the product life cycle is divided into three parts; therefore, the

total product life cycle cost has the following corresponding parts: manufacturing cost, mainte-

nance cost and EOL cost. The classification has a premium assumption, similar to the one Chung

et al. (2011) discussed: the manufacturing process is limited to Original Equipment Manufacturer

(OEM), which means company manufactures products or components that are purchased by an-

other company and retailed under that purchasing company's brand name. Eq. 3-11 shows this

measure.

𝑇𝐶 = 𝐶𝑀 + 𝐶𝑀𝑎𝑖 + 𝐶𝐸𝑂𝐿 (3-11)

where: 𝑇𝐶 means total cost in life cycle;

62
𝐶𝑀 means total manufacturing cost in life cycle;

𝐶𝑀𝑎𝑖 means total maintenance cost in life cycle;

𝐶𝐸𝑂𝐿 means EOL stage cost in life cycle.

3.3.1.1 Manufacturing Stage Cost

The manufacturing process involves assembling manufactured or purchased components;

therefore, the manufacturing cost is divided into component cost and assembly cost. For more

detail, the organizational structure of manufacturing cost is shown in Figure 3-9.

Component
Material Cost
Component Manufacturing
Cost Labor Cost
Manufacturing
Process Cost
Manufacturing Manufacturing
Cost Facility Cost
Assembly Labor
Cost
Assembly Joint
Assembly Cost
Cost
Assembly
Material Cost
Assembly
Facility Cost

Figure 3-9 Manufacturing Cost Organization

Therefore, 𝐶𝑀 = 𝐶𝐶 + 𝐶𝐴 (3-6)

where: 𝐶𝑀 is total manufacturing cost;

𝐶𝐶 is total component cost;

𝐶𝐴 is total assembly cost.

3.3.1.2 Maintenance Stage Cost

The maintenance process involves removing a failed module/component and replacing it

with a new one; therefore, the maintenance process includes both disassembly and assembly

processes, and maintenance cost is comprised of assembly cost, disassembly cost and replaced

63
new module/component costs. The maintenance cost is positively related to maintenance fre-

quency (Chung et al., 2011). Thus, maintenance frequency should be taken into account in cost

calculations. Maintenance frequency depends on module/component failure rate, thus, it could be

evaluated through failure rate. Failure rate index (FRI) is employed to evaluate the failure rate.

Therefore, the maintenance cost is calculated by Eq.3-12.

𝐶𝑀𝑎𝑖 = 𝑇𝐿𝐶 𝐾𝐹𝑅𝐼 (𝐶𝑁𝐶 + 𝐶𝐷𝐴 + 𝐶𝐴 ) (3-12)

where: 𝐶𝑀𝑎𝑖 is total maintenance cost;

𝑇𝐿𝐶 is product total life time;

𝐾𝐹𝑅𝐼 is component failure rate;

𝐶𝑁𝐶 is new component cost;

𝐶𝐷𝐴 is the disassembly cost;

𝐶𝐴 is assembly cost.

3.3.1.3 End of Life Stage Cost

By sumproducting of EOL options cost and their corresponding fuzzy weights, total EOL

options cost will be obtained. 𝐶𝑀𝐷𝐴 has to be positive because module disassembly must be paid.

Therefore, the EOL stage cost is from Eq. 3-13.

𝐶𝐸𝑂𝐿 = 𝐶𝑂𝐸 + 𝐶𝐷𝐴 (3-13)

where: 𝐶𝐸𝑂𝐿 is total EOL cost;

𝐶𝑂𝐸 is component EOL option expected cost;

𝐶𝐷𝐴 is disassembly cost.

64
3.3.2 Environmental Sustainability Model

Environmental consideration is similar to economic consideration. This study uses

environmental impact to evaluate the environmental impact. Environmental impact can be derived

from Eco-99, and the unit is mPt. Environmental impact is also measured in the manufacturing,

maintenance, and EOL processes. Eq. 3-14 shows the calculation process.

𝑇𝐸𝐼 = 𝐸𝐼𝑀 + 𝐸𝐼𝑀𝑎𝑖 + 𝐸𝐼𝐸𝑂𝐿 (3-14)

where: 𝑇𝐸𝐼 means total environmental impact in life cycle;

𝐸𝐼𝑀 means manufacturing stage environmental impact;

𝐸𝐼𝑀𝑎𝑖 means maintenance stage environmental impact;

𝐸𝐼𝐸𝑂𝐿 means EOL stage environmental impact.

3.3.2.1 Manufacturing Stage Environmental Impact

As with the manufacturing cost, manufacturing environmental impact involves component

environmental impact and assembly environmental impact. Therefore,

𝐸𝐼𝑀 = 𝐸𝐼𝐶 + 𝐸𝐼𝐴 (3-15)

where: 𝐸𝐼𝑀 is manufacturing stage environmental impact;

𝐸𝐼𝐶 is component environmental impact;

𝐸𝐼𝐴 is assembly environmental impact.

3.3.2.2 Maintenance Stage Environmental Impact

This study uses the same method to calculate maintenance environmental impact, and FRI

(Failure Rate Index) is involved.

𝐸𝐼𝑀𝑎𝑖 = 𝑇𝐿𝐶 𝐾𝐹𝑅𝐼 (𝐸𝐼𝑁𝐶 + 𝐸𝐼𝐷𝐴 + 𝐸𝐼𝐴 ) (3-16)

65
where: 𝐸𝐶𝑀𝑎𝑖 is total maintenance stage environmental impact;

𝐾𝐹𝑅𝐼 is component failure rate;

𝑇𝐿𝐶 is product total life time;

𝐸𝐼𝑁𝐶 is new component environmental impact;

𝐸𝐼𝐷𝐴 is disassembling environmental impact;

𝐸𝐼𝐴 is assembling environmental impact.

3.3.2.3 EOL Stage Environmental Impact

By sumproducting of EOL options environmental impact and their corresponding fuzzy

weights, EOL options excepted environmental impact will be derived. The calculation is from Eq.

3-17.

𝐸𝐼𝐸𝑂𝐿 = 𝐸𝐼𝑂𝐸 + 𝐸𝐼𝐷𝐴 (3-17)

where: 𝐸𝐼𝐸𝑂𝐿 is total EOL environmental impact;

𝐸𝐼𝑂𝐸 is EOL option expected environmental impact;

𝐸𝐼𝐷𝐴 is module disassembly environmental impact.

3.3.3 Social Sustainability Model

This study uses labor hiring to evaluate social impact. However, the amount of labor is dif-

ficult to measure precisely and depends on many factors. Labor time is able to replace labor hiring

as social sustainability indicator because labor time measurement is directly related to labor hiring

and the data for labor time is easier to determine. As with cost and environmental impact, Eq. 3-

13 is used to calculate total labor time.

𝑇𝐿𝑇 = 𝐿𝑇𝑀 + 𝐿𝑇𝑀𝑎𝑖 + 𝐿𝑇𝐸𝑂𝐿 (3-18)

66
where: 𝑇𝐿𝑇 means total labor time used needed in life cycle;

𝐿𝑇𝑀 means manufacturing stage labor time used;

𝐿𝑇𝑀𝑎𝑖 means maintenance stage labor time used;

𝐿𝑇𝐸𝑂𝐿 means EOL stage labor time used.

3.3.3.1 Manufacturing Stage Labor Time

Manufacturing labor time is estimated using a similar method for calculating manufacturing

cost and manufacturing energy consumption.

𝐿𝑇𝑀 = 𝐿𝑇𝐶 + 𝐿𝑇𝐴 (3-19)

where: 𝐿𝑇𝑀 is manufacturing stage labor time;

𝐿𝑇𝐶 is component labor time;

𝐿𝑇𝐴 is assembly labor time.

3.3.3.2 Maintenance Stage Labor Time

Using the same method, formulation of maintenance stage labor time will be obtained.

𝐿𝑇𝑀𝑎𝑖 = 𝑇𝐿𝐶 𝐾𝐹𝑅𝐼 (𝐿𝑇𝑁𝐶 + 𝐿𝑇𝐷𝐴 + 𝐿𝑇𝐴 ) (3-20)

where: 𝐿𝑇𝑀𝑎𝑖 is maintenance stage labor time;

𝑇𝐿𝐶 is product total life time;

𝐾𝐹𝑅𝐼 is component failure rate;

𝐿𝑇𝑁𝐶 is new component labor time;

𝐿𝑇𝐷𝐴 is disassembling labor time;

𝐿𝑇𝐴 is assembling labor time.

67
3.3.3.3 EOL Stage Labor Time

By sumproducting of EOL options labor time and their corresponding fuzzy weights, EOL

options total labor time will be derived. The module disassembly labor time is 𝐿𝑇𝐷𝐴 . The calcu-

lation is from Eq. 3-21.

𝐿𝑇𝐸𝑂𝐿 = 𝐿𝑇𝑂𝐸 + 𝐿𝑇𝐷𝐴 (3-21)

where: 𝐿𝑇𝐸𝑂𝐿 is total EOL labor time;

𝐿𝑇𝑂𝐸 is EOL option’s expected labor time;

𝐿𝑇𝐷𝐴 is module disassembly labor time.

3.4 Key Component Based Hierarchy Clustering Algorithm

In this section, key components selection and clustering algorithms are discussed in relation

to their impact on sustainability dimensions.

3.4.1 Key Component Determination

As stated earlier, key components may carry core technologies of the product, and they may

be most costly, most impactful to the environmental, or most labor-time requiring part relative to

others in the product. This study selects the top market value components as key components, and

the quantity of key components is determined by the manufacturers according to component char-

acteristics and company’s strategic direction.

Having multiple key components in the same module may cause problems (e.g., increasing

the assembly difficulty and cost during maintenance). For example, compressor and control unit

in a refrigerator are expensive to manufacture and assemble. If two components are in the same

module, during maintenance when one is maintained, the other also needs to be disassembled and

68
assembled; this makes maintenance more expensive. Therefore, one good way to handle key com-

ponents is separating them into different modules. Consequently, the quantity of key components

should equal to the number of modules.

3.4.2 Economic Sustainability Cluster Algorithm

This study develops first clustering algorithm based on economic sustainability

optimization. The proposed algorithm takes into account key component consideration within a

product. In design for assembly (DFA) techniques, the theoretical minimum number of modules

plays an essential role, especially in determining assembly complexity. These components need to

move relative to the rest of the assembly, and they have to be separable from the assembly for

disassembly access, replacement, and repair. Therefore, conceptually, theoretical minimum

number of components is similar to the number of key components.

In this MPD algorithm, ease of assembly will be considered. Ease of assembly depends on

several factors, and one of the most important of these is the quantity of part-to-part interfaces.

The more part-to- part interfaces two components have, the more difficult their assembly is.

Therefore, when determines the key components and their corresponding modules, it needs to sort

the non-key components according to the quantity of interfaces. The component with most number

of interfaces needs to be assigned to pre-determined key component module firstly; the higher the

number of interfaces a component has the higher priority it should be assigned. In addition to part-

to-part interfaces, other factors can have effects on ease of assembly, such as component weight

and component material. Figure 3-10 illustrates the detailed MPD algorithm flow chart.

69
Figure 3-10 Economic Sustainability Clustering Algorithm Flow Chart
3.4.3 Environmental Sustainability Cluster Algorithm

The second clustering algorithm is developed for environmental sustainability optimization.

Similar to the economic sustainability clustering algorithm, the objective of this MPD algorithm

is also minimizing. The only difference is that the previous algorithm concentrates on economic

sustainability (cost) minimization, and this algorithm emphasizes environmental sustainability (en-

70
vironment impact) minimization. Therefore, the only change is replacing cost terms by environ-

mental impact terms in Figure3-11. In this algorithm, in order to be consistent, key components

and non-key components sorting order that was determined previously are also used.

Figure 3-11 Environmental Sustainability Cluster Algorithm Flow Chart

71
3.4.4 Social Sustainability Cluster Algorithm

The third clustering algorithm is developed for the goal of social sustainability optimization.

Different from both economic and environmental sustainability clustering algorithms, the objec-

tive is maximizing. Figure 3-12 shows the flow chart of this algorithm. The same key components

and non-key components sorting order that was determined previously are also applied.

Figure 3-12 Social Sustainability Cluster Algorithm Flow Chart

72
3.5 Compromised Optimal Module Structure Determination

In section 3.4, three algorithms that will lead to three optimal module structures were pre-

sented. However, the module structures reach optimum only under a specific sustainability dimen-

sion. Therefore, there is a need to integrate these three optimal module structures together and

come up with a compromising module structure. For this purpose, this study develops a Module

Structure Sustainability Index (MSSI) and evaluate all possible module structures to find the most

sustainable one. The possible module structures represent all module structures with identified key

components in separate modules. The module structure which has lowest MSSI index will be the

most sustainable one. The MSSI index is shown below.

𝐿𝐶𝐶𝑖 𝐿𝐶𝐸𝐼𝑖 𝐿𝐶𝐿𝑇𝑚𝑎𝑥


𝑍𝑖 = 𝛼 𝐿𝐶𝐶 + 𝛽 𝐿𝐶𝐸𝐼 +𝛾 (3-22)
𝑚𝑖𝑛 𝑚𝑖𝑛 𝐿𝐶𝐿𝑇𝑖

where: 𝛼, 𝛽, 𝛾 are weights and determined by customers or manufacturers;

𝐿𝐶𝐶𝑖 , 𝐿𝐶𝐸𝐼𝑖 , 𝐿𝐶𝐿𝑇𝑖 are ith module structure’s life cycle cost, environmental

impact and labor time ;

𝐿𝐶𝐶𝑚𝑖𝑛 , 𝐿𝐶𝐸𝐼𝑚𝑖𝑛 , 𝐿𝐶𝐿𝑇𝑚𝑎𝑥 are optimal sustainability value w.r.t. economic,

environmental and social sustainability.

73
Chapter 4

Results
In this chapter, case studies from the literature will be used to illustrate the application of the

proposed methodology. The results using the proposed approach will also be compared to the re-

sults from the literature. Two case studies will be applied, for which the product data sets are from

the Ph.D. dissertation of Chung (2012) and supplemented by information from websites. These

case studies include a coffee maker and a refrigerator; these two correspond to a small and simple

product and a large and complex product. The product category is based on Wolf (2001)’s work.

Wolf investigated how the products covered by the WEEE Directive are processed in collection

and recycling systems at their EOL stage. His research pointed out that product size is a key factor

in determining whether the collection and recycling systems work successfully. Consumers tend

to select product EOL options requiring the least effort. Small products are easy to carry and

transport; accordingly, consumers can easily dispose of them. Hence, low collection rates for these

products are usually achieved. For large and bulky products, however, transport to a collection

point requires more effort so higher collection rates can be achieved. Wolf categorized the WEEE

products into four groups as shown in Table 4-1.

Table 4-1 Product Category (Adopted from Wolf, 2001)


Category Volume (dm3)* Weight (kg) Example Collection Rate
Very Small Prod- <1 <0.5 Mobile phone, Watch 20%
ucts
Small Products 1-15 0.5-3 Telephone, Coffee maker 35%
Middle Size Prod- 15-75 3-15 Vacuum Cleaner, PC 60%
ucts
Large Products >75 >15 Refrigerator, Washing ma- 98%
chine
*1dm=0.1m

74
4.1 Coffee Maker Case Study

The first case study in this chapter focuses on a small product as per the categorization in

Table 4-1: coffee maker, which has eleven components. The proposed methodology covers anal-

ysis of cost, environmental impact and labor time used. The product data set of coffee maker is

based on Chung (2012)’s Ph.D. dissertation and is supplemented by data from websites including

product material market price, scrap price, and landfill price: rivcowm.org, alibaba.com, earth-

worksrecycling.com, recycleinme.com, recycle.net. The coffee maker product model is Mr. Coffee

PR 15. Figure 4-1 presents Mr. Coffee PR 15 image. The coffee maker components only include

the main parts, and some small connectors, such as fasteners and screw bolts, are excluded from

the case study.

Figure 4-1 Mr. Coffee PR 15

75
4.1.1 Case Study Introduction

Coffee maker includes eleven components, and each component interacts with other com-

ponents in the product in order to generate primary product functions. Each component has its own

attributes, such as material, cost, energy, etc., which are related to coffee maker’s sustainability

characteristics. Table 4-2 shows components’ attributes, which will be used while forming the

product modules including component names, component material, component market price, man-

ufacturing price and manufacturing energy consumed.

Table 4-2 The Components' Attributes of Coffee Maker (Adopted from Chung, 2012 with MTBF
Extension)
No. Component Material Weight (g) Price Mfg. Mfg. Environ- MTBF
($) Cost ($) mental Impact (month)
(mPt)
1 Filter Basket Plastic 90.8 3 0.94 35.41 12
2 Filter Basket Plastic 101.696 3 0.94 39.66 84
Holder
3 Lid Plastic 52.664 2 0.64 20.54 24
4 Warming Plate Steel 63.56 5 1.84 5.47 36
5 Main Housing Plastic 1273.016 4 1.32 496.48 84
6 Heating Pipe Steel 227 8 3.2 19.52 36
7 Carafe Glass 348.672 10 3.3 20.22 36
8 Carafe Handle Plastic/Steel 84.444 3 0.51 27.8 36
9 Bottom Plate Steel 214.288 3 1.33 18.43 24
10 Power Cord Copper/Plastic 60.836 2 0.8 24.34 36
11 Switch Plastic/Metal 7.264 5 1.6 2.39 36

Component interactions are crucial in product module formation since they represent rela-

tionships among components. Based on the description of Pimmler and Eppinger (1994), compo-

nent interactions could be divided into four types: spatial, energy, information and material. This

study will not rank the importance of these four types of interactions; instead, it will only identify

them and use as constraints. The components interaction relationships are shown as a connection

DSM in Figure 4-2. In this figure, “1” represents two components are interacting, while “0” means

no interaction between two components.

76
Figure 4-2 Connection Relation DSM (Adopted from Chung, 2012)
Assembly and disassembly are related to both manufacturing and maintenance in product

life cycle. The numbers in Figure 4-3 represent assembly cost when assembling two components;

the unit is US dollar.

Figure 4-3 Assembly Cost DSM (Adopted from Chung, 2012)


Figures 4- 4 ~ 4-8 present assembly labor time, disassembly cost, disassembly labor time,

assembly environment impact, and disassembly environment impact DSM, separately.

77
Figure 4-4 Assembly Labor Time DSM (Unit: s, Adopted from Chung, 2012)

Figure 4-5 Disassembly Cost DSM (Unit: $, Adopted from Chung, 2012)

Figure 4-6 Disassembly Labor Time DSM (Unit: s, Adopted from Chung, 2012)

78
1 Filter Basket
2 Filter Basket Holder
3 Lid
4 Warming Plate
5 Main Housing
6 Heating Pipe
7 Carafe
1 2 3 4 5 6 7 8 9 10 11
8 Carafe Handle
1 0.00 1.23 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 9 Bottom Plate
2 1.23 0.00 0.00 0.00 8.80 0.00 0.00 0.00 0.00 0.00 0.00 10 Power Cord
3 0.00 0.00 0.00 0.00 8.48 0.00 0.00 0.00 0.00 0.00 0.00 11 Switch
4 0.00 0.00 0.00 0.00 24.09 6.68 21.75 0.00 0.00 0.00 0.00
5 0.00 8.80 8.48 24.09 0.00 27.03 0.00 0.00 26.80 59.46 23.07
6 0.00 0.00 0.00 6.68 27.03 0.00 0.00 0.00 0.00 0.00 4.22
7 0.00 0.00 0.00 21.75 0.00 0.00 0.00 12.93 0.00 0.00 0.00
8 0.00 0.00 0.00 0.00 0.00 0.00 12.93 0.00 0.00 0.00 0.00
9 0.00 0.00 0.00 0.00 26.80 0.00 0.00 0.00 0.00 0.00 0.00 No. Component
1 Filter Basket
10 0.00 0.00 0.00 0.00 59.46 0.00 0.00 0.00 0.00 0.00 1.52
2 Filter Basket Holder
11 0.00 0.00 0.00 0.00 23.07 4.22 0.00 0.00 0.00 1.52 0.00 3 Lid
4 Warming Plate
Figure 4-7 Assembly Environmental Impact DSM (Unit: mPt) 5 Main Housing
6 Heating Pipe
7 Carafe
1 2 3 4 5 6 7 8 9 10 11
8 Carafe Handle
1 0.00 1.23 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 9 Bottom Plate
2 1.23 0.00 0.00 0.00 8.80 0.00 0.00 0.00 0.00 0.00 0.00 10 Power Cord
3 0.00 0.00 0.00 0.00 8.48 0.00 0.00 0.00 0.00 0.00 0.00 11 Switch
4 0.00 0.00 0.00 0.00 24.09 6.68 21.75 0.00 0.00 0.00 0.00
5 0.00 8.80 8.48 24.09 0.00 27.03 0.00 0.00 26.80 59.46 23.07
6 0.00 0.00 0.00 6.68 27.03 0.00 0.00 0.00 0.00 0.00 4.22
7 0.00 0.00 0.00 21.75 0.00 0.00 0.00 12.93 0.00 0.00 0.00
8 0.00 0.00 0.00 0.00 0.00 0.00 12.93 0.00 0.00 0.00 0.00
9 0.00 0.00 0.00 0.00 26.80 0.00 0.00 0.00 0.00 0.00 0.00
10 0.00 0.00 0.00 0.00 59.46 0.00 0.00 0.00 0.00 0.00 1.52
11 0.00 0.00 0.00 0.00 23.07 4.22 0.00 0.00 0.00 1.52 0.00

Figure 4-8 Disassembly Environmental Impact DSM (Unit: mPt)

4.1.2 Coffee Maker EOL Strategy Analysis

Based on the description given in section 3.1, EOL strategy cost, environmental impact and

labor time can be calculated. Certain necessary data were collected from websites and the literature,

including material market price, scrap price, landfill price, are summarized in Table 4-3; the rele-

vant prices are from the U.S. market and are recorded in 2014. This study assumes that there is no

miscellaneous cost needed for this coffee maker. EOL strategy cost can be calculated according to

Table 1 of Appendix A. For example, the primary recycle cost for filter basket is

0.8×0.0908=$0.073, and its secondary recycle cost is 0.725×0.0908=$0.066.


79
Table 4-3 Material Market Price, Scrap Price and Landfill Price

Market Scrap Landfill


Mate-
Price Price Price
rial
($/kg) ($/kg) ($/kg)
Plastic 0. 8 0.725
Steel 0.904 0.882
0.041
Glass 0.814 0.801
Copper 6.015 5.33
Alumi- 1.5 1.2
num

Under the estimation of environmental impact from Eco-99 and job creation/labor time from

Tellus Institute with Sound Resource Management (TISRM) (2010), we can derive EOL strategy

environmental impact and labor time. Table 2 and Table 3 in Appendix A present EOL strategies’

environmental impact and labor time separately, in terms of components. For example, primary

recycle of plastic is -215mPt; therefore, primary recycle of filter basket is -215×0.0908=-19.52mPt.

Reuse does not need any process; therefore, the environmental impact for reuse is 0.

All EOL strategies’ sustainability values are determined above. However, in order to derive

the expected sustainability values of each component, EOL strategies’ weights are needed. Ac-

cording to section 3.1, weight determination requires EOL strategy characteristics’ fuzzy evalua-

tion, transfer equation calculation, and left-right defuzzification. Table 1 in Appendix B shows the

fuzzy evaluation data for EOL strategy characteristics and the dataset is from a skilled worker in

a Chinese coffee maker factory. For each component, by plugging the data from Table 1 in Ap-

pendix B into Table 3-5, we obtained the fuzzy set for each EOL strategy with respect to the full

sustainability dimensions. Table 4-4 represents fuzzy set for filter basket. By applying the left-

right fuzzy ranking method, each fuzzy set can be converted into a single number. As such, the

fuzzy set elements in Table 4-4 can be turned into single numbers as shown in Table 4-5. Nor-

malizing the elements in Table 4-5 yields the sustainability weights shown in Table 4-6. The fuzzy

80
set and corresponding single numbers are listed in Tables 1~2 in Appendix C. If combines the

sustainability values from Appendix A Tables 1~3 with corresponding weights, the expected EOL

sustainability value will be calculated in Table 4-7. For example, the expected EOL stage cost is

derived by: 0.21 ×2.7+0.2×0.073+0.17×0.066+0.21×0.0071+ 0.21×0.0037=$0.60

Table 4-4 Filter Basket Fuzzy Set After Transferring

Sustainability Remanufactur- Primary Secondary


Reuse Incinerate Landfill
Dimension ing Recycle Recycle
(A1) (A5) (A6)
(A2) (A3) (A4)
Economic (0.1.25, (5, 7.5,
(0,0,0) (5,7.5,10) (0,2.5,5) (0,0,2.5)
Performance 4.875) 10)
Environmen- (0,
(7.5, 10, (12.5, (0, 18.75, (0, 0,
tal Perfor- (5, 7.5, 10) 0.0208,
10) 37.5, 75) 50) 1.62)
mance 0.144)
Social Perfor- (1.25,
(0, 0, (5, 7.5,
mance (0, 0, 2.5) 1.875, (0, 0, 2.5) (0, 0, 2.5)
2.5) 10)
4.375)

Table 4-5 Filter Basket Defuzzification Numbers

Sustainability Remanufactur- Primary Secondary


Reuse Incinerate Landfill
Dimension ing Recycle Recycle
(A1) (A5) (A6)
(A2) (A3) (A4)
Economic
0 5.2 4.8 4.18 5.2 5.1
Performance
Environmen-
tal Perfor- 4.75 5.2 4.52 4.25 4.72 4.58
mance
Social Perfor-
5.1 5.1 4.75 5.2 4.81 4.81
mance

Table 4-6 Filter Basket Sustainability Value Weight

Sustainability Remanufactur- Primary Secondary


Reuse Incinerate Landfill
Dimension ing Recycle Recycle
(A1) (A5) (A6)
(A2) (A3) (A4)
Economic
0 0.21 0.2 0.17 0.21 0.21
Performance

81
Environmen-
tal Perfor- 0.17 0.19 0.16 0.15 0.17 0.16
mance
Social Perfor-
0.17 0.17 0.16 0.17 0.16 0.16
mance

Table 4-7 shows all coffee maker components’ expected EOL stage sustainability values in

terms of cost, environmental impact and labor time.

Table 4-7 Coffee Maker Expected EOL Sustainability Value


Environment
Economic
Sustainability Social Sustainability
Component Sustainability
(Environment (Labor Time: s)
(Cost: $)
Impact: mPt)
Filter Bas-
0.6 -5.51 5.20
ket
Filter Bas-
1.04 -7.54 5.85
ket Holder
Lid 0.72 -5.27 5.25
Warming
1.78 -0.43 3.65
Plate
Main
1.84 -94.20 62.90
Housing
Heating
2.68 -2.33 11.93
Pipe
Carafe 3.23 76.89 7.00
Carafe
1.00 -3.43 4.84
Handle
Bottom
1.05 -2.42 11.66
Plate
Power
0.70 -1.78 3.49
Cord
Switch 1.57 -0.55 0.69
Coffee
16.21 -46.57 122.46
Maker

According to Eqs. 3-13, 3-17, 3-21 EOL stage sustainability value can be calculated as:

EOL Stage Cost:

𝐶𝐸𝑂𝐿 = 𝐶𝑆𝐸 + 𝐶𝐷𝐴 = 16.21 + 𝐶𝐷𝐴 (4-1)

82
EOL Stage Environmental Impact:

𝐸𝐼𝐸𝑂𝐿 = 𝐸𝐼𝑆𝐸 + 𝐸𝐼𝐷𝐴 = −46.57 + 𝐸𝐼𝐷𝐴 (4-2)

EOL Stage Labor Time:

𝐿𝑇𝐸𝑂𝐿 = 𝐿𝑇𝑆𝐸 + 𝐿𝑇𝐷𝐴 = 122.46 + 𝐿𝑇𝐷𝐴 (4-3)

𝐶𝐷𝐴 , 𝐸𝐼𝐷𝐴 and 𝐿𝑇𝐷𝐴 is determined in section 4.1.5.

4.1.3 Coffee Maker Life Cycle Sustainability Model

According to Eqs. 3-11, 3-14, 3-18, coffee maker’s life cycle sustainability model can be

derived as following:

Economic Sustainability Life Cycle Model

𝑇𝐶 = 𝐶𝑀 + 𝐶𝑀𝑎𝑖 + 𝐶𝐸𝑂𝐿 =17.54+46.73+16.21+𝐶𝐷𝐴 (4-4)

Environmental Sustainability Life Cycle Model

𝑇𝐸𝐼 = 𝐸𝐼𝑀 + 𝐸𝐼𝑀𝑎𝑖 + 𝐸𝐼𝐸𝑂𝐿 = 936.32 + 1943.63 − 46.57 + 𝐸𝐼𝐷𝐴 (4-5)

Social Sustainability Life Cycle Model

𝑇𝐿𝑇 = 𝐿𝑇𝑀 + 𝐿𝑇𝑀𝑎𝑖 + 𝐿𝑇𝐸𝑂𝐿 = 120 + 1185.33 + 122.46 + 𝐿𝑇𝐷𝐴 (4-6)

4.1.4 Coffee Maker Key Component Determination

In the proposed methodology, clients and manufacturers will determine the number of key

components. Specifically, in this coffee maker case study, it pre-determines four key components

as the top four highest value components: component 4 (Warming Plate), component 6 (Heating

83
Pipe), component 7 (Carafe) and component 11 (Switch). These four key components are assigned

to four individual key component modules.

4.1.5 Coffee Maker Key Component Based Clustering Algorithm Result

The result of key component-based algorithm relies on the order of non-key components’

entry to the algorithm. Hence, it needs to arrange the order of non-key components. The sorting

according to the number of physical interfaces is shown in Figure 4-2. Some of the non-key com-

ponents have the same number of interfaces; therefore, the second consideration: component

weight, is also needed in order to get a unique order. The more weight the component has, the

more difficult it is to assemble. Consequently, the priority of assembly will be given to the higher

weighted component. By considering both interface quantity and component weight, the sorted

non-key components are listed in Table 4-8.

Table 4-8 Non-Key Component Order

Component Inter- Weight Or-


Name face # (g) der
Main Housing
7 1273.016 1
(5)
Filter Basket
2 101.696 2
Holder (2)
Power Cord
2 60.836 3
(10)
Bottom Plate
1 214.288 4
(9)
Filter Basket
1 90.8 5
(1)
Carafe Handle
1 84.444 6
(8)
Lid (3) 1 52.664 7

84
Applying three key-component based clustering algorithms (Figures. 3-10~12) to the coffee

maker case study, and deciding the order of non-key components entry algorithm, non-key com-

ponents will be allocated into key component modules as shown below.

Table 4-9 Optimal Module Structure and Corresponding Disassembly Sustainability Value

Sustainability Clustering Disassembly Sustainability


Module Structure
Algorithm Value
Economic Sustainability [1,2,3,5,6,9]; [4];
$0.29
Algorithm [7,8];[10,11]
Environmental Sustaina- [1,2,3,5,6,9,10];
83.32mPt
bility Algorithm [4];[7,8];[11]
Social [1,2,3,4,5,9,10];
88s
Sustainability Algorithm [6];[7,8];[11]

Using the values presented in Table 4-9, the optimal life cycle sustainability model with

respect to cost, environment impact and labor time will be derived,

Life cycle cost model:

𝑇𝐶 = 𝐶𝑀 + 𝐶𝑀𝑎𝑖 + 𝐶𝐸𝑂𝐿 =17.54+46.73+16.21+0.29=$80.77 (4-7)

Life cycle environmental impact model:

𝑇𝐸𝐼 = 𝐸𝐼𝑀 + 𝐸𝐼𝑀𝑎𝑖 + 𝐸𝐼𝐸𝑂𝐿 =936.32+1943.63-46.57+83.32=2916.7mPt (4-8)

Life cycle labor time model:

𝑇𝐿𝑇 = 𝐿𝑇𝑀 + 𝐿𝑇𝑀𝑎𝑖 + 𝐿𝑇𝐸𝑂𝐿 =120+1185.33+122.46+88=1515.79s (4-9)

4.1.6 Coffee Maker Compromised Optimal Module Structure

With pre-given weights to three sustainability values (0.4 to economic and environmental

sustainability and 0.2 to social sustainability), the MSSI index can be derived as follows:

85
𝐿𝐶𝐶 𝐿𝐶𝐸𝐼 1515.79
𝑍𝑖 = 0.4 × 80.77𝑖 + 0.4 × 2916.7𝑖 + 0.2 × (4-10)
𝐿𝐶𝐿𝑇𝑖

All possible key component module structures will be searched in order to find the mini-

mum 𝑍𝑖 ,; module yielding this minimum is the most sustainable module structure.

By applying Chung (2012)’s module structure feasibility codes, we identified all feasible

module structures, and then plug into coffee maker MSSI index, the compromised module struc-

ture has been identified, which is [1,2,3,4,5,9,10]; [6]; [7,8]; [11], with life cycle cost $80.94, life

cycle environmental impact 2917.65 mPt, life cycle labor time 1515.79s and MSSI index 1.00097.

4.2 Refrigerator Case Study

The second case study in this chapter focuses on a large size product as per the categorization

in Table 4-1: refrigerator, which has twenty components. The proposed methodology covers anal-

ysis of cost, environmental impact and labor time used. The product data set of the refrigerator is

based on Chung (2012)’s Ph.D. dissertation and is supplemented by data from websites including

product material market price, scrap price and landfill price: rivcowm.org, alibaba.com, earth-

worksrecycling.com, recycleinme.com, recycle.net. Figure 4-9 presents refrigerator dissection

sketch. The refrigerator components only include the main parts, and some small connectors, such

as fasteners and screw bolts, are excluded from the case study.

86
Figure 4-9 Refrigerator Dissection Sketch (Adopted from Chung, 2012)
4.2.1 Case Study Introduction

The refrigerator has been decomposed into 20 components, and each component interacts

with other components and performs a sub-function to form the product function. The 20 compo-

nents and their corresponding sub-functions are listed in Table 4-10, and other component attrib-

utes, including material, weight, price, manufacturing price, etc., are shown in Table 4-11. Man-

ufacturing environmental impact for each component was calculated according to Eco-99 excel-

based software. Figure 4-10 represents refrigerator’s connectivity graph.

Table 4-10 The Sub-Functions of the Components in the Refrigerator (Adopted from Chung,
2012)

No. Component Sub-Function


Name
1 Cabinet Houses the other component sin the refrigerator and supports the struc-
Frame ture of the refrigerator
2 Cabinet Houses foods in the refrigerator
3 Duct in Room Directs the air in the cabinet

87
4 Fan Unit 1 Provides air flow around the condenser to dissipate heat
5 Fan Unit 2 Circulates air flow in the cabinet
6 Evaporator Absorbs heat from the refrigerator and evaporates the refrigerant to cool
the refrigerator
7 Rear Board Provides rear insulation for condenser assembly at bottom
8 Compressor Forces the refrigerant into a smaller tube and propels it though the sys-
tem
9 Condenser Makes refrigerant condense and give off heat to the surroundings
10 Base Provides bottom insulation for condenser assembly
11 Door 1 Isolates the cabinet interior from the surroundings (top)
12 Door 2 Isolates the cabinet interior from the surroundings (bottom)
13 Gasket 1 Prevents air leak for door 1
14 Gasket 2 Prevents air leak for door 2
15 Door Liner 1 Provides insulation for door 1 and hold foods
16 Door Liner 2 Provides insulation for door 2 and hold foods
17 Control Unit Controls the activation timing of the compressor and heater
18 Heater Defrosts the inside of the cabinet
19 Dryer Filters the water and impurities in the refrigerant
20 Shelf Set Supports foods in the cabinet

Table 4-11 Refrigerator Attributes (Adopted from Chung, 2012)


No. Component Material Weight Price Mfg. Mfg. En- MTBF EOL Option
Name (g) ($) Cost vir. Im- (month)
($) pact
(mPt)
1 Cabinet Fe 23606 180 53.13 2596.66 312 RU/RM/PRC/SRC/L
Frame
2 Cabinet Plastic 29313 788 243.13 10845.81 192 PRC/SRC/L/I
3 Duct in Plastic 1028 40 12.65 380.36 162 PRC/SRC/L/I
Room
4 Fan Unit 1 Fe 483 26 8.51 53.13 108 PRC/SRC/L
5 Fan Unit 2 Fe 483 26 8.51 53.13 108 PRC/SRC/L
6 Evaporator Al 532 50 14.01 414.96 120 RU/RM/PRC/SRC/L
7 Rear Board Fe 986 27 8.67 108.46 336 RU/RM/PRC/SRC/L

88
8 Compressor Fe 7985 80 24.01 878.35 144 RU/RM/PRC/SRC/L
9 Condenser Fe 2669 40 12.44 293.59 204 RU/RM/PRC/SRC/L
10 Base Fe 1240 26 8.25 136.4 324 RU/RM/PRC/SRC/L
11 Door 1 Fe 2693 60 19.10 296.23 336 RU/RM/PRC/SRC/L
12 Door 2 Fe 4331 70 21.88 476.41 348 RU/RM/PRC/SRC/L
13 Gasket 1 Plastic 40 6 2 14.8 72 PRC/SRC/L/I
14 Gasket 2 Plastic 60 9 3 22.2 72 PRC/SRC/L/I
15 Door Liner Plastic 2236 52 16 827.32 168 PRC/SRC/L/I
1
16 Door Liner Plastic 4472 104 32 1654.64 168 PRC/SRC/L/I
2
17 Control Fe/Plastic 4677 332 108.08 1122.48 156 RU/RM/PRC/SRC/L/I
Unit
18 Heater Al 112 42 13.44 87.36 84 PRC/SRC/L/I
19 Dryer Cu 111 13 3.96 155.4 144 RU/RM/PRC/SRC/L
20 Shelf Set Plastic 1266 34 10.5 468.42 252 PRC/SRC/L
RU: Reuse RM: Remanufacturing PRC: Primary Recycle SRC: Secondary Recycle L: Landfill
I: Incineration

Figure 4-10 Connectivity Graph of Refrigerator (Adopted from Chung, 2012)

89
Figures 4-11 ~ 4-17 present refrigerator’s connection DSM, assembly time DSM, assembly

cost DSM, disassembly time DSM, disassembly cost DSM, assembly environmental impact DSM,

and disassembly environmental impact DSM.

Figure 4-11 Refrigerator Connection DSM (Adopted from Chung, 2012)

Figure 4-12 Refrigerator Assembly Time DSM (Unit: s; Adopted from Chung, 2012)

90
Figure 4-13 Refrigerator Assembly Cost DSM (Unit: $; Adopted from Chung, 2012)

Figure 4-14 Refrigerator Disassembly Time DSM (Unit: s; Adopted from Chung, 2012)

91
Figure 4-15 Refrigerator Disassembly Cost DSM (Unit: $; Adopted from Chung, 2012)

Figure 4-16 Refrigerator Assembly Environmental Impact DSM (Unit: mPt)

92
Figure 4-17 Refrigerator Disassembly Environmental Impact DSM (Unit: mPt)
4.2.2 Refrigerator EOL Strategy Analysis

By using methodology illustrated in Chapter 3, sustainability data set for refrigerator com-

ponents is able to be obtained in Appendix D, where Table 1 presents refrigerator component

EOL strategies cost, Table 2 shows component EOL strategies environmental impact and Table 3

component EOL strategies labor time.

Appendix E presents EOL characteristics fuzzy evaluation. By plugging into Table 3-5 trans-

fer equations, fuzzy evaluation for each of component could be calculated. Cabinet frame fuzzy

evaluation after transferring, defuzzification numbers and sustainability value weight are presented

in Tables 4-12~14 in this section. Other components relevant information will be provided in Ap-

pendix F.

Table 4-12 Cabinet Frame Fuzzy Set After Transferring

Primary Secondary
Sustainability Reuse Remanufacturing Incinerate Landfill
Recycle Recycle
Dimension (A1) (A2) (A5) (A6)
(A3) (A4)
Economic (2.66,8.6 (0,1.25,4.8 (0, 0,
(5, 7.5, 10) (0,2.5,5) (5,7.5,10)
Performance 6,20) 75) 2.5)

93
Environmen-
(7.5, 10, (6.25,25, (0, 12.5, (0, 0,
tal Perfor- (5, 7.5, 10) (0,0,0.9)
10) 56.25) 37.5) 0.45)
mance
Social Per- (7.5,10,1 (6.875,9. (0, 0,
(0.75,1.33,2) (5,7.5,10) (0, 0, 2.5)
formance 0) 375,10) 2.5)

Table 4-13 Cabinet Frame Defuzzification Numbers

Primary Secondary
Sustainability Reuse Remanufacturing Incinerate Landfill
Recycle Recycle
Dimension (A1) (A2) (A5) (A6)
(A3) (A4)
Economic Per-
5.56 5.2 4.8 4.18 5.2 5.1
formance
Environmental
4.75 5.2 5.77 5.37 5.01 5.01
Performance
Social Perfor-
4.75 4.62 4.72 5.2 5.1 5.1
mance

Table 4-14 Cabinet Frame Sustainability Value Weight

Sustainability Remanufactur- Primary Secondary


Reuse Incinerate Landfill
Dimension ing Recycle Recycle
(A1) (A5) (A6)
(A2) (A3) (A4)
Economic
0.19 0.17 0.16 0.14 0.17 0.17
Performance
Environmen-
tal Perfor- 0.15 0.17 0.19 0.17 0.16 0.16
mance
Social Perfor-
0.16 0.16 0.16 0.18 0.17 0.17
mance

Table 4-15 shows all refrigerator components’ expected EOL stage sustainability values in

terms of cost, environmental impact and labor time.

Table 4-15 Refrigerator Expected EOL Sustainability Value


Economic
Environment Sustainability Social Sustainability
Component Sustainability
(Environment Impact: mPt) (Labor Time: s)
(Cost: $)
Cabinet Frame 69.59 -431.75 1252.63
Cabinet 7.3 -2009.41 196.91

94
Duct in Room 0.25 -72.22 7.26
Fan Unit 1 0.16 -9.74 3.11
Fan Unit 2 0.14 -11.29 3.14
Evaporator 15.8 -119.27 28.95
Rear Board 8.75 -16.47 54.74
Compressor 27.77 -141.26 424.18
Condenser 12.86 -49.58 129.35
Base 7.67 -22.06 68.03
Door 1 19.45 -49.18 143.64
Door 2 23.12 -73.13 232.31
Gasket 1 0.01 -2.14 0.27
Gasket 2 0.02 -3 0.39
Door Liner 1 0.6 -160.03 15.69
Door Liner 2 1.23 -315.38 30.05
Control Unit 106.7 -206.2 252.99
Heater 0.04 -25.7 0.78
Dryer 4.09 -1.69 5.98
Shelf Set 0.33 -74.65 8.38
Refrigerator 305.88 -3794.15 2858.78

According to Eqs. 3-13, 3-17, 3-21 EOL stage sustainability value can be calculated as:

EOL Stage Cost:

𝐶𝐸𝑂𝐿 = 𝐶𝑆𝐸 + 𝐶𝐷𝐴 = 305.88 + 𝐶𝐷𝐴 (4-11)

EOL Stage Environmental Impact:

𝐸𝐼𝐸𝑂𝐿 = 𝐸𝐼𝑆𝐸 + 𝐸𝐼𝐷𝐴 = −3794.15 + 𝐸𝐼𝐷𝐴 (4-12)

EOL Stage Labor Time:

𝐿𝑇𝐸𝑂𝐿 = 𝐿𝑇𝑆𝐸 + 𝐿𝑇𝐷𝐴 = 2858.78 + 𝐿𝑇𝐷𝐴 (4-13)

𝐶𝐷𝐴 , 𝐸𝐼𝐷𝐴 and 𝐿𝑇𝐷𝐴 is determined in section 4.2.4.

95
4.2.3 Refrigerator Life Cycle Sustainability Model

The life time of refrigerator is around 16 years, 192 months (Fiol, 2013). According to Eqs.

3-11, 3-14, 3-18, refrigerator’s life cycle sustainability model can be derived as following:

Economic Sustainability Life Cycle Model

𝑇𝐶 = 𝐶𝑀 + 𝐶𝑀𝑎𝑖 + 𝐶𝐸𝑂𝐿 =623.27+337.91+305.88+𝐶𝐷𝐴 (4-14)

Environmental Sustainability Life Cycle Model

𝑇𝐸𝐼 = 𝐸𝐼𝑀 + 𝐸𝐼𝑀𝑎𝑖 + 𝐸𝐼𝐸𝑂𝐿 = 20886.11 + 21780.18 − 3794.15 + 𝐸𝐼𝐷𝐴 (4-15)

Social Sustainability Life Cycle Model

𝑇𝐿𝑇 = 𝐿𝑇𝑀 + 𝐿𝑇𝑀𝑎𝑖 + 𝐿𝑇𝐸𝑂𝐿 = 3210 + 7900 + 2858.78 + 𝐿𝑇𝐷𝐴 (4-16)

4.2.4 Refrigerator Key Component Determination

In the proposed methodology, clients and manufacturers will determine the number of key

components. In this refrigerator case study, it pre-determine four key components as the top four

highest value components: component 1 (Cabinet Frame), component 2 (Cabinet), component 16

(Door Liner 2) and component 17 (Control Unit). These four key components are assigned to four

individual key component modules.

4.2.5 Refrigerator Key Component Based Clustering Algorithm Result

By considering both interface quantity and component weight, the sorted non-key compo-

nents are listed in Table 4-16.

96
Table 4-16 Non-Key Component Order

Or-
Component Name Interface # Weight (g)
der
Evaporator (6) 4 532 1
Compressor (8) 3 7985 2
Door 2 (12) 3 4331 3
Door 1 (11) 3 2693 4
Condenser (9) 3 2669 5
Duct in Room (3) 3 1028 6
Door Liner 1 (15) 2 2236 7
Fan Unit 2 (5) 2 483 8
Heater (18) 2 112 9
Dryer (19) 2 111 10
Gasket 2 (14) 2 60 11
Gasket 1 (13) 2 40 12
Shelf Set (20) 1 1266 13
Base (10) 1 1240 14
Rear Board (7) 1 986 15
Fan Unit 1 (4) 1 483 16

Applying three key-component based clustering algorithms (Figures. 3-10~12) to the refrig-

erator case study, and deciding the order of non-key components entry algorithm, non-key com-

ponents will be assigned to key component modules as shown below in Table 4-17.

Table 4-17 Optimal Module Structure and Corresponding Disassembly Sustainability Value

Sustainability Clustering Disassembly Sustain-


Module Structure
Algorithm ability Value
[1,7,10,11,13,15];
Economic Sustainability [2,3,5,6,18,20];
$3.2
Algorithm [12,14,16];
[4,8,9,17,19]
[1,7,10,11,12,13,14,15];
Environmental Sustaina-
[2,3,5,6,18,20]; 2530.17mPt
bility Algorithm
[16];[4,8,9,17,19]
[1,7,10,11,12,13,15];
Social
[2,3,4,5,6,9,19,20]; 1468s
Sustainability Algorithm
[14,16];[8,17,18]

97
Using the values presented in Table 4-11, it can derive the optimal life cycle sustainability

model with respect to cost, environment impact and labor time,

Life cycle cost model:

𝑇𝐶 = 𝐶𝑀 + 𝐶𝑀𝑎𝑖 + 𝐶𝐸𝑂𝐿 =623.27+337.91+305.88+3.2=$1270.26 (4-17)

Life cycle environmental impact model:

𝑇𝐸𝐼 = 𝐸𝐼𝑀 + 𝐸𝐼𝑀𝑎𝑖 + 𝐸𝐼𝐸𝑂𝐿 =20886.11 + 21780.18 − 3794.15 + 2530.17 =

41402.31mPt (4-18)

Life cycle labor time model:

𝑇𝐿𝑇 = 𝐿𝑇𝑀 + 𝐿𝑇𝑀𝑎𝑖 + 𝐿𝑇𝐸𝑂𝐿 =3210 + 7900 + 2858.78 + 1468 = 15436.78s

Therefore, MSSI index is

𝐿𝐶𝐶
𝑖 𝐿𝐶𝐸𝐼
𝑖 15436.78
𝑍𝑖 = 0.4 × 1270.26 + 0.4 × 41402.31 + 0.2 × (4-19)
𝐿𝐶𝐿𝑇𝑖

4.2.6 Refrigerator Compromised Module Structure

By searching all possible module structures with given weights 0.4, 0.4, and 0.2 regarding

to economic, environmental and social aspect, the most sustainable option is [1,7, 10, 11, 13, 15],

[2, 3, 4, 5, 6, 8, 9, 19, 20], [12, 14, 16], [17, 18]. The disassembly cost, environmental impact and

labor time are $3.6, 3144.25mPt, and 1317s, respectively. Consequently, the life cycle cost, envi-

ronmental impact and labor time are $1,270.66, 42,016.39mPt, and 15,285.78s. The MSSI is

1.008034.

98
Chapter 5

Sensitivity Analysis and Comparison

5.1 Sensitivity Analysis

There are two sensitivity variables: pre-defined key component number and pre-defined sus-

tainability value weights in this study, therefore, sensitivity analysis should be conducted regarding

these two variables. Use coffee maker case study as an example, and generate two types of sensi-

tivity analysis. The first one is fixing key component number, and varying sustainability value

weights; another one is fixing sustainability value weights, and varying key component numbers.

For the first type, it pre-determines 4 key components, and vary sustainability value weights by

adding or deducting 0.1, the sensitivity analysis results are presented in Table 5-1. Figure 5-1

shows the relationship between sustainability value weights and corresponding minimum MSSI

index. For the second type, economic, environmental and social sustainability weights will be fixed

as 0.4, 0.4 and 0.2, and vary key component numbers from 1 to 11. The sensitivity analysis results

and plot are shown in Table 5-2 and Figure 5-2.

Table 5-1 Sustainability Weights v.s. MSSI Index v.s. Optimal Module Structure
Economic Sustainability Environmental Sus- Social Sustaina- Minimum Optimal Module Structure
Weight tainability Weight bility Weight MSSI Index
0.1 0 0.9 1.00021 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.1 0.1 0.8 1.000243 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.1 0.2 0.7 1.000276 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.1 0.3 0.6 1.000308 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.1 0.4 0.5 1.000341 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.1 0.5 0.4 1.000373 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.1 0.6 0.3 1.000406 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.1 0.7 0.2 1.000438 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.1 0.8 0.1 1.000471 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.1 0.9 0 1.000037 [1,2,3,5,6,9,10]; [4]; [7,8]; [11]
0.2 0 0.8 1.000412 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.2 0.1 0.7 1.000454 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.2 0.2 0.6 1.000486 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.2 0.3 0.5 1.000519 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.2 0.4 0.4 1.000551 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.2 0.5 0.3 1.000584 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.2 0.6 0.2 1.000616 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.2 0.7 0.1 1.000649 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.2 0.8 0 1.000074 [1,2,3,5,6,9,10]; [4]; [7,8]; [11]

99
0.2 0 0.8 1.000421 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.3 0 0.7 1.000631 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.3 0.1 0.6 1.000664 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.3 0.2 0.5 1.000697 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.3 0.3 0.4 1.000729 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.3 0.4 0.3 1.000762 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.3 0.5 0.2 1.000794 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.3 0.6 0.1 1.000872 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.3 0.7 0 1.000108 [1,2,3,5,9,10,11]; [4]; [6]; [7,8]
0.4 0 0.6 1.000842 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.4 0.1 0.5 1.000874 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.4 0.2 0.4 1.000907 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.4 0.3 0.3 1.00094 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.4 0.4 0.2 1.000972 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.4 0.5 0.1 1.001005 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.4 0.6 0 1.000093 [1,2,3,5,9,10,11]; [4]; [6]; [7,8]
0.5 0 0.5 1.001052 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.5 0.1 0.4 1.001085 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.5 0.2 0.3 1.001118 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.5 0.3 0.2 1.00115 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.5 0.4 0.1 1.001183 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.5 0.5 0 1.00077 [1,2,3,5,9,10,11]; [4]; [6]; [7,8]
0.6 0 0.4 1.001263 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.6 0.1 0.3 1.001295 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.6 0.2 0.2 1.001328 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.6 0.3 0.1 1.001361 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.6 0.4 0 1.000062 [1,2,3,5,9,10,11]; [4]; [6]; [7,8]
0.7 0 0.3 1.001473 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.7 0.1 0.2 1.001506 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.7 0.2 0.1 1.001538 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.7 0.3 0 1.000046 [1,2,3,5,9,10,11]; [4]; [6]; [7,8]
0.8 0 0.2 1.001684 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.8 0.1 0.1 1.001716 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0.8 0.2 0 1.000031 [1,2,3,5,9,10,11]; [4]; [6]; [7,8]
0.9 0 0.1 1.001578 [1,2,3,4,5,9]; [6]; [7,8]; [10,11]
0.9 0.1 0 1.000015 [1,2,3,5,9,10,11]; [4]; [6]; [7,8]
[1,2,3,5,9,10,11]; [4]; [6]; [7,8]
1 0 0 1 or
[1,2,3,5,6,9]; [4]; [7,8]; [10,11]
0 1 0 1 [1,2,3,5,6,9,10]; [4]; [7,8]; [11]
0 0 1 1 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0 0.1 0.9 1.000033 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0 0.2 0.8 1.000065 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0 0.3 0.7 1.000098 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0 0.4 0.6 1.00013 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0 0.5 0.5 1.000163 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0 0.6 0.4 1.000195 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0 0.7 0.3 1.000228 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0 0.8 0.2 1.000261 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]
0 0.9 0.1 1.000293 [1,2,3,4,5,9,10]; [6]; [7,8]; [11]

100
Figure 5-1 4 Dimensional Plot

According to Figure 5-1, X, Y, Z axis represent economic, environmental and social weights,

separately. The color column presents 4th dimension, MSSI index. Dark blue means low MSSI,

and dark red means high MSSI. From the plot, we are able to identify that, three corner spots are

most dark blue, which means if considers single sustainability aspect, MSSI can reach the lowest

value.

Table 5-2 Key Component Number v.s. Minimum MSSI Index


Pre-Defined Key
Key Components Optimal Module Structure MSSI Index
Components #
1 7 [1,2,3,4,5,6,7,8,9,10,11] --
2 6,7 [1,2,3,5,6,9,10,11]; [4,7,8] 1.00132
3 6,7,11 [1,2,3,5,9,10,11]; [6]; [4,7,8] 1.00109
4 4,6,7,11 [1,2,3,4,5,9,10]; [6]; [7,8]; [11] 1.000972
5 4,5,6,7,11 [1,2,3,5,9,10]; [4]; [6]; [7,8]; [11] 1.000869

101
6 4,5,6,7,8,11 [1,2,3,5,9,10]; [4]; [6]; [7]; [8]; [11] 1.000658
7 1,4,5,6,7,8,11 [1,2]; [3,5,9,10]; [4]; [6]; [7]; [8]; [11] 1.000369
8 1,2,4,5,6,7,8,11 [1]; [2]; [3,5,9,10]; [4]; [6]; [7]; [8]; [11] 1.000121
9 1,2,4,5,6,7,8,9,11 [1]; [2]; [3,5,10]; [4]; [6]; [7]; [8]; [9]; [11] 1.000096
10 1,2,3,4,5,6,7,8,9,11 [1]; [2]; [3]; [5,10]; [4]; [6]; [7]; [8]; [9]; [11] 1.000026
[1]; [2]; [3]; [4]; [5]; [6]; [7];
11 1,2,3,4,5,6,7,8,9,10,11 1
[8]; [9]; [10]; [11]

Figure 5-2 MSSI Index v.s. Pre-defined Key Component Numbers


According to Table 5-2 and Figure 5-2, each key component number has a unique MSSI

index. With the increasing key component number, MSSI will decrease, but module size/capacity

will increase. For example, 1 key component means all 11 components should be grouped in the

same module, and such module size/capacity is 11 components, and corresponding MSSI is infi-

nite; while 11 key components mean each component is a single module, and each module size/ca-

pacity is one component, and corresponding MSSI is 1, which is the lowest.

102
5.2 Comparison with DA Approach and ASCEM Approach Results

Huang and Kusiak (1998) developed the decomposition approach (DA), which covers inter-

action matrix and suitability matrix. The interaction matrix shows the pair-wise physical relation-

ship among components, and the suitability matrix shows designer’s preferences. DA approach

will form module structure by the triangulation algorithm based on the interaction matrix. ASCEM

(A supply chain evaluation methodology) was proposed by Chung (2012), which considered life

cycle economic and environmental sustainability. Therefore, ASCEM has similarities with pro-

posed methodology. However, it didn’t take into account key component in methodology devel-

opment. In order to compare ASCEM, DA and proposed approach, we assume component 5,6,7,8

as key components and they should be assigned to different modules; therefore, it assumes that

components 5,6,7,8 are strongly not preferred to be the same module in DA. The module structure

of coffee maker based on DA is shown in Figure 5-3. The detail of DA method is shown in Ap-

pendix G. The left-hand side is interaction matrix and the right-hand side is the suitability matrix.

Figure 5-3 Coffee Maker DA Module Structure

From Figure 5-3, component 1 (filter basket) and 2 (filter basket holder) are in the same

module. Component 3 (lid), 5 (main housing), 9 (bottom plate), 10 (power cord) and 11 (switch)

are in the same module. Component 4 (warming plate) and 7 (Carafe) are in the same module.

103
Component 6 (heating pipe) is in a single module and component 8 (carafe handle) is in another

single module.

ASCEM is a supply chain evaluation methodology, presented by Chung (2012). The method

focuses on the evaluation of product module structure from the view of the reverse supply chain

operations. Chung provided supply chain life cycle cost (LCC) and supply chain life cycle energy

consumption (LCEC). The steps of ASCEM are detailed in a flow chart in Appendix H. The opti-

mal LCC module structure is shown in Table 5-3. Table 5-4 will compare all these three method-

ologies for the coffee maker.

Table 5-3 ASCEM Result (Adopted from Chung, 2012)


1 2 3 4
1 0 0 1 0
2 0 0 1 0
3 0 0 1 0
4 0 0 1 0
5 0 0 1 0
6 1 0 0 0
7 0 1 0 0
8 0 0 0 1
9 0 0 1 0
10 0 0 1 0
11 0 0 1 0

From Table 5-3, component 6 (heating pipe), component 7 (Carafe) and component 8 (carafe

handle) are in different single modules. Component 1 (filter basket), 2 (filter basket holder), 3 (lid),

4 (warming plate), 5 (main housing), 9 (bottom plate), 10 (power cord) and 11 (switch) are in the

same module.

104
Table 5-4 Comparison between DA, ASCEM and Proposed Method
Decomposition Ap-
Methodology ASCEM New Methodology
proach
Key Compo- [5], [6], [7], [8] [5], [6], [7], [8] [5], [6], [7], [8]
nents
Module Struc- [1,2]; [5,3,9,10,11]; [1,2,3,4,5,9,10,11]; [1,2,3,5,9,10,1 [1,2,3,4,5,9, [1,2,3,4,5,9,10
ture [4,7]; [6]; [8] [6]; [7]; [8] 1]; [7]; [4,6]; 10,11]; [7]; ]; [7]; [6,11];
[8] [6]; [8] [8]
Life Cycle Cost $80.9 0.0099% $80.87 0.062% $80.82 -- --
higher higher
Life Cycle En- 2916.81 0.82% 2905.99 0.45% -- 2893.06 --
vironmental Im- mPt higher mPt higher mPt
pact
Life Cycle La- 1508.79s 3.25% 1503.79s 3.59% -- -- 1557.79s
bor Time lower lower

From Table 5-4, the new methodology has the best performance in terms of economic, social

and environmental sustainability. DA focuses on physical interaction and design preference; AS-

CEM emphasizes physical interaction and supply chain economic and environmental sustainability.

The new methodology highlights full-dimension sustainability as well as considers physical inter-

actions. The clustering algorithm in DA is a triangulation algorithm, which is not an optimal ap-

proach (Ma and Okudan, 2013); ASCEM applies a heuristic searching algorithm to evaluate all

possible module structures and search for an optimal module structure. The new method here ap-

plies a key component based clustering algorithm, which finds the optimum by searching heuris-

tically. DA partially specifies key components by using design preference. For example, in Fig-

ure5-3, component 5, 6, 7, 8 are set strongly un-designed in suitability matrix in order to handle as

key components separately, which means these four components must be grouped into different

modules. The identification of key components depends on designer’s preferences presented in the

suitability matrix; these preferences are subjective. The method proposed here identifies key com-

ponents not only as per designer preferences but also by comparing components, such as selecting

most costly component as the key component; therefore, it is more objective. ASCEM cannot be

105
able to identify key components. In addition, recent research shows that DA has several limitations

in clustering components (Ma and Okudan, 2013).

5.3 Response to Research Questions

The research questions in Chapter 1 are:

Can the new MPD method specify and concentrate on the key components in the design

process?

Can the new MPD method take uncertain EOL options into account in life cycle sustaina-

bility?

Can the new MPD method improve three-dimension sustainability performance with key

components specification and uncertain EOL options consideration?

From the results shown in Table 5-4, it will be able to state that the new modular product

design approach could specify the key components. Different key components will form a variety

of module structures by clustering algorithms. It can also state that the new modular product design

approach could take the uncertain EOL options into account. The methodology provides an uncer-

tain EOL options fuzzy representation approach, which replaces all EOL options by an expected

sustainability value. In addition, it can state that the new modular product design approach reaches

a sustainability optimization. The proposed three clustering algorithms are able to group compo-

nents with corresponding sustainability objective in each iteration, and checks the corresponding

sustainability value for each formed module structure until finding the optimal one. Therefore, to

summarize all three statements, the new modular product design approach could get a compro-

mised optimal module structure for key components from the perspective of life cycle sustainabil-

ity with uncertain EOL options consideration.

106
Chapter 6

Conclusion and Future Work


In this chapter, the dissertation’s conclusion and the future work direction are discussed.

6.1 Conclusion

As a response to the increasing concerns and requirements of environment friendliness, so-

cial harmony, and economic improvement, product design in manufacturing has become one of

the most important sectors influencing global sustainability, as almost all products are outputs of

the product development process. In this research, a sustainable product design approach that in-

tegrates the product life cycle is developed, encompassing three-dimension sustainability, the key

components emphasis, and the element of uncertainty related to options during the EOL stage. The

product life cycle is a composite of the manufacturing stage, the maintenance stage, and the EOL

stage. In this work, life cycle cost, life cycle environmental impact and life cycle labor time have

been employed to represent economic sustainability, environmental sustainability, and social sus-

tainability performance, respectively. Uncertainty at the EOL stage has been tackled using EOL

strategies expected values. Likewise, 15 EOL strategy characteristics, fuzzy evaluations, and trans-

fer equations were introduced to aid the determination of weights and then to derive expected

values. Key components concentration is one of the most critical objectives in developing this

methodology. Three sustainability-related clustering algorithms were developed to generate key-

component-based module structures. Coffee maker and refrigerator case studies, representing

small and large products respectively, were used to demonstrate the proposed methodology. In a

comparison with DA and ASCEM, the proposed new MPD approach has smallest cost and envi-

ronmental impact, but the largest labor time, which means it has a better sustainability performance

than either DA or ASCEM.

107
The proposed methodology is designed to improve product life cycle sustainability perfor-

mance by forming product module structures during the vague and early design stage. It is practical

and flexible because the required information is easy to determine and assess with the aid of fuzzy

linguistic variables; therefore, it can be implemented during the early design stage where many

uncertainties are present. Moreover, the method involves designers in the decision making process

related to EOL strategy by incorporating fuzzy assessment. During the early design stage, the de-

signer has significant power to shape the design intent and influence customer behaviors. In addi-

tion, the proposed methodology combines the quantitative three-dimension of sustainability eval-

uation and the qualitative designer perception analysis. This new approach thus is suitable to be

applied in product life cycle sustainability performance assessment or combined with other product

design methods in order to achieve specific design goals. This methodology can be implemented

for sustainable life cycle improvement; implementation can be made (1) during the initial design

stage of a new product; and/or (2) during the redesign of a product.

5.2 Future Work

The main drawbacks and corresponding future research directions for this proposal are three-

fold. The first is the use and adaptation of additional social sustainability indicators. Because

social sustainability is a complex concept, many elements in it hold potential for analysis. This

study only considers one indicator: labor time. Future research will evaluate social sustainability

aspects in a more comprehensive manner. The second direction involves updating transfer equa-

tions. In this work, transfer equations are drawn and developed from the work of Remery et al.

(2012) and are used to integrate fuzzy evaluation and defuzzification. When additional sustaina-

bility indicators are involved, more adequate transfer and aggregation equations will be needed.

The third future direction is key component determination. The methodology requires researchers

108
to pre-define the number of key components and to determine key components based on sorting

components. In prospective analysis, a new approach to selecting key components is needed. For

example, a theoretical minimum number of parts/components plays an essential role in determin-

ing assembly complexity in DFA (Storch, 2014) and conceptually, the theoretical minimum num-

ber of parts/components is similar to the quantity of key components. Therefore, a revised ap-

proach might use a similar method to determine the key component quantity and then identify key

components accordingly.

109
Reference

Appelqvist, P., Lehtonen, J.M. and Kokkonen, J. (2004). Modeling in product and supply chain
design: literature survey and case study, Journal of Manufacturing Technology Management, 15(7),
pp. 675-686.

Ameli, M., Mansour, S. and Ahmadi-Javid, A. (2016). A multi-objective model for selecting de-
sign alternatives and end-of-life options under uncertainty: A sustainable approach. Resources,
Conservation and Recycling, 109, pp. 123-136.

Avent, M.S. and Weigel, A.L. (2009). An application of the design structure matrix to integrated
concurrent engineering. Acta Astronautica, 66(5-6), pp. 937-949.

Baldwin, C.Y. and Clark, K. B. (2000). Design rules: The power of modularity design. Cambridge,
MA: MIT Press.

Beach, R., Muhlemann, A.P., Price, D.H.R., Paterson, A. and Sharp, J.A. (2000). A review of
manufacturing flexibility, European Journal of Operational Research, 122(1), pp.41-57.

Bellman, R.E. and Zadeh, L.A. (1995). Decisions making in fuzzy environment. Management Sci-
ence, 17(4), pp.144-164.

Bryant, C.R., Sivaramakrishnan, K.L., Wie, M.V., Stone, R.B. and McAdams, D.A. (2004). Pro-
ceedings of DETC’04 ASME 2004 Design Engineering Technical Conference and Computers and
Information in Engineering Conference: A modular design approach to support sustainable design,
Salt Lake City, Utah.

Bufardi, A., Gheorghe,R ., Kiritsis, D. and Xirouchakis, P. (2004). Multicriteria decision-aid ap-
proach for product end-of-life alternative selection, International Journal of Production Research,
42(16), pp.3139-3157.

Cebon, P., Hauptman, O. and Shekhar, C. (2008). Product modularity and the product life cycle:
new dynamics in the interactions of product and process technologies. International Journal of
Technology Management, 42(4), pp. 365-386.

Center for Remanufacturing and Reuse, (2013), A description of the design for end-of-life process,
http://www.remanufacturing.org.uk/pdf/story/1p295.pdf?session=RemanS-
ession:807618970f36920117UiJ20B9285.

Chen, S.J. and Hwang, C.L. (1992). Fuzzy Multiple Attribute Decision Making Methods and Ap-
plication. Springer, Berlin, Heidelberg.

Chiu, M.C. and Okudan, G.E. (2014). An investigation of the impact of product modularity level
on supply chain performance metrics: an industrial case study. Journal of Intelligent Manufactur-
ing, 25, 129-145.

110
Chiu, M.C., Chang, C.H., Chen, Y.T., Chiou, J.Y., & Chang, Y.J. (2016). Redesign for sustaina-
bility using particle swarm optimization method. Journal of Industrial and Production Engineer-
ing, 33(2), 103-113.
Chung, W.S. (2012). A modular design approach to improve product life cycle performance based
on optimized closed-loop supply chains. Ph.D. Dissertation, Penn State University.

Chung, W. S., Okudan, G.E. and Wysk, R.A. (2011). Proceedings of the ASME 2011 International
Design Engineering Technical Conference & Computers and Information in Engineering Confer-
ence: A modular design approach to improve the life cycle performance derived from optimized
closed-loop supply chain, Washington, DC.

Danneels, E. and Kleinschmidt, E.J. (2001). Product innovativeness from the firm’s perspective:
its dimensions and their relation with project selection and performance. Journal of Product Inno-
vation Management, 18(6), pp. 357-73.

Das, K. and Chowdhury, A.H. (2012). Designing a reverse logistics network for optimal collection,
recovery and quality-based product-mix planning. International Journal of Production Economics,
135, pp. 209-221.

Delgado, M., Verdegay, J.L. and Vila, V. (1993). Linguistic decision making models. Interna-
tional Journal of Intelligent Systems, 7 (5), pp.479–492.

Dowlatshahi, S. (1992). Purchasing’s role in a concurrent engineering environment, International


Journal of Purchasing Materials Management, 28(1), pp.21-25.

Environmental Protection Agency. (1994). Development of a Pollution Prevention Factors Meth-


odology Based on Life-Cycle Assessment: Lithographic Printing Case Study. Washington DC: Of-
fice of Research and Development.

Erdos, G., Kis, T. and Xirouchakis, P. (2001). Modelling and evaluating product end-of-life op-
tions. International Journal of Production Research, 39(6), pp.1203-1220.

Ernst, R. and Kamrad, B. (2000). Theory and methodology: evaluation of supply chain structures
through modularization and postponement. European Journal of Operational Research, 124, pp.
495-510.

Fan, B.B., Qi, G.N., Hu, X.M. and Yu, T. (2013). A network methodology for structure-oriented
modular product platform planning. Journal of Intelligent Manufacturing, 1-18.

Feitzinger, E. and Lee, H. L. (1997). Mass customization at Hewlett-Packard: The power of post-
ponement. Harvard Business Review, (Jan-Feb), pp. 116-121.

Fernandez, I. and Kekale, T. (2005). The influence of modularity and industry clockspeed on re-
verse logistics strategy: implications for the purchasing function. Journal of Purchasing and Sup-
ply Management, 11(4), pp. 193-205.

Fiol, T. (2013). The life expectancy of 7 major appliances. Accessed in Nov. 9, 2015,
http://blogs.hrblock.com/2013/10/21/the-life-expectancy-of-7-major-appliances/
111
Fixson, S.K. and Park, J.K. (2008). The power of integrality: linkages between product architecture,
innovation, and industry structure. Research Policy, 37, pp. 1296-1316.

Fujita, K., Amaya, H. and Akai, R. (2014). Mathematical model for simultaneous design of module
communalization and supply chain configuration toward global product family. Journal of Intelli-
gent Manufacturing, 24, 991-1004.

Garcia, R. and Calantone, R. (2002). A critical look at technological innovation typology and in-
novativeness terminology: a literature review. Journal of Product Innovation Management, 19(2),
pp. 110-132.

Geniescafe (2011). http://geniescafe.tumblr.com/.

Gershenson, J.K and Prasad, G.J (1997). Modularity in product design for manufacturability. In-
ternational Journal of Agile Manufacturing, 1(1), pp. 99-110.

Gershenson, J. K., Prasad, G. J, and Zhang Y. (2003). Product modularity: Definitions and benefits.
Journal of Engineering Design, 14(3), pp. 295–313.

Gershenson, J. K., Prasad, G. J. and Zhang, Y. (2004). Product modularity: Measures and design
methods. Journal of Engineering Design, 15(1), pp. 33-51.

Go, T.F., Wahab, D.A., & Hishamuddin, H. (2015). Multiple generation life-cycles for product
sustainability: the way forward. Journal of Cleaner Production, 95, 16-29.
Gonzalez, B. and Adenso-Diaz, B. (2005). A bill of material-based approach for end-of-life deci-
sion making in design for the environment. International Journal of Production Research, 43(10),
2071-2099.

Gu, P., Hashemian, M. and Sosale, S. (1997). An integrated modular design methodology for life
cycle engineering. Annals of the CIRP, 46(1), pp. 71–74.

Halstenberg, F.A., Buchert, T., Bonvoisin, J., Lindow, K. & Stark, R. (2015). Target-oriented
modularization-addressing sustainability design goals in product modularization. Proceedings of
22nd CIRP Conference on Life Cycle Engineering, 603-608.
Han, M. (2012). Proceedings of 2012 International Conference on Quality, Reliability, Risk,
Maintenance and Safety Engineering: Estimation of failure rate and its applications in reliability
engineering, Chengdu, China.

Hansen, P., and Mladenoviĉ, N. (1997). Variable neighborhood search for the p-median. Location
Science, 5, 207-226.

Hansen, P., Mladenoviĉ, N., and Perez-Brito, D. (2001). Variable neighborhood decomposition
search. Journal of Heuristics, 7, 335-350.

Hata, T., Kato, S. and Kimura, F. (2001). Proceedings of EcoDesign 2001: Second International
Symposium on, Design of product modularity for life cycle management. Dec. 11-15, pp. 93-96.

112
Huang, C. C. and Kusiak, A. (1998). Modularity in design of products and systems. IEEE Trans-
actions on Systems, Man, and Cybernetics, Part A, 28(1), pp. 66–77.

Hutchison, K., Quigley, J., Raza, M. and Walls, L. (2008). Proceedings of 2008 International En-
gineering and Engineering Management: Empirical Bayes Methodology for Estimating Equipment
Failure Rates with Application to Power Generation Plants, Singapore, pp.1359-1364.

Ijomah, W., Bennett, J.P. and Pearce, J. (1999) Remanufacturing Evidence of Environmental Con-
scious Business Practice in UK. EcoDesign '99: First International Symposium on Environmen-
tally Conscious Design and Inverse Manufacturing, Published by IEEE Computer Society Pisca-
taway NJ, Tokyo, Japan, pp. 192-196.

Ishii, K., Juengel, C. and Eubanks, C. F. (1995). Proceedings of the 1995 ASME Design Engineer-
ing Technical Conferences, 7th International Conference on Design Theory and Methodology:
Design for product variety: Key to product line structuring. Boston, MA.

Jeong, M-G., Morrison, J.R., & Suh, H-W. (2015). Product modularity to jointly address functional
and ecological requirements in the early design stage. Proceedings of IEEE International Confer-
ence on Automation Science and Engineering, 33-38.
Ji, Y.J., Chen, X.B., Qi, G.N. and Song, L.W. (2012). Modular design involving effectiveness of
multiple phases for product life cycle. International Journal of Advanced Manufacturing Technol-
ogy, pp. 1-14.

Ji, Y.J., Jiao, R.J., Chen, L. and Wu, C. (2013). Green modular design for material efficiency: a
leader – follower joint optimization model. Journal of Cleaner Production, 41, 187-201.

Jose, A. and Tollenaere, M. (2005). Modular and platform methods for product family design:
Literature analysis. Journal of Intelligent Manufacturing, 16, pp. 371–390.

Kamrani, A.K. and Salhieh, S. E. M. (2008). Modular Design. In Kamrani, A. K. & Nasr, E.A.
(eds.), Collaborative Engineering: Theory and Practice (207-227). New York, NY: Springer.

Keeney, R.L. and Raiffa, H. (1976). Decisions with multiple objectives: preference and value
tradeoffs. Chichester, Wiley.

Kim, K. and Chhajed, D. (2000). Commonality in product design: cost saving, valuation change
and cannibalization. European Journal of Operation Research, 125, pp. 602-621.

Kiritsis, D., Bufardi, A. and Xirouchakis, P. (2003). Proceeding of 2003 IEEE International Sym-
posium on Electronics and the Environment. Multi-criteria decision aid for product end of life
options selection. May. 19-22, pp.48-53.

Koga, T. and Aoyama, K. (2008). Proceeding of ASME 2008 International Design Engineering
Technical Conferences & Computers and Information in Engineering Conference: Modular design
method for sustainable life-cycle of product family considering future market changes. Brooklyn,
New York.

113
Kohn, H.F., Steinley, D. and Brusco, M.J. (2010). The p-median model as a tool for clustering
psychological data. Psychological Methods, 15, 87-95.

Kreng, V. B. and Lee, T. P. (2004a). Modular product design with grouping genetic algorithm: a
case study. Computers and Industrial Engineering, 46(3), pp. 443-460.

Kreng, V. B. and Lee, T. P. (2004b). QFD-Based modular product design with linear integer pro-
gramming—A case study. Journal of Engineering Design, 15(3), pp. 261-284.

Krikke, H., Le Blanc, I. and van de Velde, S. (2004). Product modularity and the design of closed
loop supply chains. California Management Review, 46(2), pp. 23-39.

Kuo, T.C. (2013). Mass customization and personalization software development: a case study
eco-design product service system. Journal of Intelligent Manufacturing, 24, 1019-1031.

Kusiak, A. (1987). The generalized group technology concept. International Journal of Produc-
tion Research, (25)4, pp.561-569.

Kusiak, A. and Chow, W. S. (1987). Efficient solving of the group technology problem. Journal
of Manufacturing Systems, 6(2), pp. 117–124.

Kusiak, A. and Wang, J. (1993). Efficient organizing of design activities. International Journal of
Production Research, 31(4), pp. 753-769.

Kumar, C. S. and Chandrasekharan, M. P. (1990). Group efficacy: A quantitative criterion of good-


ness of block diagonal forms of binary matrices in group technology. International Journal of
Production Research, 28, pp. 233–244.

Lai, X.X. and Gershenson, J.K. (2009). Proceedings of the ASME 2009 International Design En-
gineering Technical Conferences & Computers and Information in Engineering Conference, DSM-
based product representation for retirement process-based modularity, San Diego, CA, Aug30-
Sep.2.

Lau, A. K. W., Yam, R. C. M. and Tang, E. (2007). The impacts of product modularity on com-
petitive capabilities and performance: An empirical study. International Journal of Production
Economics, 105(1), pp. 1–20.

Lau, A.K.W. (2009). Proceeding of PICMET 2009 Portland International Conference on Manage-
ment of Engineering and Technology: Managing modular product design: critical factors and a
managerial guide. Portland, OR.

Lavigne, B.B., Bassetto, S. and Agard, B. (2014). A method for a robust optimization of joint
product and supply chain design. Journal of Intelligent Manufacturing, 1-9.

Lee, S.G., Lye, S.W. and Khoo, M.K. (2001). A multi-objective methodology for evaluating prod-
uct end-of-life options and disassembly. International Journal of Advanced Manufacturing Tech-
nology, 18, pp.148-156.

114
Lee, H.L. and Billingtong, C. (1994). Designing products and processes for postponement. Dasu,
S., Eastman, C. (Eds.), Management of Design (105-122), Kluwer Academic Publishers, Boston.

Lewis, P.K., Mattson, C.A., & Wood, C.D. (2015). Modular product optimization to alleviate pov-
erty: an irrigation pump case study. International Journal of Product Development, 20(1), 49-73.
Li, J.Z., Zhang, H.C., Gonzalez, M.A., and Yu, S. (2008). A multi-objective fuzzy graph approach
for modular formulation considering end of life issues. International Journal of Production Re-
search, 46(14), 4011-4033.

Lin, C-T, Qiu,H. and Tseng, Y-H. (2006). Agility evaluation using fuzzy logic. International Jour-
nal of Production Economics, 101, pp.353-368.

Micheletti, G.F. (1985). Application of new technologies for fully integrated robotized automobile
engine production. Robotics and Computer-Integrated Manufacturing, 4(12), pp. 141-148.

Ma, J.F. and Okudan, G. E. (2013). Proceedings of International Conference on Engineering De-
sign-ICED2013: An analysis of decomposition approach application on design engineering &
suggestions for improvement. Aug.18-22, Seoul, South Korea.

Marco, P., Eubanks, C.F. and Ishii, k. (1994). Compatibility analysis of product design for recy-
clability and reuse, Proceedings of the 1994 ASME Computers in Engineering Conference,1994.

Marshall, R., Leaney, P. G. and Botterell, P. (1998). Enhanced product realisation through modular
design: An example of product/process integration. Journal of Integrated Design and Process
Technology, 3(4), 143–150.

Means, P. and Guggemos, A. (2015). Proceedings of 2015 International Conference on Sustainable


Design, Engineering and Construction, ICSDEC 2015: Framework for life cycle assessment (LCA)
based environmental decision making during the conceptual design phase for commercial build-
ings. 118, pp. 802-812. May 10~13.

Mukhopadhyay, S.K. and Setoputro, R. (2005). Optimal return policy and modular design for
build-to-order products. Journal of Operations Management. 23, pp. 496-506.

Newcomb, P. J., Bras, B. and Rosen, D.W. (1996). Proceedings of the 1996 ASME Design Engi-
neering Technical Conferences— 8th International Conference on Design Theory and Methodol-
ogy: Implications of modularity on product design for the life cycle. Irvine, CA.

Nepal, B., Monplaisir, L. and Singh, N. (2005). Integrated fuzzy logic-based model for product
modularization during concept development phase. International Journal of Production Econom-
ics, 96(2), pp.157-174.

Nicholls, D. (2007). An introduction to the RIAC 217PLUSTM component failure rate models.
Journal of the Reliability Information Analysis Center, first quarter 2007, available for PDF down-
load from the RIAC at http://theRIAC.org.

115
Okudan, G. E, Lin, T.K. and Chiu, M. C. (2012). Proceedings of the International Workshop on
Green Supply Chain: Carbon footprint implications of modularity and projections for the reverse
logistics, Arras, France.

Pimmler, T. U. and Eppinger, S. D. (1994). Proceedings of the 1994 ASME Design Engineering
Technical Conferences—6th International Conference on Design Theory and Methodology: Inte-
gration analysis of product decompositions. Minneapolis, MN.

Rantanen, K. and Domb, E. (2002). Simplified TRIZ. Crc Press LIc.

Regazzoni, D. and Rizzi, C. (2008). Proceedings of the ASME 2008 International Design Engi-
neering Technical Conference & Computers and Information in Engineering Conference: Enhanc-
ing modular design with creativity tools. Brooklyn, NY.

Remery, M., Mascle, C. and Agard,B. (2012). A new method for evaluating the best product end-
of-life strategy during the early design phase, Journal of Engineering Design, 23(6), pp.419-441.

Rodriguez, S. I., Roman, M. S., Sturhahn, S. C. and Terry, E. H. (2002). Sustainability assessment
and reporting for the University of Michigan’s Ann Arbor campus (joint Master’s thesis). Report
No. CSS02-04. Accessible online at http://css.snre.umich.edu/css_doc/CSS02-04.pdf.

Rose, C.M. (2001), Design for environment: a method for formulating product end-of-life strate-
gies, Ph.D dissertation, Stanford University.

Sandborn, P. and Myers, J. (2008). Designing engineering systems for sustainability. Handbook
of performability engineering (Misra, K. B., ed., Chapter 7, 81-104). London: Springer.

Shehab, E.M. and Abdalla, H.S. (2001). Manufacturing cost modeling for product development.
Robotics and Computer Integrated Manufacturing, 17, pp.341-353.

Shin, J.H., Jun, H.B., Kiritsis, D., and Xirouchakis, P. (2011). A decision support method for prod-
uct conceptual design considering product lifecycle factors and resource constraints. International
Journal of Manufacturing Technology, 52, 865-886.

Singh, R. K., Kumar,S., Choudhury,A.K. and Tiwari, M.K. (2006). Lean tool selection in a die
casting unit: a fuzzy-based decision support heuristic, International Journal of Production Re-
search, 44(7), pp. 1399-1429.

Shin, J.H., Jun, H.B., Kiritsis, D. and Xirouchakis, P. (2011). A decision support method for prod-
uct conceptual design considering product lifecycle factors and resource constraints. International
Journal of Manufacturing Technology, 52, pp. 865-886.

Smith, S. and Yen, C.C. (2010). Green product design through product modularization using
atomic theory. Robotics and Computer-Integrate Manufacturing, 26, pp. 790-798.

Smith, S., Smith, G.C., Jiao, R. and Chu, C-H. (2013). Mass customization in the product life cycle.
Journal of Intelligent Manufacturing, 24, 877-885.

116
Stevels, A. L. N. (1997) Optimization of the End-of-Life System. In Brezet, J.C. and van Hemel,C.
(eds.) Ecodesign: A Promising Approach. Paris, France, UNEP Working Group on Sustainable
Product Development: pp. 346.

Steward, D.V. (1965) Partition and Testing Systems of Equations, Journal of the Society for In-
dustrial and Applied Mathematics: Series B, Numerical Analysis, 2(2), pp. 345-365.

Stone, R. B., Wood, K. L. and Crawford, R. H. (1998). Proceedings of the 1999 ASME Design
Technical Conferences—11th International Conference on Design Theory and Automation: A heu-
ristic method to identify modules from a functional description of a product. Las Vegas, NV.

Storch, R. L. (2014). Design in Manufacturing Firm. Retrieved from http:// cour ses.washing-
ton.edu/inde494/Design%20for%20X.ppt

Teitz, M. B., and Bart, P. (1968). Heuristic methods for estimating the generalized vertex median
of a weighted graph. Operations Research, 16, 955-961.

Tseng, H.E., Chang, C.C. and Li, J.D. (2008). Modular design to support green life-cycle engi-
neering. Expert Systems with Applications, 34, pp. 2524-2537.

Tellus Institute with Sound Resource Management, (2010). More Jobs, Less Pollution: Growing
the Recycling Economy in the U.S. http://docs.nrdc.org/globalwarming/files/glo_11111401a.pdf

Ulrich, K. and Eppinger, S. D. (2000). Product design and development (2nd ed.). New York:
McGraw-Hill.

Umeda, Y., Nonomura, A. and Tomiyama, T. (2000). Study on Life-cycle Design for the Post
Mass Production Paradigm. Artificial Intelligence for Engineering Design, Analysis and Manu-
facturing 14(2), pp. 149-161.

Umeda, Y., Fukushige, S., Tonoike, K. and Kondoh, S. (2008). Product modularity for life cycle
design. CIRP Annals (Manufacturing Technology), 57(1), pp. 13-16.

Umeda, Y., Fukushige, S. and Tonoike, K. (2009). Evaluation of scenario based modularization
for life cycle design. CIRP Annals (Manufacturing Technology), 58, pp. 1-4.

Walls, M., (2006), Extended Producer Responsibility and Product Design: Economic Theory and
Selected Case Studies. Resources for the Future discussion paper 06-08.

Wang, Y. and Tseng, M.M. (2013). A naïve bayes approach to map customer requirements to
product variants. Journal of Intelligent Manufacturing, 1-9.

Whitaker, R. (1983). A fast algorithm for the greedy interchange of large-scale clustering and
median location problems. INFOR, 21, 95-108.

Wikipedia (2013), Original Equipment Manufacturer, http://en.wikipedia.org/wiki/Origi-


nal_equipment_manufacturer.

117
World Commission on Environment and Development Report. (1987). From one earth to one
world: An overview. Oxford: Oxford University Press.

Yadav, O.P., Singh, N., Chinnam, R.B. and Goel, P.S. (2002). A fuzzy logic based approach to
reliability improvement estimation during product development. Reliability Engineering and Sys-
tem Safety, 80, pp.63-74.

Yan, J. H., Feng, C. H. and Cheng, K. (2012). Sustainability-oriented product modular design
using kernel-based fuzzy c-means clustering and genetic algorithm. Proceedings of the Institution
of Mechanical Engineers, Part B: Journal of Engineering Manufacture, 226 (10): pp. 1635-1647.

Yang, S.L. and Li, T.F., (2002). Agility evaluation of mass customization product manufacturing.
Journal of Materials Processing Technology, 129 (1–3), pp.640–644.

Yen, C.C., & Smith, S. (2009). Proceedings of the ASME 2009 International Design Engineering
Technical Conference & Computers and Information in Engineering Conference: Product modular
design using atomic theory, San Diego, CA.

Zadeh, L.A. (1965), Fuzzy Sets. Information and Control, pp.338-352.

Zamagni, A., Buttol, P., Buonamici, R., Masoni, P., Guinee, J.B., Huppes, G., Heijungs, R., Van
Der Voet, E., Ekvall, T. and Rydberg, T. (2009). D20 Blue Paper on Life cycle sustainability
analysis.
Accessible online at http://www.leidenuniv.nl/cml/ssp/publications/calcas_report_d20.pdf.

Zhang, W.Y, Tor, S.Y. and Britton, G. A. (2006) Managing product modularity in product family
design with functional modeling. International Journal of Advanced Manufacturing Technology,
30, pp. 579–588.

Zhang, Y. and Gershenson, J.K. (2003). An initial Study of Direct Relationships between Life-
Cycle Modularity and Life-Cycle Cost, Concurrent Engineering, 11, pp. 121-128.

118
Appendix A
Coffee Maker EOL Strategy Sustainability Value

Table 1 Component EOL Strategy Cost ($)

No. Component Re- Remanufac- Primary Recy- Secondary Incinera- Landfill


use ture cle Recycle tion
1 Filter Basket 3 2.7 0.073 0.066 0.0071 0.0037
2 Filter Basket
3 2.8 0.081 0.074 0.0075 0.0042
Holder
3 Lid 2 1.9 0.042 0.038 0.0038 0.0022
4 Warming 5 4.8 0.00003
Plate 0.057 0.056 0.0026
5 Main Hous- 4 4.5 0.086
ing 1.02 0.92 0.052
6 Heating Pipe 8 7.8 0.21 0.20 0.00011 0.0093
7 Carafe 10 9 0.28 0.28 0.0099 0.014
8 Carafe Han- 3 2.9 0.00013
dle 0.072 0.068 0.0035
9 Bottom Plate 3 2.8 0.19 0.19 0.0001 0.0088
10 Power Cord 2 1.8 0.23 0.18 0.0041 0.0025
11 Switch 5 4.5 0.0062 0.0058 0.0000001 0.0003

Table 2 EOL Strategy Environment Impact (mPt)

Re- Remanufac- Primary Re- Secondary Incinera-


No. Component Landfill
use ture cycle Recycle tion
1 Filter Basket 0 0.58 -19.52 -15.62 -1.18 0.32
Filter Basket
2 0 0.65 -21.36 -19.18 -1.32 0.36
Holder
3 Lid 0 0.34 -11.06 -19.06 -0.68 0.18
Warming
4 0 6.32 -4.45 -2.17 -2.03 0.089
Plate
Main Hous-
5 0 8.15 -267.33 -252.12 -16.55 4.46
ing
6 Heating Pipe 0 22.56 -15.89 -11.69 -7.26 0.32
7 Carafe 0 0.45 -5.23 -3.12 1.78 488.15
Carafe Han-
8 0 4.22 -11.82 -9.18 -1.86 0.21
dle
9 Bottom Plate 0 20.78 -15.00 -12.16 -6.86 0.30
10 Power Cord 0 5.63 -7.91 -5.66 -1.64 0.18
11 Switch 0 0.33 -1.02 -0.78 -1.60 0.018

119
Table 3 EOL Strategy Labor Time (s)

Remanufac- Primary Re- Secondary


No. Component Reuse Incineration Landfill
ture cycle Recycle
1 Filter Basket 13.08 15.08 1.33 1.03 0.065 0.065
Filter Basket
2 14.64 16.64 2.64 2.31 0.073 0.073
Holder
3 Lid 7.58 9.58 7.21 6.92 0.038 0.038
4 Warming Plate 9.15 11.15 1.26 0.5 0.046 0.046
5 Main Housing 183.31 186.31 18.05 3.09 0.92 0.92
6 Heating Pipe 32.69 35.69 3.31 2.16 0.16 0.16
7 Carafe 18.45 20.62 2.26 1.98 0.092 0.092
8 Carafe Handle 12.16 14.31 2.59 1.05 0.061 0.061
9 Bottom Plate 30.86 32.62 3.81 2.63 0.15 0.15
10 Power Cord 8.76 9.60 1.76 1.06 0.044 0.044
11 Switch 1.05 3.03 0.11 0.09 0.0052 0.0052

120
Appendix B
Coffee Maker EOL Strategy Characteristics Fuzzy Evaluation

Table 1 EOL Strategy Characteristics Fuzzy Evaluation


Filter
Filter Warming Main Heating Carafe Bottom Power
Basket Lid Carafe Switch
Basket Plate Housing Pipe Handle Plate Cord
Holder
Durability (N1) P VG F G VG G G G F G G
EOL Condition
VP G F F G F F F P F F
(N2)
Quantity of High-
Value Materials P P P F P F F F F G F
(N3)
Calorific Capacity
G G G P G P P F P F F
(N4)
Difficulty with the
Component’s Dis- VP P VP F VG F VP VP VP P P
assembly (N5)
Level of Integra-
VP P VP F VG F VP VP VP P P
tion (N6)
Quantity of Parts
VP VP P G VG VG G P F F P
(N7)
Difficulty of Dis-
mantling Part At- VP P VP F VG F VP VP VP P P
tachment (N8)
Amount of Differ-
G G G F G F G P G P P
ent Materials (N9)
Amount of Haz-
ardous Materials G G G F G F F G F G G
(N10)
Component
F F P P VG G G F VG F P
Weight (N11)
Regulation Sup-
G G F F F F F F F F F
port (N12)
Customer Prefer-
G G F F F F F F F F F
ence (N13)
Succedaneum
G F F F F F F F F F F
Price (N14)

121
Appendix C
Coffee Maker Fuzzy Calculation

Table 1 Coffee Maker Component Fuzzy Set


Filter Basket Holder
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (1.36, 8.66, (0, 1.25,
(5, 7.5, 10) (0, 2.5, 5) (5, 7.5, 10) (0, 0, 2.5)
mance 40) 4.875)
Environmental (12.5, 37.5, (0, 0.0208,
(7.5, 10, 10) (5, 7.5, 10) (0, 18.75, 50) (0, 0, 1.62)
Performance 75) 0.144)
Social Perfor-
(0,0,2.5) (0, 0.25, 0.88) (2.5, 5, 7.5) (5, 7.5, 10) (0, 0, 2.5) (0, 0, 2.5)
mance
Lid
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (0.89, 8.66, (0, 1.25,
(5, 7.5, 10) (0, 2.5, 5) (5, 7.5, 10) (0, 0, 2.5)
mance 60) 4.875)
Environmental (6.25,25, (0, 16.44, (0, 0,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5)
Performance 56.25) 60.04) 28.44)
Social Perfor- (1.25, 1.875,
(0,2.5, 5) (0, 0.25, 12.5) (5, 7.5, 10) (0, 0, 2.5) (0, 0, 2.5)
mance 4.375)
Warming Plate
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (0.8645, (0.625, 2.5,
(5, 7.5, 10) (2.5, 5, 7.5) (0,2.5,5) (0, 0, 2.5)
mance 8.66, 60) 7.3125)
Environmental (6.25,25,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5) (0,1.2,7.2) (0, 0, 3.6)
Performance 56.25)
Social Perfor-
(5,7.5,10) (0.42,1.13,3.25) (2.5,5,7.5) (2.5,5,7.5) (0, 0, 2.5) (0, 0, 2.5)
mance
Main Housing
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (2.66,8.66,20
(5, 7.5, 10) (0,2.5,5) (0,1.25,4.875) (5,7.5,10) (0, 0, 2.5)
mance )
Environmental (6.25,25,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5) (0,0,0.9) (0, 0, 0.45)
Performance 56.25)
Social Perfor- (6.875,9.375,
(7.5,10,10) (0.75,1.33,2) (5,7.5,10) (0, 0, 2.5) (0, 0, 2.5)
mance 10)
Heating Pipe
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (2.66,8.16,8. (0.625,2.5,7.3
(5, 7.5, 10) (2.5,5,7.5) (0,2.5,5) (0, 0, 2.5)
mance 16) 125)
Environmental (6.25,25, (0,0.0208,0.1 (0, 0,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5)
Performance 56.25) 44) 0.072)
Social Perfor-
(7.5,10,10) (0.5,1.25,3.25) (2.5,5,7.5) (2.5,5,7.5) (0, 0, 2.5) (0, 0, 2.5)
mance
Carafe
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (2.66,8.16,8. (0.625,2.5,7.3
(5, 7.5, 10) (2.5,5,7.5) (0,2.5,5) (0, 0, 2.5)
mance 16) 125)
Environmental (6.25,25, (0,0.0208,0.1 (0, 0,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5)
Performance 56.25) 44) 0.072)
Social Perfor- (1.25,1.875,4
(5,7.5,10) (0.17,0.38,1.75) (5,7.5,10) (0, 0, 2.5) (0, 0, 2.5)
mance .375)
Carafe Handle
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (2.66,8.16,8. (0.625,2.5,7.3
(5, 7.5, 10) (2.5,5,7.5) (2.5,5,7.5) (0, 0, 2.5)
mance 16) 125)

122
Environmental (6.25,25, (0, 0.0208,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5) (0, 0, 1.62)
Performance 56.25) 0.144)
Social Perfor- (0,0.625,3.12
(0,2.5,5) (0,0.13,1.25) (0,2.5,5) (0, 0, 2.5) (0, 0, 2.5)
mance 5)
Bottom Plate
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (0.625,2.5,7.3
(0,4.72,4.72) (5, 7.5, 10) (2.5,5,7.5) (0,2.5,5) (0, 0, 2.5)
mance 125)
(0,
Environmental (6.25,25, (0, 0,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5) 0.016,0.0936
Performance 56.25) 0.0468)
)
Social Perfor- (1.25,1.875,4
(2.5,5,7.5) (3.125,9.375,37.5) (5,7.5,10) (0, 0, 2.5) (0, 0, 2.5)
mance .375)
Power Cord
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (2.66,8.16,8. (1.25,1.75,9.7
(5, 7.5, 10) (5,7.5,10) (2.5,5,7.5) (0, 0, 2.5)
mance 16) 5)
Environmental (6.25,25, (0, 0.0208,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5) (0, 0, 1.62)
Performance 56.25) 0.144)
Social Perfor-
(2.5,5,7.5) (0.08,0.63,2.25) (0,2.5,5) (0,2.5,5) (0, 0, 2.5) (0, 0, 2.5)
mance
Switch
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (2.66,8.16,8. (0.625,2.5,7.3
(5, 7.5, 10) (2.5,5,7.5) (2.5,5,7.5) (0, 0, 2.5)
mance 16) 125)
Environmental (6.25,25, (0, 16.44, (0, 0,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5)
Performance 56.25) 60.04) 28.44)
Social Perfor-
(0,2.5,5) (0,0.5,2) (0,2.5,5) (0,2.5,5) (0, 0, 2.5) (0, 0, 2.5)
mance

Table 2 Coffee Maker Component Defuzzification Number

Filter Basket Holder


Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
5.4 5.2 4.8 4.18 5.2 5.1
mance
Environmental
4.75 5.2 5.8 5.6 4.51 4.91
Performance
Social Perfor-
5.1 4.58 5 5.2 5.1 5.1
mance
Lid
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
5.73 5.2 4.8 4.18 5.2 5.1
mance
Environmental
4.75 5.2 5.77 5.37 5.68 4.97
Performance
Social Perfor-
4.8 5.23 5.67 5.2 5.1 5.1
mance
Warming Plate
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
5.72 5.2 5 4.83 4.8 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.61 5.07
Performance
Social Perfor-
5.2 4.78 5 5 5.1 5.1
mance
Main Housing

123
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
5.56 5.2 4.8 4.18 5.2 5.1
mance
Environmental
4.75 5.2 5.77 5.37 5.01 5.01
Performance
Social Perfor-
4.75 4.62 4.72 5.2 5.1 5.1
mance
Heating Pipe
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
4.92 5.2 5 4.83 4.8 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.51 4.81
Performance
Social Perfor-
4.75 4.77 5 5 5.1 5.1
mance
Carafe
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
4.92 5.2 5 4.83 4.8 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.51 5.01
Performance
Social Perfor-
5.2 4.52 5.67 5.2 5.1 5.1
mance
Carafe Handle
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
4.92 5.2 5 4.83 5 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.51 4.91
Performance
Social Perfor-
4.8 4.59 4.71 4.8 5.1 5.1
mance
Bottom Plate
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
5.18 5.2 5 4.83 4.8 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.12 5.01
Performance
Social Perfor-
5 5.26 5.67 5.2 5.1 5.1
mance
Power Cord
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
4.92 5.2 5.2 5.75 5 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.51 4.91
Performance
Social Perfor-
5 4.59 4.8 4.8 5.1 5.1
mance
Switch
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
4.92 5.2 5 4.83 5 5.1
mance
Environmental
4.75 5.2 5.77 5.37 5.68 4.97
Performance
Social Perfor-
4.8 4.56 4.8 4.8 5.1 5.1
mance

124
Appendix D
Refrigerator EOL Strategy Sustainability Value
Table 1 Component EOL Strategy Cost ($)

No. Component Re- Remanufac- Primary Re- Secondary Incinera- Landfill


use ture cycle Recycle tion
1 Cabinet
180 170 21.34 20.82 0 0.94
Frame
2 Cabinet 0 0 23.45 21.25 2.08 1.17
3 Duct in
0 0 0.82 0.75 0.07 0.04
Room
4 Fan Unit 1 0 0 0.43 0.43 0 0.02
5 Fan Unit 2 0 0 0.43 0.43 0 0.02
6 Evaporator 50 42 0.80 0.64 0 0.02
7 Rear Board 27 23 0.89 0.87 0 0.04
8 Compressor 80 71 7.19 7.04 0 0.32
9 Condenser 40 30 2.40 2.35 0 0.11
10 Base 26 19 1.12 1.09 0 0.05
11 Door 1 60 51 2.42 2.38 0 0.11
12 Door 2 70 59 3.90 3.82 0 0.17
13 Gasket 1 0 0 0.03 0.03 0.0028 0.00
14 Gasket 2 0 0 0.05 0.04 0.0043 0.00
15 Door Liner 1 0 0 1.79 1.62 0.16 0.09
16 Door Liner 2 0 0 3.74 3.24 0.32 0.18
17 Control Unit 332 317 3.98 3.75 0.16 0.19
18 Heater 0 0 0.17 0.13 0 0.00
19 Dryer 13 10 0.67 0.59 0 0.00
20 Shelf Set 0 0 1.01 0.92 0.09 0.05

Table 2 EOL Strategy Environment Impact (mPt)

No. Component Re- Remanufac- Primary Re- Secondary Incinera- Landfill


use ture cycle Recycle tion
1 Cabinet
0 708.18 -1652.42 -1432.15 0 33.05
Frame
2 Cabinet 0 0 -5862.6 -4621 -556.95 114.32
3 Duct in
0 0 -205.6 -189 -19.53 4.01
Room
4 Fan Unit 1 0 0 -33.81 -29 0 0.68
5 Fan Unit 2 0 0 -33.81 -29 0 0.68
6 Evaporator 0 38.3 -383.04 -301.6 0 0.74
7 Rear Board 0 29.58 -69.02 -49.67 0 1.38
8 Compressor 0 239.55 -558.95 -444.16 0 11.18
9 Condenser 0 80.07 -186.83 -159.31 0 3.74
10 Base 0 37.2 -86.8 -69.7 0 1.74
11 Door 1 0 80.79 -188.51 -169.84 0 3.77
12 Door 2 0 129.93 -303.17 -219.6 0 6.06
13 Gasket 1 0 0 -8 -3 -0.76 0.16

125
14 Gasket 2 0 0 -12 -6 -1.14 0.23
15 Door Liner 1 0 0 -447.2 -395.63 -42.48 8.72
16 Door Liner 2 0 0 -894.4 -769.21 -84.97 17.44
17 Control Unit 0 83.72 -621 -498 -117.3 12.19
18 Heater 0 0 -80.64 -65.39 0 0.16
19 Dryer 0 3.33 -7.77 -4.66 0 0.16
20 Shelf Set 0 0 -253.2 -139.68 -24.05 4.94

Table 3 EOL Strategy Labor Time (s)

No. Component Reuse Remanufac- Primary Re- Second- Incinera- Landfill


ture cycle ary Recy- tion
cle
Cabinet
1 3400.51 3466.97 623.99 283.94 0 17.00
Frame
2 Cabinet 0 0 774.85 352.59 21.11 21.11
3 Duct in
0 0 27.17 12.37 0.74 0.74
Room
4 Fan Unit 1 0 0 12.77 5.81 0 0.35
5 Fan Unit 2 0 0 12.77 5.81 0 0.35
6 Evaporator 76.64 82.23 14.06 6.40 0 0.38
7 Rear Board 142.04 162.31 26.06 11.86 0 0.71
8 Compressor 1150.26 1162.95 211.07 96.05 0 5.75
9 Condenser 384.48 295.32 70.55 32.10 0 1.92
10 Base 178.63 190.12 32.78 14.92 0 0.89
11 Door 1 387.93 399.01 71.19 32.39 0 1.94
12 Door 2 623.89 648.21 114.48 52.10 0 3.12
13 Gasket 1 0 0 1.06 0.48 0.03 0.03
14 Gasket 2 0 0 1.59 0.72 0.04 0.04
15 Door Liner 1 0 0 59.11 26.90 1.61 1.61
16 Door Liner 2 0 0 118.21 53.79 3.22 3.22
17 Control Unit 673.74 690.36 123.63 56.26 3.37 3.37
18 Heater 0 0 2.96 1.35 0 0.08
19 Dryer 15.99 17.01 2.93 1.34 0 0.08
20 Shelf Set 0 0 33.47 15.23 0.91 0.91

126
Appendix E
Refrigerator EOL Strategy Characteristics Fuzzy Evaluation
Table 1 EOL Strategy Characteristics Fuzzy Evaluation (Make Up Data)
Duct Fan Fan
Cabinet Evapo- Rear Door Door Control Shelf
Cabinet in Unit Unit Compressor Condenser Base Door 1 Door 2 Gasket 1 Gasket 2 Heater Dryer
Frame rator Board Liner 1 Liner 2 Unit Set
Room 1 2
Durability (N1) VG VG F G P G G G F G G G VG G F G G F G P
EOL Condition (N2) G G F F VP F F F P F F F G F P F F F F VP
Quantity of High- P P P
P P F F F F F G F F P F F F G P F
Value Materials (N3)
Calorific Capacity (N4) G G G P G P P F P F F P G P P P F G P G
Difficulty with the VG VP VP
Component’s Disas- P VP F F VP VP VP P P F P F VP F P VP VP
sembly (N5)
Level of Integration VG VP VP
P VP F F VP VP VP P P F P F VP F P VP VP
(N6)
Quantity of Parts (N7) VG VP P G VP VG G P F F P VG VP G F VG F P G VP
Difficulty of Disman- VG VP VP
tling Part Attachment P VP F F VP VP VP P P F P F VP F P VP VP
(N8)
Amount of Different G G G
G G F F G P G P P F G F G F P G G
Materials (N9)
Amount of Hazardous G G G
G G F F F G F G G F G F F F G G F
Materials (N10)
Component Weight VG F F
F P P G G F VG F P G F P VG G F P G
(N11)
Regulation Support F G G
G F F F F F F F F F G F F F F F F
(N12)
Customer Preference F G G
G F F F F F F F F F G F F F F F F
(N13)
Succedaneum Price F G G
F F F F F F F F F F F F F F F F F
(N14)

127
Appendix F
Refrigerator Fuzzy Calculation

Table 1 Refrigerator Component Fuzzy Set


Cabinet
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (1.36, 8.66, (0, 1.25,
(5, 7.5, 10) (0, 2.5, 5) (5, 7.5, 10) (0, 0, 2.5)
mance 40) 4.875)
Environmental (12.5, 37.5, (0, 0.0208,
(7.5, 10, 10) (5, 7.5, 10) (0, 18.75, 50) (0, 0, 1.62)
Performance 75) 0.144)
Social Perfor-
(0,0,2.5) (0, 0.25, 0.88) (2.5, 5, 7.5) (5, 7.5, 10) (0, 0, 2.5) (0, 0, 2.5)
mance
Duct in Room
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (0.89, 8.66, (0, 1.25,
(5, 7.5, 10) (0, 2.5, 5) (5, 7.5, 10) (0, 0, 2.5)
mance 60) 4.875)
Environmental (6.25,25, (0, 16.44, (0, 0,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5)
Performance 56.25) 60.04) 28.44)
Social Perfor- (1.25, 1.875,
(0,2.5, 5) (0, 0.25, 12.5) (5, 7.5, 10) (0, 0, 2.5) (0, 0, 2.5)
mance 4.375)
Fan Unit 1
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (0.8645, (0.625, 2.5,
(5, 7.5, 10) (2.5, 5, 7.5) (0,2.5,5) (0, 0, 2.5)
mance 8.66, 60) 7.3125)
Environmental (6.25,25,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5) (0,1.2,7.2) (0, 0, 3.6)
Performance 56.25)
Social Perfor-
(5,7.5,10) (0.42,1.13,3.25) (2.5,5,7.5) (2.5,5,7.5) (0, 0, 2.5) (0, 0, 2.5)
mance
Fan Unit 2
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
(0,0,0) (5, 7.5, 10) (0,2.5,5) (0,1.25,4.875) (5,7.5,10) (0, 0, 2.5)
mance
Environmental (12.5, 37.5, (0, 0.0208,
(7.5, 10, 10) (5, 7.5, 10) (0,18.75,50) (0, 0, 1.62)
Performance 75) 0.144)
Social Perfor- (1.25,1.875,
(0,0,2.5) (0,0,2.5) (5,7.5,10) (0, 0, 2.5) (0, 0, 2.5)
mance 4.375)
Evaporator
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (2.66,8.16,8. (0.625,2.5,7.3
(5, 7.5, 10) (2.5,5,7.5) (0,2.5,5) (0, 0, 2.5)
mance 16) 125)
Environmental (6.25,25, (0,0.0208,0.1 (0, 0,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5)
Performance 56.25) 44) 0.072)
Social Perfor-
(7.5,10,10) (0.5,1.25,3.25) (2.5,5,7.5) (2.5,5,7.5) (0, 0, 2.5) (0, 0, 2.5)
mance
Rear Board
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (2.66,8.16,8. (0.625,2.5,7.3
(5, 7.5, 10) (2.5,5,7.5) (0,2.5,5) (0, 0, 2.5)
mance 16) 125)
Environmental (6.25,25, (0,0.0208,0.1 (0, 0,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5)
Performance 56.25) 44) 0.072)
Social Perfor- (1.25,1.875,4
(5,7.5,10) (0.17,0.38,1.75) (5,7.5,10) (0, 0, 2.5) (0, 0, 2.5)
mance .375)
Compressor
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (2.66,8.16,8. (0.625,2.5,7.3
(5, 7.5, 10) (2.5,5,7.5) (2.5,5,7.5) (0, 0, 2.5)
mance 16) 125)

128
Environmental (6.25,25, (0, 0.0208,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5) (0, 0, 1.62)
Performance 56.25) 0.144)
Social Perfor- (0,0.625,3.12
(0,2.5,5) (0,0.13,1.25) (0,2.5,5) (0, 0, 2.5) (0, 0, 2.5)
mance 5)
Condenser
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (0.625,2.5,7.3
(0,4.72,4.72) (5, 7.5, 10) (2.5,5,7.5) (0,2.5,5) (0, 0, 2.5)
mance 125)
(0,
Environmental (6.25,25, (0, 0,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5) 0.016,0.0936
Performance 56.25) 0.0468)
)
Social Perfor- (1.25,1.875,4
(2.5,5,7.5) (3.125,9.375,37.5) (5,7.5,10) (0, 0, 2.5) (0, 0, 2.5)
mance .375)
Base
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (2.66,8.16,8. (1.25,1.75,9.7
(5, 7.5, 10) (5,7.5,10) (2.5,5,7.5) (0, 0, 2.5)
mance 16) 5)
Environmental (6.25,25, (0, 0.0208,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5) (0, 0, 1.62)
Performance 56.25) 0.144)
Social Perfor-
(2.5,5,7.5) (0.08,0.63,2.25) (0,2.5,5) (0,2.5,5) (0, 0, 2.5) (0, 0, 2.5)
mance
Door 1
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (2.66,8.16,8. (0.625,2.5,7.3
(5, 7.5, 10) (2.5,5,7.5) (2.5,5,7.5) (0, 0, 2.5)
mance 16) 125)
Environmental (6.25,25, (0, 16.44, (0, 0,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5)
Performance 56.25) 60.04) 28.44)
Social Perfor-
(0,2.5,5) (0,0.5,2) (0,2.5,5) (0,2.5,5) (0, 0, 2.5) (0, 0, 2.5)
mance
Door 2
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (2.66,8.16,8. (0.625,2.5,7.3
(5, 7.5, 10) (2.5,5,7.5) (0,2.5,5) (0, 0, 2.5)
mance 16) 125)
Environmental (6.25,25, (0,0.0208,0.1 (0, 0,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5)
Performance 56.25) 44) 0.072)
Social Perfor-
(7.5,10,10) (0.5,1.25,3.25) (2.5,5,7.5) (2.5,5,7.5) (0, 0, 2.5) (0, 0, 2.5)
mance
Gasket 1
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (1.36, 8.66, (0, 1.25,
(5, 7.5, 10) (0, 2.5, 5) (5, 7.5, 10) (0, 0, 2.5)
mance 40) 4.875)
Environmental (12.5, 37.5, (0, 0.0208,
(7.5, 10, 10) (5, 7.5, 10) (0, 18.75, 50) (0, 0, 1.62)
Performance 75) 0.144)
Social Perfor-
(0,0,2.5) (0, 0.25, 0.88) (2.5, 5, 7.5) (5, 7.5, 10) (0, 0, 2.5) (0, 0, 2.5)
mance
Gasket 2
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (0.8645, (0.625, 2.5,
(5, 7.5, 10) (2.5, 5, 7.5) (0,2.5,5) (0, 0, 2.5)
mance 8.66, 60) 7.3125)
Environmental (6.25,25,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5) (0,1.2,7.2) (0, 0, 3.6)
Performance 56.25)
Social Perfor-
(5,7.5,10) (0.42,1.13,3.25) (2.5,5,7.5) (2.5,5,7.5) (0, 0, 2.5) (0, 0, 2.5)
mance
Door Liner 1
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (0.625,2.5,7.3
(0,4.72,4.72) (5, 7.5, 10) (2.5,5,7.5) (0,2.5,5) (0, 0, 2.5)
mance 125)
(0,
Environmental (6.25,25, (0, 0,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5) 0.016,0.0936
Performance 56.25) 0.0468)
)
Social Perfor- (1.25,1.875,4
(2.5,5,7.5) (3.125,9.375,37.5) (5,7.5,10) (0, 0, 2.5) (0, 0, 2.5)
mance .375)

129
Door Liner 2
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (2.66,8.16,8. (0.625,2.5,7.3
(5, 7.5, 10) (2.5,5,7.5) (0,2.5,5) (0, 0, 2.5)
mance 16) 125)
Environmental (6.25,25, (0,0.0208,0.1 (0, 0,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5)
Performance 56.25) 44) 0.072)
Social Perfor-
(7.5,10,10) (0.5,1.25,3.25) (2.5,5,7.5) (2.5,5,7.5) (0, 0, 2.5) (0, 0, 2.5)
mance
Control Unit
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (2.66,8.16,8. (1.25,1.75,9.7
(5, 7.5, 10) (5,7.5,10) (2.5,5,7.5) (0, 0, 2.5)
mance 16) 5)
Environmental (6.25,25, (0, 0.0208,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5) (0, 0, 1.62)
Performance 56.25) 0.144)
Social Perfor-
(2.5,5,7.5) (0.08,0.63,2.25) (0,2.5,5) (0,2.5,5) (0, 0, 2.5) (0, 0, 2.5)
mance
Heater
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (0.89, 8.66, (0, 1.25,
(5, 7.5, 10) (0, 2.5, 5) (5, 7.5, 10) (0, 0, 2.5)
mance 60) 4.875)
Environmental (6.25,25, (0, 16.44, (0, 0,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5)
Performance 56.25) 60.04) 28.44)
Social Perfor- (1.25, 1.875,
(0,2.5, 5) (0, 0.25, 12.5) (5, 7.5, 10) (0, 0, 2.5) (0, 0, 2.5)
mance 4.375)
Dryer
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor- (2.66,8.16,8. (0.625,2.5,7.3
(5, 7.5, 10) (2.5,5,7.5) (0,2.5,5) (0, 0, 2.5)
mance 16) 125)
Environmental (6.25,25, (0,0.0208,0.1 (0, 0,
(7.5, 10, 10) (5, 7.5, 10) (0, 12.5, 37.5)
Performance 56.25) 44) 0.072)
Social Perfor- (1.25,1.875,4
(5,7.5,10) (0.17,0.38,1.75) (5,7.5,10) (0, 0, 2.5) (0, 0, 2.5)
mance .375)
Shelf Set
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
(0,0,0) (5, 7.5, 10) (0,2.5,5) (0,1.25,4.875) (5,7.5,10) (0, 0, 2.5)
mance
Environmental (12.5, 37.5, (0, 0.0208,
(7.5, 10, 10) (5, 7.5, 10) (0,18.75,50) (0, 0, 1.62)
Performance 75) 0.144)
Social Perfor- (1.25,1.875,
(0,0,2.5) (0,0,2.5) (5,7.5,10) (0, 0, 2.5) (0, 0, 2.5)
mance 4.375)

Table 2 Refrigerator Component Defuzzilization Number

Cabinet
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
5.4 5.2 4.8 4.18 5.2 5.1
mance
Environmental
4.75 5.2 5.8 5.6 4.51 4.91
Performance
Social Perfor-
5.1 4.58 5 5.2 5.1 5.1
mance
Duct in Room
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
5.73 5.2 4.8 4.18 5.2 5.1
mance

130
Environmental
4.75 5.2 5.77 5.37 5.68 4.97
Performance
Social Perfor-
4.8 5.23 5.67 5.2 5.1 5.1
mance
Fan Unit 1
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
0 5.2 4.8 4.18 5.2 5.1
mance
Environmental
4.75 5.2 4.52 4.25 4.72 4.58
Performance
Social Perfor-
5.1 5.1 4.75 5.2 4.81 4.81
mance
Fan Unit 2
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
5.72 5.2 5 4.83 4.8 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.61 5.07
Performance
Social Perfor-
5.2 4.78 5 5 5.1 5.1
mance
Evaporator
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
4.92 5.2 5 4.83 4.8 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.51 4.81
Performance
Social Perfor-
4.75 4.77 5 5 5.1 5.1
mance
Rear Board
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
4.92 5.2 5 4.83 4.8 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.51 5.01
Performance
Social Perfor-
5.2 4.52 5.67 5.2 5.1 5.1
mance
Compressor
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
4.92 5.2 5 4.83 5 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.51 4.91
Performance
Social Perfor-
4.8 4.59 4.71 4.8 5.1 5.1
mance
Condenser
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
5.18 5.2 5 4.83 4.8 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.12 5.01
Performance
Social Perfor-
5 5.26 5.67 5.2 5.1 5.1
mance
Base
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
4.92 5.2 5.2 5.75 5 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.51 4.91
Performance
Social Perfor-
5 4.59 4.8 4.8 5.1 5.1
mance
Door 1

131
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
4.92 5.2 5 4.83 5 5.1
mance
Environmental
4.75 5.2 5.77 5.37 5.68 4.97
Performance
Social Perfor-
4.8 4.56 4.8 4.8 5.1 5.1
mance
Door 2
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
4.92 5.2 5 4.83 4.8 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.51 4.81
Performance
Social Perfor-
4.75 4.77 5 5 5.1 5.1
mance
Gasket 1
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
5.4 5.2 4.8 4.18 5.2 5.1
mance
Environmental
4.75 5.2 5.8 5.6 4.51 4.91
Performance
Social Perfor-
5.1 4.58 5 5.2 5.1 5.1
mance
Gasket 2
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
0 5.2 4.8 4.18 5.2 5.1
mance
Environmental
4.75 5.2 4.52 4.25 4.72 4.58
Performance
Social Perfor-
5.1 5.1 4.75 5.2 4.81 4.81
mance
Door Liner 1
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
5.18 5.2 5 4.83 4.8 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.12 5.01
Performance
Social Perfor-
5 5.26 5.67 5.2 5.1 5.1
mance
Door Liner 2
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
4.92 5.2 5 4.83 4.8 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.51 4.81
Performance
Social Perfor-
4.75 4.77 5 5 5.1 5.1
mance
Control Unit
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
4.92 5.2 5.2 5.75 5 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.51 4.91
Performance
Social Perfor-
5 4.59 4.8 4.8 5.1 5.1
mance
Heater
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
5.73 5.2 4.8 4.18 5.2 5.1
mance
Environmental
4.75 5.2 5.77 5.37 5.68 4.97
Performance

132
Social Perfor-
4.8 5.23 5.67 5.2 5.1 5.1
mance
Dryer
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
4.92 5.2 5 4.83 4.8 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.51 5.01
Performance
Social Perfor-
5.2 4.52 5.67 5.2 5.1 5.1
mance
Shelf Set
Sustainability Di- Primary Re- Secondary Incinerate Landfill
Reuse (A1) Remanufacturing (A2)
mension cycle (A3) Recycle (A4) (A5) (A6)
Economic Perfor-
5.72 5.2 5 4.83 4.8 5.1
mance
Environmental
4.75 5.2 5.77 5.37 4.61 5.07
Performance
Social Perfor-
5.2 4.78 5 5 5.1 5.1
mance

133
Appendix G Decomposition Approach
Huang and Kusiak (1998) contributed greatly in the application and method development

for decomposition approach by considering two matrixes: interaction matrix and suitability matrix.

Interaction matrix represents the physical characteristics of components, and suitability matrix

shows how other factors affect components, such as designer preferences and cost consideration.

Following are the steps of Huang and Kusiak’s classical decomposition approach (referred to as

CDA below).

Step 0. Initialization: Initialize the interaction and suitability matrices. Specify the upper

bound 𝑁𝑉 on the number of components in a module and budget 𝐵.

Step 1. Triangularization: Triangularize the interaction matrix 𝐴 into matrix 𝐴′ using trian-

gularization algorithm.

Step 2. Rearrangement: Rearrange the suitability matrix 𝐵 into matrix 𝐵 ′ so that sequence

of columns and rows in matrix 𝐵 ′ is the same as in matrix 𝐴′ .

Step 3. Combination: Combine the matrix 𝐴′ and the matrix 𝐵 ′ into the modularity matrix

(𝐴′ |𝐵 ′ ). Identify modules corresponding to the groups in 𝐴′ . In suitability matrix, A means

strongly desire, O means strongly undesired, E means desired and U means undesired.

Step 4. Deletion: Remove a component from a module if it satisfies Condition 1, and place

it in the last column of the modularity matrix. Repeat this step until no more components can be

removed.

Step 5. Duplication: Duplicate a component that satisfies Condition 2, and repeat this step

until no more components can be duplicated.

Step 6. Classification: Analyze the modularity matrix to classify the modules.

Step 7. Termination: Stop and output the results.

134
Condition 1: Remove a component 𝑘, if the following conditions are satisfied.

1) Component 𝑘 and any other component 𝑙 that appear in the same module are strongly

undesired for inclusion in the module,

2) Component 𝑘 interacts with the remaining components in the module to a lesser degree

than component 𝑙,

3) None of the sub-matrices violates constraints C1 and C2.

Condition 2: Duplicate the component if the following conditions are satisfied.

1) The component that is used and strongly desired for inclusion in two modules simultane-

ously

2) None of the submatrices violates constraints C1 and C2.

Constraint C1: Empty modules of components are not allowed,

Constraint C2: The number of components in a module cannot exceed the upper bound (𝑁𝑉 ),

and the total cost of the components duplicated cannot exceed 𝐵.

135
Appendix H ASCEM Approach (Adopted from Chung, 2012)

The ASCEM approach involves seven steps.

ASCEM Flow Chart

136
Vita
Junfeng Ma was born in Taiyuan, China, on April 15, 1985. He received bachelor degree in Me-

chanical Engineering and Automation from Xi'an Jiaotong University in China at 2007. After grad-

uation, he came to U.S.A for engineering higher education. He received master degree in Engi-

neering Management at Wilkes University, PA, in May, 2010. He joined Ph.D. program in The

Pennsylvania State University in Fall, 2010. During his stay at Penn State, he served as instructor

of EDSGN 100 (Introduction to Engineering Design) in the School of Engineering Design, Tech-

nology, and Professional Programs (SEDTAPP). He will receive Industrial Engineering and Op-

erations Research dual title doctoral degree in May, 2016. His research concentrates on implemen-

tation of operations research and data analysis in design engineering, specifically, sustainability

driven product design and manufacturing, design for product life cycle uncertainty and product

life cycle sustainability assessment and optimization. He is active member of IIE, ASME, DS and

ASEE.

You might also like