Product development knowledge-processing system

Rochester Institute of Technology

RIT Scholar Works

Theses

Thesis/Dissertation Collections

8-2005

Product development knowledge-processing

system

Jiefei Qiu

Follow this and additional works at:

http://scholarworks.rit.edu/theses

This Thesis is brought to you for free and open access by the Thesis/Dissertation Collections at RIT Scholar Works. It has been accepted for inclusion in Theses by an authorized administrator of RIT Scholar Works. For more information, please [email protected].

Recommended Citation

Rochester Institute of

Technology

PRODUCT DEVELOPMENT KNOWLEDGE-PROCESSING SYSTEM

A Thesis

Submitted inpartialfulfillmentofthe

requirementsforthedegreeof

MasterofScience in Industrial

Engineering

inthe

DepartmentofIndustrial & Systems

Engineering

Kate Gleason College of

Engineering

by

Jiefei Qiu

B.E.,Industrial _Design,Hefei

University

of_Technology,2003

DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING

KATE GLEASON COLLEGE OF ENGINEERING

ROCHESTER INSTITUTE OF TECHNOLOGY

ROCHESTER, NEW YORK

CERTIFICA TE OF APPROVAL

M.S.

DEGREE THESIS

The M.S. Degree Thesis of liefei Qiu

has been examined and approved by the

thesis committee as satisfactory for the

thesis requirement for the

Master of Science degree

Approved by:

Paul H. Stiebitz

Professor Paul Stiebitz, Thesis Advisor

Edward Hensel

Thesis/Dissertation Author Permission Statement

Title of thesis or dissertation: _ _ _ _ _ _ _ _ _ -:-_---=,---_____

-fNldud

!Jeve~ &4~ P/'4)t:'~7 -¥~

Nrumeofauthor:

~~JjL/~·e~f~~~/-·--~Ql~/u'u~---Degree: /Vl,rs ~

(21-

J'u

'e-y:.e

Program: J~f~ ~~

College: u~

C?j

~eec7

I understand that I must submit a print copy of my thesis or dissertation to the RIT Archives, per current RIT guidelines for the completion of my degree. I hereby grant to the Rochester Institute of Technology and its agents the non-exclusive license to archive and make accessible my thesis or dissertation in whole or in part in all forms of media in perpetuity. I retain all other ownership rights to the copyright of the thesis or dissertation. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.

Print Reproduction Permission Granted:

I, , hereby grant permission to the Rochester Institute Technology to reproduce my print thesis or dissertation in whole or in part. Any reproduction will not be for commercial use or profit.

Signature of Author: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Date: ______ __

Print Reproduction Permission Denied:

L

,

T/e

f

e

i'

a "

u , hereby deny permission to the RIT Library of the

Rochester Institute of Technology to reproduce my print thesis or dissertation in whole or in part.

Jiefei Qiu

y

Signature of Author: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Date:

"J/

1)6/

[2 0 t>.J

Inclusion in the RIT Digital Media I1brary Electronic Thesis

&

Dissertation (ETD) Archive

I, , additionally grant to the Rochester Institute of Technology Digital Media Library (RIT DML) the non-exclusive license to archive and provide electronic access to my thesis or dissertation in whole or in part in all forms of media in perpetuity.

I understand that my work, in addition to its bibliographic record and abstract, will be available to the world-wide community of scholars and researchers through the RIT DML. I retain all other ownership rights to the copyright of the thesis or dissertation. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I rum aware that the Rochester Institute of Technology does not require registration of copyright for EIDs.

I hereby certify that, if appropriate, I have obtained and attached written permission statements from the owners of each third party copyrighted matter to be included in my thesis or dissertation. I certify that the version I submitted is the same as that approved by my committee.

ABSTRACT

Thisthesis develops a_methodology that integrates

Quality

Function Deployment

(QFD),

Theory

of Inventive Problem

Solving (TRIZ)

and Design Structure Matrix

(DSM)toaddresstheproblem of_{analyzing, interpreting,}and_transformingtheknowledge

dataflow

during

theproductdevelopmentprocess.

A product development process is the sequence of steps an enterprise takes to

conceive, design, andcommercialize aproduct. The general process includes: _Planning,

Concept _Development, System-Level _Design, Detail _Design,

Testing

and _Refinement,

and Production Ramp-Up. One way to think about the development process is as a

knowledge-processing

system.

Priorto thisstudy,therearefewarticles

integrating

TRIZ andQFD. QFDis atool

for _"identifying the needs ofthe customer and _translating the language ofthe customer

into the language of the engineer". The QFD process also _{identifies engineering}

contradictions within the _existing system. TRIZ can provide solutions for these

contradictions. Itoffers one ofthe most efficientmeansto generate creative solutionsfor

thedesired improvementsprioritizedin QFD. Oncea creative solutionis_{developed using}

TRIZ,the_{corresponding}part characteristics canbefurtherdecomposed_{using QFD.}

If we think of_{knowledge-processing} as a work _process, then both rework and

iterative design requires additional project resources and time. Iteration is the repetition

of activities due to the input of new information. DSM is an effective tool to plan and

manage product development processes through information flow analysis

by helping

ACKNOWLEDGEMENTS

First, I thank _my advisor Paul _Stiebitz, for his continuous support in my MS

program. Paulhas beenafriend and mentor. He was alwaysthereto meet and talkabout

myideas, tosolve_anyunsolvable_{problems, to}proofreadand mark_{up my}papers. He had

confidence in me when I doubted myself and brought out the good ideas in me. More

importantly, he taught me how to work hard and _{play hard,} and how to reduce stress!

Withouthisencouragement and constant_guidance, Icould nothave finishedthis thesis. I

also thank _{my co-advisor,} Dr. Edward _Hensel, who guided _my thesis and gave me

insightful comments. He always asked me good questions to

help

me thinkthrough _my

problems.

I am also grateful to the

following

people at RIT: _Marilyn, Dr. _Mozrall, Dr.

Esterman andDr. _Thorn,for

helping

me at _anytime with various aspects ofthis thesis. I

also appreciatethecontributionfromthe Industrial and Systems

Engineering

Department

for providingmethe_{necessary computing facilities}towork onthisthesis.

All _others, who have provided direct or indirect assistance to me, are also

gratefully acknowledged. The preparation ofthis thesis would not have been possible

withoutthesupport and valuable contributions oftheseindividualsand organizations.

Last, butnot least, I thank_{my family: my}parents,

Guoqiang

Qiu, and Shui_Chen,

for _giving me life, for _educating me for unconditional support and encouragement to

pursue _my interests, even when the interests went beyond boundaries oflanguage and

geography; my elder brother Jiewen Qiu, for _sharing his experience of how to do

TABLE OF

CONTENTS

LIST OFFIGURES vi

LIST OFTABLES ix

NOMENCLATURE x

1 PROBLEM STATEMENT 1

2 INTRODUCTION 4

2.1 IntroductiontoQFD 4

2.2 IntroductiontoTRIZ ₉

2.3 IntroductiontoDSM 11

3 LITERATURE REVIEW 14

3.1 AnIntroductionto _QFD,TRIZandDSM 14

3.2 The_QFD,TRIZandTaguchi Connection: Customer-Driven Robust Innovation

(Terninko) 17

3.3 A Frameworkfor Design Process Specifications Management_(Yassine,and

Falkenburg) 18

3.4 Product Development Process Captureand

Display Using

Web-Based

Technologies_(Sabbaghian, Eppingerand_Murman) 19

3.5

Apply

theDesign Structure Matrixto System DecompositionandIntegration

Problems: A ReviewandNewDirections_(Browning) 20

3.6 DesignofObjective Functions for OptimizationofMulti-DomainSystems

(Andersson,Jochenand_Petter) 21

3.7 The Design Structure System: A Method for

Managing

theDesignofComplex

Systems(Steward "The Design Structure

System")

23

3.8

Planning

and

Managing

theDesignofSystems (Steward

"Planning

and

4 PRODUCTDEVELOPMENT KNOWLEDGE-PROCESSINGSYSTEM 27

4.1 ArchitectureofHPDeskJet 1200C 27

4.2 The NetworkofICOMs overview 29

4.3 NetworkofICOMs forQFD 31

4.4 NetworkofICOMsbetweenQFDandTRIZ 39

4.5 Networkof_{ICOMs betweenQFD}andDSM 53

5 _{RECOMMENDATIONS}ANDCONCLUSION 69

5.1 Conclusions fromthisresearch 69

5.2 Futureresearch 71

6BIBLIOGRAPHY 72

APPENDIX 1 House of

Quality

Model description 77

APPENDIX2 Customer Requirements Definitions 78

APPENDIX 3

Engineering

Characteristics Definitions 79

APPENDEX4 _RelationshipbetweenCRs andECs 85

APPENDIX5 House 1 for QFD 86

APPENDIX6 TRIZ Report using CREAX 87

LIST OF FIGURES

Figure 1 1

Figure 2 Dataflow betweenTRIZand QFD 1

Figure3 Data flowbetweenDSMandQFD 2

Figure4 Data flow in_{TRIZ, QFD,DSM (Time sequence)} 2

Figure 5 ICOM 3

Figure 6 ICOMexample 3

Figure 7 Houseof

Quality

4

Figure 8 Graph fora system ₁₂

Figure 9 Matrixforthesystem 12

Figure 10 DSM_taxonomy 15

Figure 11 Sample Task-based DesignStructureMatrix 15

Figure 12 Sample DSMwith explicitinformation flow 16

Figure 13 Multi-tiered DSM configuration sample 16

Figure 14 Design Structure Matrixforthe

landing

geardesignprocess 22

Figure 15

Relationship

matrixforthe

landing

gear system 23

Figure 16 25

Figure 17 26

Figure 18 26

Figure 19 DeskJet 1200Carchitecture (Dangelo₆₉₎ 27

Figure20 Overviewofthemedia path oftheHP DeskJet 1200Cprinter(Broder₇₂₎

28

Figure 21 ICOM 29

Figure 22 ICOMexample 29

Figure23 _HOQ Model (Adapted from Fig. 2ofLai-Kow's paper) 31

Figure 24 A graphic network ofICOMs for QFD 31

Figure 25

Using

ICOMstorepresentthesetransforms 31

Figure26 Fl transformfromthenetwork 32

Figure27 Thepair-wise comparison method 32

Figure 30 F4transformfromthenetwork 37

Figure 31 Houseof

Quality

_{(Adapted fromQFD}

Capture)

38

Figure 32 TheNetworkof_{ICOMs betweenQFD}andTRIZ 39

Figure 33 TheNetworkof_{ICOMs between}QFD andTRIZ 39

Figure 34

Using

ICOMstorepresentthese transforms 39

Figure35 F5transformfromthenetwork 40

Figure 36Content forF5transform 41

Figure 37 F6transformfromthenetwork 42

Figure 38 TRIZprocess 42

Figure 39Contentfor F6transform 46

Figure 40 F7transformfromthenetwork 47

Figure 41 Content for F7transform 48

Figure 42 F8 transformfromthenetwork 48

Figure 43 F8transformfromthenetwork 49

Figure 44 F9transformfromthenetwork 50

Figure 45 Contentfor F9transform 50

Figure46 F10transformfromthenetwork 51

Figure 47 F8toF10transform 51

Figure48 Content for F10transform 51

Figure 49 _HOQ 1 53

Figure 50_HOQ2 52

Figure5 1 Linked housesfromCustomer RequirementstoProduction Requirements

52

Figure 52The NetworkofICOMs between QFDandDSM 53

Figure 53 53

Figure 54 Fl 1 transformfromthenetwork 54

Figure55 Contentforthe originalDSM 55

Figure56 F12transformfromthenetwork 56

Figure 57 Content for Fl 2transform 57

Figure 58 F13transform fromthenetwork ₅₈

Figure 60 Relative Importance from

HOQ

1 61

Figure 61 F15transform fromthenetwork 62

Figure 62 PartitionedDSMwithTechnicalRisk & Relative Importance 63

Figure 63

Tearing

heuristics 63

Figure 64 DSMafterprocess_A,B 65

Figure 65 DSMafterprocess C 66

Figure66 DSMafter processD 67

Figure 67 DSMafterprocessE 68

Figure 68Houseof

Quality

Model 77

Figure 69 Curl 78

Figure 70

Drop

placementerrors caused

by

swath advance 79

Figure 71

Banding

caused

by

misdirection 80

Figure 72 Waittime

banding

80

Figure 73

Mottling

81

Figure 74 Bleed 82

Figure 75

Spray

83

Figure 76 Shim spray 83

Figure 77Jagged lines 83

Figure 78

Relationship

between CRandEC 85

Figure 79 House 1 for QFD (Adapted from QFD_Capture) 86

Figure 80 Linkedhousesfrom Customer Requirementsto Production Requirements

LIST

OF TABLES

Table 1 SevenCustomer Requirements forour product 33

Table 2 ImportancePriorities forCRs 33

Table 3 Eighteen

Engineering

Characteristics 35

Table 4Content forF3transform 36

Table 5 Final Importance Ratings 38

Table 6 14solutions fromTRIZ 45

Table7 Content for EC 55

Table 8 Technical Risks forthe

Engineering

Characteristics 59

Table 9

Tearing

heuristics 64

NOMENCLATURE

ICOM:

1. "The term ICOM refers _collectively to the _group of information flows

between activities which have one offour roles in the activity"

(Department

of Defense par. 5). _{I, C,} O and M represent _{Input, Control,} Output and

Mechanismrespectively.

2. I represents a matrix as the _{input of data "used} to produce the output of an

activity"

(Department ofDefensepar. 5). Orepresents a matrix as theoutput

of data "produced

by

or _resultingfromthe activity"

(Department ofDefense

par. 5). Both I andO are _usually expressed

by

two orthree capital characters

such as _CR, ECand so on.

3. Therectangle represents the function process,

including

the Mechanism that

is under constraint of Control. Control is defined as "Information that

constrains or controls an

activity"

(Department of Defense par. 5).

Mechanism is defined as

"Usually

people, machines, or systemsthatperform

theactivity"

(DepartmentofDefensepar. 5).

4. Thesubscript ontherightofI/Orepresentsthe sequence oftheinputs/outputs

throughthatfunction.

5. This thesis focuses on the data exchange

during

the product development

process, sothecontentforM andOwill notbe discussed here.

Ordered Set:

A=

(ai,a2,a3,...,aj) isan"orderedset"

withj elements,andalsoa

lxj

matrixincases

when_aj e

1R (

Jechpar.2). The definition for"orderedset"

is introducedon

http://plato.stanford.edU/entries/set-theory/primer.html#l.

B=

(bi,b2,b3,...,

bj)

is justanordered set_withj elementsincases when

bj

_g

R

C2, '12

-22

'\N

'IN

CM\ CM2 'MN

isa matrix suchthat cmn e

CRm,m=l,2, ..

., M (M=7inthiscase): Customer Requirements forour product.

IPm,m=l,l,...,M (M=7 inthiscase): Importance PrioritiesforCRs.

Fl represents the assignment of relative importance _IPm to the customer requirement

CRm,usingthepair-wise comparisonmethod.

(ECj

|

CRj depends on_ECj), i =1,2,...,M and

j

= _1,2,.

..,N: Anordered set of

Engineering

\f

Characteristics upon which a given Customer Requirement _depends, EC =

(_J

(ECj

|

/=i

CRjdepends onECj).

F2 isthe_mappingofCRtoEC.

MT: A MxNmatrix, showingtheextenttowhich each_CRmis dependentupon eachECn..

F3: Define MTmn=

0 if CRmis independentof_ECn;

MTmn=

1 if_{CRmdepends weakly}on_ECn;

MTmn=

MTmn=

9if_CRmdepends_stronglyonECn.

\/m,n m=l,2,...,Mandn=l,2,...,N.

FR: The FinalImportanceratingsforeachelementofEC.

F4: Thetransform togettheFRvalues.

FR= IPT _x MT.

CT =

(CTU

CT,,3

CTM

... CT,jN CT2,3 CT2,4 ... CT2,N CTN.,,N), showing

"1" or

"0"

for the contradiction _relationship between _every two elements of the

Engineering

Characteristics.

Where

CT^y

=_{0 if}

thereisno contradictionbetween_ECXand_ECy;

CTx,y

= _{1 if}

thereisacontradictionbetweenECX andECy.

Vx,y

x=l,2,...,N-l andy=2,...,N.

F5: It represents the contradiction selection from the

Engineering

_{Characteristics, using}

theknowledge fromengineering, physics,designandotherfields.

SL"'5' =

(SLxyiSL"0^... SL*^): Anordered set ofsolutions,only if

CT^

= _1.

F6 represents the TRIZ process, using the contradiction to provide creative solutions

(SL).

SS: Anordered set ofsolutions, selectedfrom yl

USL^U... SLxwyw.

F7 represents the selected solutions

(SS)

from SLX _y

U SL y U ... SLxwyw, based on

ouranalysis andknowledge.

NPC =

(NPCiNPC2... NPCpi+P2+...+pw)

=

(NPCi NPC2 ... NPCE), E is the total number

of part characteristics: Anorderedset shows_{corresponding}partcharacteristicsfor SS.

F8: To

identify

thepart characteristics(NPC)relatedwithSS.

F9 transforms the _engineering characteristics in House 1 into the new _engineering

characteristicsinHouse2.

F10: Oncethiscreative solution is developed _usingTRIZ, the_{correspondingengineering}

characteristicscanbe further decomposed into_{House 2.}

DM: Itshowsthe_{DSM relationship} matrix with 18 variables.

Fl 1 representstheDSMprocess.

DP: The DSM relationshipmatrix after

being

partitioned.

F12represents the_partitioningprocess.

TR: The Technical Riskvaluesfor_{corresponding engineering}characteristics.

F13: To accesstheTechnical Riskvaluesforeach ofthe_engineeringcharacteristics.

RI: The Relative Importance valuesfor corresponding engineeringcharacteristics.

F14: ToaccesstheRelative Importancevalues foreach ofthe_engineeringcharacteristics.

TS: Themoreeffectivetimesequencefortheproductdevelopmentplan.

1

PROBLEM

STATEMENT

As Fig. 1 _{shows, this thesis} develops a _methodology, which integrates

Quality

Function Deployment _(QFD),

Theory

ofInventive Problem

Solving (TRIZ)

andDesign

Structure Matrix

(DSM)

to transform the knowledge data flow

during

the product

developmentprocess.

A

**"

TRIZ

QFD

DSM

Figure 1

The process to _apply this _{methodology is described} as follow. First of_all, QFD

identifies the language of the customers and translates them into the language of the

engineers, which is referred to as _"engineering

characteristics"

(Terninko par. 1). The

QFDprocess also identifies_engineeringcontradictions withinthe_existingproduct. TRIZ

thenprovides creative solutions _using these contradictions forthe desired improvements

prioritizedin QFD. Once a creative solutionis developed using _TRIZ,the _{corresponding}

part characteristics ofthis solution canbe further decomposedinto House 2 using QFD.

Thisistheprocesstointegrate QFDand_TRIZ,whichis shownin Fig. 2 below.

Designsolutions

Eng.Char.

TRIZ

Contradictions

QFD

Secondly, these _engineering characteristics become variables in DSM.

By

reordering the _variables, DSM provides a more effective product development plan for

the complexprocess system throughinformationflow_analysis, withtheinput parameters

such asRelative Importanceand _{TechnicalRisk.Thisprocessis shownin Fig. 3 below.}

Time SequenceforEC,s

DSM QFD

EngineeringCharacteristics RelativeImportance

Figure3 Dataflow betweenDSMandQFD

Fig. 4 showsthedataflowin_QFD,TRIZandDSM usingtimesequence.

Time SequenceforPCs

Customer Needs

Relative Importance! _i-\Q*7

QFD

Eng Char

Contradiction^*

TRIZ

Designsolutions

Eng. Char.

Figure4 Data flow inTRIZ, QFD,DSM(Timesequence)

WeuseatoolcalledICOMas shownin

Fig

5 and6

during

thisprocess. _{I, C,}Oand

M represent _{Input, Control,} Output and Mechanism respectively. Input is transformed

intoOutput

by

_usingtheMechanismthatisunder constraint ofControl.Thisfunction can

be expressed as Fi. i is the sequence ofthe function. Function is defined as: "A rule of

correspondence between two sets such that there is a unique element in the second set

assignedtoeach elementinthefirstset"

(Dictionary

par.2). For_example, inFig. 5 and_6,

represents the control such as our _{manufacturing capability} and so _on; M represents the

mechanism such as QFD and so on. Based on our _{manufacturing capability,} Fl means

that we can use QFD as a tool to analyze our product and our competitor's _product, to

establishthedetails forourfutureproduct.

Control_(C)

Input_(I)

Function i

Mechanism_(M)

Output₍₀₎

[image:19.539.139.382.156.363.2]

?

Figure 5 ICOM

Suchas Manufacturing capability...

Ourproduct

Competitor'sproduct

Function 1

Our futureproduct

SuchasQFD...

Figure 6 ICOMexample

Priorto this study, there are fewarticles

integrating

_{TRIZ, QFD,} and DSM. It is

an innovative trial, but is it reasonable? In particular, the

following

research questions

will beanswer:

1. Whatnewknowledgeis gained

by

combingQFD, TRIZandDSM

during

thedesignphase of a product?

2. Whatprocessshouldbe followedtoobtainthenewknowledge?

3. What datastructure representstheknowledge?

In an effort to answer these questions, this paper uses the product development

process from a knowledge _processing perspective. The knowledge processing system is

2

INTRODUCTION

2.1 Introduction to QFD

Quality

Function Deployment

(QFD)

is astructured methodto translatethe voice of customers into the voice ofthe engineers to meet the needs (Terninko par. 1). The

House of

Quality

is a basic design tool of QFD. It is composed of several _tables,

including

the Customer Requirement _(CR),

Engineering

Characteristics _(EC), the

Relationship

Matrix between _CR&EC, the Tradeoff_{Table, Benchmarking,} and Overall Importance Table. Alloftheseparts are shownbelowin Fig. 7.

Figure7 Houseof_Quality

Theprocess for using Houseof

Quality

is explainedasbelow:

1. ListtheCustomer Requirements_(CR)onthe leftside ofthe matrix.The

Customer Requirements are _usually acquired through customer

[image:20.539.79.443.307.575.2]

First of all, the customers for a product are identified. In the consumer

market, theremightbe different kindsof customers. For_example,iftheproductis

a _{cellphone, the} customers include both the _young students and old people. The

driven desire to purchase the cellphone for these two kinds of people is _usually

different. "There are evenmultiple customervoices within a single organization:

thevoice ofthe_{procuring organization,}thevoice of_{the user,} andthevoice ofthe

supportingormaintenanceorganization"

(Crowpar. 1).

Secondly, the voice of the _customers, which is also called Customer

Requirements _(CR), should be collected. These Customer Requirements are

usually acquired from _{observations, surveys, feedbacks,} interviews and _group

meetings. "Interviews are useful because _they allow youto _effectively probe for

detail. Focus groups are productive because _they allow you to

develop

a lot of

creative ideas

by having

the participants buildupon one another's comments.

In-context customer visits allow team members to _actually observe how customers

use_existingproductsor perform _existingfunctionsand can leadtoa

dramatically

improved_{understanding}of whatthecustomer_really

needs"

(QFD Capturepar. 2).

ThislistofCRisthenenteredintotheleftside ofthematrix as shown above.

2. EvaluatetheImportance Priorities fortheCustomerRequirements.

In_{this process,} compare all ofthe Customer Requirements to each otherto

determine the customer Importance

Priority

_using a 1 to _{5 rating} for each

Customer Requirement. "Use ranking techniques and paired comparisons to

develop

priorities"

3.

Develop

the

benchmarking

evaluation between our product and

competitive products.

The development team compares ourproduct with competitive products, in

orderto betterunderstandwhatthefactorsare

leading

to thepurchase. _Therefore,

the development teamknows whatthe customer likes anddislikes. Based onthis

Customer Competitive _Analysis, the development team should determine the

Customer Competitive

Priority

_Ratings, Customer Performance Goals for the

Customer_{Requirements,}andCustomer Improvement Ratio.

4. Establish the

Engineering

Matrix _(EM), which is composed of

Engineering

Characteristics _(EC), to respond to customer

requirements.

Establish the technical requirements, which are called

Engineering

Characteristics here, to respondto each Customer Requirement. All

Engineering

Characteristics are then formulated the

Engineering

Matrix as the Fig. 7. The

Engineering

Characteristics should be measurable and practical. That is to _say,

"The measurement should be determined without extensive data collection or

testing"

(Crowpar. 3).

5. Decide the improvement direction for each of the

Engineering

Characteristics.

The development team should decide whether each

Engineering

6.

Develop

the

Relationship

Matrix between CR andEC.

The relationship between CR and EC represents how these two lists are

related with each other. _{This relationship} might be defined

by

asking, "To what

degree does this

Engineering

Characteristic predict the customer's satisfaction

withthis requirement?"

(Crowpar. 4). The developmentteam can determine the

scale usedto definetherelationship. Forexample, itcan be used with values of1

through 10. "Standard QFD practice _usually supports the values _{1, 3,} and 9"

(Crowpar. 4). Askthis question for each CR and EC _combination, and then the

answers canbe usedto

develop

the_relationshipmatrix,which isshown above.

7.

Develop

the Target Values and Final Importance Ratings for the

EngineeringCharacteristics.

Setting

Target Values shouldbe basedonthedata followed:

"The team has benchmarked the _existing products to gain a good

understanding of whatlevel of actual performanceis required inordertoproduce

thedesired levelof perceived performance.

By

evaluatingthe Tradeoffs between everytwo

Engineering

Characteristics

in order to determine what compromises _may be required and how those

compromises wouldbe

made"

(QFD Capturepar.26).

After _considering so _many data, the team sets the Target Values for each

are we _{really going} to do with respectto this product or

service?" (QFD Capture

par. 26).

BasedonthisTechnical Competitive_Analysis,thedevelopmentteamshould

determine the Initial Technical _Ratings, the Technical Competitive

Priority

Ratings, Technical Target Values for the

Engineering

_{Characteristics,} Technical

Improvement Ratio andFinalImportance ratings.

8. Use symbols to determine potential positive and negative interactions

between everytwo

Engineering

Characteristics

This can be determined

by

_{asking "If}we improve our performance forone

element of the

Engineering

_{Characteristics,} what is the impact on _every other

EC?"

(QFD Capture par. 27). The development team will determine whether

improvement ofthis

Engineering

Characteristics helps or weaken the effect of

everyother_EngineeringCharacteristics and markthem_usingpositive or negative

symbols.

9. Analyze the matrix and finalize the product development strategy and

product plans. Determinerequired actions and areas offocus.

"Finalize target values. Are target values _properly set to reflect appropriate

tradeoffs? Do target values need tobe adjusted _consideringthe

difficulty

rating?

Are _they realistic with respect to the price points, available _technology, and the

difficulty

rating? Are _they reasonable with respect to the importance ratings?

2.2 Introduction to TRIZ

TRIZ is the acronym for

"Theory

of Inventive Problem Solving" in Russian.

ARIZisthe acronymfor "Algorithm forInventive ProblemSolving"

in Russian. It is the

mainTRIZmethodfor_solvingcontradiction.

ARIZhas been developed_{for manyyears,} sothereare several versionsfor it. The

main parts ofdifferentversions are _{the same,} though _everyversionhas its ownfeatures.

For _{example, the} ARIZ presented in thebook And

Suddenly

the Inventor Appeared has

sevensteps,comparedtoanotherAmericanizedversion of_ARIZ,whichhasninesteps.

The Americanized version of ARIZ is used as example to introduce ARIZ as

below.

1.0 "AnalyzetheSystem" (Marconipar.

15)

The purpose ofthis _step is to build a model forthe problem. The 9-screen

methodisoneofthemethods usedto definetheproblem.

2.0 "Analyzethe Resources"(Marconipar.

15)

The resources include space, time, substances and fields.

By

_analyzing the

resources ofthe system,subsystem, supersystemandthe environment, itprovides

usinformationtodealwiththe_upcomingContradictions.

3.0 "Define the Ideal Final Result and Formulate the Contradiction"

Improving

onepartofa system_mayimpairother partsofthesystem. This is

called a Contradiction. The Contradictions are stated _here, which we need to

remove _{later. The Ideal Final Resultgivesusthe targetoftheproblem.}

Ideal Final Result

(IFR)

cannot be reached _sometimes, but it allows us to

buildapathto thesolution.

4.0 "Separatethe Contradiction"

(Marconipar.

16)

By

using resources which have been analyzed _{in step 2,} we separate the

Contradictioninordertoeliminateit inthelatersteps.

5.0

"Apply

the Knowledge Base: _{Effects, Standards,} and Principles"

(Marconipar. ₁₆₎

Apply

the40principles toprovidethesolutionsfortheContradiction.

6.0 "Changethe 'Mini-Problem'" (Marconipar.

16)

Sometimes, the definition and contradictions of the problem contain the

limitationsof psychological _inertia, and then this _step offers waysto revisit both

ofthem.

7.0 "Review the Solution and Analyze the Removal of the Contradiction"

(Marconi par.

17)

The solutions will be evaluated to see how well it fits the requirements of

8.0

"Develop

Maximum Usage ofthe Solution" (Marconipar.

17)

We generalize the solution into a _method, and then we would like to

maximize the usage ofthis solution. Can this solution be applied to other cases?

Seeking

this answer is the purpose ofthe step. This step cannot give us much

information forour_product,butwillgiveus useful resources inthefuture.

9.0 "Review All the Stages in ARIZ in 'RealTime' Application."

(Marconi

par.₁₇₎

Review what our actual steps were _{in using ARIZ. Compare} the steps with

the standard _process, and then _specify the _difference, add the solutions to our

knowledgedatabase.

2.3 Introduction to DSM

The definition of a system is as follow: "A system is a collection of parts and

relations betweentheparts suchthat thebehavior ofthewhole is afunctionnot _onlythe behaviorof_{the parts,}but also oftherelations_among

them"

(Steward "SystemsAnalysis"

1).

Generally

speaking, we can considertheproductdevelopmentprocess as a system.As

the system gets _larger, it becomes more complex. Therefore, one _way tounderstand the

system is to take it apart to _study the relations. As we break a system into smaller and

smaller_{parts, the}parts_{usually become}easierand easiertounderstand.

Therearetwowaystorepresentthestructure ofthesystem: graph andmatrix.

which is drawnfrom one vertex to another vertex showsthat the behaviorofthe one part

affects the behavior ofthe other part"

(Steward "Systems Analysis"

12). In Fig. _8, the

arrowfromvertex 1 tovertex2 impliesthatpart 1 affects part2.

(3; y

*0

>

Figure 8 Graphfora system

Another way of_representing the structure of a system is using a matrix called

Design Structure Matrix (DSM). It is a matrixrepresentation of a project. Thepurpose of

DSM is to findthe_{best way} to sequence activities. Forexample, thegraph in Fig. 8 can

be represented_using the matrixin Fig. 9. The cells onthe diagonal are filled. The rows

and the columns correspond to the parts of a system. Each cell with value

"1"

in the

matrix correspondstoanarcinthe graph. Thatis_{to say,} acellwithvalue

"1"

incolumni

row

j

represents anarc from vertexi tovertexj. Forexample, thecell inrow2 column 1

representsan arcfrom vertex 1 tovertex2.

12 3 4 5

1

2

3 '

4 1

5 1 1

[image:28.540.196.317.516.630.2]

"Reading

along a single row reveals all ofthe tasks whose output is required to

perform the task _{corresponding} to that

(Reading

the DSM par. 3). For example,

activity 2relies oninformation fromactivities 1.

"Reading

down a specific column reveals which task receives information from

the task_{corresponding}to thatcolumn"

(Reading

theDSMpar.3). Forexample,activity 1

delivers informationtoactivities2 and4.

TheprocesstobuildaDSMisasbelow:

1)

Interview the engineers to

identify

the appropriate system elements. List

them intheDSMas labelsfortherows and columns. Both oftheelements

intherow and column shouldbe inthesame order.

2)

Determinethe_{relationship betweenthe elements,}

including

theinputsand

outputs ofthesysteminteraction.

3)

Fillthecells onthediagonal.

4)

Fill the value of

"1"

into a cell in column i row

j

ifthere is an arc from

vertexitovertexj.

5)

Double-check thematrix with engineers and managersto_verifytheDSM.

Just as what was mentioned before, a system is _{usually broken} into smaller and

smaller parts to become easier to understand. There are two approaches to do this:

partitioningandtearing. Thesetwoprocesses willbe introducedwhen_theyare appliedin

3

LITERATURE

REVIEW

3.1 An Introduction to

QFD,

TRIZ and DSM

TRIZ istheacronym for

"Theory

ofInventive Problem Solving"in Russian.

DSM isatool to representthe task interactions of a project. It is a matrix with m

rows and columns. "Ifthere is aninformation

dependency

from task i totask_j, thenthe

cell of element

ij

is marked with an X. _Otherwise, the cell of element

ij

is left

empty"

(Yassineand

Falkenburg

224).

There aretwo main categories ofDSMs: static and time-based. Terninko gave us

thedefinitionas below:

1)

"Component-Based or Architecture DSM: Used for _modeling system

architecturesbasedon components and/or subsystems andtheirrelationships.

2)

Team-Based or Organization DSM: Used for modelingorganization structures

basedon people and/or groups andtheirinteractions.

3)

Activity-Based or Schedule DSM: Used for modeling processes and _activity

networksbasedon activitiesandtheirinformation flowand otherdependencies.

4)

Parameter-Based(orLow-Level _Schedule) DSM: Used_{for modeling}low-level

relationships between design decisions and_parameters, systems of_equations, subroutine

parameterexchanges

etc"

(292-293).

Design Structure Matrices (DSMs) I Static Time-Based

I

Component-based DSM People-based DSM Activity-based DSM Parameter-basedDSM

Figure 10 DSM_taxonomy

Sabbaghian, Eppingerand Murmanintroducedthree types oftask interactions of

DSM. "In Figure _11, Tasks 1 and2 are 'independent' since no information isexchanged

betweenthem. Thesetaskscanbe executed_{simultaneously (in}parallel). Tasks_{3, 4,} and5

are engaged ina sequentialinformationtransferand considered 'dependent'. These tasks

would_typically beperformed inseries. Tasks 7 and_{8, however,} are_{mutually dependent}

on information. These are

'interdependent'

or 'coupled' tasks often _requiring multiple

iterationsforcompletion" (par. 12).

Task S^quenc

r: r. - \\

5 r-Taskl Ta5K2 Ta*3 TS6K4 Tasks Task6 Task7 _I Task3 Tasks J

)

y

<ndep?r i-^i Taski U 1

-^fTast 3|->fTask!]_>jTask5}~> JDependent

Jilouftied ^_

TsskDependency

Figure11 Sample Task-based Design Structure Matrix

Sabbaghian, Eppinger and Murman also introduced Data-Driven Approach as

follow. "A

dependency

(marked cell) iscreated once atask'soutput isdefined asinputto

information flow. As seen, the

dependency

between tasks 5 and 7 in the DSM results

fromtask7"(par. 15).

Pfrq'-iirg-i

S3 rF^ni _>fTj

Figure 12 Sample DSMwith explicitinformationflow

DSM can be decomposed into a

hierarchy

of smaller DSMs as Figure 13. "The

modeling effort begins from the highest level activities and deliverables in a project

(called Level 1). _Next, eachhigh-level taskis further decomposed into a set of sub-tasks

forming

a series oflevel 2

DSMs"

(Sabbaghian, EppingerandMurmanpar. 19).

Level 3

Level 1

[image:32.539.179.334.106.206.2]

3.2 The

QFD,

TRIZ and Taguchi Connection: Customer-Driven

Robust Innovation

(Terninko)

Thispaper combines

Quality

_{Function Deployment(QFD), TRIZ,} andTaguchi in

the designprocess. In my _thesis, I also combine QFD and _TRIZ, but the approaches of

these two papers are different. This paper just gave me some thoughts about how to

combinethem.

After gathering the customer requirements (the voice of the customer), QFD

translates them into design requirements (the voice ofthe _{engineer) (par. 1). The TRIZ}

methodologyprovides creativesolutionsforthedesignimprovements.

Theprocess _{combining QFD}andTRIZisasfollow.

> "The _{first step} isto

identify

the customer_{requirements,} and thenrankthe

priorityofthe requirements. TRIZ allows a more aggressive attitudebecause

ofthepossibilities offered

by

thetechnicallinesof evolution.

> The second _{step is} to understand the customer's needs. The voice ofthe

customer context table should add the system and environment resources to

the"who," "what," "where," "when" and"how"informationneededin TRIZ.

> The third _step maps the subjective demanded quality information ofthe

customer into the objective measures of performance used

by

the engineer.

The matrix used is called the House of Quality. If a product is a model

upgrade, it is important to

identify

contradictions between different

> The _{fourth step} is concept generation. This is one ofthe most powerful

aspectsofTRIZ. TRIZwouldgenerate _manyalternativesforthe improvement

inthe_{reliability, manufacturability, cost,and environmentalimpact.}

> The last _step is about the _{manufacturing} process. TRIZ provides for a

search for technologies which can improve _{process, the} equipment and the

capabilityoftheircurrenttechnology"_{(par. 31-66).}

3.3 A Frameworkfor_{Design Process Specifications Management}

(Yassine,

and

Falkenburg)

This paper _mainly focuses on Design Process Specification _Management, but

thereare still someuseful approachesto_DSM, which are related with_mythesis.

Firstly, the paper introduced DSM in section 2. Then it introduced axiomatic

design in section 3. Section 4 combined them through a _{sensitivity design} structure

matrix

(SDSM)

representation. Section 5 was about the new concept of specification

management. Section 6 presented an application example ofthe new method. Section 7

proposed a generalized graphical solution. _Finally, Section 8 presented a _summary and

some_concludingremarks.

SDSM can be considered as the improved DSM. It gives us a new method to

improvethedesignprocess,

by

combing DSMand axiomatic. "Inthe SDSM_{matrix, the}

off-diagonal entries inrowiand column

j

representthepartialderivative fortheoutput of

taskito theoutput oftaskj:

-Oj/-_Oj.

Extending

_theclassical DSMrepresentation

by

the

inclusion ofthe _sensitivity coefficients allows us to understand and capture the design

Fromthis paper, Icanlearnthreemethods.

> _Sometimes, we can use mathematical knowledge to improve existing

models.

> Wecanalso combinetwomethods

(models)

inour research.

r-It is a good choice to use graphical tools in some research. From the

graphical _{perspective, the}

relationship between the parameters will become

moreobvious.

3.4 Product Development Process Captureand

Display

Using

Web-Based Technologies

(Sabbaghian,

Eppinger and

Murman)

"Today's large product development programs can be divided into hundreds of

cross-functional _teams, characterized

by

the participation of thousands of designers,

working on hundreds of thousands of tasks over a period of several

years"

(par. 5).

Automobiles,rockets, or satellitesaretypicalexamples of such products.

Now, the information needed in the developmentprocess is _usually obtained

by

tremendous interviews and cross-function _group meetings. _However, there are some

problems

during

this process. Just as what Sabbaghian, Eppinger and Murman said,

"Significant effort is spent

facilitating

interactions among users in order to resolve data

integration issues". "This synchronousdataacquisition approachis

logistically

difficultto

carry out andis therefore time

consuming"

(par. 27-28). Internet_technology was chosen

as the most suitable infrastructure for the proposed system due its ease of

deployment,

"The systemhasbeendevelopedontheWindowsNTplatform_usingJava, Active

Server Pages_(ASP), MS SQL-Server RDBMSandJDBC

middleware"

(par. 24).

3.5

Apply

the Design Structure Matrixto System Decompositionand

Integration Problems: A Review and New Directions

(Browning)

This paper reviews four DSM applications: Component-Based or _{Architecture,}

Team-Based or Organization _DSM, Activity-Based or Schedule _DSM, and

Parameter-Based DSM. Italsodiscusses researchdirectionsand newDSMapplications.

The fourth part of the paper focuses on Information Flow-Based Process

Modeling Using

the Activity-Based DSM. The knowledge here is much related with _my

thesis, so Iwould liketo focus onthispart. It is helpfulto think of productdevelopment

from an information processing perspective. _First, we use the DSM to describe the

input/outputactivitiesbasedontheinformation_flow,toshowthe

dependency

structure of

a process. _Then,theDSM isrearrangedto showtheimprovedprocess_{architecture,}which

will minimizetheunintentionaliteration.

1)Decomposetheprocessintoactivities;

2)

Documentthe_{information flow among}theactivities(their_{integration);}

3)

Analyze the _sequencing of the activities into a

(generally)

_maximally feed

forwardprocessflow.

The widely practiced initial step in analysis is called partitioning. Besides these

methods,there are several other methodstoresolve thecoupledblocksofactivitiesinto a

feasible execution sequence. The

following

two methods are the ones related with my

"Aggregation: _{While aggregating} two or more activities into a single _activityto

reduce subdiagonal marks and _simplify the _{DSM may} seem _attractive,

doing

so makes

themodelless useful

by

_'sweepingunderthe

rug'

the_veryissues it should expose.

Decomposition:

Decomposing

coupled activities can reveal ways tointermingle

thelower-levelactivitiesthateliminate feedback" (299).

3.6 Design ofObjectiveFunctions for Optimization ofMulti-Domain

Systems

(Andersson,

Jochen and

Petter)

Multi-domainsystems can be characterized as complex systems, which combine

different fieldsof_{engineering (1). The}aircraft

landing

gearinthepaperis atrueexample

of a multi-domain simulationproblem, because it is combined of_mechanical, electrical

andhydraulic sub systems.

"The

landing

gear system consists of the actual

landing

gear, the hydraulic

actuatorthatcreatestheretardation movement andthe_{hydraulic supply}system"

(1).

First ofall, thepaper used DSM toanalyze the system. As we can see from Fig.

14, the DSM method can visualize the coupling between different design tasks for the

gear design process. We also find that the parameters for the system that creates the

movement are _totally dependent on the

landing

gear parameters but not the other _way

Landinggear_leg Movementcontrol

1Actuation _Supply

A B c D E F G H 1 J K L M

A X _{Wheel diameter}

B X _{Numberof wheels}

C X X _Leglength

D X _Legdiameter

E X X _{Supportingbeam}

F X X X X XXX X X Lengthoflever

G X X X X X X X X X Angleoflever

H

1 X

X

X X

X

X X

X X

X

X Piston Area

Pistonstroke

J X Systempressure

K _XXX

X X

Valvesize

L X _Pumpsize

M XXX X Accumulatorsize

Figure 14 Design Structure Matrix forthe_landinggeardesignprocess

Secondly,itusedQFDtoanalyzethesystem.

> "Establish the characteristics ofthe

landing

gear system on the vertical

axis on the left in Figure 14. Some characteristics come from aircraft

regulations

(FAR)

that have to be _fulfilled, i.e. these are demands that the

systemhas to _meet, for instance minimum

descending

_velocity or maximum

brake distance" (3). Other characteristics come from the customer

requirements.

> List the system parameters

(engineering

_{characteristic)} on the horizontal

axis.

> As

being

shown in Fig. 15, fill the matrix to express ^he _relationship

we&t

Descendingvetocgy

FtetraeSonftne

Energycorcsunplion

Price

Crowdcarrying capacity

Loctaig tendiqgear

Brakedistance

E

Wheel

System Parameters

Landinggearteg Actuator Hvd.SuppV

Figure 15_Relationshipmatrixforthe_landinggearsystem

Then the paper used the simulation package Pro Mechanica Motion, which is a

Multi-Body-Simulation _(MBS) environment and HOPSAN package to analyze the

system.

3.7 The Design Structure System: A Method for

Managing

the Design

ofComplex Systems (Steward "The Design Structure

System")

This paper applied the techniques ofthe Design Structure System to

develop

an

effective_engineeringplanintheprocessofthedesignofa system.

System design involves the determination of interdependent variables. "Before

some variables can be determined, other variables must first be knownor assumed. For

example, A cannot be determined unless B is first known or assumed, but B cannot be

determined unless A is first known or assumed. This implies a precedence order ofthe

variables,andconsequentlyofthetasksofdeterminingthese (71).

This paper introducedtwo approaches to rearrange the DSM to generate a faster

"Partitioning

is the process of_{manipulating (i.e.}

reordering) the DSM rows and

columns suchthat thenewDSM arrangementdoes notcontain_anyfeedbackmarks."

Itis

anideal_wayto transform the DSM intoalowertriangularform. _However,it ishardto do

so forcomplex _engineering systems. _Therefore, "the variables can be reordered so as to

confine the marks in the matrix to appear either below the diagonal or within square

blocks on the diagonal." "The variables with the block each depend on all the other

variables in the same block _{and, therefore,} cannotbe determined one at a time without

making

estimates."

Then,the next_stepistobreakthecircuits tofinda placetobeginthe

iteration.

(72)

"Tearing is to choose a set of marks _representing where we might make

estimates."

We want to remove these marks so that no marks then would appear above

the diagonal. That means no additional estimates need to be made. The marks we

removed are calledtears. "The firstchoice of marksto tearisthose that theengineerfeels

would make good estimates. These are the predecessors that he _already has a good

estimate_for,or, can make a poor estimateforwithout_{significantly affecting}thevariables

theyprecede. We assigntheseahigh-levelnumber, e.g., 9. Highnumbers_{may be}usedto

indicatethebetterplacesto_tear, lowernumbers_{may be}usedto indicate where estimates

are harder to make. We usually use level numbers inthe range from 6 to 9 to indicate

these sensitivities. Ifthis _tearing leaves a reduced matrix with blocks _containing more

thanonevariable, thenwe stillhave circuitsthathavenotbeentornyet. We haveto make

furthertears to breakthesecircuits. We usuallyuse levelnumbers in therangefrom 1 to

3.8

Planning

and

Managing

the Design ofSystems (Steward

"Planning

and

Managing")

This paper presents a method for _planning the design process

by

_applying

partitioning,tearingand_usingshunt_{diagrams into DSM.}

The paper

by

the same author The Design Structure System: A Method for

Managing

the Design ofComplex Systems has introduced _partitioning and_tearing well,

so I will focus onhowtouse shunt diagramshere. Thepurpose of_tearing isto choose a

set of marks where we might makeestimates,andthenwecanbreaktheiteration.

First, there are some concepts we shall know. A principal circuit is the longest

circuit we canfindinthe block. Nowthereare fourpaths betweennodesinthisprincipal

circuit whichhave no arc or other node in common withtheprincipal circuit. These are

called shunts. Two paths between the same nodes inthe same directions are said to be

parallel. Otherwise, _they are called ant parallel. In Fig. _16, B representsthe shuntbegins

atthenode; Erepresentstheend ofthe shunt; Pshowsthenumber of shuntsthatpassthis

node neither

beginning

nor _ending here; S showsthenumber of parallel _shunts; I shows

thelargest indexof_anyoftheparallel shunts.

Index

6SS42 BEPS1

b I B I I 0 2 3 5

i 3 K I 12 125

11 IB 1 0 2 3 S

_i_ _E_ _E_.t..20 2 1 12

e 1 0 0 112

9 B I 1 0 1 2 S

Graph - _Principal

Circuit and shunts Shunt diagram Shunt diagram _summary

There is no unique or optimal answer for tearing. The paper here gives us three

methods thatcan beused. The first is to tear theprincipal circuit whereP is low. Wecan

tear eitherthe arcs _entering the node (just above the _node) or the arcs _exiting the node

(just belowthe node). The secondisto tearwhere S is_small,givinga

big

drop

inthe size

ofthe_{remainingblock,}asin Fig. 17. We can choose multipletearsintheprincipal circuit

so no single shunt crosses allthe_lines,asin Fig. 18.

Index

5 4 4 3 3 2 _d c 1 a h_

d

B B

B

1

I

E

d

c

10 1 SI

c IXI02I 1

1 1 B 1 1 B 1 1 I^X 0^1 1

a E 1 r E 1 a f X X X t_0_l i

h E E l_ h 1 X X X 1 0 1

Tear to get small block inside

Tear with low I in Shunt Diagram Summary

Figure 17

5 4 4 3 3 2

a b i

_C B_ B I E_

1 I B II B

a E I IE I

h E E I_

.d_ 1 c a b_

a i_o_t 5 i

i I 10 2 1 I

C I X l_X 0_l I

a I X X X l_0_l I

h I x x x_f_o_l

Multiple tsars in Principal Circuit

sono one shunt crosses all the tear lines in Shunt Diagram

4

PRODUCT

DEVELOPMENT

KNOWLEDGE-PROCESSING

SYSTEM

4.1 Architecture ofHP DeskJet 1200C

Inorderto explainthenew_{methodology inthis thesis clearly,}the _methodologyis

appliedintotheHPDeskJet 1200CPrintercase.

"The DeskJet 1200C made a quantum

leap

in the _{already impressive DeskJet}

capabilities"

(Printer Workspar. 1).

Firstof_{all, theprinting}mechanismisexplainedinthispart.

As the Fig. 19 belowshows, theprinting system components consist ofthe paper

advance mechanism (stepper motor, drive shafts, and _gearing system), the print

cartridges, the carriage and its _{axis, the} carriage drive (dc _{motor, belt, pulleys, etc.),} the main_heater,andthevapor removal system.

Front Sfidec

Encoder_Strip

Idler_Pulleyand

Drive Bell

DC Motor

Sintered Bronze_{Bushing j}

RodMourns

Star Wheel _PrimingArea AdjustmentScrew

[image:43.539.84.446.418.647.2]

The printingprocessisexplained asfollowed."Mediastackedinthe input_tray(1)

is

individually

picked

by

a media pick roller _{(not shown)} and driven around the curved

preheat zone

(2)

where it ispreconditioned (moisture is drivenoffandthe temperatureis

raised). Whenthepage reachesthepinch/drive rollers_(3),themaindrivesystem₍₄₎takes

overfromthepick rollerdrive (notshown). Onceintheprint zone

(5)

themediais heated

further and ink is sprayed onto the page. The heating, soaking, and

drying

causes the

mediatomove out ofits_plane, butthemediacontrol shims

(6)

help

holditflatforbetter

print quality. Thepage is then

incrementally

advanced and printed upon until the entire

page has beenprinted. _Finally,thepage isfedout intotheoutput_tray

(7)

andtheprocess

is readyto repeat"

[image:44.540.93.424.348.622.2]

(Broder 74). Thiswhole processcould beunderstood withthe

help

of

Fig. 20.

Pap.erCon*IShins

(T)Main Drive Sjttem

4.2 The NetworkofICOMs overview

Weuse atoolcalled ICOM as shownin

Fig

21 and 22

during

thisprocess._{I, C,} O

andMrepresent_{Input, Control,}Outputand Mechanismrespectively. Inputistransformed

into Output

by

_usingtheMechanismthatisunderconstraintofControl. This functioncan

be expressed as Fi. i is the sequence ofthe function. Function is defined as: "A rule of

correspondence between two sets such that there is a unique element in the second set

assignedtoeach element inthe first set"

(Dictionary

par. 2). For_example,in Fig. 21 and

22, I represents our product and the competitor's _product; O represents our future

product; C represents the control such as our _{manufacturing capability} and so _on; M

represents the mechanism such as QFD and so on. Based on our _{manufacturing}

capability, Fl means that we can use QFD as a tool to analyze our product and our

competitor's_{product, to}establishthedetails forourfutureproduct.

Control_(C)

Input₍₁₎

Function i

Output₍₀₎

t Mechanism _(M)

Figure 21 ICOM

Suchas

Manufacturing

capability.

Ourproduct

Function 1

Our futureproduct

Competitor'sproduct

f SuchasQFD.

Figure 22 ICOMexample

For the purpose of _representing the data movement in the whole product

Throughoutthe follow_discussion,thedefinitions for thesymbolsused inFigures

23

-67are asfollow:

1. "The term ICOM refers _collectively to the _group of information flows

between activities which have one offour roles in the activity"

(Department

of Defense par. 5). _{I, C,} O and M represent _{Input, Control,} Output and

Mechanismrespectively.

2. I represents a matrix as the _{input of data} "used to produce the output of an

(Department ofDefensepar. 5). Orepresents a matrix astheoutput

of data "produced

by

or_{resulting from}the activity"

(DepartmentofDefense

par. 5). Both I and O are_usuallyexpressed

by

two or three capital characters

such as _CR, ECand so on.

3. Therectangle represents thefunction process, whichtransforms input datato

unique output data. This process includes the Mechanism that is under

constraint of Control. Control is defined as "Information that constrains or

controlsan

activity"

(DepartmentofDefensepar. 5). Mechanism is definedas

"Usually

people, machines,or systemsthatperformtheactivity"

(Department

ofDefensepar. 5).

4. The subscript ontherightofI/Orepresentsthesequence oftheinputs/outputs

through thatfunction.

5. This thesis focuses on the data exchange

during

the product development

process, sothecontentforMandOwill notbe discussed here.

4.3 Network ofICOMs for QFD

Figure 23 isadescription fortheHouse of

Quality

model. It is based onFig. 2 of

Lai-Kow'spaper.

Step5

EngineeringCharacteristics (EC):

EC, EC,...ECH

CR Step1 Customer Refniremeitis (CR): CR, CR, CRh Step: Importance Priority Ratings (TP): IP. IP* IPm Step6

RelationshipMatrix(MT)Between CRsaniECs:

EC, ECj ... EC,

CR, MT MT,2 MTlH

CR2MTa WTn...MTu,

CRH MTh, MT^, ...MTujj MT

Step3 CustomerCompetitive Analysis:

Customer CompetitivePriority'

Ratings,

CustomerPerformance GoalsforCRs,

Customer ImprovementRatio

Step4 final Ratings: SeepT TechnicalCompetitive Analysis:

InitialTechnical Ratings Technical CompetitivePriorityRatings,

Target Values for EC's. Technical ImprovementRatio

Step8 Final Importance Ratings (FR):

FR,FR2...FRg

FR

Figure23_HOQModel (Adapted from Fig. 2of_{Lai-Kow's paper)}

This network ofICOMs showshowthe datamoves from one process to another

during

QFD. Thereare fourtransforms

during

this process,which are shownin

Fig

24. [image:47.539.68.457.138.373.2]

CR CR F2 IP F1 EC CR F3 MT F4 FR

Figure 24 Agraphic network ofICOMs for QFD

Allofthese transformscanberepresented_{using ICOMs}asFig.25.

CR

[image:47.539.165.360.482.554.2]

Fl Transform

Fig. _{26, 27,} Table 1 and 2 represent the 1st transform ofthe network. Fig. 26 is

derived from Fig.24. Table 1 and 2arethecontentsforthis transform.

According

to what

Clausing

says "The language of the customer tends to be

subjective, qualitative, and _{nontechnical",} seven Customer Requirements are acquired

through surveys and interviews with the customers

(Clausing

65).

They

are denoted as

CRm, m=l,2, ..., M. Inthis case, M equals 7, so CR is anordered set with 7 elements.

TheelementsofCRare shownin Table 1.

IP stands forthe Importance Priorities foreach element ofthe CR. Inthis case it

isan ordered set as well as a7x1 _{matrix, the}values of which are showninTable 2.

CR

F1

IP

Figure 26 Fl transformfromthenetwork

Fl represents the assignment of relative importance _IPm to the customer

requirement CRm.

"Relative"

here means this is an example of"ordinal and

rational"

scale. Thatis _{to say,} inFig. _27, Color quality is five times as importantas _{Archivability,}

because of IPi=5 and IP7=1. The value of _IPm can _{be determined using} a number of

attribute_weightingmethods. Inthis example, thepair-wise comparisonmethod was used

(basedonUlrich _135),as illustratedin Figure 27.

i CR, Desigr_type

reference_(egone ofCanonpunter Rating

1 Colorquality 3 5

2 Air fillquahty 3 4

3 lineand edgequality 3 4

4 _{Archivability} 3 1

5 Cockle 3 3

6 Curl 3 3

7 "Whiteness 3 2

First of_{all, the} elements ofCR are listed inthe second column ofthe matrix, as

shownin Fig. 27. _Secondly,giventheexistenceofadesign_reference, a_weightingof5 is

assigned ifthe design attribute is much better that the reference; a _weighting of4 is

assigned ifthe designattribute is better_{than the reference;} 3 ifthe designattribute isthe

same as the _reference; 2 if it is worse _{than the reference;} and 1 if it is much worsethan

the reference. A competitive product or alternate design ofknown performance _may be

used as a_reference, forexample, one type ofCanonprinter. Theresult is shown in Table

2. Inour_{example, the}purpose istoexplainhow touse _{this method,} sotheelements ofIP

might be different from the real _values, because this determination of these values

requires alotof_resources, such as real _data, enoughtimeand experience. _However, what

willbe discussed inthis thesisis suretobecorrect_usingthesevalues.

Seven Customer Requirements forour product Importance Priorities

CRi: Color quality IP,: 5

CR2: Airfill _quality IP2: 4

CR3: Lineand edge_quality Fl IP3: 4

CR4: Cockle IP4: 3

CR5: Curl IP5: 3

CRe: Whiteness IP6: 2

CR7:

Archivability

IP7: 1

F2 Transform

Fig. 28 andTable 3 representthe2ndtransformofthenetwork.

Define

(ECj

|

_CR,depends on_ECj). i=1,2,...,M and

j

= _1,2,...,N

as an ordered set

of

Engineering

_{Characteristics upon which a given Customer Requirementdepends,then}

M

EC =

(J

(EQ

|

CR depends on ECj). There are N=18

Engineering

Characteristics in

t=t

this case, all ofwhich are given on Hunt's paper (Hunt 18-20). The elements ofEC are

shown in Table 3 below and the definitions for these

Engineering

Characteristics are

given onAppendix 3.

F2 isthe_mappingofCRtoEC (see Fig. 28).

CR

F2

EC

Figure 28 F2transformfromthenetwork

Ed:

[<->]

Color Gamut

EC2:

[I] Drop

placement_accuracy

EC3:

[I]

_Banding

EC4:

[|]

Waittime

banding

EC5:

[|] Mottling

EC6:

[1]

Bleed

EC7:

[<-]

Coalescence

EC8:

[<-]

Registration

ECio:

EC,,:

EC12:

EC13:

EC14:

EC15:

EC,6:

EC17:

EC18:

I]

Spray, leathering, and

jaggy

lines

I]

Visual measurementof paperflatness

I]

Visual measurementoftheheightofthecornerson aflatsurface

-] Theamountofinkplacedonthemedia

|]Dot

density

?] The consistencyofthe

drop

volume

?] The consistencyofthe

drop

size

>]Inkpenetration rate

?] Orifice diameter

Note: "|"representstodecreasetheamount;

"j"

representsto _increase;"-"

representsit

isnot_necessaryorhardtodecidethedirection.

Table 3 Eighteen_EngineeringCharacteristics

F3 Transform

Fig. 29 andTable 4representthe3rdtransformofthenetwork

MT

>

EC

F3 CR

Figure 29 F3transformfromthenetwork

MTisthe matrixthatrepresentstheextent towhicheach _CRm is dependentupon

(Crowpar. 4). In this case, the rational scale _{1. 3,} 9

(corresponding

to weak, moderate,

and_strongdependence_{respectively)}isused.ThetransformF3 is definedas followed:

DefineMTmn=

0ifCRmisindependentof_ECn;

MTmn =

1 if_CRmdepends_weaklyon

ECn;

MTmn=_{3 if}

CRmdepends_moderatelyon_ECn;and

MTmn=

9 if_CRmdepends_stronglyonECn.

Vm,n m=l,2,...,M and n=l,2,...,N.

The fullmatrix oftheserelationships isshownin Table 3 below. For_example,the

relationshipbetweenCR, and_EC, is considered as

"strong"

thatcorresponds to thevalue

MT,,=

9. Theprocesstogetthesevalues isshownin Appendix 4.

Relationship

between

EC,

EC₂ .

ECS

CRandEC

CR,

CR2

-MT

MT2,

MT,2

MT22

MT,S

-MT2S

without values

CRXI

MTA

42 MTm- _

F3

'_r

EC,

EC

CR,

"9 0 0 0 0 0 0 0 0 0 0 0 0 0 9

CR2

0 9 3 3 3 3 3 9 0 0 0 3 3 9 0 1

Relationshipbetween

CR3

0 9 0 0 0 0 0 0 9 0 0 3