ABSTRACT
Title of Document: AUTOMOTIVE DESIGN-TO LIFE-CYCLE
CRITERIA FOR LOWERING WARRANTY
COSTS AND IMPROVING OWNERSHIP
EXPERIENCE THROUGH THE USE OF A NEW
?BINARY DECISION MODEL? AND
APPLICATION OF A ?WARRANTY INDEX?
William Ireland, Master of Science, Reliability
Engineering, 2009
Thesis directed by: Dr. Aris Christou, Chair
Professor of Mechanical Engineering
Department of Materials and Nuclear Engineering
Financial challenges facing the automotive sector require identification of new
opportunities for quality improvement. A new Design-To Life-Cycle-Cost strategy is
introduced that applies a unique ?Binary Decision Logic Model? that classifies
corrective action opportunity into Life-Cycle categories. The intended result is to lower a
manufacture?s warranty costs and improve ownership experience. This is done by setting
Design-To goals in a Life-Cycle way for Reliability and Serviceability.
The sample space for data to drive this change of process is found in an existing
warranty system with data elements consisting of failure occurrence, failure symptom,
mileage, part cost, and labor cost. One can investigate new factors, such as the
?Warranty Index,? that parses the corrective action in favor of lowering part costs or
labor costs found in a typical service event. The data considers opportunities over
mileage and time domains to improve vehicle quality over the Life-Cycle.
Title Page
AUTOMOTIVE DESIGN-TO LIFE-CYCLE CRITERIA FOR LOWERING
WARRANTY COSTS AND IMPROVING OWNERSHIP EXPERIENCE THROUGH
THE USE OF A NEW ?BINARY DECISION MODEL? AND APPLICATION OF A
?WARRANTY INDEX?
by
William C. Ireland
Thesis submitted to the Faculty of the Graduate School of the
University of Maryland, College Park, in partial fulfillment
of the requirements for the degree of
Master of Science
2009
Advisory Committee:
Professor Aris Christou, Chair
Professor Mohammad Modarres
Professor Jeffrey Herrmann
Copyright Page
? Copyright by
William C. Ireland
2009
ii
Preface
Many business models are not as effective in today?s marketplace. There exist a
need for a more effective quality improvement implementation strategy from the design
cues that come from a wide range of product usage and warranty data. In the automotive
sector, a new approach to evaluate ?Design for Maintainability? or ?Design for Service?
or more broadly ?Design-To Life-Cycle Cost? is needed to improve product quality over
traditional design objectives. There is a larger need to understand vehicle usage and
maintenance after the sale in contrast to a focus upon design functionality and efficient
assembly. Design for Assembly or just DFA, is a methodology that does not sufficiently
address the service needs or the full spectrum of a customer?s needs over the product?s
Life-Cycle.
The object of the this study is to build upon a logic definition used in the aircraft
industry known as Reliability Centered Maintenance and apply a new set of binary
decision rules with a unique use of a Warranty Index to help guide the Design to a more
effective Life-Cycle alternative. This should improve quality by reducing warranty costs
and improve the ownership experience of the customer.
iii
Foreword
A better design strategy is one that considers the full Life-Cycle of a product.
Often the techniques to develop product requirements fall short of this intended purpose.
The methods in product development are largely incongruent with the more important
goal found in a Life-Cycle approach. Proposed in this study is a binary decision
algorithm that is hierarchical and able to be integrated into a product development cycle.
When to apply these techniques to maximize the benefits in the design process has not
been apparent and can and should impact the very architecture of given design theme.
This presentation is intended to be read with a point of view that will focus upon
the nature of automotive design processes embracing the serviceability aspects of systems
created so that more control over warranty costs and the customer?s ownership
experience can be employed. The outcome can then be realized in a way that will lead to
improved customer satisfaction.
The unique approach offered is intended to differentiate on part and labor costs
that have not been addressed satisfactorily as a design criterion. Those development
programs that are looking more narrowly at costs per unit or repairs per thousand vehicles
have not been able to identify these additional quality improvement opportunities. In
order to improve upon this work it is hopeful that suggestions will be made and brought
to the attention of the author.
iv
Dedication
This thesis is dedicated to the hard working team at Ford Motor Company, family
and friends that have sacrificed to help provide the continuation of my dream to pursue
this advanced degree. It has taken a long time to accomplish and even survive the
various ups and downs at Ford and at home. At times it just did not seem possible. A
special thank you is in order:
First, a special thanks to Debby, a spouse whom freely asks the hard questions
that helps to drive a better understanding of the technical ideas in a business-wise way.
Second, a thank you for those that helped support a distance learner whose efforts
were required to see me through ? the help was so appreciated!
It has been my pleasure to participate in the program. Please count on me to
express the benefits of continuous education with the University of Maryland that does
make a difference in so many ways where the least of those ways are experienced in the
improved quality products and services we enjoy that so many graduates have a part.
Assurance technologies bring into our lives better products services and in a small but
exiting role I have had the privilege to participate.
v
Acknowledgements
The writing of this document was with the help of the Brighton Public Library
where their facility was made available with its various resources. Access to the internet,
reference materials, and facilities made it an excellent environment for study while
offering an isolated space needed to ponder the subject and develop a relevant body of
work. Being a distance learner did not afford me the usual scholastic environment of a
university nor provide access to those facilities. However, this gap was bridged
effectively by use of the local library.
It is hoped that this work will meet a need in design, one that differentiates on
Life-Cycle design strategies in a practical way. Within existing data structures there is a
means to re-visit the data and emerge with new and actionable information offering Life-
Cycle design alternatives. These opportunities include advances in vehicle longevity and
ownership experience with a window towards a wider impact in new vehicle
architectures.
It is further hoped that these unique concepts presented will be built upon in the
next generation factory to fully integrate the concept of design for the Life-Cycle and a
means to utilize concepts of maintainability with personality. A natural outgrowth of the
Life-Cycle design development may find an expression in an ?i-car.? This truly will
yield a means to balance reduction in plant complexity in favor of rational
personalization offered by the local dealer.
vi
Table of Contents
ABSTRACT........................................................................................................................ 1
Title Page ............................................................................................................................ 2
Copyright Page.................................................................................................................... 3
Preface................................................................................................................................. ii
Foreword............................................................................................................................iii
Dedication.......................................................................................................................... iv
Acknowledgements............................................................................................................. v
Table of Contents............................................................................................................... vi
List of Tables ...................................................................................................................viii
List of Figures.................................................................................................................... ix
List of Figures.................................................................................................................... ix
Chapter One: ....................................................................................................................... 1
Introduction..................................................................................................................... 1
Statement of the Problem................................................................................................ 2
Definition of Terms......................................................................................................... 2
Assumptions.................................................................................................................... 5
Significance of the Study................................................................................................ 6
Study Proposal ................................................................................................................ 6
Limitations of the study .................................................................................................. 7
Organization of the Study ............................................................................................... 7
Chapter Two........................................................................................................................ 9
Related Research........................................................................................................... 11
Life-Cycle Design..................................................................................................... 11
Cost of Quality.......................................................................................................... 12
Design-To Life-Cycle and Congruence to Quality Measures....................................... 16
Chapter Three.................................................................................................................... 18
Robustness and Reliability of Study Data .................................................................... 18
Business Processes Affected by the Study.................................................................... 19
Warranty, Price and Design Dilemmas......................................................................... 23
The Warranty Dilemma: ........................................................................................... 23
The Price Dilemma: .................................................................................................. 25
The Design Dilemma:............................................................................................... 25
Looking Closer at a Warranty System.......................................................................... 27
The Basic Properties of the ?Warranty Index? ............................................................. 29
Warranty and Service Parts Business ........................................................................... 32
A Deeper Look into Part and Labor Cost of Repair: ................................................ 33
Historical Serviceability in Prior Vehicle Generations............................................. 33
Exposing the Price Point for the Aftermarket........................................................... 36
Monopolistic Pricing................................................................................................. 37
Sense of Fairness....................................................................................................... 39
Chapter Four ..................................................................................................................... 42
Required Organizational Structure ............................................................................... 42
Project Management Structure...................................................................................... 43
vii
Process Deployment ? Target Setting on New or Existing Programs .......................... 45
Deployment - Warranty Data System Usage ............................................................ 49
Deployment - Warranty Data Filtering..................................................................... 50
Deployment ? BDLM & Warranty Index Procedure................................................ 50
Deployment - Cost Creep and Maintenance Time Control ...................................... 53
Deployment - Quality Assessment............................................................................ 54
Deployment ? Descriptive Statistics on Key Factors ............................................... 62
Qualitative Analysis Part vs. Labor against Frequency................................................ 63
Binary Decision Logic, Frequency Lists, & Event Landing......................................... 69
Binary Decision Logic, Maintainability Decision ........................................................ 72
Decision Rule ? Example Candidate for Maintainability............................................. 73
Data Dispersion Part to Labor....................................................................................... 73
Building Warrant Index Design Guidelines.................................................................. 79
Understanding a Single Data Point............................................................................... 82
Ten Elements of Quality in Vehicle Design ................................................................. 85
Ten Related Quality Measures...................................................................................... 86
Brief Review of Reliability, Maintainability and Availability ..................................... 87
Generating High Leverage Frequency Lists ................................................................. 89
Fault Avoidance and Fault Tolerance in Design........................................................... 93
The ?Warranty Index? Interrogates the Design ............................................................ 97
Chapter Five.................................................................................................................... 103
Conclusion and Findings............................................................................................. 103
Future Research .......................................................................................................... 105
Appendix......................................................................................................................... 107
Reliability Bath Tub Curve, Targets and Tasks.......................................................... 107
Glossary .......................................................................................................................... 120
Bibliography ................................................................................................................... 122
viii
List of Tables
Table 1. Warranty Claims - Qualitative Analysis............................................................ 32
Table 2. Ten Elements of Quality in Vehicle Design. ..................................................... 85
Table 3. Ten Related Quality Measures........................................................................... 86
Table 4. Availability Classifications................................................................................ 87
Table 5. Formula for Availability (total). ........................................................................ 87
Table 6. Formula for Availability (Inherent). .................................................................. 88
Table 7. High Frequency Lists from Service Event Driven Effects. ............................... 90
ix
List of Figures
Figure 1. Basic Warranty Comparison - Miles (thousands). ............................................. 9
Figure 2. Graph of Vehicle Repair Cost Labor vs. Part................................................... 21
Figure 3. Buyback Due to Lemon Law Data. .................................................................. 24
Figure 4. The OEM-Dealer-Customer Interaction........................................................... 41
Figure 5. The Typical Repair Flow Chart........................................................................ 41
Figure 6. Reliability Action Sheet .................................................................................... 46
Figure 7. Maintainability Action Sheet ............................................................................ 47
Figure 8. Cost Action Sheet.............................................................................................. 47
Figure 9. Basic Design Action Sheet................................................................................ 48
Figure 10. Warranty Database Definition........................................................................ 49
Figure 11. Excel Data Feed ? Warranty Data Filtering for Study .................................... 50
Figure 12. Depth of Warranty Target Data in Service...................................................... 57
Figure 13. Frequency ? Labor: Region I-IV. ................................................................... 64
Figure 14. Frequency ? Labor-Parts: Region I-IV........................................................... 65
Figure 15. Mean and Standard Deviation and Warranty Index: Region I-IV.................. 66
Figure 16. Frequency MTBF ? Part - Labor: Region I-IV. ............................................. 68
Figure 17. Binary Decision Logic Model ? Criteria Defined. .......................................... 69
Figure 18. Event Trigger Topology and Driver Interface Leading to Service................. 71
Figure 19. Binary Decision Logic Model, Maintainability Branch. ................................. 72
Figure 20. Design Rule Applied to Frequency Listings. .................................................. 73
Figure 21. Key Graphs Explained ? Full Vehicle............................................................. 75
Figure 22. Key Graphs Explained, Failure Code: i1,j1,k1............................................... 76
Figure 23. Key Graphs Explained ? Failure Code i1,j1,k2............................................... 77
Figure 24. Excel ? Pivot Table that Generates that Category Statistics. .......................... 78
Figure 25. Warranty Index Design Guidelines. ................................................................ 79
Figure 26. Assessment of Factor Behaviors, Mean and Standard Deviation................... 79
Figure 27. Generating Targets for Design-To Life-Cycle, Cost Category Variability.... 80
Figure 28. Target Modeling, Repair Code: P_F2G4, 3rd quartile Target........................ 81
Figure 29. Target Modeling, Repair Code: L_F2G3, 3rd quartile Target. ...................... 81
Figure 30. Data Point of a specific repair cost elements, Part and Labor........................ 82
Figure 31. Filtering Rationale, no presale data used for target setting. ........................... 83
Figure 32. Maintainability Design Actions List and Maintenance Time Categories. ..... 84
Figure 33. Weibull Analysis of Fatigue Data. .................................................................. 91
Figure 34. Binary Algorithm for Event Frequency.......................................................... 99
Figure 35. Binary Algorithm to Identify Application for Design for Serviceability. .... 100
1
Chapter One:
Introduction
One goal of most business models is to provide a quality product or service at an
affordable price that meets the need of the consumer bringing satisfaction over the
expected use of that product or service and in return receive a fair profit to continue such
provision of services or products in the market place. However straight forward this may
seem, the objective is not easily accomplished in today?s fierce competitive and difficult
economic climate. There is therefore a need to consider more intently what
improvements can be offered the customer to assure his satisfaction and to manage and
minimize quality costs effectively.
Design cues that come from a wide range of product usage and warranty data may
hold a key to continuous improvement over a wider range of customer expectation not
realized, specifically in the automotive sector. In particular, one significant item is the
high cost of commercial warranty and low customer satisfaction facing businesses, as
addressed by Roush and Webb (2001), including high value properties (p. 529). This is
certainly true in the automobile industry. One observation includes the high warranty
costs that still could amount to over $27.5 billion spent per year world wide, according to
the Newsletter Subscription Service, Warranty Week, (2006) and costs per car per year
over $500 dollars. It should be clear that these costs can be associated directly to the high
cost of repair as a factor of labor and part costs.
The high service costs affects each the manufacturer and customer. From a
customer and Life-Cycle cost of ownership view, the effects of product failure directly
2
impacts customer satisfaction with the possible loss of vehicle use and the high cost of
repair out side of warranty and to the manufacturer, the in-warranty costs as noted.
Opportunity therefore exists in current business and design processes to minimize these
warranty and ownership costs.
This needed improvement should be a natural evolution in design processes
building on the value of being ?best-in-class? that has focused strictly on a ?find-and-fix?
approach to product failure and reducing ?things-gone-wrong.? A Design-To Life-Cycle
cost approach may offer a more proactive design improvement methodology.
Statement of the Problem
Design teams are facing the law of diminishing returns with current improvement
programs. Quality has been improved and warranty has been reduced but more
opportunity is not being identified effectively, in part do to certain business practices that
do not address the two driving factors, part cost and labor cost. The motivation should be
clear that new ways to address improvement are needed in a design approach that is
recommended by this study where the concept of Design-To Life-Cycle cost is being
explored. Related approaches have been successful that have in part been adopted by
military programs and the commercial airline industry.
Definition of Terms
The key terms identified herein represent those concepts and engineering
practices needed to successfully understand and develop the Design-To Life-Cycle Cost
(DTLCC) quality improvement methodology. A practical application is provided
utilizing fundamental practices of Design Assurance and using a Binary Decision Logical
3
Model (BDLM) similar to Reliability Centered Maintenance, (RCM). Of concern within
product development teams are the needed requisite skills for practicing Design
Assurance including Maintainability engineering. Secondarily, an organization will have
to possess the aptitude to address organizational inertia to bring new players to the design
responsibility.
Design Assurance ? Includes the engineering disciplines of Reliability and
Maintainability. These technical fields require certain skill sets to be effectively
deployed in a product cycle for a durable good, such as an automobile.
Life-Cycle Cost - an acquisition strategy that considers concept, development,
production, service, and disposal costs of a product. These concepts have been used
since the early 70?s in Military programs partially due to the high cost of service and poor
reliability of fielded equipment.
PAF ? A practice to manage quality costs or the cost of quality (COQ) The
primary attention of a COQ method is the PAF model which is defined by the three
terms:
(P) Prevention: Actions taken to ensure that a ?process? provides quality
products and services.
(A) Appraisal: Actions measuring the level of quality in the given
process.
(F) Failure: Actions to correct quality issues internal - prior to the
customer and after the sale or external feedback.
4
Reliability ? A characteristic of design. The probability that a product will
perform its intended functions and under specific operating conditions. Expressed
another way, reliability is the percent of vehicles, which meet customer requirements
(without failure) at a specified time/mileage objective. As a statistical measure or a
probability of survival. As an engineering discipline, Reliability includes those
engineering activities that assure designs adequately address product duty cycle, the
stress and strength of materials, and follow and recommend design practices that avoid,
reduce and eliminate failure modes.
Maintainability ? A characteristic of design and installation which is expressed
as the probability that an item will conform to specified conditions within a given period
of time, when maintenance action is performed in accordance with prescribed procedures
and resources. Also, the measure of the ability of an item to be retained in or restored to
specified condition when maintenance is performed by personnel having specified skill
levels, using prescribed procedures and resources, at each prescribed level of
maintenance and repair. As an engineering discipline it is those engineering activities
that assure designs adequately consider human factors, the level of repair, the activities to
minimize the time to repair through proper maintenance practices and logistic concerns
and strategies.
Design of Assembly ? a practice to build features and characteristics into a design
to enhance the build and assembly of products in an effective manner that minimizes
costs of assemble and considers human factors in process to minimize variation.
5
Reliability Centered Maintenance, RCM - developed as a flight line to depot
level response strategy for assuring a repair time can be scheduled with little interference
to an operational mission. A logical decision approach that identifies a proper
maintenance strategy using product reliability and safety features that assure critical
mission factors with a goal to optimize the maintenance effectiveness and assure mission
success throughout a logistic and service concept of a fleet maintenance program.
Part Costs - Costs that include parts in the repair.
Labor Costs - Costs that include labor in repair ? Divide by rate to get time.
Assumptions
The study methodology to be developed assumes there is a data structure in an
automotive manufacturing environment that possesses certain records (and fields) typical
of such warranty systems. This data must have both part and labor costs identified in a
repair event and some level of codification as to the function and causal component part.
By causal, it is understood that it is knowledge of failure and service activities that one
can investigate that is or can be defined down to the root cause of failure if necessary.
Further, that there is a service organization that supports the warranty system and certain
design skills for design assurance are present. This may not be the case when the design
area of maintainability is discussed for which there is a gap in the automotive sector in
product development.
6
Significance of the Study
The introduction and development a new methodology that identifies specific
weaknesses in design, previously hidden from most product teams, is a welcome idea
given that there is an extremely high cost in warranty still being reported and born by the
automotive manufacturer during an economically challenging time. The study purpose is
to show how an application of a sound BDLM and the use of a ?Warranty Index? can
improve the design in a Life-Cycle manner.
Study Proposal
In this study, the development of a new Design-To Life-Cycle methodology and
the benefits of its application will be shown to establish a unique discriminating criterion
or Warranty Index that can be used to identify Life-Cycle-cost related weaknesses in
design and related business practices. The result will be to lower repair costs and/or
improve vehicle reliability by addressing vehicle design issues including vehicle
architecture. In so doing, there will be a positive impact on residual value, customer
loyalty and create an attraction for new customers.
The process to identify Life-Cycle design requirements is to consider not just
corrective related measures towards unreliability but to include goals of serviceability
and address related cost drivers that can be interpreted in terms of change opportunities
required to improve the design to a more robust Life-Cycle design. This should have
benefit to the Original Equipment Manufacturers (OEM) and customers. The new
practice would improve a customer?s broader perception of vehicle quality while
lowering warranty liability. Implementation of a new Design-To targeting method can be
7
most successful if it can be used early in a cycle plan where change opportunity can be
effectively implemented. By reducing costs of ownership for the customer and yielding
higher margins through reduced warranty costs to the company a long term benefit is
secured.
The ability to select components that have a high impact, ones that threaten
customer satisfaction in service, will position or leverage a design team to take advantage
of the ability to differentiate and prioritize the magnitude of cost savings on a Life-Cycle
basis and in an informed manner.
Limitations of the study
This study is limited in its application as it only observed trends in several car
lines within a warranty system from a domestic automaker. Further analysis could be
applied to functional component groupings where a ?measures of importance? as
discussed by Modarres et al. (1999) in combination with a ?level of effort? similar to that
applied by Roush and Webb (2001) to help allocate or prioritize the opportunity. A cost
related allocation that optimizes on the desired change opportunity. Although the BDLM
methodology shown is applied to automobiles, there is a broader application to any
fielded product that may have a warranty such as any durable goods product.
Organization of the Study
? Chapter One will provide the introduction, statement of the problem, offer the
definition of key terms, provide study assumptions, significance, limitations and
organization.
8
? Chapter Two will present and discuss the relevant literature related to the Design-
To Life-Cycle Cost methods and quality improvement strategies by using a
Warranty Data System. The concept of a binary logic decision model is explored
and how it might be modified for use as a Design-To Life-Cycle strategy for
automotive use to improve quality.
? Chapter Three will detail the methodology of the study and its purpose with a
discussion on the part pricing issues as it relates to the study
? Chapter Four will present an application of the methodology, the findings of the
study, including statistical analysis of the data obtained.
? Chapter Five will comprise a summary of the implications of the study and a
discussion of the recommendations for future research.
9
Chapter Two
Current quality measurements between the major manufacturers have tightened in
the race for quality. There are new performers that are offering 100,000 mile warranties
where three years (or 36,000 miles) used to be a standard,
Figure 1. Basic Warranty Comparison - Miles (thousands).
by Vehicle Make
0
10
20
30
40
50
60
70
Ho
nd
a
Ni
ss
an
To
yo
ta
Su
zu
ki
Ch
rys
ler
Da
ew
oo
Do
dg
e
Fo
rd
Je
ep
Me
rcu
ry
Pl
ym
ou
th
Su
ba
ru
Bu
ick
Ch
ev
rol
et
GM
C
Po
nti
ac
Sa
tur
n
Mi
tsu
bis
hi
Ma
zd
a
Ac
ura
Lin
co
ln
BM
W
Ca
dil
lac
Ja
gu
ar
La
nd
R
ov
er
Sa
ab
Po
rsc
he
Vo
lks
wa
ge
n
Ki
a
Isu
zu
Me
rce
de
s-B
en
z
Le
xu
s
Vo
lvo Aud
i
Ol
ds
mo
bil
e
Hy
un
da
i
Inf
ini
ti
Co
ve
rag
eM
ile
ag
e(
00
0)s
* data compiled from: url: http://www.carsmart.com.
The ownership experience that is influenced by what can be called the touch
points - involves the service experience including the frequency and costs. Further, the
extension of a warranty or life limit has been explored by Ireland (1984) where he points
out that there is a need to have a statistical approach (p. 72).
10
In addition, it is the author?s point of view that there has been a trend developed
in divesting technical competency and if it is continued there will likely be an inability to
respond to the challenge in the Life-Cycle quality improvement arena. Recent actions of
OEMs to disassemble the technical competency in product teams in favor of program
management based activities to manage programs has been considered by some to be a
practice that constrains the design community in a way that threatens a proper integration
of components and systems. There is a lack of interest in engineering competency in the
US as reported by Downey and Luccena, (2007) (p. 4).
The importance of this competency can be viewed from a dealer experience and
how this touch point to the customer may enhance or hurt loyalty and satisfaction. Then
it must be reconciled if an OEM is going to differentiate itself on something of substance.
The OEM will have to do it through the combined efforts of design for service and likely
through the dealer network. Therefore, a service focused Life-Cycle design strategy is a
new gateway to customer satisfaction and customer purchasing decisions as O?Connor
(2002) points out the need for design rules for maintainability (p.285). The buying
experience through resale and repurchase ? are all elements to be treated by design and
are required to wage the war to gain a customer?s trust.
Financial challenges facing the domestic automotive sector (now more than ever)
reveal an economic crisis. It is also a design crisis requiring the identification of new
opportunities for quality improvement and lower cost of ownership for the customer.
How OEMs do business structurally must be evaluated. Interestingly, by using the
typical data found in an existing warranty system that routinely records failure
11
occurrence, failure symptom, mileage, part cost, and labor cost one can investigate such
factors over mileage and time to observe trends with the purpose of learning from that
data against the backdrop of business operations. This data has not been unused with
respect to find-and-fix approaches to quality problems. However, additional ?critical to
quality? service related weaknesses have not been properly identified or effectively
improved upon. One can make this observation simply by organizational deployment
having engineering and service parts organizations separate.
Related Research
Life-Cycle Design
The study subject matter is most closely related to this paper presentation on the
subject of Life-Cycle design methodologies that have been identified and discussed and
for which there is a computer model approach used in beta at General Motors as reported
by Byran et al., (1992).
The entire research and application of the Life-Cycle design methodology
considered attributes of design respecting the typical functional requirements but also
producability, assembly, testability, serviceability, transportability and disposability.
A significant weakness was apparent in this model as to its reliance on the design
aid methodology that assists producability - Design for Assembly (DFA). It seems
incompatible if one applies a DFA methodology that involves sequential building of a
complex system in contrast to the required access needed to isolate, remove, repair or
replace a component at the site of failure. DFA respects sequence in production as a
12
convenience. Using DFA as a means to provide cost savings as if it was a significant
contributor to a successful approach for Design for Serviceability or Life-Cycle just
would not work well. This weakness is only touched on lightly by the authors. Most
importantly there were findings that there is a lack of a systematic means to apply
serviceability or maintainability guidelines. This study?s methodology takes this into
account and solves this exact weakness in the product design effort.
Cost of Quality
A meta-study conducted by Schiffauerova and Thomas (2006) considered the
literature about measuring systems to help improve quality by a costing approach
identified as CoQ or Cost-of-Quality. In the research there was evidence presented to
show that companies that provide a formal approach to CoQ are successful in reducing
quality costs and improving quality to the customer. Of particular interest has been the
most common model used - the prevention-appraisal-failure model or PAF.
Understanding PAF in terms of the study approach considers the Life-Cycle and might be
interpreted as follows:
(P) Prevention:
Actions taken to ensure that a ?process? provides quality products and services.
a) From the customer?s point of view he is purchasing, with a vehicle, a product and
service that extends after the warranty period if one continues holding a vehicle beyond
the warranty period. Here the owner holds for residual value and some time with no
13
payments! The study Life-Cycle approach supports understanding the service experience
extending the notion of prevention into the full Life-Cycle-ownership experience.
b) From the company?s point of view both elements as a business process and as a
customer focus goal the trend has been to look at the ownership experience and service
side of ownership as a profit from the sale, financing, and resale of parts if a vehicle
owner is surviving the warranty period.
This has not fostered satisfaction or loyalty. In contrast, a system of design rules
for Design for Life-Cycle cost would naturally bring a customer focus to extend beyond
the showroom and purchase event and into the dealer and personal service arena where a
customer-manufacturer-service quality event occurs instead of a dreaded customer
service event.
(A) Appraisal:
Actions measuring the level of quality in the given process
a) From the customer point of view, there is the voting with his wallet ? where he makes
his initial purchase of a product or service resulting from a given business process. Then
there is the feedback after the sale as to customer satisfaction from its use, its repair if
repairable and finally its disposal and re-purchase cycle. This can be communicated by
a web blog or through warranty event comments and ultimately by his next purchase but
seldom beyond warranty. Again, the study Life-Cycle approach supports improving
quality and ownership experience in service.
14
b) From the company point of view, management of warranty data, internal measuring,
external measuring, and paid for services that assess quality and customer sentiment or
marketing buzz to associate product or service to market share achieved. How that data
is used is not clear in the PAF model ? as to its effective use.
(F) Failure:
Actions to correct quality issues internal - prior to the customer and after the sale or
external feedback
a) From the customer point of view, he hopes that manufacturing defects are minimized
so he does not buy a car made on Monday or Friday or a new model out in its first year.
As to external there is a warranty system when new and residual upon sale. His service
experience will determine over a longer exposure than an advertisement campaign his
buying decisions. If unsuccessful, a lemon law can recoup his purchase price. If no
service can be economical, then it makes the next buying decision less favorable and
should be no surprise. So PAF supports having an excellent service network ? less
interested in profit but understood that there must be some ? just and reasonable business
case supported. The study Life-Cycle approach supports and is uniquely able to focus on
service, a missing element only referred to for future study in the research for the soft
targets. This makes this study all the more meaningful looking back at PAF systems.
b) From the company?s point of view ? it is more interested in how it is competing
against others to the oversight of the customer. Service has been generally ignored.
Resale values have been, according to the domestic automakers, compromised by lease
fleets sales mix diluting the value and resale.
15
Intangibles on failure costs are often hard to resolve according to the research.
However, it is fully resolved in successive buying decisions. This study does not relate
loss of market share and demand modeling but the shift or loss of share has a dollar value
that can be calculated and projected over the Life-Cycle behaviors and aftermarket
service experiences where there is likely a critical mass that is hard to over come and
regain customer confidence if lost.
Further it was noted in the report that the PAF model was ?supported by Modarres
and Ansari (1987)? where they expand the dimensions of the PAF model to include cost
of inefficient resource utilization and quality design cost. This observation is a good fit
for alignment to the process using the study Life-Cycle approach presented in this study
where quality is broad in scope over the Life-Cycle.
In this, the concepts of maintainability are either creating a good impression or a
bad one to a customer inside and outside of warranty. The modularity or serviceability
(a design attribute to facilitate ease of service ) in vehicle design today does not take
advantage of the power of this design philosophy when compared to the life-limit-
sustainability mind set that has established today?s themes of vehicle architecture. Life-
Cycle design strategy has the power to open up architecture not found in most cycle
plans. It would be a new field of work if automakers would be willing to address this
area of design.
Consider vehicle durability; aging and purchase decisions that have been studied
by Lanoy (2005). Basically, the service cost that equals the leasing cost is near the point
of decision. Service costs are fully developed by industry research companies to the
16
extent that the government applies this research when developing and fixing
reimbursement rates of automobile usage considering the costs of ownership in various
markets.
Related to the cost of ownership and cost of quality are the extended warranties
that may sweeten the pot towards extended ownership but it is also viewed as a profit
model for OEMs and hence may detract from the fostering of future sales due to Life-
Cycle issues experienced by the customer.
Design-To Life-Cycle and Congruence to Quality Measures
In the development and application of the Binary Decision Logic Model and
Warranty Index approach created in this study it will be shown that by common warranty
tools a design may be leveraged against specific component knowledge to achieve the
desired quality improvement.
These issues have been hidden in the automotive industry due to a profit focus
from service and segregation of business practices that effect labor and material or part
costs. Understanding the stubbornness of parts and labor costs can help in developing
strategies to overcoming these items. Flexibility needed to change these cost drivers is
seldom achieved due to how entrenched systems are that burden them. Flexibility to
apply and develop criteria to effectively identify and control these cost drivers can be
accomplished by the use of the Life-Cycle approach. However significant the
opportunity to truly adopt this approach it is evident that a dramatic wake up call is in the
market place demanding companies find new objective methodologies that can ?find and
fix? to new levels the various structural issues and failure modes in a design. There is but
little chance to survive without engaging changing of practices that have hindered higher
17
quality achievement. The confidence to proceed is coming from economic conditions
that pressures loss of market share or even OEM viability.
Finally, there are then two keys to know to define new design goals for service
and Life-Cycle cost favorable to both warranty cost reduction and a lowering of the cost
of ownership. The first key is that of the price of parts in service. Often, price is
dismissed as a warranty cost factor that cannot be adjusted downward. It has been
considered unmovable as a contractual value with supply that is governed by various
state and federal laws to some extent as reported by Zenz, (1981). The pricing structure
must be confronted as there is room to change to improve quality and profitability. Price
driven warranty cost in warranty parts distribution is not to be viewed from archaic
monopolistic offering given by the dominant controlling interests in the manufacture as
reported by Hamilton and Macauley (2008). This pushing of the monopolistic price to
service parts is what it termed ?succumbing to the Coase temptation (p.4).? Under these
circumstances such a market share pricing model can rarely develop customer
satisfaction. A gain in sales should offset any losses from trying to extract high
aftermarket sales.
The second key is realizing that a maintainability design solution for better
serviceability has largely been overlooked as a design strategy in the ownership of a new
vehicle design. Here the principals of maintainability can be exploited to reduce labor
costs in service.
18
Chapter Three
Robustness and Reliability of Study Data
To bring advancement in the design process there must be a win-win outcome to
maintain business practice viability while attempting to cross company lines while
lowering costs of ownership. At the same time an increase a vehicle?s value within the
ownership experience is expected through this process of the study. This is where
satisfaction and loyalty are fostered in an increasingly difficult and fiercely competitive
market.
With the goal in sight and an environment of data overload from every type of
information available today it is important to use caution on reliance or belief in the new
data emporium that the web offers. More importantly, when considering the factors that
support a sustained change from a data systems usage one should meet certain system
robustness criteria as a requirement.. These criteria have been evaluated by Henley and
Hiromitsu (1981) discussing stability and integrity over time for a change process to
work. If the new process demonstrates robustness for the elements in the data structures
and the data itself validated as in an existing tested database then re-mining of that data
for formerly hidden value could offer a lasting solution.
A company?s warranty system has just those properties of integrity and has been
stable in terms of assessing frequency and quality of event (inside a warranty period).
The information includes cost, event description, date and various items traceable to the
service event and its resolution. As much as a root-cause determination is lacking in such
symptom based systems the ability to trigger deeper investigations to get at cause exists.
19
This is required to achieve what is herein called ?actionality.? It is the subject of most
problem solving methodologies where one must address a primitive mechanism or root
cause to assure failure recurrence is prevented and that a problem is understood beyond
the symptom level. This is presented by DRM Technologies in their R & M Primer (not
dated) assuring the use of a failure review and corrective action system or FRACAS.
Business Processes Affected by the Study
It is not always clear how business practices may hinder improvement strategy.
However, there is a divide between the cost factors that can be controlled and the
engineering needed to fix a problem. Inspection of cost factors is hardly novel in a
warranty system as it is first a financial reimbursement system for the dealer network and
secondarily a tool to understand reported repair events for the company within the time or
usage domains assisting the find-and-fix process.
With this in mind, engineering?s ?find and fix? approach to the data does not have
impact on the direct cost elements only on the reliability or occurrence rate of a failure
factor. This improvement methodology had worked and is behind the reliability growth
models expressed by Crow (1974). Frequency and time rates of occurrence are able to
be studied to understand if such specific component failure rates within a population of
vehicles is increasing or decreasing over time allowing a means to statistically estimate
future or total cost of warranty. In this application, one must consider if there are failure-
modes that increase over time (See the Appendix ?Reliability Bath Tub Curve, Targets
and Tasks).? There are some such as Sonza and Carvalho , (2005) who have evaluated
fault tolerant and fault avoidance activities to improve reliability and service. This
20
assessment can be enhanced by various design approaches and the use of probability
analysis of the failure data. One can prioritize by the total cost of improvement to repair
costs against the volume in the market.
In this view, the Cost-per-Unit or CPU (a typical quality measure of performance)
that indexes the warranty expenditure of a specific problem across the number of vehicles
produced to get a liability per vehicle over a period of say 3 or 12 months. The CPU, if
used as a design target, is difficult to manage for engineering groups. What is not
addressed in these assessments are the part and labor costs as segregated factors since
engineering in general has considered those factors unchangeable cost factors and
reliability as the item to change in a design.
The notion that price items cannot change is from the belief that the price is
established by a financial business practice where a component being serviced is the
result of a contractual agreement determined by a commercial business unit negotiation.
Yet even the labor rate can be affected by the application of maintainability if less
complexity or skill level would be required. The next question is what can be controlled
in engineering effecting cost? The ownership of the time to repair has not been identified
as an engineering effort but relegated to service organizations to address almost as an
after thought devising how something is repaired or replaced. Engineering all but ignores
the time-to-repair as a design parameter.
The segregated responsibility is a fox guarding the chicken coop where service is
interested in making money from that service event and work package. At most, goals
are reactive only when not deemed competitive to other vehicles. This is a very rough
granularity to lack actionality unlike other quality measures and performance targets.
21
When comparing military contracting to automotive programs for maintainability
requirements, military programs are highly defined. However, this focus also concerns
itself with service needs being driven by poor reliability Gryna et al., (1960) with
alignment to the formation of the AGREE report, 1957 (p. 25) yet lessons blend the needs
of Reliability and Maintainability.
Within a company it is not unusual to find an organizational chimney
exacerbating the situation that could identify and treat service issues aggressively by
being responsible to competitive analysis in terms of serviceability but only if it is in the
design requirements definition phase of a program. Instead of having this presence in a
program serviceability has been addressed only through a series of checklists for
reporting against the competitive benchmarks - it has never been a requirement of design
with any teeth. One personal example is the 2.5 hours to replace a headlight on a
Mustang which is a high impact component subject to replacement frequencies due to
accidents and life limit characteristics.
Although this additional insight has been available in the data it has rarely been
considered methodically and proactively. If addressed in more detail, these elements of
warranty and ownership costs reveal relevant actionable opportunity. Here is the new
leverage point that is the foundation for this study to offer a new ?find and fix? to a Life-
Cycle design approach.
Figure 2. Graph of Vehicle Repair Cost Labor vs. Part.
Y-axis: Part Costs
X-axis: Labor Costs
Finding: Data clustering in part and labor near the
origin, meaning well behaved repair in general.
However, Significant higher Part variable needs to be
understood.
22
Investigating the part and labor elements of a repair event (i), one finds that these
costs are mapped into a function and to related hardware that needed to be repaired or
replaced (Part Cost (i) to Labor Cost (i) for a given repair) . The Warranty Index, as a
ratio of these two values is used to relate characteristics of the design requirements in
relation to serviceability. In an absolute sense, each term is a measure of its complexity.
A high Part Cost (i) should be designed to be easy to service if there is a concern of
frequent repair or make it so reliable and durable that it does not need service.
If the ratio of Labor Cost (i) is much greater than Part Cost (i) then the resultant
ratio is a number that can be as large as 10 or even 20, reference Figure 2. ?Graph of
Vehicle Repair Cost Labor vs. Part,? A full scale assessment of a design is needed to
develop and classify formerly unrealized improvement opportunities from specific threats
discovered in the warranty data base. The cost and frequency of repair from failure or
high incident accidents are inherit in the design and have resulted in unnecessary
warranty and costs of ownership largely unresolved. The chart above shows the
relationship of labor costs estimated from the knowledge of the material costs in an
average repair. Knowing the labor and dividing by the labor rate would yield an
understanding of the time involved or allowed for that repair, a characteristic of the
design.
The development of the Warranty Index from the type of data shown above in
Figure 2. ?Graph of Vehicle Repair Cost Labor vs. Part,? will be shown to provide
guidance for an improved ?Life-Cycle design? methodology. By using the BDLM and
23
the ?Warranty Index,? design teams can readily identify commodities that need
improvement in reliability and durability over the Life-Cycle and if needed, provision for
effective service. A review of current means to assess serviceability requirements is
found to be largely ineffective at identifying and correcting these formerly hidden design
defects in part addressed by Grzanowski et al., (2002). Therefore, the use of the BDLM
and Warranty Index will overcome today?s vehicle design practices that deny any
significant need for Maintainability as a ?Design-To? parameter and where it has been
sufficient in the past to settle for answering serviceability in a general competitive sense.
Warranty, Price and Design Dilemmas
The Warranty Dilemma:
A company doing business to offer its goods or services for profit will have an
interest in their customer?s resulting satisfaction. To continue viable, those goods and
services offered by the company then must meet or exceed customer expectations with a
fair price to value realized as mentioned by Grzanowski and Long (2002) quoting Henry
Ford. Revenues must also meet investor expectations creating a dual edged sword.
In the marketplace, the manufacturer bears a risk as does the consumer. Buyers
beware, caveat emptor, and from biblical perspective, the seller says it is everything
while the buyer is saying it is nothing when negotiating price to value (Proverbs 20:14,
KJV) and so it seems little has changed. In current market applications a seller or
manufacturer and buyer do not walk away after the sale but there is a continuing
obligation of service and maintenance with respect to the products purchased. When the
product is a motor vehicle purchased for business or personal use the risks are significant.
24
A typical data collection set for an OEM Buy Back under lemon laws offers another
stream of data to understand the impact of serviceability as a design criterion. Data
can be evaluated for chronic symptoms and service issues to reveal weakness in
designing for serviceability which can be a combination of design inadequate fault
isolation and skill level of service personnel.
Year Make Mileage Disposition
Buyback
Rationale Buyback Concern
When a lot of time is spent, the Labor content to Part content can be multiple times
part content.
The purchase of warranty contracts considers many elements of specific and implied
liability that spans fitness for use, merchantability, product defect and remedies when
things go wrong.
Today, a warranty will be offered to offset the belief the consumer may have as to
his financial risk from the selection and use of a product that may not meet expectations.
A warranty may be offered as part of a product to just be competitive in the value offered
in its goods or services and to some extent as part of the reputation of the company to
stand behind its product, refer back to Figure 1. Basic Warranty Comparison ? Miles
(thousands). With the adoption of the various local lemon laws, a risk expands the
liability of the manufacturer up to the purchase price rather than a lesser specific repair.
Figure 3. Buyback Due to Lemon Law Data.
A manufacturer is cognoscente of the fact that the concept of cost-of-ownership
broadens the scope of customer satisfaction well beyond that of acquisition costs and
value received or perceived at purchase. An inducement on its face to offer an extended
warranty from the grantor?s perspective looks at it as an additional service agreement that
can be an additional profit center against the risk of failure outside of a basic warranty.
25
On the service side, an extended warranty brings some peace of mind and control to the
buying relationship as a marketing tool and revenue stream to the company and as a
means to assure proper care is given to a vehicle requiring maintenance to operational
specifications to maximizing operating life and performance. Under these arrangements
then there is customer value and peace of mind for a covered repair as the customer does
not want to pay aftermarket prices for major items. The preventive nature to assure
service is performed under an extended warranty is a means to reduce variability in
service and encourage proper maintenance of a vehicle at the expense of losing coverage
in warranty.
The Price Dilemma:
It is apparent that a minimization of the costs of ownership and warranty will
improve the long term operating results of both the company and the consumer. One
significant conflicting approach to the aftermarket comes from the OEM sales goals. A
balance needs to be struck between warranty costs when the manufacturer pays within
warranty and when the customer pays outside of warranty. The conflicted element is
when the company considers the sale of service parts as a major profit making arm of the
company. This practice may not always be a proper business model in today?s economy
and buyer choices. With pressure to push the price up and costs down this model cannot
continue to work when there is limited resources to owners.
The Design Dilemma:
26
There are pressures in design to define a level of effort to improve quality by
various attributes. By improving the component reliability the frequency factor in the
failure event reduces the cost of warranty for that type of repair within the time and cycle
domain in the population of impacted vehicles.
However, it is felt in the dealer network when there is success in improving
reliability it also decreases the dealer?s ability to service vehicles on a volume repair basis
weather paid by the company or the customer. Lowering the cost of the part that is
replaced lowers warranty and the sense of dealer profit if outside of warranty. This is a
factor in the Dealer Business Model where this concern is voiced by MacNamara, (1998).
These pressures have continued the last ten years with continuous warranty events
shrinking with improved reliability as reported in the Detroit News by Hoffman (2008).
Looking further into the service support of the OEM-Dealer network the service-
parts-for-profit model includes technical service to dealer networks and fleets. High
frequency items of service consumables and normal wear and high failure components
require many parts to fill the pipeline to assure effective service. In this case, central
depots or dealers are on hand.
It is a normal process to prepare the service industry through past program
analysis that identifies most common replacement parts. Yet, it is uncommon to have a
maintainability model as an automotive vehicle characteristic of design. On the other
hand, it is common place to have a maintainability plan for military systems. There are
many who develop this concept to the automobile community but it has not taken hold as
it continues to be offered in the research as a new methodology for the auto industry as
by Vacante (2001) and Klyatis and Klyatis (2006) to mention just two. Klyatis proposes
27
that the four elements of quality, Reliability, Durability, and Maintainability offers the
right but complex solutions requiring accelerated testing strategies. The weakness seems
to be centered in the complexity of the approach.
A detailed review of certain warranty and aftermarket part sales can help structure
the costs in service, under warranty or outside of warranty. The costs may be
accumulated in a warranty model as a cost of repair that includes the two elements of
total part and total labor costs on an average basis. Seldom is this detail actually
considered as part of design alternatives with prices that are typically fixed due to
commercial agreements discussed. Over a life time, the individual knowledge of
repairing a certain problem for a certain cost can be understood by the frequency of the
repair event per a useful life or other measurement. At best, serviceability efforts
currently will look at labor as a comparative benchmark that says we are as good as the
other guy rather than managing improvement potential through a customer vehicle
market driver.
Looking Closer at a Warranty System
When looking at the costs of service under warranty from the manufacture?s point
of view one can disassemble the costs into its constituent parts. Fundamentally, total
warranty costs are the sum of part and labor cost for a covered repair. The frequency or
rate of occurrence of such unscheduled repair can be understood by a particular
component failure rate. Then the product of the cost and frequency of that repair during a
coverage period would be the cost of warranty for all such warranty repairs. Distribute
this across all failures and their repairs would collectively sum to total warranty.
28
The rate of occurrence of component failure or failure rate is tabulated by
summation of repairs expressed as Repairs per Thousand Vehicles in Service (Rs/1000)
as an index of quality of a particular vehicle over a given service period to serve as an
index of quality and even a design target for improvement. This data that is gathered
from a warranty system database is managed by the OEM for just this purpose.
Similarly, the data can be looked at from a Cost Per Unit (CPU) basis which is the
vehicle?s cost for service on a per vehicle basis removing the total number of units of
scale from the calculation ? again over an equal time in service as the Rs/1000. The cut-
off service periods used typically include 0 months (pre-sale issues), 1, 3, and 12 months
in service. The feedback from this system of warranty reporting brings to light the need
to identify specific failure modes as contributors and the development of action plans to
eliminate the individual failure modes ? this is in part the find-and-fix approach found at
the customer?s expense rather than in development. The effort to reduce the failure
modes is referred to as ?failure mode avoidance? and is prioritized by either CPU or
Rs/1000 or both.
These modes of failure can be understood by analysis as to the cause of failure
with a proper investigation. The macro grouping of issues is often subdivided by vehicle
functional groups or by replaceable component tracking ? at the warranty serviceable part
level. The assignment of responsibility as to being a manufacturing issue is separated
from a design controllable issue or even a supplier issue. In all cases, it is still the OEM?s
issue to consider a possible corrective resolve. The influence on Design towards failure
mode avoidance has been to improve the reliability and hence reduced warranty. This
approach has reached diminishing returns with most competitors reaching parity.
29
To distinguish amongst OEMs one can see their marketing plans that include
some variation reference, Figure 1. ?Basic Warranty Comparison ? Miles (thousands),?
with respect to warranty offerings. Since there is only symptom level data in warranty
data then without a deeper investigation or without a deeper look into specific failures as
to root cause in the warranty data base, there is a lack of visibility as to seeing deeper
causes of failures against vehicle architecture. Warranty systems do not routinely attempt
to address, in a feed-back sense, back into the development cycle plan with respect to
serviceability.
The Basic Properties of the ?Warranty Index?
The ?Warranty Index? as it is developed in this study, is the relation between the
Labor Cost and the Material or Part Cost of a given repair. The approach attacks head-on
the belief that neither of these items can be changed effectively by design. These factors
that are inherent in a warranty system?s database and with some review can be evaluated
to a component level and with appropriate assessment. These factors can become
primary targets that trigger an investigation into ?Design-To? objectives include
Maintainability and Serviceability in support of a Life-Cycle design methodology. From
these failure events or from analysis proactively where they are triggered by frequency
rankings. The frequency rankings are part of the Warranty Index application
methodology. These rankings become discriminating characteristics of design and are
applied to Life-Cycle cost design principals.
30
The first tier triggers are frequency of event, as frequency drives the
multiplication times the cost of a repair in warranty costs. Breaking down the total cost
of repair into the above factors allows one to see that there is a ratio that most repairable
parts are contrasted. The Warranty Index can be subject to a lot of manipulation
statistically but in this study relevance and simplicity can guide this approach. Then it is
the Warranty Index defined by looking at a ratio of Labor Cost to Part Cost - a
differentiation that can exist to help prioritize effort to improve design.
Understanding the need for improvement comes to each of us that have sought a
repair service. A most interesting human condition goes to the ?sense of fairness? one
has about a repair. If for instance, a part costs $5 dollars but the labor is $200 dollars ?
something is wrong! But a light on a corner module or under the dash needing the dash
removed just might cost this much.
The failure in design is the lack of access, the need for special tools or the
requirement to remove ?bonus parts? ? ones that must be removed first - adding waste to
the formula. This is a touch point that one would hope would be easy to effect as to the
degree of dissatisfaction that results when out of warranty.
Since frequency triggers the analysis into serviceability and most systems will
compile a list of ?high value parts? that are often repaired or break in use then how might
one create a hierarchical approach to setting up ?Design-To? targets? A proper response
to this listing follows the basic principals found in Reliability Centered Maintenance or
RCM as analyzed and discussed by Vacante (2001) and described in report ALM-43-
7494-C, Chapter 12.
31
Using the Warranty Index as a relation between cost of labor and parts, there is a
logical decision process utilized to guide a design team. As in any use of a logical
decision model, an effectiveness criterion should be developed when performance goals
are intended. Use of a logical model has been shown to be effective if the following three
elements of are used in its design as reported by Kellogg Foundation (2004) where
models are constructed from conditions, actions, and rules to follow. The selection of a
decision algorithm or logic model helps provide a systematic and visual way to represent
and share an understanding of the relationships among the resources defined to operate
the system and the activities planned together with the changes or results one hopes to
achieve. A prescriptive guide for Design-To Life-Cycle cost and service requires such an
orderly device to communicate the process.
The decision model is simplified differentiate which is more, labor or part cost,
where one can begins to focus attention on the relations in the data. The ratio between
labor and part cost helps focus improvement opportunity. As discussed above, higher
labor is a problem if the part is cheap. In a similar fashion, a highly integrated part that
has no serviceable parts may mean that a very large part cost is charged and the labor
being low, yielding a low ratio of Labor to Part cost. So the model would be consider
extremes on either side of unity ? very large or very small ratio with respect to unity in
the ratio being nominal ? labor and part costs being nearly equal where equal opportunity
may exist for improvement.
32
Low High
Hi
gh II. Alert
(labor cost) IV. Bad
Lo
w I. Good III. Alert
(part cost)
Labor - Part Costs - 2 x 2 Box Analysis
Part Cost
La
bo
rC
ost
An analysis in quadrants, like the one below, helps bring a three dimensional
relationship into focus. Here the cost factors of Parts to Labor relative to frequency (Part
- X axis, Labor -Y axis) versus frequency ( Z - axis), see Figure 16. Frequency MTBF ?
Part ? Labor: Region I-IV.?
Table 1. Warranty Claims - Qualitative Analysis.
Warranty and Service Parts Business
?The related service parts business in America is said to be over 7.7 Billion
dollars in 2006. Further, manufactures such as Ford Motor Company have indicated a
reduction of 1 Billion in warranty costs due to improved quality that has reduced
component failure rates. The manufacturers, dealers, and customers all are significant
stake holders in this relation? says Hoffman, (2008) in a recent news story.
33
A Deeper Look into Part and Labor Cost of Repair:
Service from either an independent or a franchised dealer has given rise to a
phrase ?out-the-door-pricing.? This terminology is supposed to give confidence to the
consumer as to the belief that there are no hidden costs associated with that dreaded trip
to the dealer or service center for parts needed when not covered by warranty. This will
also include service for consumables and items not covered under a warranty agreement
such as a tire replacement due to ?normal? wear. However, this terminology does not
change the cost of any particular service or part needed to restore a vehicle?s operation.
It does serve to understand the customer?s real expectation to want fair and complete
service for their vehicle. With the evaluation of such consumer focused organizations as
JD Powers, Consumer Reports, insurance agencies, and the Federal Government,
publishing comparative performance data there is clearly room to improve quality and
owner satisfaction.
Historical Serviceability in Prior Vehicle Generations
With an eye on serviceability when quality was lower and reliability poorer,
service solutions had to be essentially easy. Today, systems are more complex, less easy
to diagnose a problem and service the repair yet they are lasting longer. Increased
longevity is realized today to exceed up to 8 years. On an individual repair basis, there is
an increase in complexity that has pushed up the cost of repairs where there is less
attention to service effectiveness and user serviceable parts. Complexity of diagnosis
34
tools and time to isolate failure causality eludes most service groups reviewed by Ishi et
al., (1993). The serviceability effectiveness has suffered even though reliability has
improved.
In an analogous way, computer software design in the 80s under the early DOS?
or assembler for embedded systems environment required execution in very limited
memory. Applications for desktop and earliest lap tops, such as the Otrona? to run on
floppy disks or 5/14? with storage capacity of 360K bytes and no hard drive storage or at
most 10 Mega bytes (IBM Xt). Assembly Language, BASIC, FORTRAN, C, and various
languages used to create executable code to manage the task in fewer than 256K bytes of
Random Access Memory. Today, that is not the case and where computer instruction is
no longer coded in a compact way as there are Millions of Bytes of Code required just to
create a structure for programs to initiate. The analogy when comparing the auto designs
of today for longer life and longer service intervals gives less attention to service
effectiveness increasing fault diagnosis and makes servicing PCs or Cars left to skilled
technicians.
For a service center to be competitive to support franchise administration there is
a challenge when looking at the time-to-repair costs from the prevailing labor rates within
region and adding the additional margins required sustaining a business. A second time
element for a service event is design related. Just how long a service event takes is a
formidable task to decide by the OEM and how it is handled in the service center.
Traditional maintainability time elements are often defined by the allowable time
durations to assess against labor codes due to a warranty service event. These times are
spelled out in maintenance manuals and through the service system defining warranty
35
service events. In other words, they are contractual time standards to conduct most
service events. If on one hand a dealer can perform the task in less time, there is more
margin and hence more profit. Alternately, if the tasks cannot be performed to standard
there is a network that interacts with the actual field issues and resolution is still possible
where the system can update with experience to changing the standard times upwards.
By looking at the interaction of manufacture, dealer, and customer one can expect
a successful service event can be obtained because of the risk under the issuance of state
lemon laws. Most of the obligations of the interested parties lie within the OEM who
must be able to satisfy a repair need within three visits or be subject to the financial
penalties of the law. This penalty can be up to the reimbursement for the entire vehicle.
The liability that falls to the manufacture is an inducement to maximize elements of
service through design. Typical elements like diagnosis and validation methods aid in
service care. This is why within an OEM-Dealer network there is a service technician
within the OEM to aid a dealer technician to solve problems that are truly design driven
and often labor intense.
Within the specified time parameter allowed for any service event there is a
complexity in the design with respect to the ease of that maintenance. The accounting of
diagnosis, access time and any removal of ?bonus parts? explain the costs experienced.
Bonus parts are those parts that must be removed to access the causal part needing the
actual repair. Again, this drives labor costs higher and if the broken part is inexpensive
then there is high dissatisfaction.
Poor diagnosis, lack of modularity, inadequate training, need for specialized tools,
ineffective ergonomics are all failure modes of poor maintainability. All this adds to a
36
new area to target reduction potential in warranty cost and poor ownership experience.
The adverse effect of any and all of these elements can be addressed by design practices
in Maintainability engineering. Again, to a customer under a warranty there is the
aggravation of down time opposed to service or part cost. Similarly, high service costs
outside of warranty and the expectation for the cost of a repair will influence future
purchase decisions and opinions about the product that could take years to correct.
Components that cannot be restored in an adequate time or within adequate costs
limits will cause a negative bias. Through application of the BDLM and the ?Warranty
Index? service related issues between part costs and labor costs can be challenged. The
offensive component design practices that are entrenched in a vehicle?s design
architecture can now be exposed and solutions for effective serviceable parts and part
price points can be measured and managed.
Exposing the Price Point for the Aftermarket
Price determination involves both legal and economic considerations for any part
that is ultimately sold to a consumer. The OEM selling their product under terms of a
warranty has a duality of concerns in which pricing draws a distinct difference between
the original part-price built into equipment and that which is differentiated by being a
service part or a replacement part, discussed by Zenz, (1981). The cost for service parts
being higher than the same part built in producing new product originally. It is
commonly known that to build a car out of parts from a service center will cost
considerably more than the parts procured that support the OEM production and with
good reason. There can be costly logistic considerations of distribution, storage and
37
support coupled with somewhat unpredictable periodic low volume purchases for service
parts after production through the service life of the product ? support for up to 10 years
past last production.
Focusing on general pricing strategies one may include any mix of cost, market
conditions, economics, administration, controlled or even psychological strategies.
Monopolistic Pricing
The more the corporate position in the marketplace is monopolistic then the more
a price strategy may favor the administrative pricing characterized as having higher profit
potential indicated Hamilton and Macauley, (2008). This may be interpreted by the
consumer as just adding an unfair weight upon his back ? the very consumer the
company wants to sell its next product at time of disposal and replacement, a point made
by Zenz, (1981).
Further, the notion that selling original equipment manufacturer?s (OEM) parts in
the aftermarket as a profit center has likely been leveraged to the determinate of the
consumer, refer to the fact that Ford?s Field Service Division was 350th of 500 of the
fortune 500 in 2001 with $2,000,000,000 in profit. Looking at more reasonable price
point for aftermarket product is needed to heighten ownership experience and customer
satisfaction.
The marketing model that sells administratively low for an OE product can with
proprietary required consumables continue to earn through these replacement parts ? e.g.
ink to printers or software to a computer. Similarly, one could compare gas and oil
consumption to the car costs. Sustaining consumables include additional items as
38
insurance, licensing, maintenance scheduled and unscheduled and lastly interest on loans
to purchase vehicles. Having the terms of loans equal the life of a vehicle would liken
the operational and acquisition costs to a guaranteed flat rate pricing. These facts are
explored by Hamilton and Macauley, (2008) when looking at longevity, durability, and
operational costs as a price willing to be paid for maintaining a vehicle to be equal to
leasing a vehicle.
For an OEM with a resale service part seller there are controls or legal
considerations that are regulated by the Fair Trade Laws. Court action has provided for
dual pricing strategies but prevents unfair trade where lower priced OEM parts would
unfairly compete with resale and therefore it is illegal. For this reason, the differentiation
of OEM and resale must be clearly stated on purchase orders.
Where trade within a state may very between states with regards to pricing of a
resale item there is a direct impact on cost of ownership. Where all this concerns the cost
of ownership is when together, with the cost of warranty, there are clear effects to both
seller and buyer. If the buyer experiences a failure of his vehicle there may be a concern
if it is covered under warranty. If covered then it becomes only a time liability from the
buyer's point of view. To the manufacturer it is a warranty expense. If the selling price
for that part, in warranty versus out of warranty is a shock it is required of a service part
to be at service pricing and not the lower priced OEM part. Mark ups for the part
manufacture, mark up for the dealer sales, and mark up for the OEM distributor is in this
chain if the OEM considers the service part sales as a profit center. Having part
reliability survive a warranty period pushes the burden to the buyer if useful life equals
the warranty period and will adversely affect resale once this weakness in a vehicle is
39
significant. Each a time this liability is a cost liability and in his view it adds to his cost
of ownership.
Another design concept is to push reliability with longevity past the
?sustainability? life-limit ? vehicle end of life as a disposable. Under this condition, the
unscheduled maintenance would be minimized and in reality is ultraistic. However,
much of this is evaluated in engineering and test demonstrations to understand the end-
of-life of components and systems.
Sense of Fairness
It seems everyone considers a sense of fairness when faced with a repair cost. In
other words, there is an expected cost for the value of a failed part and its repair. If not
deemed fair this goes to customer dissatisfaction. Under warranty, it is the interruption of
usage. Outside of warranty the full liability of time and cost is to be born by the
consumer and can be a shock as it was subsidized prior under warranty. On the other
hand, if the price is believed to be fair then a service experience may be very positive
one. Such is the point of view of the chief editor of the Reliability Collaborative
Association, Vacante, (2001).
The regional impact of the cost of car ownership provides powerful insight into
some of the issues facing car owners. The variance is in large part due to economic and
insurance rating factors. The study of cost of ownership differences help to establish
rates for reimbursement for car usage for tax purposes, state and company policy. One
respected research organization for such data is from Runzheimer International, as
reported by a CNN Money staff writer Les Christie, (2008). The lowest yearly cost of
40
$7,399 is found in Knoxville TN whereas the highest of $11,844 is from Detroit. These
costs identified that insurance in Detroit is more than $5,000 annually from a baseline
vehicle.
The Maintenance contribution is measured in cents per mile with a low of 4.69 in
Bismarck, North Dakota to 7.35 cents per mile in San Francisco as expected for a
baseline vehicle being a Ford 500 2006 fully loaded used for 4 years at 15000 miles per
year. Many of the factors influencing the costs are also design related including fuel
economy, accident repair, and service requirements. Naturally this excludes warranty
costs which are a part of the purchase price of a vehicle considering the data.
The element of insurance that is design related is the costs of coverage for
comprehensive and collision. Additional studies of such costs are on a vehicle by vehicle
basis and the expected frequency is of an actuarial nature. It is not clear what if any cost
is factored in for deductible costs but even minor accidents are evaluated by the Insurance
institute for 5 mile per hour crash tests presented by Mayes and Wasilak (2004).
Because an incandescent light assembly should be on a high frequency list efforts
should be used to drive down labor and material costs. All customer experiences are
cumulative as to their relationship to the OEM and purchasing decisions in the future. A
simplified interaction viewed by the customer includes relating to the customer from
purchase through service, see Figures 4. The OEM-Dealer-Customer Interaction. and
Figure 5. The Typical Repair Flow Chart.
41
Figure 4. The OEM-Dealer-Customer Interaction.
Figure 5. The Typical Repair Flow Chart.
Customer
purchase and
usage
Dealer
sales and service
OEM
Manufacturer
& warranty
Repair Event
Diagnosis
Remove / Repair or
Replace / Reinstall
Verify
Administration
Complete
Start
42
Chapter Four
In this section, the development and application of the study methodology is
given together with an organizational deployment of the BDLM and ?Warranty Index?
process. Guidance is shown as to the meaning of the factors and certain graphical
analysis to both qualitatively and quantitatively create an example for a given data
structure in warranty.
Required Organizational Structure
In order to conduct the study in a target organization there is a required
establishment of policy, procedures and the definition of the roles and responsibilities for
the quality improvement processes described.
An organization that meets today?s quality standards provides the basis for that
responsibility. A new element for the automotive team is to assure there are Design
Assurance specialists that can carryout the data driven processes that take the following
form:
Basic Design Actions ? long term
Reliability Projects
Maintainability Projects
Parts Cost Reduction in Projects
Labor Time Requirements in Projects
43
There are high level organizational objectives that are cascaded to lower levels of
indenture of the work breakdown structure of systems produced. The targets that become
objectives are developed at four levels of responsibility:
1. Level 1: Organizational ? DTLCC goals at the Top Functional Level:
2. Level 2: Vehicle
3. Level 3: High Level Functional Groups.
a. Engine
b. Body
c. Electrical
d. Chassis
4. Level 4a: Lower Level Functional Level
5. Level 4b: Lower Level Part / Item Level
Project Management Structure
The project management structure defines the team, organizes the tasks and assures the
design team carries out the mission to achieve Life-Cycle wise designs. Eighteen tasks
have been developed that are summarized in a table. A Walk through those 18 steps are
given as the method used to assure this concept can be deployed.
1. Design-To LCC Decision Logic Modeling
Develop the policy, procedures and work instructions to carry out the quality
improvement program.
2. Roles and Responsibilities Defined
44
The program elements agree with procedures and identify CEO ? sets policy;
Vice Presidents and Directors Carry out the Policy and assure commodities
write procedures and work instructions; Program Managers, Chief Engineers
Identify the teams; Design Assurance Technologies Staffed; Cost and Service
Organization involved early in the program.
3. Establish Design Requirements
Design Requirements are set at four levels. The first considers past program
data and sets a tasks to reduce variance in service by the 3rd quartile in part
costs and labor in service. Reliability tasks are established based on
sufficiency to the improvement goals. Statement of Work, required with
suppliers.
4. Data Process
Validate the data systems for each project
5. Generate Frequency Lists
Use the Decision Logic to generate the Frequency Lists
a. Reliability List Generated
b. Maintainability Labor List Generated
c. Maintainability Part Cost List Generated
d. No Project Lists Generated
6. No-Action Lists Published
The No-Action lists are those components that are not high frequency threats.
7. Define Actionable Lists
a. Reliability Project team reviews and makes proposal
b. Maintainability Labor Project team reviews and makes proposal
c. Maintainability Part Project team reviews and makes proposal
8. Undefined Action Lists
Those items that have no planned actions but are on the frequency lists. These
become projects if goals are not met by assigned projects.
9. Assign and Track Projects (objectives set in program plan)
Chief Engineer Approves of Project Commitments and obtains budget.
45
10. Reliability Frequency Projects
Conducted
11. Accident Frequency Projects
(assigned in the Maintainability projects)
12. Maintainability Projects
Conducted
13. Failure Rate Improvement
Specialized Reliability Project ? Tracked by PM and CE
14. Extended Life Limits
Specialized Reliability Project ? Tracked by PM and CE
15. Part Cost Projects
Specialized Maintainability Project ? Tracked by PM and CE
16. Labor Cost Projects ? Tracked by PM and CE
Specialized Cost Project ? Tracked by PM and CE
17. Completed DTLCC
18. Monitor ? Executive Level
Process Deployment ? Target Setting on New or Existing Programs
Chief Engineer ? Product Group or Functional Group
Existing Program Application: (how to get started)
? Using the Statistical 3rd Quartile cost values (sufficiently large sample) as a guide
to begin controlling Labor and Part costs.
? Using the manager?s Maintainability Action Sheet, Identify the action plan
benefits by time category for labor factor
46
? Using the manager?s Cost Opportunity Action Sheet, Identify the Cost / Pricing
challenge.
New Program Application: - Ground Zero ? uses surrogate data and can apply
techniques earlier in a cycle plan for improved effectiveness.
? Using existing program surrogate data develop new targets for Life-Cycle Design.
? In response to the Binary ? Decision Logic Assessment:
1. Using the manager?s Reliability Action Sheet,
2. Using the manager?s Maintainability Action Sheet, and
3. Using the manager?s Cost Action Sheet,
4. Using the manager?s Basic Design Action Sheet.
The action plans have the following form for Reliability actions with noted signatures
required:
Figure 6. Reliability Action Sheet
Other Resurces Y/N Reference Project
Decision Logic Used
Maintainability
System Cost
Subsystem Architecture
Component Quadrent Part Threshold
Requested by: Quandrent labor Threshold Date
Design Relibility Actions
Action Proposed
Current
Value
Improvement
Percent How Verified Validation Approval
Life Extention
Safety Factor
Redundance
Simplification
Derating
Variation Reduction
Other Robustness Actions
Failure Mode Avoidance
Signatures Sign below Date
Chief Engineer
Commodity Specialist
Reliability Specialist
Service Engineer
Program Project No.
47
Figure 7. Maintainability Action Sheet
Figure 8. Cost Action Sheet
Other Resources Y/N Reference Project
Decision Logic Used
Maintainability
System Reliability
Subsystem Architecture
Component Quadrant Part Threshold
Requested by: Quadrant Labor Threshold Date
Design Cost Actions
Action Proposed
Current
Value
Improvement
Percent How Verified Validation Approval
Current Logistics Costs (including service strategy)
Current Production Parts Cost
Current Service Parts Cost
Current Quantity in Service
Warranty Agreements
Other Cost Factors
Commercial Issues
Signatures Sign below Date
Chief Engineer
Commodity Specialist
Purchasing Specialist
Service Engineer
Program Project No.
Other Resources Y/N Reference Project
Decision Logic Used
Reliability
System Cost
Subsystem Architecture
Component Quadrant Part Threshold
Requested by: Quadrant Labor Threshold Date
Design Maintainability Actions
Action Proposed
Current
Maint
Time
Improvement
Percent How Verified Validation Approval
Fault Isolation & Self Diagnosis
Parts Standardization
Modularization
Accessibility
Repair vs. Replacement
Proactive Maintenance Preventive
Proactive Maintenance Predictive
Signatures Sign below Date
Chief Engineer
Commodity Specialist
Maintainability Engineering
Service Engineer
Project No.Program
48
Figure 9. Basic Design Action Sheet
Other Resources Y/N Reference Project
Decision Logic Used
Maintainability
System Reliability
Subsystem Cost
Component Quadrant Part Threshold
Requested by: Quadrant Labor Threshold Date
Design Architecture Actions
Action Proposed
Current
Value
Improvement
Percent How Verified Validation Approval
Current Targets for Cost
Current Targets for Reliability
Current Targets for Maintainability
Structural Systems that Hinder Change
Next Available Time
First Program Opportunity
Other Design Issues
Signatures Sign below Date
Fundamental Design Cycle Specialist
Chief Engineer
Commodity Specialist
Purchasing Specialist
Reliability Specialist
Service Engineer
Program Project No.
49
Deployment - Warranty Data System Usage
Figure 10. Warranty Database Definition
A Warranty System is a collection of Repair Events and is used in this study as follows:
A Record is a repair Event
A field is a data association within the record.
Fields Used in this study
Time Domain:
Miles,
Days in Service
Qualifiers:
Vehicle Styles
Functional Codes (Symptom Based)
Service Part Codes (Hardware)
Labor Costs
Part Costs
Vehicle Styles
A vehicle line may have major style identifiers such as a convertible or a performance
enhanced version, a sedan or a coup and so on. Therefore, each may have a different
application that can influence a system?s stress e.g. miles driven per unit of time.
Functional Codes (Symptom Based)
Functional codes break the repair down to lowest symptom e.g. Electrical - > Radio - >
CD Changer (not working) and is similar to Effect in the Design FMEA.
Service Part Codes (Hardware)
Labor Costs ? Time to repair is not in the database directly. Labor costs could be reduced
to time by knowing the labor rate.
Part Costs ? Includes all materials.
50
Deployment - Warranty Data Filtering
Removed pre-sale warranty claims by setting filter of view days in service >0.
Part Costs > 0
Labor Costs > 0
Figure 11. Excel Data Feed ? Warranty Data Filtering for Study
Deployment ? BDLM & Warranty Index Procedure
Rationale
An automotive manufacturer?s warranty system forms a wealth of knowledge
concerning the quality (or the lack of quality) of a fielded product. The warranty system
is a managed business process that provides and aids to a dealer network in response to a
repair required. The vehicle owner obtains a needed repair and the dealer obtains
reimbursement for covered repairs.
51
In this report, when costs are noted, unless otherwise stated, it means a retail
value as viewed by a customer who pays or as reimbursed amount to the dealer which in
general are equivalent costs.
Warranty agreements are part marketing and part competition and part regulation.
In addition to a reimbursement system for service providers for warranty covered repairs,
the data also includes the ability to make assessments of product quality and as such have
been culled to effect quality improvement opportunity. A fresh review of this data has
identified a new business process developed to improve quality in design ? a design that
focuses on the Life-Cycle. This Life-Cycle approach is operationalized and examined in
this study.
Fundamental understandings of the key parameters of cost, under warranty, are
needed to develop the Life-Cycle design approach. Essentially, the cost elements of a
repair process include a claim made by the dealer for the Parts and Labor used to affect a
repair. Further, these parameters are fields in the larger data record for that repair with
their association within that and other such warranty repair records.
It is important to note that the tracking and assessing of costs of a particular repair
or type of repair are the factors collected with a warranty claim record that is required
input of the dealer actions. Repair records that allow for symptom level information and
hardware level information all of which help assess the warranty claim and assist in
understanding the reason for the claim in terms of the product function and the impact of
profits due to the dealer visit and as related to the product unreliability in the use of the
vehicle.
52
Warranty Index
A created and measurable parameter is proposed by the ratio of Part Costs divided
by Labor Costs. This statistic will be called the ?Warranty Index.? The index, as Part
Cost divided by the Labor Cost, is defined as an independent and random variable
associated with a repair event. Similarly, it can be a tabulated value based on the
collection of repair events associated by various factors in the warranty repair record.
Inquiries into the warranty data base for all similarly defined repair events becomes a
sample set of the warranty data, representing members of the sample space of all repairs.
Let?s define further a value given as the Total Cost of Repair, or TCR that is
simply the sum of the cost items that yield the total part and labor costs for a specific
repair.
Total Cost of Repair, TCR
The TCR for any repaired item is a dependent variable with independent variables
being the paired cost contributors of Part Cost, given by P(i) and Labor Costs, given by
L(i) for part and labor costs respectively of the ?ith? repair.
The cost to effect a given repair are entered by a dealer for which there is a
warranty agreement that allows certain recoverable dollar and time values in a pre-
53
defined manner as determined by the manufacturer?s service parts organization and
dealer assistance function.
The labor charged comes from a detailed compilation of standard repair
reimbursement rates that have been developed and authorized by the service organization
of the manufacturer. The labor costs are derived from the local labor rates and assessed
against the time allotment.
Each P(i) and L(i) are found in the warranty system record and retained on each
repair in warranty and possesses certain detail down to the vehicle identification number,
odometer reading, days in service, technician and customer comments to name a few. All
these elements of the record can aid the warranty data manager to assure that a properly
coded, validated repair record exists. For this study we will take a closer look at a data
types from the warranty record that will support the design approach for the Life-Cycle.
Deployment - Cost Creep and Maintenance Time Control
Repair data is captured primarily according to documented repair agreements
through a set of standards between the OEM and the service network. The warranty
system classifies the data by a uniform set of interactions. The system is on line and if a
given repair is not resolved at a dealer service center under the warranty cost recovery
rules then a resolution process exists to review the standard and the allowable
reimbursement.
As to the effect on empirical results for labor costs or parts needed over time,
there is a property of ?Labor Creep.? If a dealer?s skill, training and ability assures
repairs are completed in less time than the standard then there is no complaint made.
54
There is more profit per repair in this case. However, if the service cannot be performed
in the allotted time, then a complaint system can be used to adjust the time standards
upwards ? hence labor hours for certain coded failure repair events will grow. While the
growth or elevation is correcting a problem, the opposite does not hold where less time is
allotted for lack of complaints on allowable time.
Deployment - Quality Assessment
When large numbers of repairs are gathered of a similar nature, then one may
graphically observe a trend. One macro trend reveals that overtime labor hours are
reduced across all repairs on average. This may be attributable to initial quality in a new
model introduction where the bugs have not all been worked out sufficiently or from a
lack of experience in the field for the repairs.
To this end, a network of service technical assistants are available to
interact/train/and certify regionally dealer service centers. Over time, a maturity results
in both product and service costs.
Construct for key variables of the study
Definitions related to labor costs, L(i)
The average and standard deviation for labor costs for the ith repair is given by:
1
)(
and 1
2
1
?
?
??
??
??
n
LL
sn
L
L
n
i
avri
n
i
i
avr
Similarly, the average part costs for the ith repair is given by:
55
1
)(
and 1
2
1
?
?
??
??
??
n
PP
sn
P
P
n
i
avri
n
i
i
avr
This is done at the lowest symptom code available in the data structure.
Note that P(i) and L(i) are independent and random variables
Finally, the total cost of the ith repair is given by:
TCR(i) = P(i) + L(i)
Where the average and standard deviation of the costs of a the ith repair are given by:
1
)(
and 1
2
1
?
?
??
??
??
n
TCRTCR
sn
TCR
TCR
n
i
avri
n
i
i
avr
Finally, the reciprocal of the Coefficient of Variation (CV) is given by:
1/CV = 1/ (s/Xavr) or just
Xavr / s.
The CV has been applied as a standard measure related to variation in the data
sample. The simple meaning gives objective power in terms of the development of the
sample space in the study?s measured values. Specifically, CV answers the question of
what percent of the mean is consumed by the standard deviation.
56
If the mean were 20 and the standard deviation 5, then the CV is 5/20 or 25%.
One could then say there is a limit to the confounding as to the dispersion of the data
about the mean of the given measured value.
If a normal distribution is attributed to the measured values and its variation, then
one would know that there is 68% chance that of all values measured they will likely fall
within a standard deviation of the mean value ? or 68 out of 100 measurements and an
estimate of the mean the answer should be between 20 - 5 and 20 + 5 or between 15 and
25. These are central tendencies of the data that analysts may rely from which inference
may be made.
The choice to use 1/CV in the study considers that the reciprocal of the CV or
1/CV. We have substituted Xavr for mean in the general formula. As used in this study
is found by:
1/CV = 1 / s/Xavr = Xavr / s
The utility of the valuation and being able makes it easy to differentiate in relative
terms about the ratio being unity with respect to the data and its behavior.
In a repair process we observe the following:
If s = Xavr then 1/cv = 1
If s << Xavr then 1/cv = Xavr / s(small) >> Unity
If s >> Xavr then 1/cv = Xavr (small) / s(large) << Unity.
57
Applying the 1/cv can be a can be used in a macro sense to understand if a
measure value is orderly and if it behaves well as an estimator. Note the measure is
considered well behaved if s << Xavr or if ill behaved when s >> Xavr.
When compared to production, one wants to minimize variation so there is a
smaller loss function ? directly related to quality in manufacturing where it is desirable to
have all the parts the same to favor assembly quality. However, in a service or repair
process, variability is higher, much higher. In part due to less control operator to
operator, or technician to technician.
With the P(i) and L(i) defined a given repair can then be assigned a class of repair
to levels that differentiate symptoms and hardware at lower levels of failure mode and
cause. Here we have collectively the TCR(ijkl) failure type where a coded repair may
follow the following descriptors:
Figure 12. Depth of Warranty Target Data in Service.
I = Design Hierarchy ? Body, Chassis, Electrical and so on. Design Responsibility is
often defined organizationally along these lines of basic design.
J = Functional Subgroup - such as Brakes ? a hardware related descriptor - a subset of
Chassis in this example.
K = Symptom Code - such as Noisy, Broken, Wrong Feel, a qualifier similar to a failure
symptom.
L = Hardware Code - Part grouping that has been replaced.
58
These four codifications allow inquiry into the warranty data to allow for effected
hardware and symptom basis. This in turn allows an association of miles or days in
service to the classified repair. If an actual mechanism of failure is not readily
identifiable at the detailed part level where the point of failure occurred then physical
parts can be returned to the responsible design engineer or plant personnel for analysis.
The descriptor of K and L can be reversed with different outcomes in the search. Any or
all of the descriptors can be named for evaluation. One could request all things with the
symptom code for noisy. A sense of a quiet vehicle to a customer there has often been an
association with a halo effect on quality overall. Noise is also one of those issues related
to basic design and separate from lose fastener; it is hard to correct in service.
With this presentation, only the three descriptors are used as noted above. P(ijk)
and the associated L(ijk) for any number repairs of the same coded basis, there are
statistics of cost generated that will be utilized to show the impact in design terms of well
behaved or not and with an respect to the impact on Design-To Life-Cycle.
?Well behaved? means there is lower variability in the repair events of the coded
repair. ?Ill behaved? means there is a large variation in the repair event as is easily
viewed by the 1/CV valuation.
Combining the use of histograms with an investigation of the three cost terms,
TRC, P, or L several common observations are made.
1. The histogram is flat over a large range of cost.
The 1/CV value of the cost item will be much greater than one. The repair
cost element(s) do not have a ?well behaved? repair event. The large
59
variability for the coded repair becomes an opportunity to understand the
variability and determine what if any corrective action may reduce service
costs.
2. The histogram is bi-modal.
When the 1/CV value is approximately unity - there may be a bi-model repair
cost for the coded repair. In one case there may be two repair types that are
differentiated in the population of coded repairs and in another, there may be
improper repair or ?bonus? parts ? those replaced due to poor diagnosis before
the real problem is found. The histogram may appear somewhat Normally
distributed.
3. There is good definition in the histogram.
With the 1/CV value much less than one, a given repair?s cost elements that
follow this trend is well defined and understood to the consistent part or labor
cost of the coded repair.
The following guideline is used to provide individual assessments of a repair?s
part or labor costs. Based on the ?wellness? of the behavior the analyst can recommend
the qualitatively within a program team the use of the statistic in managing quality risk
within the Life-Cycle methodology based on the 1/CV valuation of the given cost
element together with the use of the Warranty Index in the BDLM.
60
Program Team
A discussion of the responsible organizational entities is needed to realize the
benefits of a Design-To Life-Cycle approach. The responsibilities can become more
evident if cost factors are also aligned in the design team. The product team is
challenged to addresses service cost for parts as an example. Take an exiting vehicle
needing brakes serviced. It is classified as a component subject to wear. As brakes are
applied to stop a vehicle, frictional forces result in having the brake pads forced upon the
rotor where at a micro level grind down the pads at a certain rate based on the applied
force, material properties of the contacting parts. The rotor in this case, is a harder
material than the pad by design. The rotor allows for the rejection of heat so not to boil
the brake fluid or diminish the braking forces needed to slow the vehicle. The braking
force can further be linked back through the fluid, booster using engine vacuum to
multiply braking advantage, master directing the flow of brake fluid, interacting with the
anti-lock brake unit, and driver pedal input that created the chain of energy transfer that
would cause deceleration.
In this situation, there are also undesired outcomes such as brake dust ? the
evidence of wear. This continues to a point where the pad needs replacement. There is
provision to make it easy to service the pads due to their limited life. If all goes well, the
materials have lasted long enough, but not over a Life-Cycle ? so it is by the virtue of the
wear-out rate, serviceable ? by easy of access and available replacement parts that are
inexpensive, say $80 dollars a pair to replace and out the door ? a consumable that may
not be under warranty.
61
Variability in the repair may be involved with many repairs. As a process of
service in this case we might have the following:
1. document the repair request and estimate
2. logistics time to set up for repair
3. skill level in place
4. tools in Place
5. specific Training / assistance in place
6. scheduling
7. enter the bay
8. hoist the vehicle
9. diagnose complaint to a serviceable repair item
10. access the brakes by removing the tire
11. inspect the pad
12. inspect the rotor
13. grind the rotor if needed
14. replace the rotor if needed
15. repair the caliper if needed
16. replace the caliper if needed
17. remove and install the new pads
18. re-install the tire and wheel
19. remove from the bay
20. verify the repair
21. logistics to document the repair
22. receive payment
23. restore to the customer
Clearly, if the rotors and calipers need replacing, the costs can be variable for the
simple symptom of worn brake pads and be $80 dollars or $800 dollars.
When considering all this, from the service items definition, one might suggest full
life pads, maybe twice as thick to last twice as long if that takes one through the Life-
Cycle and no longer would it have to be made serviceable, take time at the dealer, and so
on. When considering the extra secondary damage that results from waiting too long to
replace the pads ? rotor and caliper replacement, then what is the part cost of these items
and could they be made (sold) at a lower price in service.
62
These types of decisions, on a Life-Cycle basis, are not on the day to day evaluation
of a program for a vehicle. The considerations are bounded by the warranty period, if it
is a consumable and not under warranty as in the brake replacement and what a Life-
Cycle design rule might be in this situation.
To turn this data into actionable design goals and projects to improve the Life-Cycle
objectives, one also has to develop a responsive system to prioritize the actions and to set
targets for each of the coded repair types. This is done by the generation of frequency
lists, the number of repair types are translated into three frequency lists.
Deployment ? Descriptive Statistics on Key Factors
It is required to validate the key indicators or factors to understand the BDLM
with the Warranty Index and its application. To develop the descriptive statics the related
repair event data for part and labor costs are studied:
Vehicle Function Codes
Select Most Detailed Symptom Code
Identifies Major Functional Subsystem in vehicle architecture and then to
a specific symptom that can be associated with a Design FMEA effects code or even to
the Failure Mode level of description, e.g. CD Changer Inoperable.
Index of measured values in a general way is shown below:
Part, Labor, Warranty Index - using descriptive statistics:
Measured value = Xi Cost data ? labor or part costs. They are paired.
63
1
)(
and 1
2
1
?
?
??
??
??
n
XX
sn
X
X
n
i
avri
n
i
i
avr
Indication of strength of the predictor can be given by:
Indexed value: Xavr / s or 1 / C.V. Or alternately C.V.(*)
? used as a measure of variance relative to the mean ? for expedience - the smaller
the better or more well behaved x(i)?s as an indicator. Used in terms of > or < or
<< or >> than 1. It is the reciprocal of the Coefficient of Variation (C.V.) which
expresses what percent of the mean is the standard deviation. As a design rule,
one might want a 2 or a 3 which says a given repair is consistent across the
service network in terms of part cost or labor. For each Functional Symptom
Code ? a tally is built for each Labor and Part Cost for the ?ith? repair.
Qualitative Analysis Part vs. Labor against Frequency
When the analysis using the binary decision logic drives the flow to consider a
need for maintenance then the ?Warranty Index? analysis directs that flow to consider
part cost or maintainability practices to mitigate the cost on a Life-Cycle basis.
What is depicted in the following 2 x 2 charts is guidance to develop criteria to be
used for weighting the service event as ?good,? to ?very very bad,? reference Figure 13.
Frequency ? Labor: Region I-IV. The extremes of the ratings qualitatively, reflect the
degrees of impact comparing part to labor costs. Further, the impact to satisfaction has a
larger impact when the part costs are somewhat lower in comparison to labor and not just
high costs for each alone or together.
64
One could similarly look at the box with ?good? as the low cost part and low cost
labor solution with a low frequency occurrence and likely not impact any satisfaction of
the user in or out of warranty.
The charts look at increasingly higher rates of occurrence for the single event that
the charts represent. The frequency trigger could be unreliability or part failure, accident
frequency risk, or life limit issues. The lowest failure rate is of concern if the part and
labor cost are both high. Also shown is the guidance provided by the Warranty Index
that shows preference to the action that should be recommended given either (P) part cost
reduction tasking or (L) labor time reduction.
Figure 13. Frequency ? Labor: Region I-IV.
Frequency ? Labor: Region I. ?Good?
Frequency is related directly to service count rate of a given repair type. To fix the
service rate generally attributed to failure rate, a reliability improvement may be
considered or even a life-limit extension. However, with the frequency low and labor
cost low there are no recommended changes due to the assessed factors.
Frequency ? Labor: Region II or III ?Alert?
Low High
Hi
gh II. Alert
(cost) IV. Bad
Lo
w I. Good III. Alert
(frequency)
Frequency - Labor - 2 x 2 Box Analysis
Frequency
La
bo
rC
ost
65
When either frequency or cost for labor high alone then it may not generate a
need for improvement actions. The threshold of action must be derived by design rules
and can be associated with knowledge of best in class competitor or continuous
improvement philosophy. However, if the cost of labor is high even with frequency low,
it may indicate a need to improve fault identification and verification. There are no part
costs being considered in this repair.
Frequency ? Labor:. Region IV ?Bad?
Claims that have high labor cost and high frequency are typically identified for
corrective action. In general, the frequency of repairs per thousand vehicles produced
will have a trigger to initiate at least a review of high failure rate events.
Figure 14. Frequency ? Labor-Parts:
Region I-IV.
Part ? Labor: Region I. ?Good?
This is a target area where costs for
service meet customer expectations. Design-To goals can be established that focus on
part complexity in service
Part ? Labor: Region II or III ?Alert?
Low High
Hi
gh II. Alert
(labor cost) IV. Bad
Lo
w I. Good III. Alert
(part cost)
Labor - Part Costs - 2 x 2 Box Analysis
Part Cost
La
bo
rC
ost
66
High Part cost is fixed as a service part in or out of warranty and much higher cost
than an OEM part for production. Actions are commercial in nature. Pressures under
monopoly pricing and a desire to make significant profit in retail drives this cost element.
At times, there is a reduction in price but it is unusual.
If labor cost is high, this can equate to time or labor hours per repair rather than
labor rate. Design solutions are available but seldom pursued.
Part ? Labor:. Region IV ?Bad?
If a service part cost is high and labor cost is high then this is a design risk
component if frequency is high. Corrective measures are seldom taken as architecture is
fixed. The top service items by frequency can be benchmarked.
Figure 15. Mean and Standard Deviation and Warranty Index: Region I-IV.
The evaluation of the relation of the
mean and standard deviation of a sample
is explored in terms of the reciprocal of
the CV. Its utility is asking if the number
is greater than one or less than one or
near one.
Low High
Hi
gh II
1/CV > 1
IV
1/CV ~ 1
Lo
w I
1/CV ~ 1
III
1/CV < 1
If the standard deviation relative to the mean of either
Part Cost or Labor Cost of a given repair type, then it is a
candidate for reduction in variation.
Mean / Standard Deviation
2x2 Box Analysis - 1/ CV
Sta
nd
ar
d
De
via
tio
n
Mean
67
Once the statistics for Part Cost and
Labor Cost for a give repair are
understood, then the application can be
recommended for use in the decision
algorithm where a warranty index greater
than one will favor changing the part
cost. The first thing to consider is if the
generic goal of reducing part costs by the
3rd quartile will satisfy the goal.
In this chart, all the values of the
warranty index are expressed
qualitatively.
Low High
Part Alert
WI < 1
1. Warranty Index - showing preference
to Part Cost Reduction if WI > 1 and Labor Cost
Reduction if WI < 1.
Warranty Index - 2 x 2 Box Analysis, 1
Part Cost
Low High
Hi
gh II(L > P)
WI < 1
IV
(High High)
( l ~ P)
Lo
w I(Low , Low)
(L ~ P)
III
(L < P)
WI > 1
Warranty Index P/L - 2 x 2 Box Analysis, 2
Part Cost
La
bo
rC
ost
If the Warranty Index, the ratio of Part to Labor cost is
less than one, then the
68
Figure 16. Frequency MTBF ? Part - Labor: Region I-IV.
These three charts show the relationship of the
warranty index and the interpretation as an
increase in frequency impacts the decision
process.
This chart is the lowest failure rate or highest
Reliability and given a corporate strategy on the
target setting, this chart is indicating that the low
likely not be put on a project list, but on the No-
Project list if low part cost and low labor cost and
lower failure rate. The do not touch list is a
powerful tool to resist change for the sake of
change.
Here the assessment of poor failure rate puts us
into a need for change in service if there is no
Reliability project to fix this situation.
Frequency Trigger ( Reliability, Life-Limit, Accident)
Frequency Low (High MTBF)
Low High
High II. AlertL < P : WI > 1 IV. BadL=P : WI=1
Low I. GoodL=P : WI=1 III. AlertL > P : W I < 1
Low Frequency - 2x2 Box Analysis
Labor Cost
Pa
rt
Co
st
Frequency Trigger ( Reliability, Life-Limit, Accident)
Frequency High (Low MTBF)
Low High
High II. Very Bad IV. Very Very Bad
Low I. Bad III. Very Bad
High Frequency - 2x2 Box Analysis
Labor Cost
Pa
rtC
os
t
69
Binary Decision Logic, Frequency Lists, & Event Landing
The use of binary decision logic is founded in the RCM process discussed prior.
The maintenance effectiveness strategy replaces the mission critical view found in the
RCM flight ? service tool. One omission in this study is the attention in a design Failure
Modes Analysis or DFMEA. The ratings collectively or individually can relate the
factors of Severity (safety or economic) alone, or Severity times Occurrence (failure rate
related) or all three such risk identifiers used in the DFMEA as a product of Severity,
Occurrence and Detection (raking with or with out warning or monitoring to the
triggering event).
Figure 17. Binary Decision Logic Model ? Criteria Defined.
Failure Event
1.1 Failure
without
warning
1.2 Failure with
warning
1.3 Condition
Monitored
1.4 Preventive
Maintenance
(Scheduled)
1.1.1 Stranded
1.1.2 Severly
Degraded
Performance
1.1.3 Limp
home mode
1.2.1 Time
Effect
1.2.2 Warning
Effect
1.2.1.1 Time
Sufficient to
seek service
1.2.1.2 Time
insufficient to
seek service
1.2.2.1 Dash
Indicator
Monitored
1.2.2.2 Visual
or Audible
Non-Monitored
1.3.1 Service
Needed
Immediately
1.3.2 User
Servicable
Soon
1.3.3Seek
Service Soon
1.3.2.1 Driver
Action
Restores
1.3.2.2 Driver
Services -
Adds fluids
70
These are techniques that rank failures into various risk associations and are
familiar tools in the automotive industry managing design and failure mode avoidance
management. These categories are treated without safety called out in the Event Trigger
Topology and Driver Interface Leading to Service classifications.
When one has developed the BDLM with clear functional and operational
characteristics, as shown in Figure 17 ?Binary Decision Logic Model ? Criteria Defined?
then the concept of designing a fault tolerance could be made that might be called ?The
Event Landing.? The description of the event and the frequency that generates a related
service action would then influence the design strategy. The design team when
considering a candidate change recommendation can decide which branch to Design to in
the topology that so service is arrived at by design in a certain manner. For instance, the
designer specifies that he would rather fail with warning and specify that there will be an
indicator on the dash (element 1.2.2.1).
The service engineer would then write the ?service landing? mode into the service
manuals for the vehicle and then Reliability engineer would evaluate the design for the
frequency status and would evaluate what rate of failure is to the standard and if
maintenance requirements need to be evaluated in a Life-Cycle manner. If acceptable
frequencies then the job is done, if not then the cost engineer or the Maintainability
engineer would then address the two cost elements in the design for service.
The design in our case is interrogated for 1) Reliability, 2) Life-Limit, and 3)
Accident threat wherein the team is asked to evaluate if a measure of frequency is
sufficiently low occurrence that there is little need to improve that design element. If
insufficient a rate of occurrence that an improvement is recommended, the team is given
71
a priority order as to first ask to improve reliability, then life-limit. If the design is still
insufficient or not challenging the reliability then to do to consider the economic issues of
cost effective maintenance through a logic transfer to the maintainability interrogation.
If the conclusion is that the interrogation is sufficiently reliable, sufficiently not
prone to accident replacement and sufficiently not life-limited wearout and the cost to
repair is sufficiently low(not a high risk on its own no matter what the frequency), then
the assessment is done. The program receives a list of items from the assessment that
essentially directs the traffic through the design requirements definition. The list of times
that are historically non-offensive go on a do not touch or no action list.
Figure 18. Event Trigger Topology and Driver Interface Leading to Service.
1.0 Relia
Is the component sufficiently
Reliable?
1.1 (Y)
Is the item a high frequency
accident item?
1.2 (N)
Can Current Reliability be
improved?
1.1.1 (Y)
Accident Frequency List
1.2.1 (Y)
Reliability Improved
(LL, HR)
Go to 1.1, Accident Frequncy
List Branch
1.2.2 (N)
Can the Architecture be
changed?
1.1.2 (N)
Is the Item a Limited Life
Component?
1.2.2.1 (Y)
Change Architecture
(LL, HR)
1.2.2.2 (N)
Low Rel Frequency List
Go to 2.0 Maint
(Low Rel FL)
1.1.2.1 (Y)
Life Limit Frequency List
1.1.2.1 (N)
Done
Goto 1.1 Accident Frequency
List Branch
go to 2.0 Maint
(Life Limit FL)
Go to 2.0 Maint
(Accident FL)
M
A
A
Part (i) - (n) for
analysis --->
M
A
72
Binary Decision Logic, Maintainability Decision
During this interrogation, the logic model checks effectiveness of the service cost
and if effective and sufficient then the component is done, no further action is needed. If
not, the maintenance index is used to identify the best opportunity from a cost perspective
what should be evaluated. In both Parts I and II architecture of the design is challenged.
In each case there is a feedback loop to re-evaluate if needed. Also, the part pricing is
challenged if part price is high. This is novel in the industry today. As a practical matter,
there would be a need to develop a maintainability skill set in the OEM product
development team.
Figure 19. Binary Decision Logic Model, Maintainability Branch.
2..0 Maint
Is Total Cost to Service
Acceptable?
2.1 (Y)
Done
2.2 (N)
Is the Maint Index < 1
(L>P)
2.2.1 (Y)
Can Maint Time be
reduced?
2.2.1.1 (Y)
Improve Maint, (L)
2.2.1.2 (N)
Can Basic Design
be improved?
2.1.2 (N)
Can Part Price be
reduced?
2.2.2.1 (Y)
Improve Maint (P)
2.2.1.2 (N)
Can Basic Design
be improved?
Go to 2.0 Maint
(re-evaluate)
Go to 2.0 Maint
(re-evaluate)
M
Warranty Index:
(Part Cost / Labor Cost)
B
M
B
73
Decision Rule ? Example Candidate for Maintainability
Figure 20. Design Rule Applied to Frequency Listings.
item Repair Part Service Summary Reliability Life Limit Accident Comment -Accessibility Part Labor WarrantyIndex
1 Front Pads Disc pads - front remove and install -remove and replace both sides X Wear - mustremove tire Low Low =
2 Rear Pads/Shoes Disc pads/Drum Shoes - rearRemove and Replace, both sides X Wear - mustremove tire Low Med < 1
3 Front Rotor Front Rotor Remove, Turn orReplace, Both Sides X X
Warp - might
have secondary
damage - pads
High Med > 1
22 Windshield Remove and Replace Windshield X Threat ForeignObject Damage High Med > 1
Frequency List Items
Note: If exact costs are not known in terms of Part Costs, P(i) and Labor Costs, L(i) their
relative values may direct effort through the Warranty indexing noted above.
In this example, a listing of the frequency driven items become consideration for
maintainability actions since they are high impact from Part I of the algorithm for Binary
Decision using the ?Warranty Index? Method. The three triggers for being on a
frequency list is given as reliability, Life-limit and Accident initiated. Event description
of the service needed for part and labor are provided. An opinion as to the type of
maintainability strategy price strategy is offered. The status of the index is given as a
ratio of Part Cost / Labor Cost in relative terms as in the quadrant analysis. More than
one trigger could exist and be treated separately for causal impact.
Data Dispersion Part to Labor
The analysis process that helps to create the frequency lists are displayed in a set
of five graphs. The detail of the data query selects data based on the filtering discussed
and the level of symptom codes - up to four levels deep ? referred to as the (ijkl)
granularity. One level can be interchanged (k and l) to reveal more hardware or symptom
74
resolution into the data set. Only the (ijk) approach is used. The object is to classify a
certain repair data and its variance from historical data and assess the rate of occurrence
to be used in the target setting process for quality improvement for the Life-Cycle. The
key parameters include Reliability and Service.
Reliability looks at targets for life limited components and basic reliability or
failure rate while service considers part cost and labor times (or costs). The five graphs
represent a proposed data standard that can help identify visually how the frequency
factory is working and the variation of part to labor costs look over time or mileage.
From an introduction to the data on the whole vehicle, a manager can identify the
overall relationship or cost of repair for each repair and can be considered a serviceability
standard in the historical record of the vehicle. Each design change can ask how this
might impact the vehicle as costing more or less with respect to cost of repair over time
in and out of warranty.
This value can be compared to other vehicles and selected repair inquiry by the
responsible design community to determine existing best practices ? ones that cost the
least in warranty spending and lowest cost of ownership in the Life-Cycle.
Here, graph 1, there is a rate of usage or miles per day. The graph uses miles driven on
the Y-axis and days in service on the X-axis.
75
Figure 21. Key Graphs Explained ? Full Vehicle.
Y-axis: Mileage Data Ref: Full Vehicle
X-axis: Days in Service
Purpose: Application gross duty cycle in miles per day.
Approximately 60 miles per day. A Life-Cycle estimate
of oil changes in a population can be significant costs in
oil and dollars by increasing the interval.
Y-axis: Labor Costs
X-axis: Mileage
Purpose: Understand the labor costs over time. Service
has more variability early in a vehicles service period.
Learning may explain reduction over time. Care not to
have unequal populations per interval with respect to
density.
Y-axis: Part Costs
X-axis: Labor Costs
Purpose: Data that demonstrates the density of the
average repair is focused in a cluster. There are flyers
that have broken away from the pack and represent
possible need to understand why.
Y-axis: Total Cost of Repair (per event)
X-axis: Labor Costs
Purpose: A relation between labor and total cost of
repair. Here there is an estimate that total cost of a
repair is 1.4 times the labor content plus ~ $16. This
could be used as a cost estimating relationship in
planning. It also reveals that there is never a part cost
without a labor cost.
Y-axis: Total Cost of Repair (per event)
X-axis: Part Costs
Purpose: A relation between part and total cost of
repair. Here there is an estimate that total cost of a
repair is 1.4 times the part content plus ~$70. dollars.
This could be used as a cost estimating relationship in
planning. It also reveals that there may be a part costs
can be zero indicating a soft fix issue exists on new
vehicles.
76
Figure 22. Key Graphs Explained, Failure Code: i1,j1,k1.
Y-axis: Mileage Data Ref: (i1,j1,k1)
X-axis: Days in Service
Finding: Frequent occurrence with many days in service
but low miles accumulated at first.
Y-axis: Labor Costs
X-axis: Mileage
Finding: There is a spike at first with low mileage and
very excessive labor costs. Much lower labor costs with
learning. What lessons were learned and how were the
problems early solved.
Y-axis: Part Costs
X-axis: Labor Costs
Finding: Data clustering in part and labor near the
origin, meaning well behaved repair in general.
However, Significant higher Part variable needs to be
understood.
Y-axis: Total Cost of Repair (per event)
X-axis: Labor Costs
Finding: A relation between labor and total cost of
repair. Here there is an estimate that total cost of a
repair is 1.4 times the labor content plus ~ $16. This
could be used as a cost estimating relationship in
planning. It also reveals that there is never a part cost
without a labor cost.
Y-axis: Total Cost of Repair (per event)
X-axis: Part Costs
Finding: The relation between part and total cost of
repair indicates there is significant part variability as
compared to labor.
77
Figure 23. Key Graphs Explained ? Failure Code i1,j1,k2
Y-axis: Mileage Data Ref: (i1,j1,k2)
X-axis: Days in Service
Finding: Wide spread in the data. Likely to occur at
any mileage or days in service, favors neither cycle
domain.
Y-axis: Labor Costs
X-axis: Mileage
Finding: Weak association labor to mileage with respect
to lowering of labor over time.
Y-axis: Part Costs
X-axis: Labor Costs
Finding: Highly correlated groupings. 5 discrete part
cost levels identified. Higher costs need to be
investigated. A lot of labor cost variability within
groups. Possible faulty diagnosis that has systematic
higher part costs if find that they are all the same issue.
Y-axis: Total Cost of Repair (per event)
X-axis: Labor Costs
Finding: no findings with respect to labor, no
stratification in the data observed. Labor is about the
same cost but shows higher part costs dominate as the
component of TRC less Labor.
Y-axis: Total Cost of Repair (per event)
X-axis: Part Costs
Finding: Clearly actionable data in part costs.
78
Figure 24. Excel ? Pivot Table that Generates that Category Statistics.
Pivot Table - Showing F1 and
G1 level 1 & 2 Symptom
Codes, Individual Repair
Events, and Summary with
Mean and Std dev
BOD_CAB_STL (All)
LEVEL_1 F1
LEVEL_2 VIN_CD Data Total
G1 1FBNE31L06HA15985 ODOMETER 1424.00
DAYS IN SVCE 14.00
1. LABOR_COST 21.60
2. PART_COST 15.28
WARRANTY_INDEX 0.71
1FBNE31L06HB19067 ODOMETER 4504.00
DAYS IN SVCE 147.00
1. LABOR_COST 75.82
2. PART_COST 4.50
WARRANTY_INDEX 0.06
1FBNE31L06HB31171 ODOMETER 11239.00
DAYS IN SVCE 74.00
1. LABOR_COST 75.33
2. PART_COST 27.74
WARRANTY_INDEX 0.37
G1 Average of MILEAGE 6072.03
G1 Average of DAYS_IN_SVCE 102.75
G1 Average of LABOR_COST 135.59
G1 Average of PART_COST 113.78
G1 Average of WAR_IDX 1.04
G1 StdDev of MILEAGE 6774.51
G1 StdDev of DAYS_IN_SVCE 80.44
G1 StdDev of LABOR_COST 269.65
G1 StdDev of PART_COST 186.02
G1 StdDev of WAR_IDX 1.36
79
Building Warrant Index Design Guidelines
Figure 25. Warranty Index Design Guidelines.
Figure 26. Assessment of Factor Behaviors, Mean and Standard Deviation.
Study Data
Vehicle
Line
Level 1
Symptom Code
Level 2
Symptom Code
Part
Average
Part
Standard Dev Behavior
Labor
Average
Labor
Standard
Dev
Behavior Action Direction
All F1 G1 113 186 1.65 135 269 1.99 Labor: Bad-Needs attention immediately
Part: Needs attention immediately
All F1 G2 85 48 0.56 59 48 0.81 Labor: Needs attention soon/watch
Part: Good
All F2 G3 111 217 1.95 134 99 0.74 Labor: Good
Part: Needs attention immediately
All F2 G4 94 182 1.94 139 97 0.70 Labor: GoodPart: Needs attention immediately
Mean Relation Std Dev
Indexed
Mean/Std Dev
1/(CV)
Comment
Mean (i) >> Std Dev (i) <<1 Very Gooddefinition
Mean (i) > Std Dev (i) <1 Gooddefinition
Mean (i) = Std Dev (i) 1 Poor definition
Mean (i) < Std Dev (i) >1 Bad definition,Variability
Mean (i) << Std Dev (i) >>1 Very Bad,Variabilility
80
Figure 27. Generating Targets for Design-To Life-Cycle, Cost Category Variability.
In the above graph there are four regions defined. They are used to prescribe to
the program teams what the improvement targets are in terms of part and labor costs. By
identifying variation in the warranty data system, and making an assessment concerning
the data scatter for the given repair level for the Part and Labor components then one can
set upper limits on the data as an improvement challenge.
By a detailed descriptive statistical study as found in this failure code [ F2G3]
(refer to Figure 28 and 29 respectively), one can set the new target at the 3rd quartile limit
? as a design rule. This has been selected as the targeting threshold for the cost factors in
service. Zone II is the part cost ?threshold? with Zone III being the labor ?threshold? and
collectively in Zone IV, the must fix cost area. Program teams are tasked to get the costs
under control in this manner.
81
Figure 28. Target Modeling, Repair Code: P_F2G4, 3rd quartile Target.
Figure 29. Target Modeling, Repair Code: L_F2G3, 3rd quartile Target.
82
Recall that the warranty system is primarily a reimbursement system administered
through a dealer network to honor the warranty contract with its sale of a vehicle to a
customer. The warranty system and network of dealers is like an insurance policy and of
course is subject to abuse. Therefore many checks and balances are applied to assure
integrity for the most part reveals. A byproduct of this system is the joint opportunity to
take system of failure information that initiates the dealer visit under warranty and the
associated information that is captured to correct the subject defect and those repair
actions that restore vehicle function.
Understanding a Single Data Point
Figure 30. Data Point of a specific repair cost elements, Part and Labor.
Format [ 4 ] - Labor vs Part Costs
0.00
200.00
400.00
600.00
800.00
1,000.00
1,200.00
0 500 1000 1500 2000 2500 3000
Part (or Material) Costs
La
bo
rC
os
ts
Each data point is a repair event representing
constituent cost factors.
Total$ = ( x(Part$), y(Labor$)) ~ ( $825, $381) = $1,206
Where ( x + y ) = Total Repair Costs
This data is based on a selection of a certain functional code and a
certain vehicle. Permanent filters are on, excluding specific data.
83
Data that shows warranty claims from a Pre-Service or sale date is excluded from the
study. The specific case refers to the number of repair or claims events prior to the
introduction into service. This occurs when a vehicle goes to a third party for a major
modification as in a performance vehicle. This is an interesting area to study the effects
of such delay and function while under warranty rules in the pre-sale condition, as
vehicle would not start, battery too low ? battery replaced. It does not seem reasonable to
have these warranty claims under this usage.
Figure 31. Filtering Rationale, no presale data used for target setting.
From this, one can learn through detailed analysis and physical review of part
returns the root cause of most failures. This data and frequency data come together to
prioritize ?find and fix? approaches to quality improvement and warranty improvement
programs. Year on year the domestic automotive companies have gained ground on
quality and are now reported to be equal to the imports through these efforts. There are
two related lists I have gathered from my experience showing ten aspects of design
quality and quality measurement that have driven this improvement and a brief review is
MILEAGE vs Days in Service
y = 34.434x
R2 = 0.25
-10000
-5000
0
5000
10000
15000
20000
25000
30000
-400.00 -200.00 0.00 200.00 400.00
Days Warr in Service
Mi
lea
ge MILGE
Linear (MILGE)
MILEAGE vs Days in Service
y = -1.2473x
R2 = -0.0508
0
1000
2000
3000
4000
5000
6000
7000
-250.00 -200.00 -150.00 -100.00 -50.00 0.00
Days Warr in Service
Mi
lea
ge MILGE
Linear (MILGE)
84
needed to understand the opportunity uncovered by the use of the ?Warranty Index? and
decision process to improve the quality, reduce the warranty and improve the Life-Cycle
ownership experience. In each of the two lists provided the last or tenth item is related to
the study and has little depth of development in practice.
Research into Reliability Centered Maintenance, RCM practices can be developed
for integration into the design process. Many measures and data elements are shown in
this section and the key ones are defined when used throughout the study and within the
glossary at the end of the report. The ?Warranty Index? methodology serves as a tool to
identify and explore the next level of innovation in improved quality of service and
Ebling (1997) indicates a list of program planning elements to incorporate into a
Maintainability Design Action List (p.225) and Maintenance Time Categories List
(p.190), Refer to the Figure below listing such activities included in design maintenance
action types for improving maintainability and tracking time to repair elements of a given
repair process.
Figure 32. Maintainability Design Actions List and Maintenance Time Categories.
In work by Stadler and Brideau (2004), human factors are discussed that form part
of the skill set needed for maintainability engineers and service engineers. If the
automobile companies have not formalized this process it would be important to develop
this competency. Similarly, Woolford (1995) develops a full strategy for design for
Design Maintainability Actions Maintenance Time Categories
Fault Isolation & Self Diagnosis Supply delay
Parts Standardization Maintenance delay
Modularization Access
Accessibility Diagnosis
Repair vs. Replacement Replacement / Repair
Proactive Maintenance Preventive Verification
Proactive Maintenance Predictive
85
maintainability. These types of requirements are developed by O?Conner, (2005) where
he has related a program?s needs in terms of serviceability (p. xxx). Again, this goes to
developing competency in the design for service arena.
Ten Elements of Quality in Vehicle Design
Table 2. Ten Elements of Quality in Vehicle Design.
1. Design Considerations ? customer based
a. Mission
b. Duty Cycle
c. Expectations
d. Satisfaction
2. Business Considerations
a. Goals
b. Targets
c. Initiatives
d. Communication
e. Contracts
3. Reliability
a. FMEA
b. 150,000 miles
c. 10 year life
4. Durability
a. Loading Studies
b. Durability Testing
5. Safety
a. Regulation
b. Due Care
6. Survivability
a. Fit for use
b. De-rating
7. Design Philosophy
a. Sustainability
b. Useful Life
8. Design Standards
a. Warranty Agreement
b. Warranty Standards
c. Test Standards
9. Assembly Standards ? Design for Assembly
10. Design for Serviceability
a. Guide Lines
i. Common Tools
ii. Access
iii. Modularity
iv. Maintenance Concept, RCM
a. Depot
b. Flight Line
c. In-Flight
86
Ten Related Quality Measures
Table 3. Ten Related Quality Measures.
1. Rs/1000
2. Months In Service
a. High Time In Service
b. Low Tim In Service
3. Cost Per Unit
a. Factors of Labor
b. Factors of Material
4. Cost per Repair
a. Statistics
b. Car Lines
5. Scheduled Maintenance
6. Unscheduled Maintenance
a. Normal Usage
b. Accident
7. Warranty - Parts & Labor
a. In Warranty
b. Out of warranty
c. Extended Warranty
8. Data Collected
a. FCPS
b. Consumer Reports ? Annual Survey
c. JD Powers IQS/CSI ? TGWs/PPH
d. GQRS - Things gone wrong 1MIS?, 3MIS, 12 MIS, 36MIS
e. Warranty
f. Extended Warranty
9. Dealer Service
10. Service Actions
a. Labor Rate
b. Labor Hours ? Empirical
c. Labor Hours - Design
d. Top Problem lists
e. Service Center
f. Supplies Training
g. Manual ? in vehicle - Tools
h. Corporate Centers ? Trouble Shoot Typical Wear Items
i. Top 100
i. Characteristics
ii. Goals ? %improvement YOY
j. Benchmarking
i. Internal
ii. External
87
Brief Review of Reliability, Maintainability and Availability
Maintainability is often grouped within the broader subject of engineering called
design assurance. To capture the related tasks within design assurance it can be referred
to RAMS which stands for ?Reliability, Availability, Maintainability, and Safety or
Supportability.? One article on the subject of RAMS presented by Beugel-Kress, (2006),
refers to a table out of the NASA standard [[8729.1]] that identifies several formulas for a
function related to an equipment?s ?availability? - - a relation between uptime and
downtime. Within the variations of the Availability function are the progressive elements
of downtime classifications that contribute to the total downtime of equipment.
Availability = Uptime /(Uptime + Downtime)
The definition of Downtime is further classified into constituent elements that
give rise to variations of Availability, shown in the table below, with the Mean Time
Between Failures, MTBF as the rate at which unscheduled events are expected to occur
in usage.
Table 4. Availability Classifications.
Uptime
Availability MTBF CorrectiveMaintenance PreventiveMaintenance AdministrativeDowntime LogisticsDowntime
Inherent X X
Achieved X X X
Operational X X X X
Total X X X X X
Downtime Classifications
Table 5. Formula for Availability (total).
MTBF + CorrectiveMaintenance + PreventiveMaintenance + AdministrativeDowntime + LogisticsDowntime
Availability (total) =
MTBF
88
What is important to the design of a system is to know that these parameters may
be understood in terms of the as built consequences of specific design practices. This is
pointed out by Ebling (1997), in that ?Availability Inherent is shown to be based on the
failure and repair time distributions and therefore can be viewed as an equipment design
parameter? (p. 257).
Table 6. Formula for Availability (Inherent).
A(inh) =
MTBF
MTBF + MTTR
It is possible to set up design goals and their appropriate validation in
probabilistic terms. If the design team has achieved its goal for MTTR, a demonstration
would confirm that with a 1-alpha confidence that a certain repair (time to repair, ttri) and
(time to failure, ttfi) would be equal to or better than the goal, given by:
?? ??
??
?
??
? ???
? 1/Pr 1,n
avr ttr
ns
MTTRttr
And similarly for illustration only similarly for MTBF, the designers might
demonstrate a given Reliability using ttfi by:
?? ??
??
?
??
? ???
? 1/Pr 1,n
avr ttf
ns
MTBFttf
89
Note:
? ?1, ?nttf? comes from a probability density defined by the student-t
distribution with parameters ? and (n-1) with n, being the number of observations or
repair / failure events of the ith failure type. A comparative study on variations of the
Availability formula will define a similar breakdown of the time factors as may be
appropriate for management and control of the time it takes to restore equipment to
service. This is the time needed to meet objectives of a system?s intended effectiveness
through adequate design and service requirements. From a retail automobile owner?s
point of view, these time elements are indistinguishable under maintenance on a vehicle
covered by warranty. However, what is experienced by the vehicle owner is how long
transportation is unavailable and in what manner it became unavailable which can affect
owner experience and satisfaction.
Being stranded with road-side service, with rental service or with a free loaner are
conditions of purchased warranty, insurance, or dealer policy that can ease dissatisfaction
of being denied use when the vehicle requires scheduled or unscheduled service.
Generating High Leverage Frequency Lists
To explore the vehicle ownership point of view and service experience further a
traditional event driven model can be constructed into four downtime initiations:
90
Table 7. High Frequency Lists from Service Event Driven Effects.
1. Accident or failure without warning
a. Effect essential to transportation
i. Stranded, requires towing / repair at event site.
ii. Severely degraded performance, may require towing.
iii. Limp home, able to continue operation with slightly
degraded performance, repair needed of convenience.
b. Feature lost or attributes fail but not essential to transportation.
2. Failure or condition yielding a warning or cue to the driver
a. Timing Effect
i. Time sufficient to seek service
ii. Time insufficient to seek service
b. Warning Type
i. Dash Indicator (typical ? manual will instruct driver as to
meaning)
1. Information (blue or white)
2. Warning (yellow or amber)
3. Serious (red)
ii. Visual or audible cues prior to failure ? Brake Pads worn ?
whistle / grinding or even observed symptom.
3. Condition, Monitored, Threshold Monitored
a. User Serviceable
i. Cycle the key to clear codes and restore function
ii. Add Fluids Oils, Water, Washer, Fuel, air,
b. Seek Service Soon ? Amber light
c. Seek Service Immediately ? Red light
4. Preventive Maintenance, a recommended maintenance manual scheduled
on time or cycles.
With respect to item 3, Condition Monitor ? high or low conditions of
temperature, pressure, time, vibration, cycles and actionable service condition that would
precede a fault ? even the popular On-Star service found on General Motor?s models /
employ condition monitoring to enhance ownership experience. This is related to
reliability and the degree to which an operational diagnosis and a prescriptive solutions
can foresee avoiding a more serious downtime consequence allowing the user to get his
vehicle repaired effectively in a more event driven but service friendly way ? see
91
Vacante, (2001). Reliability Centered Maintenance or RCM is used to guide design of
systems through a support strategy typically related to fleet operations and maintenance
level effectiveness.
Take the example of evaluating and establishing design strategy that considers
failure effects or life-limited wear-out event as it is a common practice to lay against a
population failure distributional effects and how that product knowledge might be
managed. Set ?Design-To? objectives and targets can then be evaluated early in a design
by a method of interrogation of features and characteristics with the purpose to maximize
uptime and hence customer satisfaction or mission effectiveness.
Figure 33. Weibull Analysis of Fatigue Data.
Analyzing data provided by Johnson, (1964) provides one example that shows a
part that has failed due to high cycle fatigue that results in a Weibull shape (or slope)
parameter of 1.96. The slope/shape factor greater than one, as evaluated in MiniTab?
92
12, a statistical software package, depicts an wearout or life limiting failure rate with
additional usage. This is considered an increasing failure rate over cycles of usage. If the
data showed that 10 percent correlated to 100,000 miles then the likelihood of failure in
service would be past the warranty level of 5 years or 50,000 miles. The failure would
likely occur out of warranty and be born by the customer. If it were an axel it might have
a high price and in comparison a lower labor content to fix and seem fair as to cost.
However, the effect of an axel failure could have serious satisfaction issues as it is during
a time that the customer may be expecting to have it last the duration of his Life-Cycle.
The life-limit at a 10 percent life might be interpreted to say it needs to have scheduled
maintenance to replace the item and would be in the customer?s cost of ownership
window.
In the ?Warranty Index? method, the interrogation would consider improving the
reliability by lengthening the life-limit outside of the sustainability limit of the vehicle or
warranty challenge the basic architecture for a more fundamental change needed. Else, it
would be transferred to the maintainability element of the assessment and interrogate
effectiveness of a scheduled maintenance scenario (element 1.4 in the flow chart ref,
Figure 6) rather than a failure event like one with warning cue that can be heard by the
driver such as a ?clunk sound? or worse, failure without warning. Note ? a design failure
modes effects analysis or DFMEA risk assessment might capture the undesired effect.
To say a component has a life-limit within the window of Life-Cycle ownership then it
may need to be improved as in the fatigue life or make the component more easily
replaced with respect to the failure.
93
If the interrogation could lead to a review of design attributes of system
architecture - then the system might be able to respond to foundational elements that most
products are built upon and are hardest to change in a design cycle. As in the automotive
field, the vehicle cycle plan may be conceived many years before a derivative vehicle
even gains program status in a development stage relying on past structural design
elements and surrogate studies of related systems. Therefore, mitigation of this type of
design element is not realizable in the standard product development cycle practiced by
the OEM and was similarly proposed by Bryan, et al., (1992), but without follow through
within their computer model using an approach as an expert system design approach to
serviceability.
Examples of this are found in platform choice to build a top hat or upper portion
of a vehicle but commonizing on the chassis components of another exiting vehicle
structure that may be too ridged to give a smoother ride desired.
This lack of flexibility carries over into many design elements of the structure. A
two-year-to-market needed for a new model builds on known architectures and surrogate
data that cannot easily adapt to a desired concept of supportability or serviceability. The
consequence expressed in time-to-market is an added delay if one is to improve or fix
vehicles with poor serviceability. The consequence of not being able to fix these design
requirements will be a continuation of poor quality, warranty losses and customer
dissatisfaction and likely loss of market share.
Fault Avoidance and Fault Tolerance in Design
94
To the above events the treatment in design may consider fault avoidance and
fault tolerance to avoid a service event. However, if a component or system is prone to
join the low reliability or high frequency of failure classification proposed for use in the
binary decision algorithm then the questioning ?if reliability can be improved?? is
answered by designing systems with improved Fault Tolerance and application of Fault
Avoidance. Each method is distinct in approach to increase reliability of systems to
overcome challenges or loads that may be greater than strength or capability of a vehicle
in the intended using environment as reported by Souza et al., (2005).
What is important in this discussion is not so much what one can do to improve
reliability of a component or a system but having systems and design rules that will
adequately address systems that will not have improvements for these weaknesses that
have historically failed and are therefore candidates for maintenance and strategies of
maintainability for individual or fleet owners of automotive vehicles. How this is
managed and designed into a vehicle?s architecture and will provide a sense of durability
and satisfaction when such failures occur or in the contrary levels of dissatisfaction and
loss.
A proper design strategy can then take the frequency of occurrence knowledge
and the cost of repair (being a function of labor rate, labor time, part costs, and any
related other costs) over a warranty period and any out-of-warranty period into the cost of
ownership experience or company risk exposure and have a process to minimize this
exposure.
95
If we conclude that the characteristics of Reliability and Maintainability are
characteristics of design then one can encompass a broad range of design rules to
overcome the notion that maintainability is not a factor in automotive design.
While this paper will treat the objectivity within a maintainability program plan for
automotive companies it is sufficient to list the various tools to apply to improve or rather
minimize the costs of a repair factor in or out of warranty. Consider the following tools
also cited by Souza et al., (2005).
? Redundancy ? physical or software; Standby or active.
? Coverage ? the property of a system to tolerate failures.
? Diagnosis through ? Built in Test and Evaluation, or BITE;
? Failure Detection and Isolation which can be used to activate tolerance features as
well as be used by vehicle activities such as in ON-Star.
Clearly, if a failure over time is characterized by increasing increments of time the
risk of failure increases. This aging aspect of a design is found in moving components or
those that age or degrade with integrity over time and to this end, our modeling for
application in our decision logic is to add items like this to the high frequency of failure
list but specifically called them life-limited frequency list.
If one wants to improve the reliability, the approach is to add strength to increase
fatigue life for example to allow a longer cycle life in comparison to the cycles expected
within some safety factor. However, things like tire wear are known and accepted life-
limits. These items become maintenance interval items with replacement
96
recommendation to allow maximum performance until they should be discarded or
serviced in some way to restore specified performance criteria.
If one considers an engine and its useful life one also considers the lubricant and
proper fuels and applications of coolant and its maintenance or change rules to assure
optimum performance per a preventive schedule. The obvious lengthening of service life
on these items or making an engine more robust to in the presence of noise (dirty oil) the
longer life and sense of durability will.
Assessments commonly done that reveal the condition of parts to act in a certain
manner when failed or failing is covered in analysis methods such as:
Design Failure Modes and Effects Analysis ? a single failure effect understood to be
caused by any one of possible several single mechanisms of failure.
Fault Tree Analysis ? able to identify multiple concurrent effects of failures to lead to
certain failures not identified in a DFMEA ? e.g. chance of unintended ignition ( spark
and fuel rich air).
Reliability Block Diagram - to consider budgeting of failure rates and interfaces and
location within a structured system with other methods to estimate and or identify
concerns of a reliability performance nature and how to enumerate and validate a
prediction for a part or system.
In each case one applies strategies to mitigate the occurrence or changes Design
to even minimize the effect. Clearly stated is the finding that fault tolerance methods are
97
applicable to automotive applications due to the complexity to have many micro
processor driven systems in a vehicle.
Even systems such as antilock brakes interpret data that is smoothed in averaging
while trying to interpret a signal is representing a failed state or just spurious noise yet
being able to respond quick enough to offer stability control responses in real time.
These design practices prevent a user from seeking service under these less than exact
sensing systems where codes set and recover with a key cycle that may have
intermittently indicated a fault ? this would be the application of fault tolerance.
The ?Warranty Index? Interrogates the Design
The proposed ?Warranty Index? provides a way to allow an interrogation of the
design through a binary decision algorithm and would require a part or work breakdown
structure assessment that is respecting architecture and function. The proposed
interrogation process asks in a logical and hierarchical way to determine if an outcome
referred to as a dependability (D) factor is implemented in a design ? which is the ?Green
Box? (low freq and cost) showing a design is satisfactory and is not on a high frequency
list. The designation of three frequency lists comprehensively differentiates design
challenges ? one for accident caused another for life-limited components and one for low
Reliability. The challenges that are failure caused are considered a company risk when a
repair is covered under warranty and upon the customer when outside of warranty. The
risk due to failure of components due to accidents is significant to the customer especially
when not covered fully under collision or comprehensive. Accident causal items are
often overlooked beyond the industry standard that evaluates costs of repair benchmarks
98
for public review such as that given for 5 mile per hour crash test cost comparisons given
by Mayes and Wassilak, (2004).
A review of the literature by Klyatis and Klyatis, (2006), showed that there is no
time in the traditional product development cycle to solve architecture issues.
Additionally, this is clearly where data, design concept and true gains as maintenance
costs can be optimized. This is especially true concerning the maintained frequency lists
from older surrogates. Reliance on surrogate data is a weakness in establishing Design-
To-goals for service with no real-time data developed to assess future cycle plans for
serviceability. Even a review of service manuals are no longer used to assess design
standards for service actions. The assessments, if timely, would influence warranty
reduction by design and customer satisfaction thereby improved.
One leverage point to the OEM today is to improve warranty and then offset with
an incremental increase in warranty period for same dollar exposure. This ability to stand
behind the product reduces consumer risk and is likely to increase consumer buying
decisions and is a factor of Durability and longevity.
Additionally, service and support philosophies can be more actively defined that
will, by design, have a direct effect and an economic impact on the vehicle warranty,
ownership costs and experience.
To capture much of the design concepts in customer ? vehicle interface (reference
Figure 7) and what drives a service event and what may be included in a service manual.
Design efforts and applications of design analysis such as design failure modes and
99
effects analysis considers system effect from the customer point of view but rarely
service. Seldom will a design team add service objectives to their design goals and it is
rather a specialized team that will try to leverage limited resources to make a design more
serviceable with respect to known problems.
The binary algorithm approach with the ?Warranty Index? methodology considers
the frequency of occurrence of a service driven event model. Once the design has been
interrogated in this way then the maintainability interrogation takes place based on the
prioritization of the Warranty Index to prescribe the opportunity in service to improve
ownership costs and warranty expense.
Figure 34. Binary Algorithm for Event Frequency.
1.0 Relia
Is the component sufficiently
Reliable?
1.1 (Y)
Is the item a high frequency
accident item?
1.2 (N)
Can Current Reliability be
improved?
1.1.1 (Y)
Accident Frequency List
1.2.1 (Y)
Reliability Improved
(LL, HR)
Go to 1.1, Accident Frequncy
List Branch
1.2.2 (N)
Can the Architecture be
changed?
1.1.2 (N)
Is the Item a Limited Life
Component?
1.2.2.1 (Y)
Change Architecture
(LL, HR)
1.2.2.2 (N)
Low Rel Frequency List
Go to 2.0 Maint
(Low Rel FL)
1.1.2.1 (Y)
Life Limit Frequency List
1.1.2.1 (N)
Done
Goto 1.1 Accident Frequency
List Branch
go to 2.0 Maint
(Life Limit FL)
Go to 2.0 Maint
(Accident FL)
M
A
A
Part (i) - (n) for
analysis --->
M
A
100
Here the Part II ? Binary Algorithm Model with Warranty Index can interrogate
the design for service opportunity. If the total cost of service is low or acceptable, the
frequency is not sufficient to drive a change.
Figure 35. Binary Algorithm to Identify Application for Design for Serviceability.
In 2002, Ford Motor Co. published a manual entitled ?Designing for Improved
Serviceability ? NEO? Grzanowski and Long, (2002), and conspicuously the title
indicates that it needed improving serviceability in design and that is commendable. This
was a revamp from a 1990s version that included reliability in its title. The objectives
were to reduce Life Time Costs and improve customer satisfaction and charged the
design community with the responsibility. The concept of Life-Cycle cost was also
2..0 Maint
Is Total Cost to Service
Acceptable?
2.1 (Y)
Done
2.2 (N)
Is the Maint Index<1
(L>P)
2.2.1 (Y)
CanMaint Time be
reduced?
2.2.1.1 (Y)
Improve Maint, (L)
2.2.1.2 (N)
Can Basic Design
be improved?
2.1.2 (N)
CanPart Price be
reduced?
2.2.2.1 (Y)
Improve Maint (P)
2.2.1.2 (N)
Can Basic Design
be improved?
Go to 2.0 Maint
(re-evaluate)
Go to 2.0 Maint
(re-evaluate)
M
WarrantyIndex:
(Part Cost / Labor Cost)
B
M
B
101
introduced to focus on customer satisfaction rather than warranty cost reduction. The
enablers expressed in the training included an understanding of certain check lists to
evaluate any ?new design.?
The buying motives discussed showed a focus on customer?s needs and included
not just the sales experience but a list of service items:
? Quality repair work.
? Repair when promised.
? Quick service response.
Then the design interrogation for Serviceability would indicate a need to remove
labor out of the design. Then the question is asked by how much and what are the
obstacles ? it appears obvious but access would be one design challenge to consider.
Ergonomics might be another. Few can explain why these approaches in the design are
so deep but it goes to the interface analysis and the Design-To targets that have not been
communicated or benchmarked against best practices.
If one would think that Design for Assembly would cover this as an inverse
solution in reverse just consider building from start to finish in sequence to finding the
needle in the haystack and trying to break into that orderly building of a system. If one
built in modules and it could come at anytime into a series of logical building blocks with
few obstacles in the route to take. It is as if you would need a Garmin to do the job well
102
due to the lack of training or skill level that threatens self service. This automated
approach may be a good idea for future study to help interactively service of vehicles.
In summary there are 4 areas addressed:
? Scheduled Maintenance ? to develop the manuals.
? Cost of Ownership, a life time cost assessment that attempts to understand part
and labor costs with the goal or reducing those costs. Part costs are often
contractual and labor becomes the biggest opportunity.
? Damageability,
? Serviceability ? support in the field and parts availability.
103
Chapter Five
Conclusion and Findings
A summary of the implications of the study and a discussion of future research
show that the study objectives have been satisfied. In addition there was one surprise
finding not planned by the study. The study focus was an OEM in an automotive sector.
The study sought to:
? Identify a need for quality improvement
? Utilize existing data structures
? Develop a robust Binary Decision Logic Model to interrogate the data
? Operationalize the change process
? Challenge business process for cost of parts in service
? Establish a means to develop design rules for Design-To Life-Cycle
? Interrogate the logic model with a created index to guide the design
? Create a path to change vehicle architecture or basic design
? Exercise the methodology so it could be repeated by others.
? Identify the scope of the process.
These findings jointly support the objective to:
? Lower warranty costs and
? Lower costs of ownership
104
The surprise in the study was the confirmation that there is a technology gap or
lack of a skill set for the Design Assurance Field within product engineering. However,
this might be easily filed by available military program / commercial contracting firms.
With the increase in cost and warranty pressures, Design-To Life-Cycle Costs can
satisfy the need to find improved ways to reduce warranty costs and lower the cost of
ownership ? a period outside of warranty - by leveraging the same cost factors, Parts and
Labor in service.
Demonstrating the application of the Algorithm in the Binary Decision Logic
Model allows for future study to see how much warranty improvement is possible. The
model also introduced that a branch can take the form of a design glide path ? where a
new function should find its application ? e.g. branch 1.2.3 reference the Failure Event
Tree, Figure 17. Binary Decision Logic Model ? Criteria Defined, where the designer, in
response to a failure mode that has no warning can plan to limp home by design rather
than be stranded (e.g. current state of a design). The designer selects a branch for the
highest degree of safety and customer satisfaction possible.
The approach to identify variability in a given repair code effectively targets the
design for improvement. Also, identifying the 3rd quartile upper limit as a threshold ?find
and fix? is one approach to task a team for improvement as shown in Figure 27.
Generating Targets for Design-To Life-Cycle, Cost Category Variability. It has long
been understood to reduce variation improves quality. This is another source of variation
that can be measured, is developed from a robust data structure, and has common
meaning to stakeholders within an organization and therefore can be controlled.
105
The application of the Warranty Index is a simple tool where it acts as a
weighting factor to consider the more expensive element of service costs for the given
repair. Using the same factor to understand variability of those cost elements is also open
for more investigation to help a design team leverage prospects to reduce variability.
Design weaknesses in the Life-Cycle challenge part pricing strategies and design
rules for serviceability. These are activities that are not familiar to the product
development community. By use of a fairly simple but hierarchal structure, a binary
algorithm being applied, more organizational responsibility can be a part of the design
process.
Future Research
It would be a significant advancement if two things could follow this work:
An OEM applies the technology
A durable goods provider with a warranty system applies the technology.
The support for a truly focused DTLCC model is long over due for the industrial
markets and building on the RCM methods and strong data sets in this approach is a new
green field.
On a more futuristic note, it is the author?s findings that the idea of bringing
?Maintainability by Design? deeper into the auto industry one can then lead to a new
generation of auto-industry profits with a new customer-focused model. Today, many
plant capacities are under-utilized due to lower sales volumes. While this may not
continue and demand may surge, there is more reason to focus further on the capacity
106
issue where economic pressure has been mounting on dealers with their service claims
shrinking due to increased reliability almost universally across the car lines. This
decrease in activity needs to be bolstered by increase in sales which is not happening with
the current dealer density. Dealer density is likely over capacity and will need some form
of consolidation to remain profitable.
This may appear to be the downfall of the historical dealer model. However, a
?Maintainability by Design? initiative as revealed by this study will serve better the
customer and evolve the dealer network to be an ?i-car? personalization dealer or
customizer for the customer. With a concept coined herein as the ?i-car? -
personalization dealers will offer a new view of the dealer service center ? one that not
just orders options that are a challenge for OEM profitability ? but provides a new and
vital profit while offering exceptional personalized service. Here the designs for
complexity are reduced by modular design concept and the dealer showroom would be a
personalization effort. Pricing would be OEM based and not as a service part cost
discussed earlier.
This would change the complexity issue with respect to take-rate that has been
causing unmanageable configuration control within plants today which do not work out
with current sales and marketing methodologies. Removing the complexity would save
the system in favor of leaner manufacturing concepts. Modular car themes are not new as
they have been displayed at various auto shows but no one has perfected the modularity
required for personalization. Commonization, as a cost reduction strategy, would benefit
if passed along when complexity is reduced and modularity is developed further into the
architecture of the vehicle design.
107
Appendix
Reliability Bath Tub Curve, Targets and Tasks
The bath tub curve, as a graphical rendering of the failure rate behavior over time,
has been used as a training aid in Reliability education. The generic explanation shows a
simplification with the Y-axis representing the failure rate, ?, expressed as a function of
time or cycles along the X-axis.
The failure rate as a function of time, ?(t) as it is more frequently defined as the
hazard rate, h(t), that finds its expression in a two parameter Weibull distribution
discussed in detail by Roush and Webb (2001):
h(t) = f(t) / R(t) where
f(t) = dR(t) / dt =
Here, h(t) is just the probability of failure per unit time (failure rate) for those
items that have survived to time t, (pp 73-159). Also, the probability density function is
given by f(t) and the Probability of Survival to time t is given by the cumulative density
function:
F(t) where,
R(t) = 1 ? F(t).
108
Considering how h(t) might vary with time yields three regions of interest that
define a useful understanding dealing with aging effects and how one may go about
improving the reliability or failure rate behavior. The generalized interpretations as to the
cause of the failure rate variation are explained in simplified terms:
Description Region I:
This region has been associated with a range of time or cycles at initial usage at t1
= 0 to t2, where the failure rate has been observed to decrease with usage. This has often
been associated with manufacturing defects due to lack of consistency part to part and
assembly to assembly.
Description Region II:
Here the usage ranges from t2 to t3 where the failure rate is essentially level where
it has been characterized by failures of random and chance. This period has been named
the useful life period.
Description Region III:
109
The period from t3 to the end of life of an item is where the failure rate is
characterized by a wearout condition. Here is where the failure rate increases with each
additional unit of time. The onset of this condition can be sudden or gradual as in crack
propagation or brake pad wear respectively. What is often referred to as the knee of the
curve is the change point where failure likelihood progresses rapidly.
Conveniently, the probability of failure that each of these regions depict is
supported by experience and intuition has been modeled by the Weibull cumulative
probability density function for which Reliability is expressed as:
R(t) = 1 - F(t)
or given most simply as the two parameter Weibull model:
Where R, is the Reliability and is the probability of surviving to a mission time or
cycles.
Given:
t, the mission duration in cycles or time.
 ?, the Weibull Shape parameter or Weibull Slope.
 ?, the failure rate, in the same units as the mission duration
e, the Euler number with approximate value of which is
2.718281828.
110
Given this landscape of the bath tub curve depiction then each of the three regions
can be explored for actionable quality improvement.
Reliability Projects - Region I:
The changing failure rate or hazard rate as a function of time decreasing with
shape parameter less than unity, one might infer from data modeled in this way that by
inspection a flattening of the curve is desirable. Just how to do this given the nature
being classified as early life failures can be verified from warranty data that the
decreasing failure rate items are made up of start up issues, lack of experience, training,
manufacturing defects, and identify that the variability is therefore controllable by the
plant directly or indirectly and are not design issues.
Since these types of failures can reveal themselves by in plant data collection, test
and inspection and by early fleet or vehicle use it is important to have an early response
organizational structure or a focused launch team as processes like burn-in are not
typically employed where process controls are of more concern to better quality.
111
This would change the effect of the force of mortality over the interval and
therefore lower the failure density and improve the overall reliability and lowering
warranty costs.
Reliability Projects - Region II:
In region II, the hazard rate is flat or constant and does not vary significantly with
time. The Weibull Reliability model shape parameter, being equal to unity in this region
reduces the two parameter Weibull model to the exponential model:
112
The idea that the failure rate is constant over this interval embraces the notion that
failures are the result of random loads against random strengths and when the load
exceeds the strength as a challenge, a failure results. To correct issues related to loading
greater than expected requiring an engineering solution.
One could address the intrinsic reliability of the constituent parts involved and
influence event occurrence by adding strength or reducing the load through a transfer
function or masking the fault or adding redundancy to mention several reliability actions
that can be defined at the mechanism of failure level. One investigation shown by
Collins (1993) expresses the use of fracture mechanics by use of critical strength
113
parameters to assure failure predictors can reveal a strength that is greater than stress or
loading (p 62) by design.
Reliability Projects - Region III:
The mechanisms of failure in region III are those that require periodic service or
replacement over the Life-Cycle of the product. The design being subject to
accumulating damage or increasing degradation with each additional increment of usage
cycles needs to be managed in terms of Life-Cycle design practices. With the hazard rate
increasing with time or cycles, there is greater risk of failure on the effected systems and
components. Items of wear may require maintaining an operating condition such as clean
lubrication where the oil itself is a life limited item.
114
Clearly, if the oil change rate can move from 3,000 mile intervals to 6,000 mile
intervals, to even 10,000 mile intervals, the reduction in service events, labor and
supplies can be dramatic when figured about 10,000,000 cars being driven 50 miles a day
over a 10 year life with 1 gallon of oil per change as it would amount to 67,000,000 oil
changes and as many gallons of oil to support. When looking at the cost of $25 dollars
per event then that would be a cost of $1.67 billion dollars at the 3,000 mile interval rate.
Responsibility for Assurance Actions in the Life-Cycle Design:
In order to more effectively design for the Life-Cycle, a quality organization
would need to staff engineering with the needed assurance disciplines. Within in the
Design Assurance activities of product design is a skill set that can apply the elements of
Reliability, Availability, and Maintainability in a balanced way towards lowest warranty
costs and lowest operating costs while maximizing availability.
One needs to ask what is the right value for ?(t) and ?t3? for the ?knee? of the bath
tub curve and address if there is a means to push the curve further out or flatten it. Each
manipulation of the design is reducing the density of failure events per unit of time or
cycles. Changes that move the failure mechanism and its activation energy might be
found in chemical and material properties between interfacing parts to hopefully change
the aging effects. Failure mode avoidance may be abated through complex design
changes and validation methods.
One other situation may occur if the Weibull shape factor can be increased rather
than decreased and be more highly reliable over the Life-Cycle. Here, by more highly
115
organized more highly homogeneous material fabrication it may then possess physicals
properties with fewer imperfections that reduce the density of an initiation site of a
failure. Instead of failures over a wider area in wearout there is a longer life that does not
degrade appreciable until a later in life sharp knee exists.
The type of changes required to affect improvement wear failure mechanisms,
latent in a design, require maintenance and maintainability design practices engaged.
These latent or hidden life limit failure types required special testing to identify these
weakest links in the associated design. One tool shown to be effective as a validation
method given by Hobbs (1997) is Halt (Highly accelerated life-testing) or Hass (Highly
accelerated stress screening) that looks at validating a design as being able to survive the
Life-Cycle (p. 138). In part, testing techniques are applied through overstress tests and
then relating them to specific cross sections of application stressors in a multivariate way
to help timely identification of a wearout failure mode/mechanism and demonstrate
robustness or conformance to a life-limit. Depending on the severity of a given failure,
there is significant motivation to avoid potentially and the very high costs of a recall -
that goes well beyond fixing a service point.
116
The Warranty Data Base and Reliability Prediction and Targets.
It can be asked, ?What enough is or what is sufficient reliability?? The approaches
taken can look at a combination of three processes of improvement:
? Benchmark the Competition and Close Gaps.
? A Process of Continuous Improvement drives change.
? Assess the needs and wants of the Customer and develop requirements
Turning to the warranty system to implement improvements one needs to identify
specific and contrasting alignments to strategy ? that can be complex. One could average
data in terms of Repairs per 1000 vehicles sold or one could try to look at component
reliability and how requirements have been specified and verified ? what ever approach is
taken, it must be objective and specific to be actionable.
Towards the objective approach in target setting, one key understanding with respect
to the population of service events in a warranty data base is the knowledge that each
record represents an independent failed sample. Parts and systems that do not fail are
operating as intended. It is often assumed that the rate of usage in warranty is equal to
the rate of usage not in warranty.
Once an assessment of the ?cadavers? (warranty system records) is conducted and
then add to this the population knowledge of how many product are produced
representing the sample space one can then project the project population statistics (using
Weibull times to failure modeling). The steps are provided below:
117
One approach that is closely associated with the repair per 100 vehicles is to
consider the production of 100 sequential vehicles and observe the times to first failure
per each group. Then through methods developed by Johnson, (1964) the Weibull shape
parameter is preserved and an rank order statistic is used to slide the curve with the same
shape (slope on a linear graph) by the rank ordering of 1/100 to the full population with
failure rate, ? and shape parameter ?, (on a linear graph the Weibull slope).
118
The object here is to align programs reliability goals to say a b.01 life in the Life-
Cycle period or interval of the fielded system. The b.01 life is that point at which 1%
percent of the population will have failed. By taking the first failures per 100 a
representation like this in an interval, then one has the characteristic life (scale parameter)
point where 63.8% of the population failed at the stated ?cadaver? sample life from
warranty and then the population
projected directly to the 1% level.
The failure sample data set at the
63.8% level equals the population
statistic at 1% level - the scale
parameter of the cadaver
population. From this graphical interpretation that can be made with little knowledge of
the entire population that would be prohibited in testing, but having data analyzed in the
119
area of interest (such as the b.01 life) rather than waiting for data over the whole life
spectrum. This can help develop testing schemes that are less costly.
To get a sense of the scale of failures in an automotive system it is not unusual
that each vehicle on average will come back at least once in 12 months. That would be
100 returns per 100 vehicles per year or an index of 1 return per vehicle in a year ? every
vehicle returned once in warranty in a first year of service.
Given the usage rate of 60 miles per day and a 10 year period as the Life-Cycle
design point, then the average mission is given as 200,000 miles. This would be the
durability desired, the point where a sustainable design, one without failure would not
have a ?knee? until after that point. The variability of the mean life time might strain the
Design-To goal of the 95%tile user as the standard. This just goes to show the nature of
setting standards for reliability being sufficient and if not make sure it is maintainable.
120
Glossary
? Degradation: Allowable fading of colors, wear, fit, and finish over time.
? Life-Cycle Costs ? The process of determining all relevant costs from
conceptual development through production, utilization, and phase-out.
? Maintainability (1) ? a characteristic of design and installation which is
expressed as the probability that an item will conform to specified conditions
within a given period of time, when maintenance action is performed in
accordance with prescribed procedures and resources.
? Maintainability (2) - is the measure of the ability of an item to be retained in
or restored to specified condition when maintenance is performed by
personnel having specified skill levels, using prescribed procedures and
resources, at each prescribed level of maintenance and repair.
? Maintenance ? is the operations ? related activities undertaken after a system
has failed in the field to keep it operational or to restore it to operational
condition.
? Reliability: The probability that the product will perform its intended
function over time/mileage under specific operating conditions. Expressed
another way, reliability is the percent of vehicles, which meet customer
requirements (without failure) at a specified time/mileage objective.
? Reliability Demonstration: Testing for a useful life period; definition of
failure; test conditions (noises); required performance (safety/dependability,
high confidence, irritation, cost of ownership); and Sample size.
? Robustness: The ability of a product to meet the expectations of the
customers (which includes assembly and service as well as the end customers)
throughout the range of the noise factors (manufacturing variation,
environmental effects, changes over life, customer usage and system
interactions). Note: Full Service Suppliers understand these terms rather than
to just make parts to specification.
? Service Contract: An extended warranty purchased by a consumer to repair
an item, reimburse the consumer, and may include a deductible.
? Useful Life: 10 years / 150,000 Miles based on 10% of customers will use
their vehicles more than 150,000 miles in 10 years.
121
? Warranty Cost Sharing: Financial accountability of the supplier ? to reward
extraordinary efforts to reduce warranty below the agreed to target or
otherwise share in the cost of excess beyond the agreed target.
? Warranty: is a contractual guarantee to the buyer concerning product
performance. Failure and repair costs are allocated between the manufacture
and the buyer. The warranty serves to limit the manufacturer?s liability by
specifying consumer responsibilities and operating and servicing conditions.
Terms may include replacement, repairing, reimbursement.
? Reliability Centered Maintenance, (RCM) - developed as a flight line to
depot level response strategy for assuring a repair time can be scheduled with
little interference to an operational mission.
122
Bibliography
Beugel-Kress, Peter. "RAMS - Usable and Useful Concept for Automotive." SAE
Technical Papers Series. 2006-01.2006-01-1615 (April 03 2006): 1-14.
Bryan, C., Eubanks, C.F., Ishii, K. (1992). Design for Serviceability Expert System.
ASME Computers in Engineering. August 1992, San Francisco, CA. Vol.1. ISBN
0-7918-0935-8. pp. 91-98. Bryan - Expert System Design for Serviceability.
Christie, Less. "Costliest_car_ownership." CNN Money. 8 Mar. 2008.
<http://money.cnn.com/2005/07/29/sutos/costliest_car_ownership_cities/index.htm>.
Collins, Jack A., Failure of Materials in Mechanical Design. New York, New York: John
Wiley & Sons, 1993.
Crow, L.H. Reliability Analysis for Complex, Repairable Systems, in Reliability and
Biometry, ed. by F. Proschan & R.J. Serfing, (pp. 379-410). Philadelphia: SIAM,
1974.
Downey, Gary and Lucena, Juan. ?Globalization, Diversity, Leadership, and Problem
Definition in Engineering Education.? International Conference on Research in
Engineering Education, Honolulu, HI, 2007 pp 1-10.
DRM Technologies, Inc. R& M Primer, Reliability & Maintainability Handbook, undated
publication, 2001.
Ebeling, Charles E. An Introduction to Reliability and Maintainability Engineering. Eds.
Eric M. Munson and John M. Morris. Boston, Massachusetts: McGraw-Hill,
1997.
Frank Gryna, Naomi McAfee, Clifford Ryerson, Stanley Zwerling. Reliability Training
Text, 2nd Ed., The Institue of Radio Engineers, Inc. 1East 79th Street, NY 21, NY,
1960.
Grzanowski, Marion, and Long, Mary. "Designing for Improved Serviceability - NEO
Participant Guide." Introduction Methodology Design Manual Resources.
Dearborn Michigan: Fairlane Training and Development Center, January 2002. 1-
90. 90 p. January 2002
<http://www.mso.ford.com/fesd/vsp/upstream/dfis/dfis1.htm>.
Hamilton, Bruce W., and Molly K. Macauley. "Competition and Car Longevity."
www.rff.org. 2 Apr. 2008. <http://www.rff.org/documents/RFF-DP-98-20.pdf>.
123
Henley, Ernest J., and Hiromitsu Kumamoto. Reliability Engineering and Risk
Assessment. 1st ed. Ed. Lori Opre. Englewood Cliffs, New Jersey: Prentice-Hall,
1981.
Hobbs, G. K., ?What HALT and HASS Can Do for Your Products,? Evaluation
Engineering, November 1997, p. 138.
Hoffman, Bryce, ?Quality gain has hitch for Ford dealers.? The Detroit News, Detroit
Michigan, 2008.
Ireland, William. ?Cost Savings by Establishing Life Limits?, Proceedings: Annual
Reliability and Maintainability Symposium, San Francisco, CA. 1984 p 72-77.
K. Ishii, C. Eubanks, and M. Marks, ?Evaluation Methodology for Post-manufacturing
Issues in Life-Cycle Design,? Vol 1 Issue 1. CERA Journal, 1993.
Johnson, Leonard G. The Statistical Treatment of Fatigue Experiments. New York, New
York: Elsevier Scientific Publishing Company, 1964.
Johnson, Leonard. "How Profits are Related to Quality and Reliability." Statistical
Bulletin. 17.Bulletin 2 (May 1987): 1-7.
The Kellogg Foundation. ?Logic Model Development Guide? Publication by The WK
Kellogg Foundation, One East Michigan Avenue East, Battle Creek, Michigan,
URL: http://www.wkkf.org. 2004.
Klyatis, Lev M., and Eugene L. Klyatis. "The Stategy to Improve Engineering Concepts
of Automotive Reliability, Durability, and Maintainability Integrated with
Quality." SAE Technical Paper Series. 2006-01.2006-01-1337 (March 03 2006):
1-8.
Lannoy, Andre', Henri Procaccia, and Marc Procaccia. "What is Aging." Proceedings of
the International Workshop of Age-Dependent Reliability Models and Analysis of
Operational Data. EUR2186EN (October 05 2005): 1-11. April 2008
<http://www.jrc.nl/publications/scientific_publications/2005/EUR21862EN/>.
Mays, Jay, and F. J. Wassilak. "Styling for Damageability." Collage of Creative Studies.
Ford Motor Company. 16 Mar. 2004.
Mohammad Modarrres, Mark Kaminskiy and Vasilliy Krivtsov. Reliability Engineering
and Risk Analysis: A Practical Guide. John Wiley and Sons, New York, NY
2001.
O?Connor, Patrick. Practical Reliability Engineering. 4th ed. UK, John Wiley and Sons,
2002.
124
Schiffauerova, Andrea and Thomson, Vince. ?A Review of Research on Cost of Quality
Models and Best Practices.? The International Journal of Quality and Reliability
Management, Vol. 23, Issue 6. Bradford, 2006.
de Oliveira e Souza, Marcelo L., and Terezinha R. de Carvalho. "The Fault Avoidance
and the Fault Tolerance Approaches for Increas." SAE Technical Paper Series.
2005-01.2005-01-4157 E (November 22 2005): 1-16.
Roush, Marvin, and Willie Webb. Applied Reliability Engineering. II. 2nd ed. College
Park, Maryland: Center for Reliability Engineering, University of Maryland,
2001.
Stadler, Jacob J., and Lisa D. Brideau. "Human Factors and Maintainability in the Plant
Research Unit." SAE Technical Papers. 2004-01.2004-01-2583 (July 19 2004): 1-
6.
Vacante, Russell A. "RCM & Customer-Supplier Partnerships: A Reliability Enabler."
Partnership in RMS Standards. 5.3 Ed. Russell A. Vacante. College Park,
Maryland: University of Maryland, July 2001. 1-7. 18 p. 20 April 2008
<http://www.enre.umd.edu/rms>.
Warranty Week. "Auto Warranty vs. Quality:." Warranty Week ISSN 1550-9214. June
20, 2006. < http://www.warrantyweek.com/archive/ww20060620.html>.
Woolford, Barbara. "DESIGN FOR MAINTAINABILITY." Man-Systems Integration
Standards. 1995. <http://msis.jsc.nasa.gov/sections/section12.htm>.
Zenz, Gary J. Purchasing and the Management of Materials. 5th ed. New York, New
York: John Wiley & Sons, 1981.