Product Design Techniques for Robustness, Reliability and Optimization

Instituto Tecnologico y de Estudios Superiores de Monterrey

Campus Toluca

Product design:techniques for robustness, reliability and

optimization

Class Notes

Dr. Jose Carlos Miranda V.

v. Fall 2004

Copyright c©2004 Dr. Jose CarlosMiranda. Todos los derechos reservados.

Preface

It is widely recognized that to develop successful products, systems or services itis extremely important to follow a structured product development process. Al-though each company follows a process tailored to its specific needs, in general thestart of a product development process is the mission statement for the product.It identifies the target markets for the product, provides a basic functional descrip-tion of the product, and specifies the key business goals of the effort. The end ofthe development effort occurs when the product is launched and becomes availablefor purchase in the market place. The different activities that take place during the

product development process can be grouped into five phases: Concept development,system-level design, detail design, testing and refinement, and production ramp-up.During the detailed design and the testing and refinement phases, product optimiza-tion, robustness and reliability becomes critical. As many powerful techniques haveappeared to make a product more optimal, robust and reliable, it is necessary toknow how they work and how can they be applied to design products that exceedcustomer expectations and minimize costs.

The present notes have been prepared for the courses of Design Methodologies andProduct Design that I teach. Although these notes are far from complete and there-fore may contain many mistakes and inaccuracies, they evolve term after term andwith the help and suggestions of my students are continuously improved. Oncethese notes are mature, it is my desire to publish them to reach a wider audienceand receive further comments.

If you have any feedback, suggestions or have detected any mistakes, or simply wouldlike to assist me or contribute in this effort, please do not hesitate to contact me. Iwill be very happy to hear from you.

Jose Carlos MirandaResearch Center for Automotive Mechatronics

[email protected]


Part I

The product design process


CHAPTER 1

The Engineering Design Process

1.1 Definition of design

The word design has had different meanings over the last decades. Whilesometimes a designer is considered to be the person drafting at the drawingboard or in the computer, the word design really conveys a more engineeringand analytical sense. Design is much more than just drafting.

Suh (1990) defines design as the creation of synthesized solutions in the formof products, processes or systems, that satisfy perceived needs through themapping between functional requirements and design parameters.

In the scope of the previous definition, functional requirements (FRs) respondto the question of what a product must do or accomplish. On the other hand,design parameters (DPs) respond to the question of how the functional require-ments will be achieved. What relates the domain of functional requirements tothe domain of design parameters is design (see figure 1.1). It should be notedthat although design parameters should fulfill the functional requirements, themapping between them is not unique. For a set of functional requirementsmay be several design parameters that fulfill those functional requirements.

Another, less technical, definition of design is the one promulgated by ABET(Accreditation Board for Engineering and Technology):


3 1.2 The design process

List ofFunctional

Requirements

WHAT?

List ofDesign

Parameters

HOW?

design

Figure 1.1: Design is the process of mapping functional requirements to designparameters.

“Engineering design is the process of devising a system, component,or process to meet desired needs. It is a decision making process(often iterative) in which the basic sciences, mathematics and engi-neering sciences are applied to convert resources optimally to meeta stated objective. Among the fundamental elements of designprocess are the establishment of objectives and criteria, synthesis,analysis, construction, testing and evaluation. . . It is essential toinclude a variety of realistic constraints such as economic factors,safety, reliability, aesthetics, ethics and social impacts.”

Although several definitions of design may be found, the last one highlightsone of the main difficulties associated with design: its truly multidisciplinarynature. Design involves several, if not all, different departments in a givencompany (see figure 1.2). Design engineers should always be aware of thiscondition, involving in the design process the expertise of people of differentdisciplines.

1.2 The design process

There are many different maps or models of the design process. Some of thesemodels describe steps and their sequence as they occur in the design process.Some other models try to define or prescribe a better or more appropriatepattern of activities. Cross (1994) describe some of these models.

Copyright c©2004 Dr. Jose Carlos Miranda. Todos los derechos reservados.

1.2 The design process 4

ProductDesign

Purchasing

ManufactureEngineering

ElectronicEngineering

IndustrialDesign

MechanicalEngineering

Marketing

Figure 1.2: Engineering design core disciplines.

1.2.1. Design process

models

Probably the most simple model of the design processis the one shown in figure 1.3, where only four generalstages are outlined.

Another relatively simple model is presented by Ullman (1992) who suggest toview the design as problem solving. When solving a given problem, five basicactions are taken:

1. Establishment of need or realize there is a problem to be solved.

2. Understanding of the problem.

3. Generation of potential solutions for it.

4. Evaluation of the solutions by comparing the potential solutions anddeciding on the best one.

5. Documentation of the work.

While it is possible to see design as problem solving, it is important to realizethat most analysis problems have one correct solution whereas most designproblems have many satisfactory solutions.

A more detailed model, which involves all steps of the design process, is pre-sented in figure 1.4. As shown, this model divides design process in 5 phases:Concept development, System-Level design, Detail design, Testing and refine-ment and Production. Each phase has one or more steps. It is importantto realize that this model is general and may be necessary to follow differentpaths in one or more phases depending on the project at hand.



Exploration

Generation

Evaluation

Communication

Figure 1.3: A simple model of the design process with 4 stages.

Independently of the model, it is generally agreed that the design processshould always start with the recognition of a need. After the need has beenrecognized it is necessary to consider alternatives for its solution, which is doneduring the concept development phase. Here the statement of the problem istaken and broad solutions to it are generated. This phase presents the greatestchance for improvements and hence is specially imperative to be objective,open to new ideas and recognize when changes are needed.

Once the best ideas have been selected, preliminary design may start to furtherevaluate those ideas. In this phase testing may be of great help to differentiategood ideas from regular ones. After a design has been finally selected, detaileddesign begins to incorporate every feature that the design may need to incor-porate. Hence, a very large number of small but essential points should bedecided. After the detailed design has been re-evaluated and tested, produc-tion planning may be started and final products tested for final acceptance.

In what follows the different steps in the design process are discussed more indepth.

1.2.2. Identifying customer

needs

The need to design a new product may comefrom different sources: consumers, organizationsor governments. The need may also sometimes



Recognition of need

Conceptualization

Feasibilityassessment

Phase 1:ConceptDevelopment

Preliminary design

cost analysis / redesign

Developmenttesting

Detailed designQualification testing

Production planningand Tooling design

ProductionAcceptance testing

Phase 5:Production

Phase 2:System-LevelDesign

Phase 4:Testing andRefinement

Phase 3:DetailDesign

Figure 1.4: Detailed model of the design process.

be substituted for an idea of a product with possibilities of becoming commer-cially successful.

Eide et al. (1988) state that in industry, it is essential that products sell forthe company to survive. Inasmuch as most companies exist to make a profit,profit can be considered to be the basic need. Hence, a bias toward profitand economic advantage should not be viewed as a selfish position becauseproducts are purchased by people who feel that they are buying to satisfy aneed which they perceived as real. The consumers are ultimately the judgesof whether there is truly a need.

Identifying the needs of the costumer is one of the most important steps inthe design process and is, at the same time, one of the most difficult since isnot unusual to find that the customer does not know exactly what featuresthe product must have. Once the needs have been specified together with thecostumer, this information is used to guide the design team in establishingdesign parameters, generating concepts and selecting the best one of them.

According to Ulrich & Eppinger (2000) the process of identifying customerneeds includes five steps:

1. Gather raw data from customers.



Metric Value

The product must be . . .

easy to install Average time for installation < x seg.

durable Must withstand 10x cycles

easy to open Opens with a force of max. x newtons.

able to resist impacts Withstand drops from x meters.

able to work in cold weather Operation possible at -x◦ C.

Table 1.1: Examples of metrics and their value.

2. Interpret the raw data in terms of customers needs.

3. Organize the needs into a hierarchy of primary and secondary needs.

4. Establish the relative importance of the needs.

5. Reflect on the results and the process.

1.2.3. Establishing the

design requirements

As was briefly discussed above, when the design en-gineer is first approached with a product need, it isvery unlikely that the customer will express clearly

what is needed. In most occasions it is only know what is wanted in a verygeneral way without idea of the particularities involved.

Hence, the starting point for a design engineer is to turn an ill-defined problemwith vague requirements into a set of requirements that are clearly defined.This set of product requirements may change as the project advances, so it isconvenient to clarify them at all stages of the design process.

For the product requirements to be helpful, they must be translated to tech-nical specifications that are precise, easily understood and can be measure bymeans of one or more design variables. Ulrich & Eppinger state that “A spec-ification consists of a metric and a value.” Table 1.1 shows some examples ofmetrics and their values.

Several tools can be used to establish product specifications. Although simpleto apply, the objectives tree and decision tree methods offer a clear and usefulstarting format for such a statement of requirements and their relative impor-tance. As will be discussed later, other more sophisticated and more usefulmethod is Quality Function Deployment (QFD).



1.2.4. Concept generation According to French (1985) in this phase the state-ment of the design problem is taken and broad

solutions are generated in the form of schemes. It is the phase that makes thegreatest demands on the designer, and where there is more scope for strikingimprovements. It is the phase where engineering science, practical knowledge,production methods and commercial aspects need to be brought together. Itis also the stage where the most important decisions are taken.

In the scope of design, a concept is an abstraction, an idea that can be rep-resented in notes and/or sketches and that will eventually become a product.It is generally recognized that, for a given product, several ideas (sometimeshundreds of them) should be generated. From this pool of ideas, a couple ofthem will merit serious consideration for further evaluation and development.

The concept generation stage can be divided into 4 steps:

1. Clarification of the problem.

2. Gathering of information.

3. Use and adaptation of design team’ s knowledge.

4. Organization of team’s thinking.

Although concept generation is an inherently creative process, it is possible touse some techniques to improve it like functional decomposition and genera-tion of concepts from functions. Although sources for conceptual ideas comeprimarily from the designer’s own expertise, it can be enhanced through theuse of books, experts, lead engineers, patent search, brainstorming and currentdesigns.

1.2.5. Concept selection The purpose of concept selection is assessing thefeasibility of concepts to ensure that they are achiev-

able technically and economically. The feasibility of the concept is based onthe design engineer’s knowledge. As in the generation of concepts, the designengineer can rely in tools –like the decision-matrix method– to compare andevaluate concepts.

The importance of the concept selection phase cannot be understated. It isknown that decisions made during the design process have the greatest effecton the cost of a product for the least investment. In figure 1.5, the costof design and its influence in manufacturing cost for an automotive project



Design

Material

Labor

Overhead

0

20

40

60

80

100

FinalManufacturingCost

Influence on

CostFinal Manufacturing

Perc

ent

Figure 1.5: Design influence on manufacturing cost (After Ullman, 1992).

is shown. From the figure it can be stated that the decisions made duringthe design process have the greatest effect on the cost of a product for theleast investment. Typically, around 70% to 80% of the manufacturing cost iscommitted by the end of the conceptual phase of the design process. Hencethe importance of concept evaluation.

Also, the generation and evaluation of concepts have a great effect on the timeit takes to produce a new product. Figure 1.6 shows the number of designchanges made by two automobile companies with different design strategies.

Company A made many changes during the early stages of the design processas a result of the iterative process of generation and evaluation of concepts.Company B made just a few changes in the initial stage, but was still makingchanges later in the process, even when the product was released for produc-tion. The advantage gained by company A is clear since changes made late inthe process are far more expensive than changes made in early stages.

The evaluation of concepts to find its viability may occur not only duringconcept development, but throughout the design process. This will lead to theso called Design process paradox (Ullman, 1992). The design process paradoxstates that during the design process, the knowledge about the design increasesas the project runs in time and the design team gains understanding of theproblem at hand. Hence, the knowledge of the design team is at its top when



Company A

Company B

Release forproduction

Time

Des

ign

chan

ges

Begindesign

Figure 1.6: Engineering changes in automobile development (After Ullman,1992).

the design process is at its end. Although this seems natural, it is importantto realize that, by the end of the process, most decisions have already beenmade.

This increased knowledge at the end of the project tempt most design teamsto feel the need of re-doing the project now that they fully understand it.Unfortunately, economics almost always drive the design process, and secondchances rarely exist.

Figure 1.7 shows the dilemma above. At the beginning of the process, thedesign team has the most freedom since no decisions have been made. As timegoes by, knowledge increases as a result of the design time efforts, but freedomis lost since decisions have been made and changes are increasingly expensiveto perform.

1.2.6. Concept testing Concept testing is closely related to concept selection.It is used to gather opinions and information from

potential customers about one or more of the selected concepts that may bepursued. It can also be used gather information about how to improve anspecific product and to estimate the sales potential of the product.

Ulrich & Eppinger (2000) suggest to divide the concept testing into 6 steps:



Knowledge aboutthe design problem

freedomDesign

0

20

40

60

80

100

Perc

ent

Time into design process

Figure 1.7: The design process paradox (After Ullman, 1992).

1. Definition of the purpose of the concept test.

2. Choosing of a survey population.

3. Choosing of a survey format.

4. Communication of the concept.

5. Measurement of customer response.

6. Interpretation of results.

Both concept selection and concept testing are used to narrow the possibleconcepts under consideration. Concept selection relies in the work and judg-ment of the development team. Concept testing is based in data gathereddirectly from potential customers.

1.2.7. Preliminary design The preliminary design stage or embodiment de-sign stage, fills the gap between design concept

and detailed design. According to French, in this phase the schemes are workedup in greater detail and, if there is more than one, a final choice between themis made. There is (or should be) a great deal of feedback from this phase tothe conceptual design phase.

Is during this stage of the design process that the overall system configura-tion is defined. Extensive engineering documentation in the form of schemes,diagrams, layouts, drawings, notes or other types of documents is generatedto provide control over the project and to ensure better communication and



integration between different engineering disciplines involved on the designeffort.

The preliminary design helps to obtain more precise design requirements in-volving analysis, benchmarking, literature search, experience, good judgmentand, if necessary, testing. The refinement of the project also helps to have abetter estimate of the project cost and required time for completion.

1.2.8. Detailed design After the preliminary design stage has been carriedout, it is necessary to go into the details of the design

in order to better understand the concepts. Detail design is mostly concernedwith the design of the subsystems and components that make up the entiredesign. Because of the latter, this stage is sometimes divided into two inde-pendent parts, System-level design and the detail design itself.

In the system-level design the product arquitecture is defined and decomposi-tion of the product into subsystems and components takes place. These com-ponents may be integrated circuits, resistors, shafts, bearings, beams, plates,handles, seats, etc., depending on the nature of the product under develop-ment. Here, geometric layouts of the product and functional specifications foreach subsystem are stated.

The detail design phase includes the complete specification of each independentpart such as geometry, materials and tolerances and identifies all those partsthat will be purchased from suppliers. In this stage, the control documentationof the product is generated, including technical drawings, part productionplans and assembly sequences.

1.2.9. Production planning This stage initiates with the identification of themachines, tooling and processes required to man-

ufacture the designed product. Technical data such as dimensions, tolerances,materials and surface finishes among others are evaluated to determine theappropriate assembly sequence for the manufacturing operations. Accordingto Ertas and Jones (2000), typical tasks included in the production planninginclude:

1. Interpretation of design drawings and specifications.

2. Selection of material stock.

3. Selection of production processes.



4. Selection of machines to be used in production.

5. Determination of the sequence of operations.

6. Selection of jigs, fixtures, tooling and reference datum.

7. Establishment of tool cutting parameters, such as speed, depth and feedrate.

8. Selection of inspection gauges and instruments.

9. Calculation of processing time.

10. Generation of process documentation and numerically controlled ma-chine data.

Once the production planning has been made and all the decisions regardingproduction have been taken, a production ramp-up is made using the intendedproduction system. The purpose of the production ramp-up is to evaluate thecorrectness of the production plan, the tooling and the assembly sequences tofollow as well as to identify possible flaws before going to a full-scale produc-tion.

1.2.10. Documentation Engineers feel most of times burdened with the ideaof documenting their designs. The preparation of

documents describing the design process and the reasons behind decisionstaken is oftenly seen as as an activity that does not directly contribute tothe design. Other times documentation is seen as an unattractive task thatdoes not involve any challenge at all.

Nevertheless, documentation is as important as any other in the task in thedesign process. Product documentation is important not only in terms ofinstructions to user, maintainers or others, but is imperative for purposes likelegal protection or future product redesign.

Hence, keeping track of the ideas developed and decisions made in a designnotebook is essential. It is advisable to keep, for patent or legal purposes,a notebook with dated pages that is sequentially numbered and signed. Inthis notebook, all information related to the design such as sketches, notes,calculations and reasons behind decisions should be included. The notebookdoes not have to be neat, but certain order has to be kept. When designinformation like plots, photocopies, drawings or results of analyses are toolarge or bulky to keep in the notebook, a note stating what the document is,a brief summary of its contents and where it is filed should be written.


1.3 Quality Function Deployment 14

DesignTime

Eff

ort

ProductionProcessDetails

Traditional approach

With QFD

Figure 1.8: Traditional vs. QFD design approaches (After Ouyang et al.

When the design effort has concluded, standard drawings or computer datafiles of components showing all the information necessary for the productionof the product have to be generated. This drawings usually include writtendocumentation regarding manufacturing, assembly, quality control, inspection,installation, maintenance and, retirement.

1.3 Quality Function Deployment

It is not uncommon that designers find themselves working a problem onlyto find out later that they were solving the wrong one. An efficient designermust try by all possible means to define the correct problem at the beginningor discover the problem at earliest possible moment. The Quality FunctionDeployment technique provides a methodological way to do it.

Quality Function Deployment (QFD) originated in Japan as a help to trans-late customer requirements into technical requirements throughout the devel-opment and production of a product. It originated in Japan in the 1970’s asthe Kobe supertanker company wanted to develop the logistics for buildingcomplex cargo ships. Professors were asked to create a technique that wouldensure that each step of the construction process would be linked to fulfillinga specific customer requirement.

Using this technique, Toyota was able to reduce the costs of bringing a new car


15 1.3 Quality Function Deployment

DesignRequirements

Req

uire

men

tsC

usto

mer

Req

uire

men

tsD

esig

nParts

Requirements

ProcessRequirements

Req

uire

men

tsPa

rts

ProductionRequirements

Req

uire

men

tsPr

oduc

t

HO

WSW

HA

TS

Design

1

WH

AT

S

Details

2

WH

AT

SProcess

3

Production

4WH

AT

S

HO

WS

HO

WS

HO

WS

Figure 1.9: The four phases of QFD. From customer requirements to clientsatisfaction. The hows on each House of Quality becomes the whats in thenext.

model to the market by 60 percent and to decrease the time required for itsdevelopment by one third. As shown in figure 1.8, QFD requires more efforton the design stage, but as most design flaws are catched early in the designprocess, later stages are less prone to fail or require adjustments or redesigns.

According to Ouyang et al., Qualify Function Deployment has four distinctphases: design, details, process and production. As shown in figure 1.9, in theDesign phase, the customer helps to define the requirements for the productor service. In the Details phase, design parameters (hows) carried over thedesign phase become the functional requirements (whats) of individual partdetails. In the Process phase, the processes required to produce the product aredeveloped. Once more, the design parameters of the details phase become thefunctional requirements of the process phase. Finally, in the Production phase,the design parameters of the process phase become functional requirements forproduction.

As discussed above, QFD can be applied all the way through the design pro-cess from concept to production using the same principles on each phase. Itis generally agreed that the QFD technique is most valuable at the early de-sign stages where customer requirements have to be translated to engineeringtargets.


1.3 Quality Function Deployment 16

The QFD technique uses six steps to do this translation:

1. Identifying the customer(s).

2. Determining customer requirements.

3. Determining relative importance of the requirements.

4. Competition benchmarking.

5. Translating customer requirements into measurable engineering require-ments.

6. Setting engineering targets for the design.

Each step will be reviewed in more detail, but before going any further isconvenient to highlight that:

• No matter how well a design team thinks it understand a problem, itshould employ the QFD method.

• Customer requirements must be translated into clear engineering targetsinvolving measurable quantities.

• The QFD technique may be applied to the whole design as well as tosubsystems or subproblems.

• It is important to first worry about what needs to be designed and, oncethe problem is fully understood, to worry about how it will be designed.

1.3.1. Identification of

costumers

Sometimes is not only not clear what the customerwants, but also who the customer is. Furthermore,is very common to find that there is more than one

customer to satisfy.

Independently of how many customers may be, it is essential to realize that thecustomer, and not the engineer, is the one driving the product developmentprocess. Many times the engineer has a mental picture of how the productshould be like and how it should perform, picture that may be very differentfrom what the customer really wants. On the other hand, may products havebeen poorly received by the customers simply because the engineer failed toidentify accurately the customers’ desires.


17 1.3 Quality Function Deployment

1.3.2. Determination of

costumers requirements

The determination of customer requirements shouldbe made through customers surveys or evaluation ofsimilar existing products. Customer requirements

should be made in the customer’s own words such as “fast”, “easy”, “durable”,“light”, “strong”, etc. As much as possible, customer requirements should bestated in positive terms.

In order to facilitate understanding, requirements may be grouped in types likeperformance requirements, appearance requirements, safety requirements, andso on. If the customer has specific preference for one given type, determiningthe relative importance of different requirements may be easier to do.


relative importance of the

requirements

Not all requirements will be regarded as equallyimportant to customers. For example, “easy touse” may be more important for the customerthan “easy to maintain”, and “easy to maintain”

may be regarded as more more important than “good looking”. On the otherhand, some requirements like “safe to use”, may be regarded as absolute re-quirements rather than relative preferences.

In order to design effectively, the design team should know which attributesof their product design are the ones that most heavily affect the perceptionof the product. Hence, it is necessary to establish the relative importance ofthose attributes to the customers themselves.

1.3.4. Competition

benchmarking

Sometimes customers often make judgment about prod-uct attributes in terms of comparisons with other prod-ucts. One screwdriver, for example, may have better

grip than others or another screwdriver may seem more durable. Given thatcustomers are not generally experts, they may compare different attributes byobservation of what some products achieve.

If the product is to be well positioned in a competitive market, the designteam must ensure that its product will satisfy customer requirements betterthan competitor products. Therefore, the performance of the competition ofthose product attributes that are weighted high in relative importance shouldbe analyzed.


1.4 Some important design considerations 18

1.3.5. Conversion of

customer needs into

engineering requirements

Once a set of customer requirements have beenselected due to its importance, it is necessary todevelop a set of engineering requirements that aremeasurable.

Some of these engineering requirements, or design specifications, may be cleareddefined from the beginning. One example is the weight that a chair must with-stand. Others, may be more difficult to characterize as will be measurable bydifferent means. In the case of a chair that is to be “easily assembled” by thecustomer, “easily” may be measured in terms of the number of tools neededfor the assembly, the number of parts to be assembled, the number of stepsneeded for the assembly or the time needed for the assembly.

In this step, every effort should be made in order to find all possible ways inwhich a customer requirement may be measured.

1.3.6. Setting engineering

targets

The last step in the process is setting the engi-neering targets. For each engineering measure de-termined in the previous step, a target value will

be set. This target values will be used to evaluate the ability of the productto satisfy customer requirements. Two actions will be needed, to examine howthe competition meets the engineering requirements, and to establish the valueto be obtained with the new product.

Best targets are established using specific values. Less precise, but still usable,are those targets set within some range. Another type, extreme values, aretargets set to a minimum or maximum value. Although extreme type targetsare measurable, they are not the best since they give no clear information ofwhen the performance of a new product is acceptable. Here, evaluation of thecompetition can give at least some range for the target value.

1.4 Some important design considerations

When designing products, several considerations must be taken into account.For the inexpert designer, this considerations may or may not be obvioussources for requirements, parameters and targets. In what follows, three designconsiderations, whose importance may depend on the project at hand, arebriefly discussed.


19 1.4 Some important design considerations

1.4.1. Product distribution Most of the times, when designing a new prod-uct, the design team does not pay much atten-

tion in how the product will be distributed. Decisions regarding packaging,transportation and shelf stocking are taken after the product has been de-signed. Nevertheless, design features that could be avoided may increase thedistribution cost due to the need of special packaging, transportation or shelfs.Design teams must do everything at their hands to avoid this situations thatunnecessarily increase the cost of the product.

Taking into account the distribution of the product is specially important whenredesigning a product. Generally speaking, companies looking for an existingproduct of better features are unwilling to make extensive modifications to theexisting distribution infrastructure. In this cases, product distribution will bea major source of requirements.

1.4.2. Design for after life It is normally assumed for most engineering prod-ucts, that after it has completed its useful life, the

product will be removed from its original installation, retired and dispose of.Nevertheless, in many occasions the product is put to some second use that isdifferent from its original purpose. Consider for example, an empty 20 lts. (4Gal.) bucket that is used as a step.

The problem arises as this second use was not included in the initial designspecifications and is therefore not accounted for in the design process. Theresult may be failure and personal injury leading to product liability litigation.The fact that a certain product was used in a way never intended by theoriginal design may not be of importance on the court. Courts seem to focuson whether the failure was foreseeable and not whether there was negligenceor ignorance. The best the design team can do is to try to foresee both useand misuse an make provision in the design for credible failures.

1.4.3. Human factors in

design

Almost every product that is designed will interactwith humans whether during manufacture, opera-tion, maintenance, repair or disposal. Operation is

probably the most important since it will involve the largest span of interac-tion.

Considering operation, a good product will be the one that becomes an ex-tension of the user’s motor and cognitive functions. To achieve this, human–machine interaction features should be included as parameters in the design


1.5 Good design practices 20

process from early stages.

In order to translate functional requirements into design parameters, the studyof ergonomics has produced a body of anthropometric (human measure) datathat can be used in designing anything that involves interaction between ahuman and a product. As anyone will agree, humans bodies come in a varietyof shapes and sizes, which makes somewhat difficult to design a product to fitabsolutely everybody. Nevertheless, human measure can be well representedas normal distributions.

This last feature makes it possible to define parameters to fit, let say, 90%percent of the population. In many occasions, to be able to design for such ahigh percentage of the population it is required to include adjustable featuresto the product. One typical example is the way in which seat and steeringwheel positions can be altered in many cars to adapt the height and size of thedriver.

Other three ways in which humans may interact with products is as a sourceof power (for example when opening a door), as a sensor (for example readinga dashboard) or as a controller (for example the operating a CD player).

In the first case, information about the average force that a human can provide(or is expected to provide) is vital toward a successful product. In the secondcase, if the human is expected to be able to read information is important thatthe person has only one way to interpret the data. In the third way, a productmust be designed so there is no ambiguities in the form in which the productoperates. For the product to be easy to interact with, there must be only oneobviously correct thing to do for every action that is required.

1.5 Good design practices

1.5.1. Good design versus

bad design

The goal for the introduction of models for thedesign process is to provide a guideline to helpthe engineer/designer to achieve a better prod-uct through the use of good design practices. As

experience would tell, in most occasions it is not difficult to tell either as en-gineer or consumer, a good design from a bad design. Table 1.2 show somegeneral characteristics of good design versus bad design.


21 1.5 Good design practices

Good Design Bad Design

1. Works all the time 1. Stops working after a short time

2. Meets all technical requirements 2. Meets only some technical requirements

3. Meets cost requirements 3. Costs more than it should

4. Requires little or no maintenance 4. Requires frequent maintenance

5. Is safe 5. Poses a hazard to user

6. Creates no ethical dilemma 6. Fulfills a need that is questionable

Table 1.2: Characteristics of good design versus bad design. After Horenstein(1999).

1.5.2. Good design

engineer versus bad design

engineer

Horenstein (1999) highlights the traits of gooddesign engineer and bad design engineers. Ac-cording to Horenstein, a good engineer:

• Listens to new ideas with an open mind.

• Considers a variety of solution methodologies before choosing a designapproach.

• Does not consider a project complete at the first sign of success, butinsists on testing and retesting.

• Is never content to arrive at a set of design parameters by trial and error.

• Use phrases such as “I need to understand why” and “Let’s consider allthe possibilities”.

A Bad Engineer:

• Thinks he/she has all the answers; seldom listens to the ideas of others.

• Has tunnel vision; pursues with intensity the first approach that comesto mind.

• Ships the product out the door without thorough testing.

• Use phrases such as “good enough” and “I don’t understand why it won’tworks; so-and-so I it this way.”


1.5 Good design practices 22

• Equates pure trial and error with engineering design.

Green (1992) summarizes skills that seem to mark the expert designer in do-mains of routine design.

Supplying context. The requirements seldom provide enough information tocreate a design. This occurs in part because the client himself does not knowprecisely what he/she wants. However, another problem is that the statedrequirements imply several other, unstated, requirements. The expert can“read between the lines” and supply context that reduces the search space.

Decision ordering. Strategic knowledge is a major part of the designers’ ex-pertise. The expert designer is able to make decisions in the correct orderto avoid spending much time in backtracking and revising. Decision orderingis important because it rank constraints. The expert’s decision ordering setconstraint values in some optimal sequence.

Heuristic classification. Although the overall design problem may be ill-structured,it usually contains some well-structured components. Some decisions fell intothe heuristic classification paradigm (here, heuristic means problem-solvingtechniques that utilize self-education techniques, as the evaluation of feed-back, to improve performance). The designer begins by listing requirements,both stated and unstated, and maps them to design parameters which enableshim/her to choose a set of design classes.

Parameter abstraction. Much of routine design requires to simultaneously man-age a large collection of variable values. This can be a very complex cognitivetask since it requires the expert to maintain a large amount of information inworking memory. Experts are able to reduce the complexity of the problem byabstracting only the most important parameters, treating related parametersas single entities whenever possible.

References

1. Cross, N. (1994) Engineering Design Methods, John Wiley & Sons.2. Eide, A., Jenison, R., Mashaw, L. & Northup, L. (1998) Introduction toEngineering Design. McGraw-Hill.3. Ertas A. & Jones, J. (1996) The Engineering Design Process, second ed.,


23 1.5 Good design practices

John Wiley & Sons.4. Horenstein, M. (1999) Design Concepts for Engineers, Prentice-Hall.5. Otto, K. & Wood, K. (2001) Product Design - Techniques in ReverseEngineering and New Product Development, Prentice-Hall.6. Ouyang, S., Fai, J., Wang, Q. & Johnson, K. Quality Function Deployment.University of Calgary Report.7. Pugh, S. (1990) Total Design, Addison Wesley.8. Suh, N. (1990) The Principles of Design. Oxford University Press.9. Ullman, D. (1992) The Mechanical Design Process, McGraw-Hill.10. Ulrich, K. & Eppinger, S. (2000) Product Design and Development. IrwinMcGraw-Hill.


CHAPTER 2

Identifying customer needs

If a new or redesign product is to be successful, it should fulfill the needs of thecustomer. Unfortunately, the process of finding which are the real needs to befulfilled is not a straightforward one. According to Ulrich & Eppinger (2000),the goals of a method for comprehensively identifying a set of customer needsshould be:

1. Ensure that the product is focused on customer needs.

2. Identify latent or hidden needs as well as explicit needs.

3. Provide a fact base for justifying the product specification.

4. Create an archival record of the needs activity of the development pro-cess.

5. Ensure that no critical customer need is missed or forgotten.

6. Develop a common understanding of customer needs among members ofthe development team.

The main purpose of identifying customer needs is to create a direct informa-tion link between customers and developers. The involvement of members ofthe design team (specially engineers and industrial designers) results essentialas they must have a clear view of how the product will be used by the end


25 2.1 Customer satisfaction

user. This direct experience will help the design team not only to discover thetrue needs of the customer, but also to create better concepts and to evaluatethem in a more accurate form.

In this chapter, the next 5 steps to effectively identify customer needs will bediscussed:

1. Gather raw data from customers.

2. Interpret the raw data.

3. Organize the needs into a hierarchy.

4. Establish the relative importance of the needs.

5. The review of the process and its results.

2.1 Customer satisfaction

In order to satisfy customers, a given product must fulfill customer expecta-tions about it. Even when finding which features are wanted by the customeris a difficult task since customers usually not mention them directly, customersatisfaction translates to the implementation in a given product as much de-sired features as possible. In order to better understand this relationship, theKano diagram may be of help.

2.1.1. The Kano diagram The Kano model shown in figure 2.1, shows therelationship between customer needs and satisfac-

tion in an easy to appreciate diagram ranking the customer satisfaction fromdisgusted to delighted.

The lower curve in Kano’s diagram is called the basic performance curve orexpected requirements curve. It represent the essentially basic functions orfeatures that customers normally expect of a product or service. They areusually unvoiced and invisible since successful companies rarely make catas-trophic mistakes. However, they become visible when they are unfulfilled.

The upper curve in Kano’s diagram is called the delighted performance curve orexciting requirements curve. They are a sort of “out of the ordinary” functionsor features of a product or service that cause “wow” reactions in customers.


2.1 Customer satisfaction 26

Disgusted

Delighted

Expected Perf

ormance

Curve

Absent

Delighted Performance Curve

Basic Performance Curve

Sati

sfac

tion

Function

Fully Implemented

Cus

tom

er

Figure 2.1: Kano diagram of customer satisfaction. After Otto & Wood (2001).

They satisfy customers when fulfilled. But they do not leave customers dis-satisfied when left unfulfilled. And they are invisible to customers since theyare not even known.

The center line of the Kano diagram is called the one-to-one quality or lin-ear quality line. It represents the minimum expectation of any new productdevelopment undertaking. It is related also to performance type issues suchas “faster is better.” These represent what most customers talk about. Thus,they are visible to the company and its competitors. The expected require-ments and exciting requirements provide the best opportunity for competitiveadvantage. Hence, ways to make hem visible and then deliver on them areneeded.

Kano’s diagram is often interpreted simply as a relationship model of expectedquality vs. excited quality. What is really important, however, is that the tar-get of customer satisfaction can not only invisible but also moving. Customerexpectations increase over time. This calls for a more complex analysis anddeeper market understanding.


27 2.1 Customer satisfaction

2.1.2. Types of customer

needs

According to Otto & Wood (2001) customer needsmay be profitably considered in general categoriesbased on how easy the customer can express them

and how rapidly they change. They can be classified in three categories: first,direct and latent needs which consider observability, second, constant andvariable needs which consider technological changes and finally, general andniche needs which consider variance in the consumers.

Direct needs These are the needs that, when asked about the productcustomer have no trouble declaring as something they are concerned about.These are easily uncovered using standard methods as the one that will bedescribed hereafter.

Latent needs These are the needs that typically are not directly expressedby the customer without probing. Customer typically do not think in modesthat allow themselves to express these needs directly. Latent needs are bettercharacterized as customer needs, not of the product, but of the system withinwhich the product operates. Other products, services or actions currentlysatisfy the needs directly. Yet, these needs might be fulfilled with a developingproduct, and doing so can provide competitive advantage.

Constant needs These needs are intrinsic to the task of the product andalways will be. When a product is used, this need will always be there. Suchneeds are effective to examine with customer needs analysis, since the cost canbe spread over time.

Variable needs These needs are not necessarily constant; if a foreseeabletechnological change can happen, these needs go away. These needs are moredifficult to understand through discussions with the customer, since the cus-tomer may not understand them yet.

General needs These needs apply to every person in the customer popula-tion. It is necessary for a product to fulfill these needs if it is to compete inthe existing market.

Niche needs These needs apply only to a smaller market segment withinthe entire buying population.


2.2 Gather data from customers 28

2.2 Gather data from customers

In order to obtain information from customers, several methods are available:interviews, questionnaires, focus groups, observing the product in use andfinally, be the customer oneself. In what follows, a brief description of eachone together with pros and cons is given.

Interviews One or more members of the design team interview a number ofcustomers, one at a time. Interviews are generally carried out in the environ-ment of the costumer where the product is used. They typically last for oneto two hours.

Questionnaires A list of important concerns, questions and criteria is pre-pared by the design team and sent to selected customers. Although this typeof survey is quite useful at later stages of the design process, at this stage theydo not provide enough information about the use environment of the product.It is also important to notice that not all needs may be revealed using thismethod.

Focus groups A group of 8 to 12 customers participate in a discussionsession facilitated by a moderator. Focus groups are typically conducted in aspecial room equipped with a two-way mirror allowing several members of thedevelopment team to observe the group. It is desired for the moderator to bea professional market researcher, but a member of the development team canalso perform as moderator.

Observing the product in use When watching a customer using an ex-isting product or perform a task for which a new product is intended, detailsabout customer needs can be reveled. Observation may be passive, leaving thecustomer to use the product without any direct interference or can be carriedout along with one of the design team members allowing the development offirsthand experience about the use of the product.

Be the customer In many situations, members of the design team mayperform as users of existing competitor products or, in later stages of thedesign process, of prototypes. Although this method is very cost effective andrelatively easy to perform as no persons outside the design team are involved,it posses two main problems. First, members of the design team may nothave the required skills or experience to accurately evaluate the product, andsecond, they may feel biased towards certain characteristics of the product.


29 2.2 Gather data from customers

Occasional User

Frequent User

Heavy−duty User

Lead Users UsersServiceCenters

Retailer orSales Outlet

Figure 2.2: Customer selection matrix. After Ulrich & Eppinger (2000).

From the above methods, research carried out by Griffin and Hauser (1993)reports that conducting interviews is the most cost and effort effective method.According to their report, one 2-hour focus group reveals about the samenumber of needs as two 1-hour interviews. They also report that interviewingnine customers for one hour each will obtain over 90% of the customer needsthat would be uncovered when interviewing 60 customers. These figures whereobtained when a single function product was being considered, and may changewhen considering multi-function products. According to Ulrich & Eppinger,as a practical guideline for most products, conducting fewer than 10 interviewsis probably inadequate and 50 interviews are probably too many.

2.2.1. Selecting customers Selecting customers is not always a straightfor-ward activity as many different persons may be

considered a “customer”. Consider, for example, all those products that arepurchased by one person and used by another. In all cases, it is important togather information from the end user, and then gather information from othertype of customers and stake-holders.

A customer selection matrix like the one shown in figure 2.2, is useful forplanning exploration of both market and customer variety. It is recommendedthat market segments be listed on the left side of the matrix while the differenttypes of customers are listed across the top. The number of intended customercontacts is entered in each cell to indicate the depth of coverage.



2.2.2. Conducting

Interviews

Ulrich and Eppinger provide some general hints for effec-tive customer interaction. First, they suggest to sketchan interview guide that help to obtain an honest expres-

sion of needs. This can not be stressed enough, the goal of the interview is toobtain customer needs, not to convince the customer of what he or she reallywants. Some helpful questions and prompts to use are:

• When and why do you use this type of product?

• Walk us through a typical session using the product.

• What do you like about the existing products?

• What do you dislike about the existing products?

• What issues do you consider when purchasing the product?

• What improvements would you make to the product?

Second, they suggest the following general hints for effective interaction withcustomers:

• Go with the flow. If the customer is providing useful information, do notworry about conforming to the interview guide. The goal is to gatherinformation data on customer needs, not to complete the interview guidein the allotted time.

• Use visual stimuli and props. Bring a collection of existing and com-petitors’ products, or even products that are tangentially related to theproduct under development. At the end of a session, the interviewersmight even show some preliminary product concepts to get customers’early reactions to various approaches.

• Suppress preconceived hypotheses about the product technology. Frequentlycustomers will make assumptions about the product concept they ex-pect would meet their needs. In these situations, the interviewers shouldavoid biasing the discussion with assumptions about how the productwill eventually be designed or produced. When customers mention spe-cific technologies or product features, the interviewer should probe forthe underlying need the customer believes the suggested solution wouldsatisfy.


31 2.2 Gather data from customers

• Have the customer demonstrate the product and/or typical tasks relatedto the product. If the interview is conducted in the use environment, ademonstration is usually convenient and invariably reveals new informa-tion.

• Be alert for surprises and the expression of latent needs. If a customermentions something surprising, pursue the lead with follow-up questions.Frequently, an unexpected line of questioning will reveal latent needsimportant dimensions of the customers’ needs that are neither fulfillednor commonly articulated and understood.

• Watch for nonverbal information. The design process is usually aimedat developing better physical products. Unfortunately, words are notalways the best way to communicate needs related to the physical word.This is particularly true of needs involving the human dimensions of theproduct, such as comfort, image or style. The development team mustbe constantly aware of the nonverbal messages provided by customers.What are their facial expressions? How do they hold competitors’ prod-ucts?

2.2.3. How to document

interactions

There are four main methods for documenting in-teractions with customers:

Notes Handwriting notes are the most common method of documenting aninterview. If a person is designated as notetaker, other person can concentratein effectively questioning the customer. The notetaker should try to capturethe answers of the customer in a verbatim form. If the notes from the intervieware transcribed inmediately after it, a very close account of the interview canbe obtained.

Audio recording Audio recording is probably the easiest way of docu-menting and interview. Unfortunately, many customers feel intimidated by it.Another disadvantage is that transcribing the recording into text is very timeconsuming.

Video recording Video recording is the usual way of documenting focusgroup sessions. It is also very useful for documenting observations of thecustomer in the use environment and the performance of existing products.

Still photography Even when dynamic information cannot be captured byit, still photography can be used to capture high quality images. It also has



Question

Type of user:Willing to do follow up?Address:Customer:Customer Data: Project/Product Name

Currently uses:Date:Interviewer(s):

Customer Statement Interpreted Need Importance

Figure 2.3: Customer data template. After Otto & Wood (2001).

the advantage of being inexpensive and easy to do.

One useful aid in the collection of data from a customer interview is a customerdata template. A customer data template, like the one shown in figure 2.3,helps to record questions, answers and comments. The template can be filledduring the interview or inmediately afterwards.

In the first column, the question prompted is recorded. In the second column,a verbatim description of the answer and comments given by the customer isrecorded. In the third column, the customer needs implied by the raw data arewritten. Special attention must be given to clues that may identify potentiallatent needs like humorous remarks, frustrations or non-verbal information. Inthe last column, linguistic expressions of importance that the customer mayhave used are recorded. The importance may be expressed in terms of wordslike must, good, should, nice or poor.

According to Otto & Wood, a must is used when a customer absolutely musthave this feature, generally when it is a determining criterion in purchasing


33 2.3 Interpret raw data

the product. Must ratings should be used very sparingly; only a few must ’sper customer interview is a good rule. A very important customer need shouldhave a good importance rating. Needs that are presumed should have at least ashould rating. If the customer feels the product should satisfy this requirement,it is important enough for the design team to consider it. The nice categoryis for customer needs that would be nice if the product satisfied them but arenot critical.

2.3 Interpret raw data

At this point, customer needs are expressed in terms of verbatim written state-ments. Every customer comment or observation as expressed in the customerdata template may be translated into any number of customer needs. It hasbeen found that multiple analysts may translate the same interview notes intodifferent needs, so it is convenient for more than one team member to beinvolved in this task.

Ulrich & Eppinger provide five guidelines for writing need statements. Theyrecognize the first two as fundamental and critical to effective translation, andthe remaining three as guidelines to ensure consistency of phrasing and styleacross all team members. Table 2.1 shows examples to illustrate each guideline.

• Express the need in terms of what the product has to do, not

in terms of how it might do it. Customers often express their pref-erences by describing a solution concept or an implementation approach;however, the need statement should be expressed in terms independentof a particular technological solution.

• Express the need as specifically as the raw data. Needs can beexpressed at many different levels of detail. To avoid loss of information,express the need at the same level of detail as the raw data.

• Use positive, not negative, phrasing. Subsequent translation of aneed into a product specification is easier if the need is expressed as apositive statement. This may not apply in those occasions when thestatement is expressed more naturally in negative terms.


2.4 Organization of needs 34

Guideline Customer Statement Need Statement - Right Need Statement - Wrong

“What” not “how” “Why don’t you put The screwdriver battery is The screwdriver battery

protective shields around protected from accidental contacts are covered by a

the battery contacts?” shorting. plastic sliding door.

Specificity “I drop my screwdriver all The screwdriver operates The screwdriver is rugged.

the time.” normally after repeated

dropping.

Positive not negative “It doesn’t matter if it’s The screwdriver operates The screwdriver is not

raining; I still need to work normally in the rain. disabled by the rain.

outside on Saturdays.”

An attribute of “I’d like to charge my The screwdriver battery can An automobile cigarette

the product battery from my cigarette be charged from an lighter adapter can charge

lighter.” automobile cigarette

lighter.

Avoid “must” “I hate it when I don’t know The screwdriver provides an The screwdriver should

and “should” how much juice is left in indication of the energy provide an indication of

the batteries of my level of the battery. the energy level of the

cordless tools.” battery.

Table 2.1: Examples illustrating the guidelines for writing need statements fora cordless screwdriver (After Ulrich & Eppinger, 2000).

• Express the need as an attribute of the product. Wording needsas statements about the product ensure consistency and facilitates sub-sequent translation into product specifications.

• Avoid the words must and should. The words must and shouldimply a level of importance for the need.

After all the customer comments have been translated into need statements,the design team ends up with a group of maybe tens or even hundreds of needstatements. At this point, some may be similar, other may not be technologicalfeasible, and others may express conflicting needs. In the following section,methods for organizing and classifying these needs are presented.

2.4 Organization of needs

2.4.1. Classification of

needs

In order to work effectively with all the customerneeds, it is necessary to classify them in groups ofequal or similar statements. Each group may be sub-sequently sorted out in a list according to the relative


35 2.4 Organization of needs

importance of each need in the group. Each group list typically consists of a setof primary needs, each one of which will be characterized by a set of secondaryneeds and if needed, tertiary needs.

This process of sorting and classification is normally performed by the designteam. Nevertheless, it also exists the possibility of leaving this task to a groupof selected customers. According to Otto and Wood, this approach preventsthe customer data from being biased by the development team.

The classification of needs can be done without many difficulties following thenext steps:

1. Write each need on a small card.

2. Group similar needs eliminating redundant statements.

3. Choose a descriptive name for each group.

4. Review the process and consider alternative ways of grouping the state-ments.

When working with different customer segments, cards with different colorlabels can be used to distinguish between them. The sorting and classifica-tion process can also be done separately for each customer segment observingdifferences in both the needs themselves and their organization. The latterapproach is best suited when the segments are very different in their needsand when there is the question if just one product may suit the needs of allsegments.


relative importance of

needs

As of now, the classification of needs does not pro-vide any information regarding the relative impor-tance that the customer place on different needs.Each customer need has an importance expressed

by the own customer during the interview. It is expected that different cus-tomers will feel different regarding the importance of features according totheir own use of the product.

An elementary approach to establish the relative importance of needs is tofirst construct a set of normalized weightings by comparing the number ofsubjects who mention a need versus the total number of subjects. Hence, the


2.5 Design brief 36

Importance Ranking 1 Ranking 2

Must 9 1.0

Good 7 0.7

Should 5 0.5

Nice 3 0.3

not mentioned 0 0

Table 2.2: Two different ranking systems for the importance of needs.

interpreted importance rank of the ith customer need can be obtained from

Ri =number of times mentioned

number of subjects(2.1)

It is important to have in mind that the above method may raise inconclusiveresults as it mainly measures the obviousness of the need as opposed to itsimportance. Therefore, needs that may be obvious but not important may beranked high as opposed to important needs that may not be obvious.

A more correct approach, is to include in the ranking the importance state-ments given by the customer during the interview. In order to do so, it isnecessary to convert the subjective importance ratings into numerical equiva-lents. A typical transformation is shown in table 2.2.

Once the mapping has been carried out, the importance assigned to each cus-tomer need can be calculated as:

Ri =average rating× number of times mentioned

number of subjects(2.2)

Although a better method of ranking customer needs, the previous methodhas also its own flaws as it still may hide important needs that were reveledby only few customers but were not seen by the rest.

2.5 Design brief

After grouping and ranking customer needs, a better idea of the design problemis at hand. To keep a clear idea of the direction of the design process, the design


37 2.6 Clarifying customer needs

team may issue what is called a design brief or mission statement. A designstatement includes a brief description of the product, key business goals, targetmarkets, assumptions and constraints and stakeholders:

• Description of the product A brief description typically includes thekey customer benefits of the product avoiding implying a specific productconcept.

• Key business goals. These goals generally include goals for time, cost,quality and market share. Other goals may be added as deem appropri-ate.

• Target markets. Identifies the primary as well as secondary marketsthat should be considered during the design process.

• Assumptions and constraints. In some projects is necessary to makeassumptions in order to keep a project of manageable scope and size. Inother occasions, time, cost or even feature constraints are known fromthe beginning of the product.

• Stakeholders. It is always convenient to list all the stakeholders inorder to handle subtle issues that may appear during the developmentprocess. Stakeholders are all the groups of people who are affected bythe success or failure of the product. The list usually begins with the enduser and the customer who makes the buying decision about the product.Stakeholders also include the customers residing within the firm such asthe sales force, the service organization and the production departments.

2.6 Clarifying customer needs

One step further in the determination of customer needs is to try to clarifyall the customer need that were grouped, classified and prioritized. In fact, itis very helpful to have the clearest possible idea of the customer needs at allstages of the design process. These customer needs, that will guide the designprocess, should be expressed in a form which is easily understood and whichcan be agreed by both, client and designer.


2.6 Clarifying customer needs 38

2.6.1. The objectives tree

method

The objectives tree method offers a clear and use-ful format for such a clarification of customerneed statements in form of objectives. It also

shows in a diagrammatic form the ways in which different objectives are re-lated to each other and the hierarchical pattern in which they are organized.As with many methods in the design process, the objectives tree is not asimportant as the procedure for arriving at it.

One way to start making vague statements more specific is to try to simplespecify what it means. Consider the following example provided by Cross(1994) where an objective for a machine tool must be ‘safe’. This objectivemight be expanded to mean:

1. Low risk of injury to operator.

2. Low risk of operator mistakes.

3. Low risk of damage to work-piece or tool

4. Automatic cut-out on overload.

These different statements can be generated simply at random as the designteam discusses about the objective. The types of questions that are useful inexpanding and clarifying objectives are simple ones like ‘why do we want toachieve this objective?’ and ‘what is the problem really about?’.

Some authors also include questions like ‘How can we achieve it?’ starting togive some insight about how the objectives may be accomplished. This givesway to statements like ‘automatic cut-out on overload’ which are not objectivesby themselves but means of achieving certain objectives.

Nevertheless, it is difficult to avoid making concessions reducing the scopeof the possible solutions that may be generated in later stages of the designprocess. For this reason, in the approach followed here, everything related tothe ‘how to’ accomplish objectives will be left to the concept generation stage.

As the list of objectives is expanded, it becomes clear that some are at higherlevels of importance than others. This relative importance may be representedin a hierarchical diagram of relationships as shown in figure 2.4.

In some cases, the relative position of each statement in the diagram may bea source of disagreement between the different members of the design team.



Machine mustbe safe

Low riskof injury tooperator

Low riskof operatormistakes

Low risk ofdamage to

workpiece or tool

Automaticcut−out onoverload

How

Why

Figure 2.4: Hierarchical diagram of relationships. After Cross (1994).

However, exact precision of relative levels is not important, and most peoplecan agree when only a few levels are being considered.

At this point, it is important to notice that the level of importance of thestatement should not be confused with the level of importance of the customerneed. Here, importance is related to the statements written to try to clarifyone objective, which correspond to one customer need.

In many cases, different people will draw different objectives trees for the sameproblem or the same set of objective statements. The tree diagram simplyrepresents one perception of the problem structure. It is only a temporaryrepresentation, which will probably change as the design process proceeds.

One more elaborated example of an objective tree is shown in figure 2.5 wherethe objectives tree for the design of a car door is shown.

The procedure of building an objectives tree can be summarized using thefollowing steps:



Provideprotection

Againstinjury

Whenclosing

Whenopen

Whenclosed

Correctamount

Safedirection

Safeforce

Safeinterior

Provideslatch

Securehandle

Resistimpact

Resistdamage

securelyLatches

Stronglatch

Inaccessiblelock

Againsttheft

Providesafety

Povideseal

Closedoor

Push/pulldoor

Pivotdoor

Keepweather out

Opendoor

Provideopening

Enablein/out

Figure 2.5: Objectives tree for a car door. After Pugh (1991).



1. Prepare a list of design objectives.

2. Order the list into sets of higher-level and lower-level objectives.

3. The expanded list of objectives and sub-objectives is grouped roughlyinto hierarchical levels.

4. Draw a diagrammatic tree of objectives showing hierarchical relation-ships which suggest means of achieving objectives.

2.6.2. The functional

decomposition method

From the objectives tree method, it is clear that de-sign problems can have different levels of generalityor detail. Hence, the level at which the problem is

defined is crucial and it is always appropriate to question the level at whichthe design problem is posed. On the other hand, focusing too narrowly on acertain level may hide a more radical or innovative solution.

In any way, it is useful to have means of considering the problem level at whicha design team is to work. It is also very useful if this can be done consideringthe essential functions that a solution will be required to satisfy. This approachleaves the design team free to develop alternative solution proposals that satisfythe functional requirements.

The function decomposition method offers such means of considering essentialfunctions and the level at which the problem is to be addressed. The essentialfunctions are those that the device, product or system to be design must satisfy,independently what physical components might be used to fulfill them.

The starting point of this method is to clarify what is the main purpose of thedesign. As it has been up to now, it is important what has to be achieved bythe new design and not how is going to be achieved. The most simple way ofrepresenting this main purpose is to draw a ‘black box ’ which converts certaininputs into desired outputs (see figure 2.6). This black box contains all thefunctions which are necessary for converting inputs into outputs.

At this point, it is preferable to try to make this overall function as broadas possible, avoiding to start with a narrow function that limits the range ofpossible solutions. In order to establish in an accurately way the requiredinputs and outputs as well as the ‘system boundary ’ which defines the functionof the product or device, questions like where do the inputs come from?, what



FunctionInputs Outputs

Black Box

Figure 2.6: A ‘black box’ system model. After Cross (1994).

Subfunction

Subfunction Subfunction

Subfunction

Transparent Box

FunctionInputs Outputs

Figure 2.7: A ‘transparent box’ model. After Cross (1994).

are the outputs for?, what is the next stage of conversion?, etc. can be madeto the customer.

Usually the conversion of the set of inputs into the set of outputs is a complexset of tasks occurring inside the black box. This complex set of tasks mustbe broken down into sub-tasks or sub-functions which linked together by theirinputs and outputs satisfy the overall function of the product or device beingdesigned. As this necessary sub-functions are establish, the black box is redrawas a ‘transparent box ’ (see figure 2.7).

According to Pahl and Beitz (2001), anyone setting up a function structureought to bear the following points in mind:

1. First derive a rough function structure with a few sub-functions fromwhat functional relationships you can identify in the requirements list,and then break this rough structure down, step by step, by the solutionof complex sub-functions. This is much simpler than starting out with



more complicated structures. In certain circumstances, it may be helpfulto substitute a first solution idea for the rough structure and then, byanalysis of that first idea, to derive other important sub-functions. It isalso possible to begin with subfunctions whose inputs and outputs crossthe assumed system boundary. From these it is possible to determine theinputs and outputs for the neighboring functions, in other words, workfrom the system boundary inwards.

2. If no clear relationship between the sub-functions can be identified, thesearch for a first solution principle may, under certain circumstances,be based on the mere enumeration of important sub-functions withoutlogical or physical relationships, but if possible, arranged according tothe extent to which they have been realized.

3. Logical relationships may lead to function structures through which thelogical elements of various working principles (mechanical, electrical,etc.) can be anticipated.

4. Function structures are not complete unless the existing or expected flowof energy, material and signals can be specified. Nevertheless, it is usefulto begin by focusing attention on the main flow because, as a rule, itdetermines the design and is more easily derived from the requirements.The auxiliary flows then help in the further elaboration of the design, incoping with faults, and in dealing with problems of power transmission,control, etc. The complete function structure, comprising all flows andtheir relationships, can be obtained by iteration, that is, by looking firstfor the structure of the main flow, completing that structure by taking theauxiliary flows into account, and then establishing the overall structure.

5. In setting up function structures it is helpful to know that, in the conver-sion of energy, material and signals, several sub-functions recur in moststructures and should therefore be introduced first. Essentially, the gen-erally valid functions are described next.Conversion of energy:

• Changing energy – for instance, electrical into mechanical energy.

• Varying energy components – for instance, amplifying torque.

• Connecting energy with a signal – for instance, switching on elec-trical energy.

• Channeling energy – for instance, transferring power.



• Storing energy – for instance, storing kinetic energy.

Conversion of material:

• Changing matter – for instance, liquefying a gas.

• Varying material dimensions – for instance, rolling sheet metal.

• Connecting matter with energy – for instance, moving parts.

• Connecting matter with signal – for instance, cutting off steam.

• Connecting materials of different type – for instance, mixing orseparating materials.

• Channelling material - for instance, mining coal.

• Storing material - for instance, keeping grain in a silo.

Conversion of signals:

• Changing signals – for instance, changing a mechanical into an elec-trical signal, or a continuous into an intermittent signal.

• Varying signal magnitudes – for instance, increasing a signal’s am-plitude.

• Connecting signals with energy – for instance, amplifying measure-ments.

• Connecting signals with matter – for instance, marking materials.

• Connecting signals with signals – for instance, comparing targetvalues with actual values.

• Channelling signals – for instance, transferring data.

• Storing signals – for instance, in data banks.

6. In the case of mechanical devices, table 2.3 can be a good starting pointto identify functions.

7. For the application of micro-electronics, it is useful to consider signalflows as shown in figure 2.5. This results in a function structure thatsuggests clearly the modular use of elements to detect (sensors), to acti-vate (actuators), to operate (controllers), to indicate (displays) and, inparticular, to process signals using microprocessors.



Technicalsystem

Operate

Detect

Indicate

Activate

Process(control)

User

Figure 2.8: Basic signal flow functions for modular use in micro-electronics.After Pahl and Beitz (2001).

8. From a rough structure, or from a function structure obtained by theanalysis of known systems, it is possible to derive further variants andhence to optimize the solution, by:

• braking down or combining individual sub-functions;

• changing the arrangement of individual sub-functions;

• changing the type of switching used (series switching, parallel switch-ing or bridge switching); and

• shifting in the system boundary.

Because varying the function structure introduces distinct solutions, thesetting up of function structures constitutes a first step in the search forsolutions.

9. Function structures should be kept as simple as possible, so as to lead tosimple and economical solutions. To this end, it is also advisable to aimat the combination of functions for the purpose of obtaining integratedfunction carriers. There are, however, some problems in which discretefunctions must be assigned to discrete function carriers, for instance,when the requirements demand separation, or when there is a need forextreme loading and quality.



Absorb/remove Dissipate Release

Actuate Drive Rectify

Amplify Hold or fasten Rotate

Assemble/disassemble Increase/decrease Secure

Change Interrupt Shield

Channel or guide Join/separate Start/stop

Clear or avoid Lift Steer

Collect Limit Store

Conduct Locate Supply

Control Move Support

Convert Orient Transform

Couple/interrupt Position Translate

Direct Protect Verify

Table 2.3: Typical mechanical design functions. After Ullman (2003).

The procedure to follow to establish the required functions and the systemboundary of a new design can be stated using the following steps:

1. Express the overall function for the design in terms of the conversion ofinputs and outputs.

2. Break down the overall function into a set of essential subfunctions.

3. Draw a block diagram showing the interaction between subfunctions.

4. Draw the system boundary. The system boundary defines the functionallimits for the product or device to be designed.

In order to effectively apply the functional decomposition method, the follow-ing guidelines should be followed:

1. Document what not how.

2. Use standard notation when possible.

3. Consider logical flows.



Tea beginBREWED

Hot tea

Tea leaves(waste)

Cold water

(measured quantity)

(measured quantity)

Tea leaves

Figure 2.9: Black box model of the tea brewing process. After Cross (1994).

4. Match inputs and outputs in the functional decomposition.

5. Break the function down as finely as possible.

One simple example that can be used to illustrate the process of functionaldecomposition is that of a tea maker (Cross, 1994). The fundamental processto be achieved by such a machine is to convert cold water and tea leaves intohot tea as illustrated in figure 2.9.

Some transparent box models of the tea maker are shown in figure 2.10. Thesemodels represent three alternative processes by which the overall function canbe achieved. After considering them, the designer settled on the first pro-cess where various necessary auxiliary functions became apparent, speciallyregarding the control of the heating and brewing processes.

References

1. Cross, N. (1994) Engineering Design Methods, John Wiley & Sons.2. Otto, K. & Wood, K. (2001) Product Design - Techniques in ReverseEngineering and New Product Development, Prentice-Hall.3. Pahl, G. and Beitz W. (2001) Engineering Design - A systematic Approach.Second Ed. Springer. 4. Ullman, D. (2003) The Mechanical Design Process.Third Ed. McGraw-Hill.5. Ulrich, K. & Eppinger, S. (2000) Product Design and Development. SecondEd. Irwin McGraw-Hill.



Water isheated

Water andtea united

Tea isinfusing

Tea and waterare separated

Water isheated

Tea leavesare immersed

Water isheated

Tea leavesare wetted

Concentrate andwater are united

Energy

Water

Water

Energy

Water

Energy

Tea leaves

Tea leaves

Tea

Leaves

Tea

Tea

Leaves

(a)

(b)

(c)

Figure 2.10: Three alternatives to the transparent box model for the tea brew-ing process. After Cross (1994).


CHAPTER 3

Benchmarking and Product Specifications

Benchmarking the competition as an activity in the product development pro-cess overlaps many of the other activities as it generates data that is importantto understand a product and forecast its future development. This activ-ity cannot be understated, product developers must learn from competitors.Companies must avoid the Not-Invented-Here (NIH) syndrome that presentswhen engineers at a company choose not to use technology developed outsideit as it is considered to not be of any good. This may cause a product to fail,as it leaves the design teams and companies behind as new technology emergesat the marketplace.

Design teams must understand the importance of newly introduced technologyby competitors and be ready to respond. Benchmarking allows to meet thisgoal. It is also an important step in establishing engineering specifications.

3.1 Benchmarking

There are two main purposes for studying existing competitive products: first,creates an awareness of what products are already available, and second, revealopportunities to improve what already exists. Design teams must be awarenot only on what other products offer, but also how other competitors providesimilar products. As Otto & Wood (2001) clearly state, when engineers think


3.1 Benchmarking 50

they understand their product by mere self-inspection, they are closing doorsto a wide array of alternative possibilities.

A famous example of understanding the competition is that of Xerox Corpo-ration. When in 1979 Xerox marketshare in the copy machines segment wasrapidly decreasing, its engineers pondered the following question: “How in theworld could the Japanese manufacture in Japan, ship it over to the UnitedStates, land it, sell it to a distributor who sells it to a dealer who marks up thecost to the final customer, and the price the customer pays is about what itwould cost us to build the machine in the first place?” (Jacobson and Hillkirk,1986). Even when at the time Xerox was not able to analyze and understandcompetitor’s product, production and distribution, they have now competitivebenchmarking activities. These activities allows them to focus on how to besuccessful, rather than how competitors can be better than them.

In order to understand the competition, design teams must tear down and an-alyze competitive products. This activity must be done periodically, not onlysupporting new design efforts but also developing a continuous understandingof trends and directions in technology development. Many large companieshave entire departments devoted only to benchmarking activities. These de-partments provide insight not only on new technological developments, butalso in the position of the company’s products in the marketplace in terms ofquality, value and performance.

Benchmarking activities are vital at all stages of the product development asthey:

• provide a way to understand what needs other products are satisfying

• provide means to establish product specifications ensuring that productsgoals superpass existing competition

• help in the concept generation stage providing best-in-class concepts

• help to incorporate in the detailed design new and improved design fea-tures of the best-in-class products

• help to find the best-in-class components and suppliers

According to Otto & Wood (2001), product benchmarking can be carried outfollowing the next steps:


51 3.1 Benchmarking

1. Form a list of design issues

2. Form a list of competitive or related products

3. Conduct an information search

4. Tear down multiple products in class

5. Benchmark by function

6. Establish best-in-class competitors by function

7. Plot industry trends

3.1.1. Form a list of design

issues

A list of issues must be developed for compara-tive benchmarking. Further, this list should becontinually revised and updated. With a focus

for benchmarking efforts, an efficient exploration path may be pursued. Theresult is a reduction in wasted time and resources.

3.1.2. Form a list of

competitive or related

products

Considering the design issues and product functionin product development, the next step is to exam-ine retailer stores and sales outlets for products thatdemonstrate these issues. For a product, it is neces-

sary to list all competitors and their different product models. In addition,all related products in their portfolio should be listed. If the competitorshave a family of products under a common platform (they use identical com-ponents for some aspects of each product but different components for nichedemands), detailed information about this should be included as it indicatethe competitor’s preferred market segments and compromises made for othermarket segments.

This step should only be an identification of the competitors in the form ofcompany names and product names. With a complete set of different products,vendors and suppliers to examine, the list should be screened by highlightingthe particular competitors that appear most crucial for the design team tofully understand. This step serves as basis for the next step, conducting aninformation search.


3.1 Benchmarking 52

3.1.3. Conduct an

information search

The information search is a step of great importance. Inorder to benchmark a piece of hardware, the design teammust gather as much information about the product as

possible. Any printed article that mentions the product, its features, its ma-terials, the company, manufacturing locations or problems, customers, marketreception or share, or any other information will be useful. Because of theproliferation of computerized databases and the World Wide Web, a good li-brary is essential. There is a generous amount of information available aboutall business operations. Before starting any design activity, a team must un-derstand the market demand for product features and what the competitionis doing to meet it. A design team should gather information on

• the products and related products

• the functions they perform

• the targeted market segments

All keywords associated with these three categories should be formed and usedin informational searches.

Sources of information can be quite varied. Most businesspersons are perfectlyhappy to discuss the market and noncompetitive business units. Althoughmost businesspersons will not provide strategic information about their owncompanies, many people are happy to tell all about their competitors. Sup-pliers will usually discuss their customers as they can, if it appears that therequester might provide an additional sale. The key is always to be open,honest and ethical when questioning for information. Once people understandthat a design team is designing a new product or redesigning an existing one,they naturally want to get involved with new orders and will help the team asfar as they legally can. Pursuit of information beyond that point is unethicaland not necessary. Most people are happy to share information, and so simplehonesty and a friendly attitude can get team members along way.

Sources of information can be divided in two main groups: public sources thatare freely accessible, and market research databases that are accessible througha fee.

Public sources of product information include:


53 3.1 Benchmarking

Libraries University libraries are filled with technical engineering modelingreferences. Many libraries that does not have a large book count, have accessto other larger libraries where information may be found and retrieved throughinter-library loans.

Thomas Register of Companies This set of documents is a “yellow pages”for manufacturing-related business. The Thomas Register list vendor by prod-uct name (http://www.thomasregister.com).

Consumer Reports Magazines These magazines survey and test a numberof common consumer products. Useful data are available for customer needs,qualitative benchmarking, engineering specifications, and warranty andmain-tenance information. If a given product is not covered in the magazines, otherproducts can provide analogies as a starting point.(http://www.consumerreports.com/,http://www.profeco.gob.mx/new/html/revista.htm).

Trade Magazines Consumer trade magazines such as Car and driver, Byte,Consumer Electronics, JD Powers and Associates, and others provide com-parative studies of products within a field. Such studies are very useful tounderstand how a given product compares with the competition and to under-stand important customer and technical criteria.

Patents After examining trade journals and uncovering which competitorshave new innovations, gathering the patents on these new innovations explainsmuch. Patent searches based on company names are difficult since companiestypically “bury” their patents by filing them under the individual names ofdesigners. Uncovering the individual patents is usually through refined top-ical searches, and hence, as much information as possible should be at handwhen doing the research. Patent information may be obtained from the Clas-sification and Search Support System (CASSIS) of from Web sites such ashttp://www.patents.ibm.com/.

Market Share Reporter Published every year by International ThomsonPublishers, this book summarizes the previous market research of Gale Re-search, Inc. It is composed of market research reports from the periodicalsliterature. It includes corporate market shares, institutional shares and brandmarket shares.

National Bureau of Standards This U.S. government branch provides,among other things, national labor rates for all major countries. This infor-


3.2 Setting product specification 54

mation proves very useful for determining competitors manufacturing costs.

Census of Manufactures Taken every 5 years by the U.S. Department ofCommerce, this census includes statistics on employment, payroll, inventories,capital expenditures, and selected manufacturing costs. Also, the supplementalCurrent Industrial Reports lists production and shipment data on industriesand some products.

Moody’s Industry Review Taken every 6 months, this survey provides keyfinancial information, operating data, and ratios on about 3,500 companies.Companies as an industry group may be compared with one another groupand against industry average.

3.1.4. Some comments

about benchmarking

Even when benchmarking can help to understand themarket, forecast trends and identify key innovationsand technology, one complaint about it is that always

provide lagging information. Hence, it is argued that market leaders can findlittle or no information at all through this practice.

Nevertheless, it should be realized that very few market leaders constantlyproduce leading technology in a market. Markets are always evolving andthe opportunity for a competitor to produce new exciting technology is alwayslatent. One way market leaders can benefit from benchmarking is from focusingit on components rather than in products. Components benchmarking mayallow them to introduce new technology in components that are not directlydeveloped by them.

One problem is commonly associated with benchmarking is the “chasing thecompetition” syndrome. This problem presents when benchmarking is onlyused to see what the competition is doing rather than to help the developmentof new competitive products.

3.2 Setting product specification

After benchmarking, one next step is to use the information gathered up to thispoint to set targets for a new product development effort. Specifications for anew product are quantitative, measurable criteria that the product should bedesigned to satisfy. In order to be useful, each specification should consist of a


55 3.2 Setting product specification

metric and a value. This value can be a specific number or a range. Examplesare: 50 Hz, 30-40 N, > 10 dB, etc.

In general terms, specifications fall into two categories, functional requirementsand constraints. As discussed before, functional requirements or engineeringdesign specifications are statements of the specific performance of a design,what the device should do. On the other hand, constraints are external fac-tors that limit the selection of the characteristics of the system or subsystem.Constraints are not directly related to the function of the system, but applyacross the set of functions for the system. In many situations, constraintscan drive the design process of a product and should be established only aftercritical evaluation.

Setting specifications is generally not a straightforward task, and specificationsare usually checked several times during the design process. Several conceptsmay be derived from a customer requirement giving rise to different engineer-ing specifications. Take for example a lid that can be either screwed or pushedto close a container. Both solutions will give way to different engineering spec-ifications since in the first case to screw is related to torque and in the secondone to push is related to force. In this case, early concept-independent criteriasuch as “opening ease” may be refined later into performance specifications forthe selected concept. In those specifications that are not expected to changeduring the design process, margins in target values of ±30% at the beginningof the design process are commonly expected.

In any case, it is primordial for each specification should be measurable, andtesting and verification of it should be possible at any stage. If for any reason,a specification is not testable and quantifiable, it is not a specification.

Ulrich and Eppinger (2000) suggest to consider a few guidelines when con-structing the list of specifications:

• Specifications should be complete. Ideally each customer needwould correspond to a single specification, and the value of that specifica-tion would correlate perfectly with satisfaction of that need. In practice,several specifications may be necessary to completely reflect a single cus-tomer need.

• Specifications should be dependent, not independent, variables.As do customer needs, specifications also indicate what the product mustdo, not how the specifications will be achieved. Designers use many types



of variables in product development; some are dependent, such the massof a product and other are independent, such as the material used tomanufacture the product. In other words, designers cannot control massdirectly because it arises from other independent decisions the designerwill make, such as dimensions and material choices. Metrics specify theoverall performance of a product and should therefore be the dependentvariables in the design problem. By using dependent variables for thespecifications, designers are left with the freedom to achieve the specifi-cations using the best approach possible.

• Specifications should be practical. It does not serve the team todevise a specification for a given product that can only be measuredby a scientific laboratory at a cost of several thousand dollars. Ideally,specifications will be directly observable or analyzable properties of theproduct that can be easily evaluated by the team.

• Some needs cannot easily be translated into quantifiable spec-

ifications. Needs like “the product instills pride” may be critical tosuccess, but are difficult to quantify. In this cases the team simply re-peats the need statement as a specification and notes that the metric issubjective and would be evaluated by a panel of customers.

• The specifications should include the popular criteria for com-

parison in the marketplace. Many customers in various markets buyproducts based on independently published evaluations (see examples ofsources in the previous section). If the team knows that its product willbe evaluated by the trade media and knows what the evaluation crite-ria will be, then it should include specifications corresponding to thesecriteria.

3.2.1. Specification Lists With the above guidelines, a specification list likethe ones shown in tables 3.1 and 3.2 can be gen-

erated. In order to help with the search for relevant design specifications, anapproach known as Specification List Generation can be of some help.

Specification List Generation uses the decomposition method to obtain a listof general specifications from latent needs such as safety, regulations and en-vironmental factors. Each specification can be labeled as a required demandor a desirable wish to communicate its level of importance.



To identify specifications, the table 3.3 devised by Franke (1995) provides agood starting point. In order to apply Franke approach follow the next steps:

1. Compile specifications and constraints and label them accordingly. Startwith specifications and follow with constraints.

2. Determine if each of the functional requirements and constraints is ademand or wish.

3. Determine if each of the functional requirements and constraints are log-ically consistent. Check for obvious conflicts. Check that specificationsare technically and economically feasible.

4. Quantify wherever possible.

5. Determine detailed approaches for ultimately testing and verifying thespecifications during the product development process.

6. Circulate specifications for comment and/or amendment inside and out-side the development team.

7. Evaluate comments and amendments.

3.2.2. Quality function

deployment

Up to this point, several pieces of information areavailable to the design team. Without proper guid-ance, the team may feel that is “lost in a see of infor-

mation”. One technique that is commonly used to help in the design processis Quality Function Deployment (QFD). One of the main advantages of theQFD method is that it is organized to develop the major pieces of informationnecessary to understand a design problem:



M. N. Metric Imp Units

1 1,3 Attenuation from dropout to handlebar at 10 Hz 3 dB

2 2,6 Spring preload 3 N

3 1,3 Maximum value from the Monster 5 g

4 1,3 Minimum descent time on test track 5 s

5 4 Damping coefficient adjustment range 3 N-s/m

6 5 Maximum travel (26 in. wheel) 3 mm

7 5 Rake offset 3 mm

8 6 Lateral stiffness at the tip 3 kN/m

9 7 Total mass 4 kg

10 8 Lateral stiffness at brake pivots 2 kN/m

11 9 Headset sizes 5 in

12 9 Steertube length 5 mm

13 9 Wheel sizes 5 List

14 9 Maximum tire width 5 in

15 10 Time to assemble to frame 1 s

16 11 Fender compatibility 1 list

17 12 Instills pride 5 Subj.

18 13 Unit manufacturing cost 5 US

19 14 Time in spray chamber without water entry 5 s

20 15 Cycles in mud chamber without contamination 5 cycles

21 16,17 Time to disassemble/assemble 3 s

22 17,18 Special tools required for maintenance 3 list

23 19 UV test duration to degrade rubber parts 5 hours

24 19 Monster cycles to failure 5 cycles

25 20 Japan Industrial Standards test 5 binary

26 20 Bending strength (frontal loading) 5 kN

Table 3.1: List of metrics for a mountain bike suspension. The relative impor-tance of each metric and the units for the metric are shown. “M.” and “N.”are abbreviations for the number of specification and the need it comes from.“Subj.” is an abbreviation indicating that a metric is subjective. (Adaptedafter Ulrich & Eppinger, 2000).



Date Demand/ Design specification Test/

Wish Verification

Functional Requirements

1/25 D Provide thrust for maximum height Bernoulli and conservation

(velocity > (20 m/s) of momentum analysis

1/25 D Maintain stable vertical flight path (less Flight tests with prototype

than 0.25 m deviation from vertical path) Design of experiments

Constraints

1/25 D Rocket length ≤ 15 cm Verify with engr. drawings

during concept generation,

detail design, etc.

1/26 D No detachable part less than Verify with dimensional

5 cm in diameter check of engr. drawings

Table 3.2: Specification sheet template, example of a toy rocket product (par-tial). Adapted after Otto & Wood (2001).

Specification category Description

Geometry Dimensions, space requirements, . . .

Kinematics Type and direction of motion, velocity, . . .

Forces Direction and magnitude, frequency,load imposed by, energy type,

efficiency, capacity, conversion, temperature

Material Properties of final products, flow of materials, design for manufacturing

Signals Input and output, display

Safety Protection issues

Ergonomics Comfort issues, human interface issues

Production Factory limitations, tolerances, wastage

Quality Control Possibilities for testing

Assembly Set by DFMA or special regulations or needs

Transport Packaging needs

Operation Environmental issues such as noise

Maintenance Servicing intervals, repair

Costs Manufacturing costs, material costs

Schedules Time constraints

Table 3.3: Categories for searching and decomposing specifications (AfterFranke, 1995).



1. The specifications of the product.

2. How the competition meets the goals.

3. What is important from the point of view of the customer.

4. Engineering specifications to work toward.

There are two points that are worth considering before applying QFD to adesign problem. First, no matter how well it is taught that a design problemis understood, the design team should employ the QFD method for all originaldesign or redesign projects. Second, the QFD technique can be applied to anentire product and its sub-systems.

To apply the QFD methodology, the following steps should be followed:

1. Identify the customers.

2. Determine the requirements of the customers.

3. Determine the relative importance of the requirements.

4. Perform a benchmarking activity to determine how competition satisfythe customers.

5. Generate engineering specifications.

6. Set engineering targets.

7. Relate the requirements of the customers to engineering specifications.

8. Identify relationships between engineering requirements.

Applying the above steps builds what is known as the house of quality. Thishouse provides in a single picture all the pieces of information gathered by thedesign team and their relationships. As shown in figure 3.1, the house hasmany rooms, each containing valuable information.



RequirementsWHAT

Customer

TargetsHow Much

Impo

rtan

ce R

atin

g

Cus

tom

er T

arge

tsan

d R

atin

gsN

ow v

s. W

hat

RelationshipMatrix

What vs How

SpecificationsHOW

Engineering Design

CorrelationMatrix

How vs How

Figure 3.1: Template for the House of Quality.



The first step for documenting information in the house of quality is to deter-mine the customer requirements and its relative importance. This informationcan be registered in the first room of the house: customer requirements. Thisroom relates to what the customers want.

The next step is to write down the information regarding the benchmarkingactivities carried out in the second room of the house: Customer targets andratings. This room relates to now vs. what or how the customer are currentlybeing satisfied.

In this step, each competing product must be compared with the requirementsof customers, rating each existing design on a scale of 1 to 5:

1. The product does not meet the requirement at all.

2. The product meets the requirement slightly.

3. The product meets the requirement somewhat.

4. The product meets the requirement mostly.

5. The product fulfills the requirement completely.

The benchmarking step is very important as it shows opportunities for bothproduct improvement and gain in market share. If all the competition rank lowon one requirement, that is clearly an opportunity, specially if the customerranked that specific requirement as essential.

After the engineering specifications have been generated, each one can be writ-ten in the third room of the house: engineering design specifications. This roomrelates how customer requirements will be measured to ensure satisfaction.

Hand in hand with the previous room is the targets room, which specify howmuch should be achieved. In this room all the target values related to eachone of the engineering design specifications are stated. In many cases, extremevalues for the delighted and disgusted states of customer satisfaction are alsoincluded for each specification.

After the previous steps have been carried out, only two more steps are missing,to relate the requirements of the customers to engineering specifications andto identify relationships between engineering requirements.

To relate the requirements of the customers to engineering specifications, theroom at the center of the house, the relationship matrix, is used. In this matrix,



Indicator Meaning Strength

� Strong relationship 9

© Some relationship 5

4 Small relationship 3

Blank Indicates no relationship 0

Table 3.4: Symbols used to indicate the level of relationship between customerrequirements and engineering design specifications.

Indicator Meaning Strength

⊕ Strong positive correlation 9

+ Some positive correlation 3

- Some negative correlation -1

Strong negative correlation -3

Table 3.5: Symbols used to indicate the level of correlation between engineeringdesign specifications.

each cell represents how an engineering specification relates to a customerrequirement. Although many parameters can measure more than one customerrequirement, the strength of the relationship can vary. The strength of therelationship is represented through the specific symbols shown in table 3.4.

To finish with the procedure, the roof of the quality house, the correlationmatrix is filled. Here, the relationship between different engineering specifica-tions is shown. The idea of the roof is to show that as one works to meet onespecification, you may be having a positive or negative effect on others. Forthis purpose, the symbols shown in table 3.5 may be used.

As the above steps are completed, the house of quality fills up. Figures 3.2 and3.3 show two different examples of houses of quality for two different products.

3.2.3. Comments on QFD

and the house of quality

One hint for effectively using the House of Qualityis that the matrix should not grow too large. Ifthe house is larger than 50 rows and columns,

then the design team should operate at different levels in the product. Anotheris to devote QFD as much time as needed. It may appear that QFD slows



down the design process, but it does not. Time spent developing informationis returned in time saved later in the process. Finally, it should be kept alsoin mind that QFD is a tool to build consensus. It is a tool to ensure that avariety of specifications from different areas converge to a successful product.

References

1. Jacobson, G. & Hillkirk, J. (1986) Xerox: American Samurai.2. Franke, H. J. (1975) Methodische Schritte beim Klaren konstruktiver Auf-gabenstellungen. Konstruktion. 27, 395-402.3. Otto, K. & Wood, K. (2001) Product Design - Techniques in Reverse Engi-neering and New Product Development, Prentice-Hall.4. Ullman, D. (2001) The Mechanical Design Process. Third Ed. McGraw-Hill.5. Ulrich, K. & Eppinger, S. (2000) Product Design and Development. SecondEd. Irwin McGraw-Hill.



9

5

5

3

3

2

2

ft3

Object Target values

Measurements Units

Technical Difficulty

Contain steam

Brew larger amount

Easy to store

Easy to clean

Easy to add tea

Stronger teaM

r. C

offe

e Ic

ed T

ea M

aker

Wes

t Ben

d C

offe

e M

aker

Old

Fas

hion

Way

Pow

dere

d T

ea

Wes

t Ben

d Ic

ed T

ea M

aker

4

3

3

3

4

3

4

1

2

3

3

3

2

4

5

2

2

2

5

4

2

5

2

2

4

5

4

5

1

3

3

3

3

3

1

West Bend

Mr. Coffee

WB Coffee Maker

Old Fashion Way

Absolute

Relative

Powered Tea

Objective

Measures

Importance

Technical

Tim

e w

ater

is in

con

tact

with

tea

Tem

pera

ture

of

wat

er in

sle

epin

g ba

sket

Tem

pera

ture

of

exiti

ng h

ot te

a

Vol

ume

of w

ater

in ta

nk

Tim

e ne

eded

to a

dd te

a

Tim

e to

cle

an p

rodu

ct

Tot

al V

olum

e

Lar

gest

siz

e of

bre

w

Hot

test

tem

pera

ture

out

side

con

tain

er

3 4 1 3 3 2 3 3 2

C sec cup C sec sec sec qt

98

98

?

98

99

na

83

8.0

<<

8

na

5

na

81

3

4.75

4.75

0

na

na

63

44

88

?

?

hot

25

45

0

5

1

5

5

5

45

30

20

20

20

240

60

27

0.2

0.4

0.2

0.4

na

na

36

3

2

2

1.5

na

na

27

?

100

?

?

100

25

18

1 2 3 4 4 6 5 6 8

Easy to add ice

Figure 3.2: House of quality for iced tea maker (partial). After Otto & Wood(2001).



Env

iron

men

tA

djus

tabi

lity

Perf

orm

ance

Rec

umbe

nt

Ene

rgy

tran

smitt

ed o

n st

anda

rd r

oad

Max

acc

eler

atio

n on

sta

ndar

d st

reet

Max

acc

el o

n 2.

5 cm

sta

ndar

d po

thol

e

Am

ount

of

chan

ge in

dam

ping

Am

ount

of

chan

ge in

spr

ing

rate

# of

tool

s to

adj

ust

# nu

mbe

r of

tool

s to

adj

ust

Rid

er h

eigh

t ran

ge

Rid

er w

eigh

t ran

ge

Rid

ers

that

not

ice

pogo

ing

Max

acc

el o

n 5.

0 cm

sta

ndar

d po

thol

e

Bik

eE C

T

Mou

ntai

n B

ike

Target (disgusted)

Target (delighted)

Recumbent

Mountain Bike

BikeE CT

Units

1

1

5

5

5

1

5

5

5

3

2

2

3

3

1

4

4

55

3

4

1

2

3

3

4

4

No noticeable water effect

No noticeable dirt effect

No noticeable temperature effect

Easy to adjust ride hardness

Easy to adjust for different heights

Easy to adjust for different weights

No pogoing

Eliminate shock from bumps

Smooth ride on streets

% gs gs gs % lbs in # # % %

95 0.4 1.6 3.0 0.0 6.0 0.0 0.0 0.0 0.0

35 0.1 0.4 0.5 20 30 4.0 2.0 1.0 0.0 0.0

50 0.1 0.7 0.9 40 40 6.0 1.0 1.0 0.0 0.0

30 0.1 0.4 0.5 6.0 0.0 1.0 0.0 0.0

50 0.2 0.7 1.0 50 50 3.0 1.0 1.0 0.00.0

100 100

100

Figure 3.3: House of quality for suspension system (partial). Adapted fromUllman (2003).


CHAPTER 4

Concept Generation

Up to now, all energies have been focused to understand the design problemand to develop its specifications and requirements. The goal now is to generateconcepts that will lead to a quality product. A concept may be defined asan idea that is sufficiently developed to evaluate the physical principles thatgovern its behavior (Ullman, 2003). Hence, concepts must be refined enoughto evaluate their form and the technologies needed to realize them. Conceptscan be represented in rough sketches, flow diagrams, set of calculations ortextual notes. In any case, each concept must contain enough details so thefunctionality of the idea can be ensured.

Sometimes, design begins with a concept to be developed into a product. Thisis considered to be a weak philosophy and generally does not lead to qualitydesign. In order to minimize changes later in the process, it is normally ex-pected that the concept generation scheme should take 20-25 percent of thedesign process.

It is very important for the design team to generate as many concepts aspossible, following the old advice:

Generate one idea, and it will probably be a poor one.

Generate twenty ideas, and you may have a good one.

It is a natural tendency to generate concepts as the design process progress,as it is naturally to associate ideas with things that we already known. This


4.1 Brainstorming 68

Concept generation methods The Morphological Chart

Basic Methods

TRIZ (TIPS)

Axiomatic designLogical methods

6−3−5 Method

Brainstorming

Figure 4.1: Some concept Generation methods.

creates a tendency for designers to take their first idea and start to refine ittoward a product. This is considered a very weak practice.

In order to avoid the previous problems, various methods aimed to generateconcepts are presented here. Some may be more complicated to follow or havetheir particular value. In any case, it is a decision of the design team which tofollow. Figure 4.1 shows the methods that will be discussed in this chapter.

It is important to recall once more the importance of information gathering.Considered the first method for concept generation, this activity should bereally the starting point of any concept generation method. This activity willinclude the search for documented ideas on solving problem functions that willincrease the scope of possible solution generated by the design team.

In the last chapter several sources of information were reviewed in the scopeof the benchmarking activity. This sources are also valid here, and figure 4.2shows a summary of them.

4.1 Brainstorming

Brainstorming is a group-oriented technique aimed to generate as many con-cepts as possible. The procedure is quite simple and has the advantage thata committed team can create a large number of ideas from different points ofview. The guidelines for brainstorming are as follows:


69 4.1 Brainstorming

Publishedmedia

InformationSources Textbooks

Product Information

Patents

Journals

Consumer Products Periodicals

Goverment Reports

Benchmarking

Analogies

Nature

Product Function

Product Architecture

Experts

Customers

Professionals in Field

People

Figure 4.2: Information sources for concept generation. Many of this informa-tion can be found through the World Wide Web.

1. Form a group with 5 to 15 people.

2. Designate a person to work as facilitator to prevent judgments and en-courage participation by all. Although some authors state that the fa-cilitator should also be a contributor. Other suggest that the facilitatorshould only guide and record the session avoiding further participation.The latter allows the designation of the most participate person of theteam as facilitator to encourage other team members to participate ac-tively.

3. Brainstorm for 30-25 minutes. The first 10 minutes are generally devotedto introduce the problem at hand. The next 20-25 minutes sees the mostgeneration of ideas, and during the last 10 minutes a sharp decline inideas may happen.

4. Do not allow the evaluation of ideas, just the generation of them. Thisis very important. Ignore any comments about the usefulness, validityor value of any idea.


4.2 The 6-3-5 method 70

5. Avoid confining the group to experts in the area, that may limit theintroduction of new ideas.

6. Avoid hierarchically structured groups. Bosses, supervisors and man-agers should not be included in many of the sessions.

Some hints may be used to stimulate new thinking and the generation of newideas:

• Make analogies, think what other devices solve a related problem, evenif they are applied to an unrelated area of application.

• Wish and wonder. Think wild. Sometimes silly, impossible ideas, giveway to useful ones.

• Use related and unrelated stimuli. First, use photos of objects or devicesthat are related to the problem at hand. Next, use photos of objectsunrelated to the problem. This activity usually gives way to new ideas.

• Set quantitative goals. Set a reasonable number of concepts and donot leave the session until you have achieve them. For a group session,individual targets of 10 to 20 concepts is reasonable.

4.2 The 6-3-5 method

One of the two main disadvantages with brainstorming is that first, all ideasare conveyed by words. Second, the generation of ideas can be dominated byone or two team members. The 6-4-5 method forces equal participation by all.

The guidelines for the 6-3-5 method are also very simple:

1. Arrange teams around a table. Although 6 members are optimal, anumber between 3 and 8 should suffice.

2. Establish a specific function of the product to work with.

3. Ask each member to draw in a sheet of paper two lines in order to createthree columns. After that, ask each member to write, 3 ideas, one oneach column, about how the function could be fulfilled. Ideas can becommunicated by words, sketches or both.


71 4.3 TRIZ

4. After 5 minutes of working in the concepts, pass the sheets of papers tothe right.

5. Give the participants another 5 minutes to add other three ideas to thelist.

6. After completing a cycle stop to discuss the results and find the bestpossibilities.

It is important to mention that there should be no verbal communication inthis technique until the end. This rules forces interpretation of the previousideas only from what it is on the paper.

4.3 TRIZ

The Teoriya Resheniya Izobreatatelskikh Zadatch (TRIZ) or Theory of Inven-tive Problem Solving (TIPS), was developed by Genrikh S. Altshuller in theformer U.S.S.R. at the end of the 1940’s. The TRIZ theory is based on theidea that many of the problems that engineers face contain elements that havealready been solved, often in a completely different industry for a totally unre-lated situation that uses an entirely different technology to solve the problem.Based on this idea, Altshuller collaborated with an informal collection of aca-demic and industrial colleagues to study patents and search for the patternsthat exist.

After spending 1500+ person-years studying at first around 400,000 patents(today the database extend up to 2.5 million patents), Altshuller discoveredthat they could be classified into five categories:

1. basic parametric advancement

2. change or rearrangement in a configuration

3. identifying conflicts and solving them with known physical properties

4. identifying new principles

5. identifying new product functions and solving them with known or newprinciples.


4.3 TRIZ 72

LevelDegree of

inventiveness

Percent of

solutions

Source of

knowledge

Approximate

number of

solutions to

consider

1 Apparent 32% Personal 10solution Knowledge

2 Minor 45% Knowledge 100improvement within company

3 Major 18% Knowledge 1000improvement within industry

4 New 4% Knowledge 100,000concept outside industry

5 Discovery 1% All that is knowable 1,000,000

Table 4.1: Levels of Inventiveness.

The first two categories were designated as “routine design”, meaning that theydo not exhibit significant innovations beyond the current technology. The lastthree categories represent designs that included inventive solutions. He alsonoted that as the importance of the innovation increased, the source of thesolution required broader knowledge and more solutions to consider before anideal one could be found. Table 4.1 summarizes this idea.


73 4.3 TRIZ

PROBLEMTO

SOLVE

1234

n

InventiveApply

PrinciplesSOLUTION

ContradictionMatrix

InventivePrinciples

Findcontradictions

TRIZ

Figure 4.3: TRIZ methodology.

Based on his studies, Altshuller observed some trends in historical inventions:

• Evolution of engineering systems develops according to the same pat-terns, independently of the engineering discipline or product domain.These patterns can be used to predict trends and direct search for newconcepts.

• Conflicts and contradictions are the key drivers for product invention.

• The systematic application of physical effects aids invention, since a par-ticular product team does not know all physical knowledge.

In this regard, Altshuller noticed that almost all invention problems involvedin one way or another the solution to a contradition. By contradition it isunderstood a situation in which the improvement of one feature means de-tracting another. The quality of the invention was in most occasions relatedto the quality of the solution to the contradiction.

Based on this premise, Altshuller devised TRIZ. The goal of using TRIZ isto find those contradictions that makes the design problem hard to solve.


4.3 TRIZ 74

1 Weight of movable object 21 Power

2 Weight of stationary object 22 Waste of energy

3 Lenght of movable object 23 Loss of substance

4 Lenght of stationary object 24 Loss of information

5 Area of movable object 25 Waste of time

6 Area of fixed object 26 Quantity of substance

7 Volume of movable object 27 Reliability

8 volume of stationary object 28 Measurement accuracy

9 Speed 29 Manufacturing precision

10 Force 30 Harmful action at object

11 Stress or pressure 31 Harmful effect caused by object

12 Shape 32 Ease of manufacture

13 Stability of the object’s composition 33 Ease of operation

14 Strength 34 Ease of repair

15 Durability of a moving object 35 Adaptation

16 Durability of a stationary object 36 Device complexity

17 Temperature 37 Measurement or test complexity

18 Illumination intensity 38 Degree of automation

19 Use of energy by moving object 39 Productivity

20 Use of energy by stationary object

Table 4.2: TRIZ 39 design parameters.

Then, use the 40 inventive principles of TRIZ to generate ideas to overcomethis problem. The 40 inventive principles were found by Altshuller to be theunderlying principles behind all patents. This procedure is depicted in figure4.3.

Applying TRIZ principles allows the innovation without having to wait for in-spiration. Practitioners of the TRIZ theory have a very high rate of developingnew, patentable ideas.


75 4.3 TRIZ

13. S

tabi

lity

of th

e ob

ject

’sco

mpo

sitio

n

13. Stability of the object’scomposition

19. Use of energy by movingobject

20. Use of energy by stationaryobject

9. Speed

8. Volume of stationary object

7. Volume of movable object

6. Area of stationary object

5. Area of movable object

3. Length of movable object

2. Weight of a stationary object

10. Force

11. Stress or pressure

12. Shape

14. Strength

18. Illumination intensity

17. Temperature

4. Lenght of a stationary object

1. Weight of a movable object

15. Durability ofa movable object

16. Durability of astationary object

2. W

eigh

t of

a fi

xed

obje

ct

3. L

engt

h of

fix

ed o

bjec

t

4. L

engh

t of

a st

atio

nary

obj

ect

5. A

rea

of m

ovab

le o

bjec

t

6. A

rea

of f

ixed

obj

ect

7. V

olum

e of

mov

able

obj

ect

8. V

olum

e of

fix

ed o

bjec

t

9. S

peed

10. F

orce

1. W

eigh

t of

a m

ovab

le o

bjec

t

11. S

tres

s or

pre

ssur

e

12. S

hape

What should be improved?

15 8

29 34

29 17

38 34

29

40 28

2 82

15 38

8 10

18 37

10 36

37 40

10 14

35 40

1 35

19 39

10

29 35

1 35 30

13 2

5 35

14 2

8 10

19 35 10 18

13 29 13 10

29 14

26 39

1 40

8 15

29 34

15 17

4

7 17

4 35

13 4

8

17 10

4

1 8

35

1 8

10 29

1 8

15 34

35 28

40 29

17 7

4010

35 8

2 14

28 10 14

35

13 14

15 7

39 37

35

2 17

29 4

14 15

18 4

7 14

17 4

29 30

4 34

19 30

35 2

10 15

36 28

5 34

29 4

11 2

13 39

30 2

14 18

26

399

7

1

1 18

35 36

10 15

36 37

2 38

2 26

29 40

1 7

174

29 4

38 34

15 35

36 37

356

36 37

1 15

29 4

1028

1 39

Wha

t is

det

erio

rate

d?

35 10

19 14

19 14 35

2 14

8 2 18

37

24 35 7 2

35

34 28

35 40

2 28

13 38

13 14

8

29 30

34

7 29

34

13 28

15 19

6 18

38 40

35 15

18 34

28 33

1 18

8 1

37 18

18 13

1 28

17 19

9 36

28 10 19 10

15

1 18

36 37

15 9

12 37

2 36

18 37

13 28

15 12

18 21

11

10 35

40 34

35 10

21

10 36

37 40

2913

1810

35 10

36

35 1

14 16

10 15

36 28

10 15

36 37

6 35 35 24 6 35

36

36 35

21

8 10

29 40

15 10

26 3

29 34

5 4

13 14

10 7

5 34

4 10

14 4

15 22

7 2

35

35 15

34 18

35 10

37 40

21 35

2 39

26 39

1 4

13 15

1 28

37 2 11 28 10

19 39

34 28

35 40

33 15

28 18

10 35

21 16

1 8

40 15

40 26

27 1

1 15

8 35

15 14

28 26

343

40 29

9 40 10 15

14 7

9 14

1517

8 13

26 14

10 18

3 14

19 5

34 31

2 19 3 17 10 2

19 30

3 35 19 2

6 27

19 16

1 40 35 34

36 22

6 38

22 35 15 19 15 19 3 35

39 18

35 38 34 39

40 18

35 6 2 28

36 30

35 10

213

19 35 19 32 19 32 13 10 13

19

26 19

13

10

39

28

9 19 5 16

3835

32 9 9 4

610

2

26163232

12 18

2 318

12 28

1 2

15 19

25

35 13

18

8 35

35

16 26

21 2

19 9

6 27

36 37

35 4

1015

35 33

2 40

34 15

1410

33 1

18 4

2 35

40

22 1

18 4

10 3

18 40

10 30

35 40

13 17

35

19 3

27

14 26

28 25

13 3

35

39 3

35 23

35 39

19 2

14 22

19 32

1 35

32

32 30 32 3

27

23

25

14 12 2

29

19 13

17 24

27 4

29 18

Figure 4.4: TRIZ contradiction matrix.Continued.


4.3 TRIZ 76

19. E

nerg

y ex

pens

e of

mov

able

obj

ect

16. D

urat

ion

of f

ixed

obj

ect’

s o

pera

tion

15. D

urat

ion

of m

ovin

g ob

ject

’s o

pera

tion

20. E

nerg

y ex

pens

e of

fix

ed o

bjec

t




9. Speed







10. Force


12. Shape

14. Strength


17. Temperature






Wha

t is

det

erio

rate

d?

14. S

tren

gth

17. T

empe

ratu

re

18. I

llum

inat

ion

21. P

ower

22. W

aste

of

ener

gy

23. L

oss

of s

ubst

ance

24. L

oss

of in

form

atio

n

25. W

aste

of

time

26. Q

uant

ity o

f su

bsta

nce

28 27

18 40

5 34

31 35

6 29

4 38

19 1

32

25 12

34 31

13 36

18 31

6 2

34 19

5 35

3 31

10 24

35

10 35

20 28

3 26

18 31

28 2

10 27

2 27

19 6

28 19

32 22

35 19

32

18 19

28 1

15 19

18 22

18 19

28 15

5 8

13 30

10 15

35

10 20

35 26

19 6

18 26

8 35

29 34

19 10 15

19

32 8 35

24

1 35 7 2

35 39

4

23 10

29 1 15 2

29

3529

14

28 26

1

35

3 35

38 18

3 25 12 8 6 28 10 28

24 35

24 26

24

30

14

29

3

40

15

14

6 3 2

16

15 15 32

19 13

19 32 19 10

32 18

15 17

30 26

10 35

2 39

30 26 26 4 29 30

6 13

40 2

19 30

10 35 39

38

17 32 17 7

30

10 14

18 39

30 16 10 35

4 18

2 18

40 4

9

15 7

14 6 35

4

34 39

10 18

10 13

2

35 35 6

13 18

7

13

15

16

36 39

34 10

2 22 2

34 10

29 30

7

9 14

1517

35 34

38

35 6

4

30 6 10 39

35 34

35 16

32 18

35

8

26 14

3

6

3

3 19

35 5

28 30

36 2

10 13

19

8

35 38

15 19 35

38 2

14 20

19 35

10 13

28 38

13 26 10 19

29 38

35 10

14 27

19 2 35 10

21

19 17

10

1 16

36 37

19 35

18 37

14 15 8 35

40 5

10 37

36

14 29

18 36

9 18

3 40

19

27

3 35 39

19 2

14 24

10 37

3510

14

2 36

25

10 36

373

37 36

4

141036

30 14

4010

14 26

9 25

22 14

19 32

13 15

32

2 6

34 14

4 6

2

14 35 29

3 5

14

10

34

17

36 22

17 9

15

13 27

10 35

39 3

35 23

35 1

32

32 3

1527

13 19 27 4

29 18

3532

27 31

14 2

39 6

2 14

30 40

35 27 15 32

35

27

26

3 30 10

40

35 19 19 35

10

35 10 26

35 28

35 35 28

31 40

29 3

28 10

29 10

27

27 3

10

19 35

39

2

354

19 28 6

35 18

10

35 38

19 28 27

3 18

10 1020

28 18 10 40

3 35

19 18

36 40

16 27

18 38

16 10 28 20

10 16

3 35

31

10 30

22 40

19 3

39

19 18

36 40

32 30

21 16

19 15

3 17

2

17 25

14 21 17

35 38

21 36

29 31

35 28

21 18

3

30 39

17

35 19 2 19

6

32

35

19 32 1

19

32 35

1 15

32 19 16

1 6

13 1 1 6 19 1

26 17

1 19

5 19

9 35

28 35

6 18

19 24

3 14

2 15

19

6

37 18

19 12 22

15 24

35 24

18 5

35 38

19 18

34 23

16 18

35 19 2

35 32

28 27

18 31

3 35

31

Figure 4.5: TRIZ contradiction matrix. Continued.


77 4.3 TRIZ




9. Speed







10. Force


12. Shape

14. Strength


17. Temperature






Wha

t is

det

erio

rate

d?

27. R

elia

bilit

y

28. M

easu

rem

ent a

ccur

acy

29. M

anuf

actu

ring

pre

cisi

on

30. H

arm

ful a

ctio

n at

obj

ect

32. E

ase

of m

anuf

actu

re

33. E

ase

of o

pera

tion

34. E

ase

of r

epai

r

35. A

dapt

atio

n

36. D

evic

e co

mpl

exity

38. D

egre

e of

aut

omat

ion

31. H

arm

ful e

ffec

t cau

sed

byob

ject

37. M

easu

rem

ent o

r te

stco

mpl

exity

39. P

rodu

ctiv

ity

3 11 1

27

28 27

35 26

28 35

26 18

22 21

18 27

22 35

31 39

27 28

1 36

35 3

2 24

2 27

28 11

29 5

15 8

26 30

36 34

28 29

26 32

26 35

18 19

35 3

24 37

10 28

8 3

18 26

28

10 1

35 17

2 19

22 37

35 22

1 39

28 1

9

6 13

1 32

2 27

28 11

19 15

29

1 10

26 39

25 28

17 15

2 26

35

1 28

15 35

10 14

29 40

28 32

4

10 28

29 37

15

17 24

17 15 1 29

17

15 29

35 4

1 28

10

14 15

161

1 19

26 24

35 1

26 24

17

24

26

16

14

29

4

28

15 29

28

32 28

3

2 32

10

18

1

1 15 17

27

2 25 3 1 35 1 26 26 30 14

7 26

29 9 26 28

32 3

322 22 33

28 1

17 2

18 39

13 1

26 24

15 17

13 16

15 13

10 1

15 30 14 1

13

2

26 18

36 14 30

28 23

10 26

34 2

32 35

440

26 28

32 3

2 29

18 36

27 2

39 35

22 1

40

40 16 16 4 16 15 16 1 18

36

2 35

1830

23 10 15

17 7

14 1

40 11

25 26

28

25 28

2 16

22 21

27 35

17 2

40 1

29 1

40

15 13

30 12

10 15 29 26 1 29 26

4

35 34

16 24

10 6 2

34

2 35

16

35 10

25

34 39

19 27

30 18

35 4

35 1 1 31 2 17

26

35 37

10 2

11 35

27 28

28 32

1 24

10 28

32 25

1 28

35 23

2 24

35 21

35 13

8 1

32 28

13 12

34 2

28 27

15 10

26

10 28

4 34

3 34

27 16

10 18

3 35

13 21

35 10

23 24

28 29

37 36

1 35

40 18

13 3

36 24

15 37

18 1

1 28

3 25

15 1

11

15 17

18 20

26 35

10 18

36 37

10 19

2 35 3 28

35 37

10 13

19 35

6 28

25

3 35 22 2

37

2 33

27 18

1 35

16

11 2 35 19 1

35

2 36

37

35 24 10 14

35 37

10 40

16

28 32

1

32 30

40

22 1

2 35

35 1 32

17 28

1 32 15

26

2 13 1 1 15

29

16 29

281

15 13

39

15 1

32

17 26

34 10

13 18 35 24

18 30

35 40

27 39

35 19 32 35

30

2

10

15

16

35 30

34 2

35 23 35

40 3

1

35

822

39 23

2 35

22 26

11 3 3 27

16

3 27 18 35

37 1

15 35

22 2

11 3

10 32

32 40

28 2

27 11

3

15 3 2

3

2 13

28

27 3

15 40

15 29 35

1410

11 2

13

3 3 27

4016

22 15

33 28

21 39

16 22

27 1

4

12 27 29 10

27

1 35

13

10 4

3529

19 29

39 35

6 10 35 17

14 19

34 27

6 40

10 26

24

142521135 1022117

40 33 6 35

1 20 10

3816

19 35

3 10

32 19

24

24 22 33

35 2

22 35

2 24

26 15 2835

226

1619

3

3135

27172

16

182

27

104

16

27 26 27

11 15

32

3 32 15 19 35

32 39

19 19 35

28 26

26

19

28 15 17

1613

15 1

19

6 32

13

32 15 2 26

10

2 25

16

19 21

2711

3 1

32

1 35

276

2 35

6

28 26

30

19 35 1 15

17 28

15 17

13 16

2 29

27 28

35 38 32 2 12 28

35

10 36

23

10 2

22 37

19 22

18

1 4 19 35

16 25

1 6



4.3 TRIZ 78

13. S

tabi

lity

of th

e ob

ject

’sco

mpo

sitio

n

2. W

eigh

t of

a fi

xed

obje

ct

3. L

engt

h of

fix

ed o

bjec

t

4. L

engh

t of

a st

atio

nary

obj

ect

5. A

rea

of m

ovab

le o

bjec

t

6. A

rea

of f

ixed

obj

ect

7. V

olum

e of

mov

able

obj

ect

8. V

olum

e of

fix

ed o

bjec

t

9. S

peed

10. F

orce

1. W

eigh

t of

a m

ovab

le o

bjec

t

11. S

tres

s or

pre

ssur

e

12. S

hape


Wha

t is

det

erio

rate

d?

21. Power

22. Waste of energy

23. Loss of substance

24. Loss of information

25. Waste of time

26. Quantity of substance

27. Reliability

28. Measurement accuracy

29. Manufacturing precision

30. Harmful action at object

31. Harmful effect caused by the object

32. Ease of manufacture

33. Ease of operation

34. Ease of repair

35. Adaptation

36. Device complexity

37. Measurement or test complexity

38. Degree of automation

39. Productivity

8 36

38 31

19 26

17 27

1 10

35 37

19 38 17 32

13 38

35 6

38

30 6

25

15 35

2

26 2

36 35

22 10

35

29 14

2 40

35 32

15 31

15 6

19 28

19 6

18 9

7 2

6 13

6 38

7

15 26

17 30

17 7

30 18

7 18

23

7 16 35

38

36 38 14

39

2

6

35 6

23 40

35 6

22 32

14 29

10 39

10 28

24

35 2

10 31

10 18

3139

1 29

30 36

3 39

18 31

10 13

28 38

14 15

18 40

3 36

37 10

29 35

3 5

2 14

30 40

10 24

35

10 35

5

1 26 26 30 26 30 16 2 22 26 32

10 20

37 35

10 20

26 5

15 2

29

30 24

14 5

26 4

5 16

10 35

17 4

2 5

34 10

35 16

32 18

10 37

36 5

36 37

4

4 10

34 17

35 3

22 5

35 6

18 31

27 26

18 35

29 14

35 18

15 2

17 40

14353610

314

1435

3

2935

2834

2015

29

182

440

1415

29

3 8

10 40

3 10

8 28

15 9

14 4

15 29

1128

17 10

14 16

32 35

40 4

3 10

14 24

2 35

24

21 35

11 28

8 28

10 3

10 24

35 19

35 1

16 11

32 35

2826

28 35

2625

28 26

5 16

32 28

3 16

26 28

32 3

26 28

32 3

32 13

6

28 13

32 24

32 2 6 28

32

6 28

32

32 35

13

28 32

13 18

28 35

27 9

10 28

29 37

2 32

10

28 33

29 32

2 29

18 36

32 28

2

25 10

35

10 28

32

28 19

34 36

3 35 32 30

40

30 18

30 18

22 21

27 39

2 22

13 24

17 1

39 4

1 18 22 1

33 28

27 2

3539

22 23

37 35

34 39

19 27

21 22

35 28

13 35

39 18

22 2

37

22 1

353

35 24

19 22

15 39

35 22

1 39

17 15

16 22

17 2

39

22 1

40

17 2

40

30 18

35 4

35 28

3 23

35 28

1 40

2 33

27 18

35 1 35 40

27 39

28 29

15 16

1 27

36 13

1 29

13 17

15 17

27

13 1

26 12

18

16 40 13 29

1 40

35 35 13

8 1

35 12 35 19

1 37

1 28

2713

11 13

1

25 2

13 15

6 13

1 25

1 17

13 12

1 17

13 16

1618

15 39

1 16

35 15

4 18

39 31

18 13

34

28 13

35

2 32

12

15 34

29 28

32 35

30

2 27

35 11

2 27

35 11

1 28

10 25

3 18

31

15 13

32

16 25 25 2

35 11

1 34 9 1 11

10

13 1 13

2 4

2 35

1 6

15 8

19 15

29 16

35 1

29 2

1 35

16

35 30

29 7

15 16 15 35

29

35 10

14

15 17

20

35 16 15 37

1 8

35 30

14

26 30

34 36

2 26

35 39

1 19

26 24

26 14 1

13 16

36 34 26

6

6 1 16 34 10

28

26 16 19 1

35

29 13

28 15

2 22

17 19

27 26

28 13

6 13

28 1

16 17

26 24

26 2 13

18 17

2 39

30 16

29 1

4 16

2 18

26 31

3 4

16 35

36 28

40 19

35 36

37 32

27 13

1 39

11 22

39 30

28 26

18 35

28 26

35 10

14 13

28 17

23 17 14

13

35 13

16

28 10 2 35 13 35 15 32

1 13

18 1

35 26

24 37

28 27

15 3

18 4

28 38

30 14

26 7

10 26

34 31

10 35

17 7

2 6

34 10

35 37

10 2

28 15

36

10 37

14

10 10

34 40

35 3

22 3910



79 4.3 TRIZ


Wha

t is

det

erio

rate

d?

21. Power

22. Waste of energy



25. Waste of time


27. Reliability







34. Ease of repair

35. Adaptation




39. Productivity

14. S

tren

ght

17. T

empe

ratu

re

18. I

llum

inat

ion

21. P

ower

22. W

aste

of

ener

gy

23. L

oss

of s

ubst

ance

24. L

oss

of in

form

atio

n

25. W

aste

of

time

26. Q

uant

ity o

f su

bsta

nce

15. D

urat

ion

of m

ovin

g ob

ject

’s

16. D

urat

ion

of f

ixed

obj

ect’

s

19. E

nerg

y ex

pens

e of

mov

able

obje

ct

oper

atio

n

oper

atio

n

20. E

nerg

y ex

pens

e of

fix

edob

ject

26 10

28

19 35

10 38

16 2

17 25

14 16 6

19

16 6

19 37

10 35

38

28 27

18 38

10 19 35 20

10 6

4 34

19

26 19 38

7

1 13

32 15

3 38 35 27

2 37

19 10 10 18

32 7

7

25

18

35 28

31 40

28 27

3 18

27 16

18 38

21 36

39 31

1 6

13

35 18

24 5

28 27

12 31

28 27

18 38

35 27

2 31

15 18

35 10

6 3

10 24

10 10 19 10 19 19 10 24 26

28 32

24 28

35

29 3

28 18

20 10

28 18

28 20

10 16

35 29

21 18

1

26 17

19 35 38

19 18

1 35 20

10 6

10 5

18 32

35 18

10 39

24 26

28 32

35 38

18 16

14 35

34 10

3 35

10 40

3 35

31

3 17

39

34 29

16 18

3 35

31

35 7 18

25

6 3

10 24

24 28

35

35 38

18 16

11 28 2 35

253

34 27

6 40

3 35

10

11 32

13

21 17

27 19

36 23 21 11

26 31

10 11

35

10 35

29 39

10 28 10 30

4

21 28

40 3

28 632

28 6

32

10 26

24

6 19

28 24

6 1

32 32

3 6 3 6

32

26 32

27

10 16

31 28

24 34

28 32 32

62

3 27 3 27

40

19 26 3 32 2 32 2 13 32

2

35 31

10 24

32 26

28 18

32 30

18 35

37 1

22 15

33 28

17 1

40 33

22 33

35 2

1

32 13

19

32

1 24

6 27

10 2

22 37

19 22

31 2

21 22

35 2

33 22

19 40

22 10

2

35 18

34

35 33

29 31

15 35

22 2

15 22

33 31

21 39

16 22

22 35

2 24

19 24

39 32

2

6

35 19 22

18

2 35

18

21 35

22 2

10 1

34

10 21

29

1 22 3 24

39 1

1 3

10 32

27 1

4

35 16 27 26

18

28 24

27 1

28 26

27 1

41 27 1

12 24

19 35 15 34

33

32 24

18 16

35 28

34 4

35 23

1 24

32 40

3 28

29 3

8 25

1 16

25

26 27

13

13 17

1 24

1 13

24

35 34

2 10

2 19

13

28 32

2 24

4 10

27 22

4 28

10 34

12 35

1 11

2 9

11 29

2728

1 4 10

13

115 15 1

28 16

15 10

32 2

15 1

32 19

2 35

34 27

32 1

10 25

2 28

10 25

35 3

32 6

13 1

35

2 16 27 2

3 35

6 22

26 1

19 35

29 13

19 1

29

1518

1

15 10

2 13

35 28 3 35

15

2 13

28

10 4

28 15

2 17

13

24 17

13

27 2

29 28

20 19

30 34

10 35

13 2

35 10

28 29

6 29 13 3

27 10

27 3

15 28

19 29

25 39

25 34

6 35

3 27

35 16

2 24

26

35 38 19 35

16

19 1

16 10

35 3

15 19

1 18

10 24

35 33

27 22

18 28

32 9

3 27

29 18

25 13 6 9 26 2

19 19

8 32 2 32

13

28 2

27

2823 35 10

18 5

35 33 24 28

35 30

35 13

29 28

10 18

35 10

2 18

20 10

16 38

35 21

28 10

26 17

19 1

35 10

38 19

1 35 20

10

28 10

29 35

28 10

35 23

13 1

235

35 38



4.3 TRIZ 80


Wha

t is

det

erio

rate

d?

21. Power

22. Waste of energy



25. Waste of time


27. Reliability







34. Ease of repair

35. Adaptation




39. Productivity27

. Rel

iabi

lity

28. M

easu

rem

ent a

ccur

acy

30. H

arm

ful a

ctio

n at

obj

ect

34. E

ase

of r

epai

r

35. A

dapt

atio

n

36. D

evic

e co

mpl

exity

38. D

egre

e of

aut

omat

ion

39. P

rodu

ctiv

ity

29. M

anuf

actu

ring

pre

cisi

on

31. H

arm

ful e

ffec

t cau

sed

byth

e ob

ject

32. E

ase

of m

anuf

actu

re

33. E

ase

of o

pera

tion

37. M

easu

rem

ent o

r te

stco

mpl

exity

19 24

26 31

32 15

2

32 2 19 22

31 2

2 35

18

26 10

34

26 35

10

35 2

10 34

1719

34

20 19

30 34

19 35

16

28 2

17

28 35

34

11 10

35

32 21

35 2

22 21 35

2 22

35 32

1

2 19 7 23 35 3

15 23

2 28 10

29 35

10 29

39 35

16 34

31 28

35 10

24 31

33 22

30 40

10 1

34 29

15 34

33

32 28

2 24

2 35

34 27

15 10

2

35 10

28 24

35 18

10 13

35 10

18

28 35

2310

10 28

23

22 10

1

10 21

22

32 27 22 35 33 35 13 23

15

10 30

4

24 34

28 32

24 26

28 18

35 18

34

35 22

18 39

35 28

34 4

4 28

10 34

32 1

10

35 28 6 29 18 28

32 10

24 28

35 30

18 3

28 40

2

28

3 3033 35 33

29 31

3 35

40 39

29 1

35 27

35 29

10 25

2 32

10 25

15 3

29

3 13

27 10

273

29 18

358 13 29

3 27

32 3

11 23

11 32

1

27 35

2 40

35 2

40 26 29 38

35111 13

27

4027

28

3513

1

3513

248

1111727

40

11

23

28 24

2622

3 33

39 10

6 35

25 18

1 13

17 34

1 32

13 11

13 35

2

27 35

3410

26 24

32 28

28 2

10 34

10 34

28 32

11 32

1

5

1

26 28

10 36

4 17

34 26

1 32

35 23

25 10 26 2

18

10 18

32 39

2826

2318

27 24

2 40

28 33

23 26

26 28

10 18

24 35

2

2 25

28 39

35 10

2

35 11

22 31

22 19

29 40

23 19

29 40

33 3

34

22 31

13 24

24 2

40 39

3 33

26

4 17

34 26

19 1

1

3 2 21

27 1

2 22 35

18 39

1 35

12 18

24 2 2 5

13 16

35 1

11 19

2 13

15

27 26

1

6 28

11 1

8 28

1

35 1

10 28

17 27

8 40

25 13

2 34

1 32

35 23

2 25

28 39

2 5

12

152612

321

34

1 16

32 25

1712

1 34

12 3

15 1

28

11 10

1 16

10 2

13

25 10 35 10

2 16

1 35

11 10

1 12

26 15

7 1 4

16

35 1

1113

34 35

7 13

1 32

10

35 13

248

35 5

1 10

35 11

32 31 31

1 15 34

1 16

1 16

7 4

15 29

37 28

27 34

35

35 28

6 37

12 17

28

35 18

5 12

35 26

5 12

35 26

34 21

15 1

24

35 18

27 2

34 27

25

15 10

37 28

15 10

37 28

15

10

24

12 17

28 24

1 35

28 37

27 4

1 35

1 15

29 15

28 37

1 13

13

1 35

12 26

1 1 32

10 25

28

7 19

1 12

34 3

2 5

9 26 24

26 27

13 1

27

1 13

5 28

11 29

1 26

13

28

2 24

3535 22

18 39

2

2

19 1

21

22 19

4029

22 19

29 28

2 33

22 35

13 24

32 1

18 10

28 26

18 23

26 24

32

2 26

10 34

26 24

2832

28 26

10 34

1 10

34 28

1 35

10 38

11 27

32

27 40

28 8

13 35

1



81 4.3 TRIZ

4.3.1. 39 design

parameters

From the patents studied, Altshuller extracted 39 designparameters that cause conflict. These 39 parameters arelisted in table . To effectively use these parameters, it is

necessary to find those two that are cause of conflict in a given design. Considerfor example that a non-moving mechanical component needs to be lighter butremain as strong. From the 39 design parameters, find the principle thatneeds to be changed, in this case, “Weight of stationary object” (principle #2).Then find the parameter that is negatively affected, in this case, “Strength”(principle #14). Then, using the contradictions matrix shown in figure 4.9,find those inventive principles that are candidates to solve the conflict. Fromthe contradictions matrix, the inventive principles that solve the contradictionbetween “Weight of a stationary object” and “Strenght” are 2, 10, 27 and 28.

4.3.2. Forty inventive

principles

Once the matrix has been used to find those inventiveprinciples candidates to solve the engineering contra-diction, they can be applied to generate solutions for

the problem at hand. These inventive principles can also be used indepen-dently of the contradiction matrix as a source of ideas to solve conflicts. Theforty TRIZ design principles to solve engineering conflicts are:

1. Principle of segmentation

• Divide an object into independent parts that are easy to disassem-ble.

• Increase the degree of segmentation as much as possible.

Examples:

• Sectional furniture, modular computer components, folding woodenruler, food processor.

• Garden hoses can be joined together to form any length needed.Drill shafts.

2. Principle of removal

• Remove the disturbing part or property of the object.

• Remove the necessary part or property of the object.

Examples:


4.3 TRIZ 82

• To scare birds from buildings and airports, reproduce the sound ofa scare bird using a tape recorder.

• Hovercraft.

3. Principle of local quality

• Change the structure of the object or environment from homoge-neous to non-homogeneous.

• Have different parts of the object carry out different functions.

• Place each part of the object under conditions most favorable forits operation.

Examples:

• Fuselage skin of commercial airplanes.

• Stapler. A pencil and an eraser in one unit.

4. Principle of asymmetry

• Make an object asymmetrical.

• Increase the object asymmetry.

Examples:

• Eccentric weight on motor creates vibration.

5. Principle of joining

• Merge homogeneous objects or those intended for contiguous (ad-jacent) operations.

• Combine in time homogeneous or contiguous operations.

Examples:

• TV/VCR, Cassette tape heads.

• The working element of a rotary excavator has special steam nozzlesto defrost and soften frozen ground in a single step.

6. Principle of universality

• Let one object perform several different functions.


83 4.3 TRIZ

• Remove redundant objects.

Examples:

• Sofa which converts from a sofa in the daytime to a bed at night.Fingernail clipper.

7. The nesting principle

• Place one object inside another, which in turn is placed in a third,etc.

• Let an object pass through a cavity into another.

Examples:

• Telescoping antenna, stacking chairs.

• Mechanical pencil with lead stored inside.

8. Principle of counterweight

• Compensate for the weight of an object by joining it with anotherobject that has a lifting force.

• Compensate for the weight of an object by interaction with an en-vironment providing aerodynamic or hydrodynamic forces.

Examples:

• Boat with hydrofoils, hot air balloon.

• Rear wings in racing cars to increase the pressure from the car tothe ground.

9. Principle of preliminary counteraction

• Perform a counter-action to the desired action before the desiredaction is performed.

Examples:

• Reinforced concrete column or floor. Reinforced shaft.

10. Principle of preliminary action

• Perform the required action before it is needed.


4.3 TRIZ 84

• Set up the object such that they can perform their action immedi-ately when required.

Examples:

• Cutter blades ready to be snapped off when old.

• Correction tape.

11. Principle of introducing protection in advance

• Compensate for the low reliability of an object by introducing pro-tections against accidents before the action is performed.

Examples:

• Fuses, electric breakers. Shaft couplers.

• Shoplifting protection by means of magnetized plates in products.

12. Principle of equipotentiality

• Change the conditions such that the object does not need to beraised or lowered.

Examples:

• Pit for change oil, Loading dock, airport gate.

13. Principle of opposite solution

• Implement the opposite action of what is specified.

• Make a moving part fixed and the fixed part mobile.

• Turn the object upside down.

Examples:

• Abrasively cleaning parts by vibrating the parts instead of the abra-sive.

• Lathe, Mill.

14. Principle of spheroidality

• Switch from linear to curvilinear paths, from flat to spherical sur-faces, etc.


85 4.3 TRIZ

• Make use of rollers, ball bearings, spirals.

• Switch from direct to rotation motion.

• Use centrifugal force.

Examples:

• Computer mouse.

• Screw lift.

15. Principle of dynamism

• Make the object or environment able to change to become optimalat any stage of work.

• Make the object consist of parts that can move relative to eachother.

• If the object is fixed, make it movable.

Examples:

• A flashlight with flexible neck.

• Bicycle drivetrain and derailer.

16. Principle of partial or excessive action

• If it is difficult to obtain 100% of a desired effect, achieve somewhatmore or less to greatly simplify the problem.

Examples:

• Raincoats, snowboards.

17. Principle of moving into a new dimension

• Increase the degrees of freedom of the object.

• Use a multi-layered assembly instead of a single layer.

• Incline the object or turn it on its side.

• Use the other side of an area.

Examples:


4.3 TRIZ 86

• A computer mouse where a 2D screen is transformed into a hori-zontal mouse pad.

• A composite wing where loads are in only one direction per layer.

18. Use of mechanical vibrations

• Make the object vibrate.

• Increase the frequency of vibration.

• Use resonance, piezovibrations, ultrasonic, or electromagnetic vi-brations.

Examples:

• Vibrating casting molds.

• Quartz clocks.

19. Principle of periodic action

• Use periodic or pulsed actions, change periodicity.

• Use pauses between impulses to change the effect.

Examples:

• Hammer drill.

• Emergency flashing lights.

20. Principle of uninterrupted useful effect

• Keep all parts of the object constantly operating at full power.

• Remove idle and intermediate motions.

Examples:

• Steam turbine, mechanical watch.

21. Principle of rushing through

• Carry out a process or individual stages of a process at high speeds.

Examples:


87 4.3 TRIZ

• Cutting thin wall plastic tubes at very high speeds so cutting actionoccurs before deformation.

22. Principle of turning harm into good

• Use harmful factor to obtain a positive effect.

• Remove a harmful factor by combining it with other harmful factors.

• Strengthen a harmful factor to the extent where it ceases to beharmful.

Examples:

• Medical defibrillator. Use of high frequency current to heat theouter surface of metals for heat treatment.

23. The feedback principle

• Introduce feedback.

• If feedback already exists, reverse it.

Examples:

• Air conditioning systems.

• Noise canceling devices.

24. The go between principle

• Use an intermediary object to transfer or transmit the action.

• Merge the object temporarily with another object that can be easilytaken away.

Examples:

• Gear trains.

25. The self service principle

• The object should service and repair itself.

• Use waste products from the object to produce the desired actions.

Examples:


4.3 TRIZ 88

• Nail resistant tires.

26. The copying principle

• Instead of unavailable, complicated or fragile objects, use a simpli-fied cheap copy.

• Replace an object by its optical copy, make use of scale effects.

• If visible copies are used, switch to infrared or ultraviolet copies.

Examples:

• Rapid prototyping. Crash test dummies.

• Measure shadows instead of actual objects.

27. Cheap short life instead of expensive longevity

• Replace an expensive object that has long life with many cheapobjects having shorter life.

Examples:

• Inkjet printer heads embedded in ink cartridges. Cardboard box.

28. Replacement of a mechanical pattern

• Replace a mechanical pattern by an optical, acoustical or odor pat-tern.

• Use electrical, magnetic or electromagnetic fields to interact withthe object.

• Switch from fixed to movable fields changing over time.

• Go from unstructured to structured fields.

Examples:

• CD player.

• Microwave oven. Crane with electromagnetic plate.

29. Use of pneumatic or hydraulic solutions

• Replace solid parts or an object by gas or liquid.

Examples:


89 4.3 TRIZ

• Power steering. Bubble envelopes.

30. Using flexible membranes and fine membranes

• Replace customary constructions with flexible membranes and thinfilm.

• Isolate an object from outside environment with thin film or finemembranes.

Examples:

• Dome tent. High Altitude Balloon.

31. Using porous materials

• Make the object porous or use porous elements.

• If the object is already porous, fill the pores in advance with someuseful substance.

Examples:

• Running shoe soles. Air filters.

32. The principle of using color

• Change the color or translucency of an object or its surroundings.

• Use colored additives to observe certain objects or processes.

• If such additives are already used, employ luminescence traces.

Examples:

• Transparent bandage. Roadway signs.

33. The principle of homogeneity

• Interacting objects should be made of the same material, or materialwith identical properties.

Examples:

• Shaft and bushing.

34. The principle of discarding and regenerating parts


4.3 TRIZ 90

• Once a part has fulfilled its purpose and is no longer necessary, itshould automatically be discarded or disappear.

• Parts that become useful after a while should be automatically gen-erated.

Examples:

• Multistage rockets. Bullet castings.

35. Changing the aggregate state of an object

• Change the aggregate state of an object, concentration or density,the degree of flexibility or its temperature.

Examples:

• Heat packs. Light sticks.

36. The use of phase changes

• Use phenomena occurring in phase changes like change of volumeand liberation or absorption of heat.

Examples:

• Fire extinguisher.

37. Application of thermal expansion

• Use expansion or contraction of materials by heat.

• use materials with different thermal expansion coefficients.

Examples:

• Thermometers. Bimetallic plates.

38. Using strong oxidation agents

• Replace air with enriched air or replace enriched air with oxygen.

• Treat the air or oxygen with ionizing radiation.

• Use ionized oxygen or ozone.

Examples:


91 4.4 The morphological chart

• Metal forming ovens. Torch cutting.

39. Using an inert atmosphere

• Replace the normal environment with an inert one.

• Carry out the process in a vacuum.

Examples:

• Aluminum cans for beverages. Arc welding.

40. Using composite materials

• Replace a homogeneous material with a composite one.

Examples:

• Steel belted tires. Tennis racquets. High performance aircraftwings.

4.4 The morphological chart

The aim of the morphological chart is to generate a complete range of alter-native design solutions for a product widening the search for potential newsolutions. It is based on the use of identified functions to foster ideas andhas two parts. First, to generate as many concepts as possible. Second, tocombine the individual concepts into overall concepts that meet all functionalrequirements.

The procedure to create and use a morphological chart is quite simple and canbe summarized as follows:

1. List the features or functions that are essential to the product. Eachfunction will be a row of the chart. The list of functions should containall the features that are essential to the product, at an appropriate levelof generalization. If a QFD procedure has been already performed, thelist of customer requirements can be used as the list of functions.


4.4 The morphological chart 92

Rotary motor withtransmission

Spring

Simgle impact

Moving mass

Multiple impacts Push nail

Solenoid Rail gunLinear Motor

ApplyTranslationalEnergy to Nail

AccumulateEnergy

ConvertElectrical Energy to

Translational Energy

Figure 4.10: Morphological table for a hand-held nailer. Adapted from Ulrich& Eppinger (2000).

2. For each feature or function, list the means by which it may be achieved.These lists will be the columns of the chart. Lists might include newideas as well as known solutions.

3. After the chart has been filled out, identify feasible combinations of sub-solutions. Each combination will be a possible solution to identify.

Figure 4.10 shows an example of a morphological chart for a hand nailer pre-sented by Ullrich & Eppinger (2000). The rows in the table correspond tothe functions identified by the design team: convertion of electrical energy totranslational energy, acummulation of translational energy and application oftranslational energy to the nail. The entries in each column correspond topossible solutions for the function at hand.

It is important to notice that in order for the chart to be most useful, theitems in the list of functions should all be at the same level of generality, andthey should be as independent of each other as possible. The list should notbe too long, however, and no more than 10 functions should be considered.Some authors advice to use no more than 4 functions at a time. If some func-tions are to be disregarded for this matter, the development team must clearlyunderstand the risks and tradeoffs of not taking them into consideration.

It is also advisable to arrange solution principles so that the columns createlogical grouping, for example, of mechanical type, of electrical type, etc. Also,sketches should be used whenever possible to convey as much information aspossible. Finally, consideration should be given only to solutions that meetthe estimated engineering specifications.



Once the chart is filled with solutions to all the specific functions listed, thenext step is to consider combinations from the range of all possible solutions.Usually a large number of combinations is possible, although restrictions applyas not all combinations of solutions are possible, for example, combinationsthat have intrinsic incompatibilities should be discarded.

It is essential to analyze very carefully each option before rejecting it. Thedesign team must have in mind that, initially many combinations may notseem to provide a practical solution to the problem at hand, specially to theinexperienced designer.

In the example shown in figure 4.10, 24 combinations can be found from theconcepts generated ( 4×2×3). Figures 4.11, 4.12 and 4.13 show the sketch offour possible solutions arising from the combination of concepts. The first so-lution, shown in figure 4.11, is due to the combination the concepts “solenoid”,“spring” and “Multiple impacts”. The second solution, shown in figure 4.12,results from the combination of “rotary motor with transmission”, “spring”and “multiple impacts”. The third, fourth and fifth solutions, shown in figure4.13, arise from the combinations of concepts “rotary motor with transmis-sion”, “spring” and “single impact”.



� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � ��

� � � � � ��


Spring

Simgle impact

Moving mass




AccumulateEnergy



Figure 4.11: Concept 1. Solenoid compressing a spring which is then releasedrepeatedly in order to drive the nail with multiple impacts. Adapted fromUlrich & Eppinger (2000).

��

� � � � �

� � � � �

� � � � �

� � � � �

� � � � �

� � � � �


Spring

Simgle impact

Moving mass




AccumulateEnergy



Figure 4.12: Concept 2 showing a possible combination of a motor with atransmission, a spring and multiple impacts. The motor repeatedly winds andreleases the spring, storing and delivering energy over several hits. Adaptedfrom Ulrich & Eppinger (2000).



��

��

��

��

��

��

� ��

� ��

� ��

� � � � � � � ��

CAM

MO

TO

R

TRIGGER


Spring

Simgle impact

Moving mass




AccumulateEnergy



Figure 4.13: Concept 3, 4 and 5 showing possible combinations of a motorwith a transmission, a spring and a single impact. The motor winds a spring,accumulating potential energy which is then delivered to the nail in a singlehit. Adapted from Ulrich & Eppinger (2000).



Gears andshafts

Turningwheels

Hydraulicram

Seated atfront

Chain orrope hoist

Airthrust

Airthrust

Reversethrust

Rack andpinion

Seatedat rear

Movingcable

Aircushion

Linearinduction

Drivenwheels

Remotecontrol

Bottledgas

Flexiblecable

Track

Petrol

Slides

Walking

Pedipulators

Steam

Means

Standing

Ratchet

Chains

Diesel

HydraulicTransmission

Steering

Stopping

Lifting

Operator

Power

Propulsion

Support

Feature

Wheels

Electric

Belts

Brakes

Screw

Rails

Figure 4.14: Morphological chart for a forklift truck, with one possible combi-nation of sub-solutions picked out by the dashed line (After Cross, 1994).

Two examples of morphological charts are presented by Cross (1994). The firstone is concerned with finding alternative versions of the conventional forklifttruck used for lifting and carrying loads. In the second one, alternatives forthe design of a welding positioner are explorer.

Regarding the finding of alternative versions of a lifting truck, the essentialfeatures of the truck are:

1. Means of support which allows movement.

2. Means of moving the vehicle.

3. Means of steering the vehicle.

4. Means of stopping the vehicle.

5. Means of lifting loads.

6. Location of the operator.

The morphological chart generated for the above functions is shown in figure4.14 where one possible solution is highlighted. It is interesting to note that



there are 90 000 possible combinations in the chart, although, some of themobviously, are not possible or include incompatible concepts.

In the second example, alternatives for the design of a welding positioner,a device used to support and hold a workpiece and locating it in a suitableposition are explored. Figure 4.15 shows the morphological chart for thiscase where several concepts are described by means of sketchs and text. Onepossible combination of concepts is indicated by the zig-zag line through thechart.

References

1. Altshuller, G. S. (1984) Creativity as an exact science. Gordon and BreachScience Publishers, New York, U.S.A.2. Cross, N. (1994) Engineering Design Methods, John Wiley & Sons.3. Otto, K. & Wood, K. (2001) Product Design - Techniques in ReverseEngineering and New Product Development, Prentice-Hall.4. Pahl, G. and Beitz W. (2001) Engineering Design - A systematic Approach.Second Ed. Springer.5. Ullman, D. (2001) The Mechanical Design Process. Third Ed. McGraw-Hill.6. Ulrich, K. & Eppinger, S. (2000) Product Design and Development. SecondEd. Irwin McGraw-Hill.



� ��

� ��

��

throughlocking(ratchet)

gear wheelpair

� � � �

� � � �

� � � �

� � � �

� � ��

straight lineguidance

sliding journalbearing

Rotationalguidance

Forminterlocking

hold directlyby hand,weight ofworkpiece

throughdrivemechanism

workpiece

ENABLEconnectionwith

ENABLErotationalmovement

ENABLEtiltingmovement

ENABLEheightadjustment

LOCKstate

CONTROLofmovement

DRIVE(by hand)

Showposition

line scale pointer scale

hole − pin

screw thread

sphere

rolling bearing

1 2 3 4 5 6

Action Principles / Families of Function carriers

mechanical

Force locking (friction)

screw or bolt

form interlocking force locking (friction)

with mechanical advantage device

mechanical

wedge pneumatic hydraulic magnetic

fulcrum pinposition

hang fromabove

levermechanism

optical mechanicalstop

electronic

ratchetmechanism

withinguidance

screwscrew with

washerwedge,brake block

rack andpinion

helical gears(crossed)

worm andworm−wheel

band,rope,chain

eccentriclever

cam

bearing, sliding or rolling

cylinder

Direct

Partial Functions

Figure 4.15: Morphological chart for a welding positioner, with one possiblecombination of sub-solutions picked out by the zig-zag line (After Cross, 1994).


CHAPTER 5

Concept Selection

In the previous chapter some methods to generate potential solutions for adesign problem where reviewed. Normally, a design team should generate tensor even hundred of ideas. Clearly not all ideas will lead to a successful product.However, at this point in time, with few information at hand, it is not possibleto say which concept is best. In concept selection the goal is to expend theleast amount of time and resources on deciding which concepts have the bestchances to become a successful product.

Concept selection requires the evaluation, of concepts with respect to somecriteria comparing their relative strengths and weaknesses in order to selectone or more concepts for further evaluation and testing. Here, evaluationshould be understood as the process of comparison and decision making. Theconcept selection phase usually requires at least three steps:

1. Estimate the technical feasibility of the concepts. All those con-cepts that are regarded as not feasible or ill-conceived are quickly dis-carded.

2. Concept screening. Concepts are compared roughly in relative termsagainst a common reference concept. Those that do not offer any advan-tage or fail to fulfill the requirements of the customer are discarded.

3. Concept scoring. A more detailed comparison is carried out includingmore information about the concepts for finer resolution.


5.0 100

DevelopmentandTesting

BriefDesign

ScreeningandScoring

Generation JudgmentFeasibility

GenerationandScreening

ScreeningandScoring

FinalConcept

Concepts

Testing

Figure 5.1: Concept selection is an iterative process that goes through differentphases and is closely related to concept generation and concept testing.

The last to steps are usually done in an iterative fashion, where conceptsare discarded and some other are combined to generate new concepts. Aftersufficient iterations have been carried out, the team goes to the next stage:concept testing. Figure 5.1 depicts the above procedure.

The iterative behavior of the design cycle is perhaps better represented by thedesign-build-test cycle shown in figure 5.2. In the diagram, two different designcycles are represented by the inner and outer loops. The first loop representthe design cycle when new or complex technologies are being use. In this case,building a physical model and testing it is the only approach possible.

The outer loop represent a more common approach where no physical de-vices are build until the very end of the process. Here, the time and expenseof building physical models is eliminated by developing analytical models andsimulating the concept before anything is build. All the iterations occurs with-out building any prototypes as all ideas are represented by means of analyticalmodels and graphical representations, usually with the help of computers.

Regardless of which design path is chosen, several benefits arises when a struc-tured approach is followed to select concepts. Probably the most most im-portant is that because concepts are compared against customer needs, theselected concept is likely to be focused on the customer. Other benefits mayinclude a reduced time to product introduction and effective decision making.


101 5.1 Estimating Technical Feasibility

Build finalproduct

Analytical modelsand graphical drawings

to refine concept and product

Simulatabletechnology

Designprototypes

Build prototypeswith each closer

to the final product

Test physicalprotoypes

Iterate

Iterate

BUILD

DESIGN

TEST

Figure 5.2: Design evaluation cycles. After Ullman (2003).

5.1 Estimating Technical Feasibility

When concepts are generated, members of the design team may experiencefeelings about the idea that can be grouped in three main reactions:

1. It will never work

2. It may work depending on something else

3. It is an idea worth considering

The above judgments regarding technical feasibility are based on the expe-rience of the design team and the individual engineers and their ability toestimate correctly. In general, it is safe to say that the more experience, themore chances the decision will be reliable at this point. Fortunately, estimatingis a skill that any person can learn and cultivate to a very good degree.

According to Otto and Wood (2001), the estimating skill of an engineer isdependent mainly on familiarity with dimensional units and with the differ-ent values along the dimensions. This familiarity takes place in two differentlevels of abstraction. First, perceived units like length or mass usually repre-sent no problem as everyone can associate their dimensions with day-to-dayexperiences. On the other hand, derived units like energy or power, usually


5.1 Estimating Technical Feasibility 102

represent a problem to estimate as different people will give values that canvary by orders of magnitude.

What allows an engineer to become more familiar with derived dimensions isto associate them with known values. For example, one might realize that2000W is 3hp, the common power for a lawnmower; or that 0.1MPa is 1atm,the atmospheric pressure at sea level. It is said that a skilled engineer willhave at least three readily understood reference levels for every dimensionalunit such as power, energy, pressure, force, acceleration, etc. Table 5.1 showssome approximated values for different units to be used as reference.

The “gut feeling” reactions to the generated concepts are worth exploring sincethey have different implications and may induce the individual or design teamto discard potentially good ideas or adapt potentially dangerous ones.

It will never work. Before discarding concepts that appears to be infeasibleor unworkable, consider it briefly from different points of view before reject it.As a guideline, before rejecting the concept answer the following questions:

• Why it is not technologically feasible?

• Does it meet the requirements of the customer?

• Is the concept different from the rest?

• Is the concept an original idea?

To answer the first two cases, where more attention is deserved, the methodsdescribed later in this chapter will be of help. In the case of the last two, itis worth analyzing if the reaction is a product of resistance to change or the“not invented here” syndrome.

It may work depending on something else. This reaction is product ofa doubt in the design team due to internal or external requirement that maybe judged to be either non-existent or not ready for consideration. Typicalquestion to be made in order to get insight of this reaction are:

• Is the technology needed available?

• Is the technology ready for production?

• Is all information needed readily available?

• Is is dependent on other parts of the product?


1035.1

Estim

ating

Tech

nical

Feasib

ility

Power Energy Mass Force Pressure Acceleration Velocity Length

(W) (J) (kg) (N) (MPa) (m/s2) (m/s) (m)

10−6 Ant crawling up Moving 5g snail: 1” × 1” piece Electrostatic attraction Moon surface: 0.13 × Centripetal acceleration of Tip speed of a wrist watch Human hair

a wall at 0.56µJ kinetic of paper: 40 between electron and 10−9 MPa a regular clock hour hour hand: 20µm/s. 10’ thickness:

1cm/s: 33µW energy ×10−6 kg proton in hydrogen hand: 0.3µm/s2 snow accumulation rate 30µm

atom: 0.08µN over 2 winter months:

0.1µm/s

10−3 LED: 40mW Bee in flight: Grape: 10g Piece of paper, weight: Blood pressure: 16 × Centripetal acceleration of Tip speed of a wrist watch Book cover

2 mJ kinetic Penny: 3g 0.04N 10−3MPa. Mars a regular clock minute minute hand: 1mm/s thickness: 2mm

energy atmosphere: 0.8× hand: 1 mm/s2 Speed of tide rising from

10−3MPa low to high: 0.1 mm/s

1 Small flashlight: A small apple Small meal or a Small apple, weight: 1N 10m underwater: 0.10 MPa Fast car acceleration: Falling body after 1/10s: Person’s height:

10W lifted 1m in large snack: Finger force for 1atm = 0.10MPa 3m/s2. Hard braking 1m/s. Walking speed: 2m

gravity: 1J. 1kg appliance buttons: 7N Piston engine car: 7m/s2. Earth gravity 1.5m/s

A small apple compression pressure: at sea level: 9.8m/s2

falling 1m: 1J 1.3 MPa

kinetic energy

100 Bright light 140km/hr fast ball: Average Bag of potatoes, weight: Piston engine firing Humans black out: 40 Highway speed: 30m/s Soccer field

bulb: 100W 114J kinetic person: 100N pressure: 3.5MPa m/s2. Belly flopping Jetliners: 250m/s length: 100m

Typical energy 70kg hard in water from a 10m Height of the

household diving board jump, Statue of

appliance: causing broken bones: Liberty: 93m

100-1000W 100m/s2.

103 Small lawn Energy effectively Mid-sized car: Two small people, Pressure to create a Marble dropped from 1m 3 times the speed of Width of a small

mower extractable 1300kg weight: 1.5kN diamond: 5GPa stopping in sand, sound: 1km/s town: 5km

engine: from a AA Elephant: Deep ocean trench: Head-on car collision

2000W battery: 1kJ 5000kg 0.11GPa occupant deceleration:

D-sized battery: 10km/s2. Bullet fired

80kJ from a rifle: 60km/s2

106 Electrical power Car @ 130km/h: A 747 fully Boeing 747: 1MN Center of the Earth: 0.40× Projectile fired from a rail Voyager 1 traveling in Dallas, TX to

to a small 1 MJ kinetic loaded: thrust. 106MPa gun: 800km/s2 outer space: 17km/s Denver, CO or

town: 1MW energy 300,000kg Boston MA to

Automotive Ocean liner: Pittsburgh PA:

Battery: 5MJ 107×106kg 1000km

109 Electrical power USS Nimitz Aircraft carrier: Saturn V or Space Center of the sun: Centrifugal acceleration of Speed of light in Earth to moon:

plant: 1GW 91,400 tons @ 0.5×109kg Shuttle: 20 ×109MPa light trapped in a black vacuum: 3.84×109m

30 knots: 9.9 0.2 GN thrust hole: 2×1013m/s2 3×108m/s

GJ kinetic energy

Table 5.1: Approximate reference values on different dimensions (adapted). After Otto & Wood (2001).

Copyrig

ht

c©2004

Dr.

Jose

Carlo

sM

iranda.

Todos

los

derech

os

reservados.

5.2C

oncep

tscreen

ing

104

Inth

efirst

two

cases,th

etech

nology

need

edsh

ould

be

carefully

exam

ined

tosee

ifit

isavailab

lean

dif

itis

already

matu

reor

canbe

by

the

time

the

pro

duct

reachpro

duction

.In

the

third

case,seriou

scon

sideration

shou

ldbe

givento

decid

eif

itis

worth

waitin

gfor

more

inform

ationto

becom

eavailab

le.Fin

ally,in

the

lastcase,

itsh

ould

be

pon

dered

ifoth

erparts

ofth

epro

duct

could

be

modifi

edto

accomm

odate

the

inten

ded

design

with

out

causin

ga

delay

inth

edesign

pro

cess.

Itis

an

idea

worth

consid

erin

g.

This

isgen

erallyth

ehard

estcase

toevalu

ate,sin

ceknow

ledge

and

experien

ceare

anim

portan

tpart

forth

eevalu

-ation

ofits

feasibility.

The

meth

ods

describ

edlater

inth

ech

apter

will

help

todevelop

adeep

erknow

ledge

abou

tth

econ

cept

inord

erto

evaluate

it.

5.2

Concept

screenin

g

Con

ceptscreen

ing

isa

techniq

ue

based

ona

meth

od

develop

edby

Stu

artP

ugh

and

isalso

know

nas

Pugh

concept

selectionm

ethod(P

ugh

,1990).

The

main

idea

ofth

etech

niq

ue

isto

narrow

the

num

ber

ofcon

cepts

quick

lycom

parin

gth

econ

cepts

betw

eenth

emselves

based

oncom

mon

criteriaan

dto

improve

the

concep

tsw

hen

everpossib

le.

Con

cept

screenin

gis

based

onth

efollow

ing

steps:

1.C

hoose

the

criteriafor

comparison

.

2.C

hoose

which

concep

tsw

illbe

evaluated

.

3.D

ecide

ona

reference

concep

tto

be

used

asa

datu

m.

4.P

repare

the

selectionch

art.

5.R

ateth

econ

cepts.

6.R

ank

the

concep

ts.

7.C

ombin

ean

dim

prove

the

concep

ts.

8.Select

one

orm

orecon

cepts.

Ste

p1:

Choose

the

crite

ria

for

com

pariso

n.

To

start,it

isnecessary

toknow

the

basis

onw

hich

the

diff

erent

concep

tsw

illbe

compared

with

each

Copyrig

ht

c©2004

Dr.

Jose

Carlo

sM

iranda.

Todos

los

derech

os

reservados.

105 5.2 Concept screening

other. During QFD, an effort was made to develop a set of customer require-ments. This requirements are generally well suited to be used as a criteria forcomparison. In some cases, when the concepts are well refined, engineeringtargets may be used instead.

Step 2: Choose which concepts will be evaluated. After concept gen-eration several options where available. These options where narrowed downdiscarding those concepts that were not technically feasible. From the optionsleft, choose the group to be evaluated. If more than 12 concepts are to beconsidered, the design team can vote to select the 12 concepts that will becompared.

Step 3: Decide on a reference concept to be used as a datum. Toselect a reference concept or datum, the design team can follow several ap-proaches. If the company already has a current product, it may serve well asa well understood concept. Other option is to use a competitive product thatthe team wish to superpass. Pugh (1990) recommends using as a datum theconcept that the team vote best.

Step 4: Prepare the selection chart. Once the criteria for comparison, theconcepts that will be evaluated and the datum all have been chosen, the nextstep is to prepare the selection chart. For that purpose the template shown infigure 5.3 can be of help.

Step 5: Rate the concepts. To rate the concepts, compare them against thedatum using a very simple scale. It is recommended to use a + if the conceptis better than the datum for the current criterion, a − if the concept is worsethan the datum and a 0 or an S (“same”) if the concept is judged to be aboutthe same as the datum or there is some ambivalence. If the decision matrix iscarried out in a spreadsheet use +1, 0 and −1 for scoring. It is advisable toto rate every concept on one criterion before moving to the next.

Step 6: Rank the concepts. After all the concepts have been rated foreach one of the criterion, four scores are generated: the number of +’s, thenumber of −’s, the number of 0’s and the net score. The net score is obtainedsubtracting the number of −’s from the number of +’s. To rank the concepts,simple use the one with the best net score as 1, the next one as 2 and so forth.

Step 8: Combine and improve the concepts. After the concepts havebeen rated and ranked, the design team should verify the validity of the results.Some recommendations for the interpretation of results are pointed out by


5.2 Concept screening 106

BConcept

CConcept

DConcept

EConcept

FConcept

Concepts

Datum

Continue?

Sum −’s

Sum 0´s

Sum +’s

Selection Criteria

Criterion 1

Criterion 7

Criterion 6

Criterion 5

Criterion 4

Criterion 3

Criterion 2

Net Score

Rank

Figure 5.3: Template for the Pugh selection chart.

Ullman (2003):

• If a concept or group of similar concepts has a good overall total scoreor a high + total score, it is important to notice what strengths they ex-hibit, that is, which criteria they meet better than the datum. Likewise,groupings of − scores will show which requirements are especially hardto meet.

• If most concepts get the same score on a certain criterion, examine thatcriterion closely. It may be necessary to develop more knowledge in thearea of the criterion in order to generate better concepts.

In many occasions, concepts can be combined to improve them. Here, to helpvisualize if concepts can be combined, Ullrich and Eppinger suggest to answerthe following questions:

• Is there a generally good concept which is degraded by one bad feature?


107 5.2 Concept screening

• Can a minor modification improve the overall concept and yet preservea distinction from the other concepts?

• Are the two concepts which can be combined to preserve the “betterthan” qualities while annulling the “worse than” qualities?

If any improved concepts arose from combination, these are added to the se-lection chart and ranked along the original concepts.

Step 9: Select one or more concepts. Once the above steps have beencarried out, and the design team is satisfied with their understanding of eachconcept, its strengths and weaknesses, it is time to decide which conceptsshould be selected for further refinement and analysis. The design team shouldalso clarify if issues need to be investigated further before a final decision canbe made. In addition, decisions should be made if the screening matrix hasprovided enough resolution and if another round of concept screening should beperformed. If concept screening has not provided enough resolution, conceptscoring should be applied next. la An example of a Pugh chart for the redesignof a coffee grinder is shown in figure 5.4. In this example, presented by Otto &Wood (2001), the goal was to evaluate different concepts all restricted to theuse of a chopper. Several ideas were developed to improve the grinder, focusingon cleaning functions. The criteria for the redesign evaluation gathered directlyfrom customer needs and engineering specifications are as follows:

• Cost: unit manufacturing cost (development and delivery costs were notconsidered). Measured in $.

• Store grinder: facility to put away in a cabinet out of sight. Measuredin cm3.

• Put in beans: Time elapsed between the beans are in a bag until thechopper switch can be activated. Measured in seconds.

• Take out coffee: Time elapsed between removing all grounds until allthe coffee is poured into a coffee maker. Measured in seconds.

• Power setup: Time elapsed between the grinder is plugged in until theswitch can be activated. Measured in seconds.


5.2 Concept screening 108

RemovableUnit

RemovableChamber

RemovableBlade

ScraperWashable

Cost

Development Risk

Cleanable

Power Setup

Take Out Coffee

Put in Beans

Store Grinder

Sum +’s

Sum 0´s

Sum −’s

Net Score

Rank

0

0

0

0

0

0

0

+

+

+

0

0

_

_

_

_

_

_

+

0

0

+

+_

_

_0

+

_

0

0

0

0

0

0

0

0

7

0

2

2

3 1

2

4

3

1

3 1

6

00 1 −3 −1

1 22 34

Selection Criteria

Figure 5.4: Pugh chart for coffee mill redesign concepts regarding cleanability.Adapted from Otto & Wood (2001).

• Cleanable: Time or steps needed from the point where the coffee hasbeen taken out until the point of being spotless. Measured in number ofsteps or seconds.

• Development risks: Difficulty getting a working alpha prototype. Mea-sured in number of potential faults or difficulties.

From the chart some conclusions could be drawn. First, the power setupcriteria does not distinguish between concepts as all of them were about thesame. Therefore, although it was an important criterion for the product, itdid not impact cleanability, and was dropped from further discussion. Next,the removable blade concept was clearly ahead of the rest and was a natural


109 5.3 Concept scoring

candidate for further development.

5.3 Concept scoring

Concept scoring is a technique very similar to concept screening and it is usedwhen increased resolution will better differentiate among concepts. In thismethod, the teams weight the relative importance of the selection criteria andfocuses on more refined comparisons with respect to each criterion. The stepsto use the method are as follows:

1. Choose the criteria for comparison.

2. Choose which concepts will be evaluated.

3. Decide on whether only one concept will be used as a datum or, if dif-ferent concepts will be used as reference for different criteria.

4. Prepare the selection chart and decide the weight for each criterion.

5. Rate the concepts.

6. Rank the concepts.

7. Combine and improve the concepts.

8. Select one or more concepts.

As most of the steps are identical to the concept screening ones, only thosedifferent will be discussed next.

Step 3: Decide on whether only one concept will be used as a datum

or if different concepts will be used as reference for different criteria.Although a single reference concept can be used for the comparative ratings ofall criteria as in the screening method, this is not always appropriate. Unlessby pure coincidence the reference concept is of average performance relativeto all criteria, the use of the same reference concept for the evaluation ofeach criterion may lead to what is known as the “scale compression effect”.Consider, for example, that the reference concept to be used as datum isbetter than the rest in 1 criterion. If this is the case, all the concepts couldbe evaluated only as “same as” or “worse than”, effectively compressing theevaluation scale to 2/3. This effect applies independently of the scale used, as


5.3 Concept scoring 110

WeightedScore

WeightedScore

WeightedScore

WeightedScore

Rating Rating Rating Rating

B C D

Selection Criteria Weight

Total Score

Rank

Continue?

Concepts

Datum

Dose metering accuracy

Ease of handling

Readability of settings

Ease of use

Durability

Ease of manufacture

Portability

Figure 5.5: Template for a concept scoring matrix.

it will be seen later. For this reason, many times different concepts are usedas reference for different criteria.

Step 4: Prepare the selection chart and decide the weight for each

criterion. The selection charts for the scoring method is very similar to Pughcharts with two exceptions. First, for each criterion, it includes its weight.Second, the chart includes two columns per concept: rating and weighted score.A template for an scoring chart is shown in figure 5.5. The weight for eachcriterion is usually defined as the percentage of importance that the criterionhas relative to the other criteria. Each percentage is defined such that the sumof all different percentages is 100%. An example illustrating the use of weightsis shown in figure 5.6 where three different cars are compared in base to fourdifferent criteria: fuel consumption, cost of spare parts, simplicity of servicingand comfort. Each criterion has its own weight defined by some chosen rules:fuel consumption weight is 50%, the cost of spare parts has a weight of 20%,easy to maintain 10% and finally, comfort 10%. It is easy to see in this examplethat the sum of all weights is 100%.

In many ocassions, the selection of the right weighting factors can be a cum-bersome task, specially if many different criteria have to be taken into account.One alternative is to use a objectives tree that includes weighting for each cri-terion. To show how objectives trees are constructed, consider the objective



Low fuel consumption

spare partsLow cost of

High comfort

Easy to maintain


50%

20%

10%

20%

Miles pergalon

Parameter

Cost of 5typical parts

Simplicityof servicing

Comfortrating

Car A Car B Car C

Score ScoreRating RatingValueValueRating ScoreValue

1.0

1.4

0.5

0.42

5

33

simple

£18

Very

Poor

7

40

£22

Com−plicated

Verygood

36

£28

Good

42

5

2

5

3

2

Average 3

4

2.0

1.0

0.2

1.0

1.5

0.4

0.3

0.8

Total Score 3.3 4.2 3.0

Rank 12 3

Figure 5.6: Scoring matrix for three alternative motorcars. Adapted fromCross (1994).

tree in figure 5.7. Each criterion in the objectives tree is represented by a circleor box with three numbers on it. At the top of each box, a number representsthe level of the criterion. For example, the set of criteria is level 1, representinga weight of 100% or 1. If there are three main criteria, then the first wouldbe represented by the number 11, the second by the number 12, the third bythe number 13, an so on. If the second criterion, number 12 has two criterionsthat must be considered, then the first one will be identified by the number121 and the second one by the number 122. If the criterion identified by thenumber 121 has to be divided into two different criteria, then the first wouldbe 1211 and the second one 1212. The objectives three can have as many levelsas necessary. The second number, at the lower left side of the box, indicatesthe weight of the factor to whom it belongs. The third number, at the lowerright end, is result of the multiplication of the weight of the criterion timesthe weight factor of its parent box. This product gives the contribution of thecriterion to the total 100%.

This procedure is better explained through an example. The above objectivestree was constructed to aid in the selection of weighting factors for the se-lection of a mechanical component. The main factors for the selection of thecomponent were specified as how safe the component was (criterion 11), itsmechanical behaviour (criterion 12) and and its cost of manufacturing (crite-rion 13). Since these three factors must add 100%, or 1 for short, then 1.0have to be divided between these factors. It was decided that safety accounted



1

12

121 122

1211 1212

+

13

131 132

11

1.0 = 0.25 + 0.22 0.22 0.15+ + 0.05 + 0.11

1.0 1.0

0.25 0.25 0.60 0.60 0.15 0.15

0.75 0.45 0.25 0.15 0.30 0.05 0.70 0.11

0.50 0.22 0.50 0.22

Figure 5.7: An objetive tree with weighting factors.

for 0.25, mechanical behavior for 0.60 and cost of manufacturing for 0.15.

Mechanical behavior was divided in two criteria, first, strenght (121) whichaccounted for 0.75 of the original 0.60 specified for mechanical behavior, andsecond, freedom from resonance (122) which accounted for 0.25 of the original0.60. It is important to stress at this point that the sum of weights of factorswhich have the same parent and are at the same level must be 1.0 or 100% (here0.75+0.25 = 1.0). It was considered to divided strenght further into two morespecific criteria, both with the same weight, stiffness (1211) and maximumallowable stress in the component (1212). Carrying out the products, the finalweight for stiffness is 0.22, which is the same value for the maximum stresscriterion.



WeightedScore

WeightedScore

WeightedScore

WeightedScore

Rating Rating Rating Rating

B C D

Weight

Safety

Stiffness

Max. allowable stress

Freedom from resonance

Cost of tooling

Cost of materials

Total ScoreRank

Continue?

0.25

0.22

0.22

0.15

0.05

0.11

ReferenceSCORING MATRIXBOTTOM−RIGHT HINGE

TEST RIG CONCEPTS

Selection Criteria

Figure 5.8: An objetive tree with weighting factors.

Finally, the cost of manufacturing was divided into two more specific costs:cost of tooling (131) and cost of materials (132). Cost of tooling had a weightof 0.30 whereas the cost of materials had a weight of 0.70. Hence, cost oftooling final weight is 0.05 and the final weight for cost of materials is 0.11.

As final observation about objectives trees, it is important to notice that thesum of weights of each level is always 1 as shown in figure 5.7 for the finallevel. In figure 5.8 the final template for the scoring matrix for this problemis presented.

Step 5: Rate the concepts. Similar to the procedure followed in the screen-ing method, here each concept is rated using a simple comparative scale. Asmore detail is needed, a more detailed rating scale is generally used. A commonoption is to use a 5 levels scale:

Relative performance Rating

Much worse than reference 1

Worse than reference 2

Same as reference 3

Better than reference 4

Much better than reference 5



DoNothing

FreshnerAir

Renuzit BakingSoda Chips

CedarWalls

VentedCarbon

Activated

Performance(olfactory distance − ft)

Cost


Rank

ScoreTotal

RatingRating ScoreScore Rating Score Rating Score Rating Score Rating Score

50%

25%

ReplacementFrequency of

Ease of replacement

12%

13%

0

100

100

100

0

25

12.5

12.5

70

52

70

0

35

13

9

0

70

76

90

50

35

19

11

6

80

84

67

40

40

21

5

8 100

100

20

100

10

25

13

13

90

20

20

67

45

5

3

8

50 57 71 74 61 61

12 3 456

Figure 5.9: Concept scoring matrix for the selection of odor control alterna-tives. Adapted from Otto & Wood (2001).

Sometimes a 10 or more levels scale is used, but its use is discouraged as itrequires more time and effort. For example, figure 5.10 presents a scoring ma-trix for an outpatient syringe where four different concepts are been evaluated.Note that in this example a 5 levels scale is being used and that different con-cepts serve as datum for different criteria. This example is different from theone presented in figure 5.6 where a 10 levels scale is used.

Step 6: Rank the concepts. Once the ratings have been specified for eachcriterion, the weighted score is obtained multiplying the rating for the weightof the criteria. The total score for each concepts is simply the sum of allweighted scores. Finally, each concept is given a rank corresponding to itstotal score.

Another example showing a more elaborated scoring matrix is presented byOtto & Wood (2001). The matrix, shown in figure 5.9, helped the designteam to evaluate between different alternatives to control the odors in a catlitter box. The careful reader should notice that the rating scale used inthis example goes from 0 to 100, which at first sight may look unnecessary.Nevertheless, in this case a 0-100 scale was chosen as the team had high-qualityinformation about the relative performance of each concept regarding each oneof the criteria. This high-quality information is usually gathered directly fromtesting and experimentation, and it is not influenced by the opinion of thedesign team members.


115 5.4 Concept Testing

5.4 Concept Testing

In the concept selection process, it is very likely that some form of customer’sresponse will be needed in order to further discuss the possibilities of theproposed concepts. In order to communicate the idea of the concept and tomeasure the response of the customer, in some cases simple verbal descriptionsor drawings will suffice. In other cases, there is no other choice but to createphysical prototypes of the product. This testing will give a better idea on thefeasibility of the concepts and the sales potential of the product.

Concept testing is carried out to facilitate decision-making during final conceptselection stages, generally after some detailed design has been done. Concepttesting is not necessary when:

• time required to test the concept is large relatively to the product lifecycle.

• cost of testing is large relative to the cost of actually launching the prod-uct.

Ulrich & Eppinger (2000) presents a 6 steps methodology for testing productconcepts:

1. Definition of the purpose of the concept test.

2. Choosing of a survey population.

3. Choosing of a survey format.

4. Communication of the concept.

5. Measurement of customer response.

6. Interpretation of results.

5.4.1. Concept testing

purpose

In this initial step, the design team should clarifywhat questions they want to answer with the test. Itis essential to define what the test or experiment is

for. Some typical questions are:


5.4 Concept Testing 116

• Which of several alternatives should be pursued?

• How the concept may be improved to better meet customer needs?

• Approximately how many units are likely to be sold?

• Should development be continued?

5.4.2. Choosing a survey

population

It is necessary to define the number of possiblecustomers to survey and in what market segmentsthey will be. This selection is carried out in a

similar fashion as the selection of customers in the “Identifying customer needsphase”. It is important, however, to have in mind that concept testing is amuch more expensive activity.

The most important question to answer is how large the survey populationshould be. Some useful guidelines are outlined next.

Factors favoring a smaller sample size:

• Test occurs early in the concept development process.

• Test is primarly intended to gather qualitative data.

• Surveying potential customers is relatively costly in time and money.

• Required investment to develop and launch the product is relativelysmall.

• A relatively large fraction of the target market is expected to value theproduct.

Factors favoring a larger sample size:

• Test occurs later in the concept development process.

• Test is primarily intended to assess demand quantitatively.

• Surveying customers is relatively fast and inexpensive.

• Required investment to develop and launch the product is relatively high.



• A relatively small fraction of the target market is expected to value theproduct.

Concept tests can be done in the early stages of the development process tosolicit feedback on the basic concept.

5.4.3. Choosing a survey

format

The following formats are commonly used in con-cept testing:

• face-to-face interaction

• telephone

• postal mail

• electronic mail

• internet

It is important to realize that each of these formats presents risks of samplebias.

5.4.4. Communicating the

concept

The way in which the concept will be surveyed,is closely related to the way in which the conceptwill be communicated. Communication of the

concept can be carried out by the following means:

• verbal descriptions

• sketch

• photos and renderings

• storyboard

• video

• simulation

• interactive multimedia

• physical appearance models

• working prototypes


5.4 Concept Testing 118

5.4.5. Measure customer

response

Most concept test surveys first communicate theproduct concept and then measure customer re-sponse. Although is good practice to include ques-

tions to measure customer response to product concepts, in many occasionsconcept test generally attempt to measure purchase intent. A useful scale tomeasure purchase intent may be:

• Definitely would buy

• Probably would buy

• Might or might not buy

• Probably would not buy

• Definitely would not buy

5.4.6. Interpreting results Usually interpretation is straightforward if thedesign team is just comparing two or more con-

cepts. It is important, though, to be sure that customers understood the keydifferences among concepts.

References

1. Cross, N. (1994) Engineering Design Methods, John Wiley & Sons.2. Otto, K. & Wood, K. (2001) Product Design - Techniques in ReverseEngineering and New Product Development, Prentice-Hall.3. Pugh, S. (1990) Total Design, Addison Wesley.4. Ullman, D. (2003) The Mechanical Design Process, Third Edition. McGraw-Hill.5. Ulrich, K. & Eppinger, S. (2000) Product Design and Development. IrwinMcGraw-Hill.



WeightedScore

WeightedScore

WeightedScore

WeightedScore

Rating Rating Rating RatingSelection Criteria Weight

Total Score

Rank

Continue?

Concepts

Dose metering accuracy

Ease of handling

Readability of settings

Ease of use

Durability

Ease of manufacture

Portability

A B C D

Master Cylinder Lever Stop Swash Ring Dial Screw

10%

20%

15%

25%

10%

15%

5% 0.2

2

2

0.3

0.6

0.3

0.75

0.2

0.45

0.15 3

4

3

5

3

3 0.3

0.6

0.75

0.75

0.3

0.6

0.15

3

2

4

2

5

4

4

0.3

0.4

0.6

0.5

0.5

0.6

3

2

3

5

3

4

0.3

0.4

0.45

0.75

0.5

0.45

0.2

3

3.05

2

3.103.45

14

2.75

No Develop No No

3

3

3

3

3

3

3

Figure 5.10: Concept scoring matrix for an outpatient syringe. The referencepoints for each criterion are signified by bold rating values. Adapted fromUlrich & Eppinger (2000).


CHAPTER 6

Embodiment design

As a design task, concept embodiment is perhaps the one that is most identifiedwith engineers as in this phase of the design process the choice of components,interfaces, materials, dimensions, shapes, tolerances, surface finishes, unionmethods, manufacturing and assembly processes, etc., are carried out.

In order to make wise choices, engineers should be able to understand throughlythe design, its functionality, objectives and constraints. Is in this stage

where engineers apply their skills in mathematics and basic science.Regardless of size, complexity or cost, products must be effectively modeled,tested and, whenever possible, refined. Methods for concept embodiment mustaid in this process.

6.1 Product Architecture

Product architecture is the the definition of the layout of systems, sub-systemsand components according to their functional purposes. This definition ofthe layout of the product must also deal with what interfaces are necessarybetween components, sub-systems and systems. Product architecture allowsthe design to be divided so building blocks can be assigned to individuals,teams or suppliers in order to permit parallel detailed design, testing andrefinement.


121 6.1 Product Architecture

Handheld vegetable peeler

Product

Wooden pencil

Kitchen knife

Swiss Army Knife

PC computer

Black and Decker cordless

drill

Tinkertoys

Fisted vegetable peeler

#1, #3 lead pencils

Kitchen knife set

Complex knives

More RAM, devices

VersaPak line of power

tools

Theme sets

Fixed unsharing

Modular platform

Modular platform

Modular platform

Adjustable for purchase

Modular platform

Adjustable for use

Common motor

Standard interfaces

Component variety

Common battery

Standard interfaces

Expanding width

Parametric handle size

Differing lead hardness

Shares no common components

Type

Integral

Modular

Derivatives Type Characteristics

for the Product and Common DerivativesProduct Architecture

Product Architecture

Figure 6.1: Product architecture examples. After Otto & Wood (2001).

Product architecture is also related what is called portfolio architecture. Port-folio Architecture relates to a group or family of products, where design strat-egy revolves around how to share components or subsystems across productsin the portfolio. Figure 6.1 shows some examples of product and portfolioarchitecture.

6.1.1. Types of product

architecture

In general terms, product architecture can be di-vided in two main types: integral architecture andmodular architecture. Each type has its own ad-

vantages and disadvantages as shown in figure 6.2.

A product has an integral architecture when no attempt is made to dividefunctions into components or systems resulting in on a very small number ofphysical elements carrying out all functions of the product. Integral architec-ture is common for high-volume products where cost is not reduced throughsharing components but through easiness of assembly. This results in productswith fewer components but much function sharing.

Modular product architecture is the result of dividing product functions intoa similar number of blocks or modules that perform a limited set of func-tions. Ideally, a one-to-one correspondence between modules and functions isachieved. In practice, modularity is not strict, and generally speaking, prod-ucts are neither fully modular or fully integral. Rather, a given product willpresent more or less modularity than another comparative product.

According to Ulrich (1995), modular architecture can be classified in three


6.1 Product Architecture 122

Improves device reconfigurability

Increases the device variety and speed of

introduction for new devices

Improves maintainability and serviceability

of device

Decouples development tasks (and

manufacturing to some extent)

Harder for competitors to copy design

Tighter coupling of team with less interface

problems

Increases system performance

Possible reduction in system cost

Pros

Makes imitation of device easier by competitors

May make devices look too similar

Reduces device performance

Modular design may be more expensive than

integral design

Hinders change of design in production

Reduces the variety of devices that can be

produced

Cons

Integral architecture

Modular architecture

Figure 6.2: Comparison of modular and integral architectures. After Otto &Wood (2001).

types: slot, bus and sectional. These three types, shown in figure 6.3, areexplained next.

Slot-modular architecture. Each of the interfaces between modules in aslot-modular architecture is of a different type from the others, so that thevarious modules in the product cannot be interchanged. An example of thistype of architecture are products that are to be assembled by the customerand are constructed in such a way that any given module can fit in only oneplace.

Bus-modular architecture. In a bus-modular architecture, there is a com-mon bus to which the other modules connect via the same type of interface.An example of bus-modular architecture are the floppy drive, DVD, CDRWand battery that connects to a bay in a laptop using the same interface.

Sectional-modular architecture. In a sectional-modular architecture, allinterfaces are of the same type, but there is no single element to which all theother modules attach. The assembly is built up by connecting the moduleseach other via identical interfaces. An example may be modular office furniturethat can be arrange in different ways depending on the modules used.

Slot-modular architectures are the most common type of modular architecturebecause for most products, different modules require different interfaces to



Slot−ModularArchitecture

Bus−ModularArchitecture

Sectional−ModularArchitecture

Figure 6.3: Three types of modular architectures. After Ulrich & Eppinger(2000)

accommodate unique interactions with the rest of the system.

6.1.2. Implications of the

architecture

Even when the architecture of a product is ini-tially defined, at least informally, since the con-cept generation stage, formal decisions are made

during the embodiment design phase. Product architecture is one of the devel-opment decisions that plays a major impact in the ability to deliver a variety ofproducts with standard components that allows better product performance,manufacturability and maintenance.

Product change

Modules are the building blocks of the product and the architecture defines howthis blocks relate to the function of the product and how the blocks interactwith each other. If each module is responsible for certain isolated functions, itwould be possible to replace or change any given module without affecting therest of the product. The contrary is true for an integral architecture, wherechanging one part of the product may have an influence in the functions carriedout by the rest of it. Some of the motives for product change are:

• Upgrades

• Add-ons

• Adaptation

• Wear

• Consumption

• Flexibility in use and reuse.



Product variety

Product variety refers to the amount of different products that any given com-pany can manufacture over a period of time. Product variety generally respondto market needs, as consumers want distinctive products. Product architecturecan help to achieve a large product variety for a minimum overhead in its cost.An example is Swatch watches, where hundreds of different combinations canbe achieved choosing different components during assembly.

Component standardization

Component standardization is the use of the same components or modules inmultiple products. This standardization allows the manufacturer to minimizecost and increase quality through the production of larger volumes and therefined design of such common components. An example are cars within sameor sister companies that share many parts and subsystems.

Product Performance

Product performance is related to how well the final product meets customerrequirements in terms of intended functions. Some examples of typical per-formance measures are speed, acceleration, efficiency, life, accuracy and noise.Here, architecture can facilitate the optimization of performance characteris-tics by means of integration and function sharing. Function sharing refers tothe implementation of multiple functions through a single module or compo-nent. Function sharing can help to optimize a design, but trade offs in theadvantages of modular architecture have to be considered by the design team.

Manufacturability

As discussed above, product architecture can influence the manufacturing costthrough product variety and component standardization. In addition, manydecisions regarding the architecture of a product influence the easiness of man-ufacturing as many complicated modules can be produced in larger volumesto reduce cost or many functions can be implemented in a single module toreduce either parts or manufacturing operations.



6.1.3. Establishing the

architecture

Due to the importance of product architecture insubsequent steps of the product design process, itshould be throughly discussed by the team design

and established in a cross-functional fashion. At the end of this step, an ap-proximate geometric layout of the product, description of the major modulesand documentation about the interaction between modules should be obtained.Ulrich & Eppinger (2000) suggest to follow a four steps approach that will beillustrated using a DeskJet printer as an example. The four recommendedsteps are:

1. Create a schematic of the product.

2. Cluster the elements of the schematic.

3. Create a rough geometric layout.

4. Identify the fundamental and incidental interactions.

Step 1: Create a schematic of the product. A schematic is a diagramshowing the constituent elements of a product as understood by the designteam. It is important that the schematic reflects the best understanding of theteam, although great detail is not necessary at this step. For the purpose ofestablishing product architecture, some authors recommend to aim for fewerthan 30 elements in the schematic. It also should be realized that there is nota unique schematic for any given product, so the team should generate severalalternatives to select from.

In figure 6.4 The schematic for a DeskJet printer is shown. Many elements inthe schematic represent physical components such as the print cartridge, whileother elements represent functional elements such as store output. Functionalelements are those that have not yet been reduced to physical concepts orcomponents, requiring further discussion by the design team in order to achievea final decision about how they will be implemented. On the other hand,components that have been reduced to physical components are generally thosethat are central to the basic product concept that the team has generated andselected.

Step 2: Cluster the elements of the schematic. In this step, the objectiveis to assign the different elements in the schematic into specific modules. Asin the previous step, the assignment of elements into modules is not unique,and the design team will be faced with different viable alternatives that canrange from the few to the hundreds.



EnclosePrinter

ProvideStructuralSupport

PrintCartridge

PositionPaper

in Y−Axis

"Pick"Paper

StoreOutput

StoreBlankPaper

ControlPrinter

AcceptUser

Inputs

DisplayStatus

SupplyDC

Power

CommandPrinter

CommunicatewithHost

Connectto

Host

PositionCartridgein X−Axis

Flow of material

Flow of forces or energy

Flow of signals or data

Figure 6.4: Schematic of the DeskJet printer. Note the presence of bothfunctional elements (e.g., “Store Output”) and physical elements (e.g., “PrintCartridge”). For clarity, not all connections among elements are shown. AfterUlrich & Eppinger (2000)



EnclosePrinter

ProvideStructuralSupport

PrintCartridge

PositionCartridgein X−Axis

PositionPaper

in Y−Axis

"Pick"Paper

StoreOutput

StoreBlankPaper

ControlPrinter

AcceptUser

Inputs

DisplayStatus

SupplyDC

Power

CommandPrinter

CommunicatewithHost

Connectto

Host

PrintMechanism

Power Cordand "Brick"

Host DriverSoftware

Enclosure

Chassis

Paper Tray

User Interface Board

Logic Board

Flow of material

Flow of forces or energy

Flow of signals or data

Figure 6.5: Clustering the elements into modules. Nine modules make up thisproposed architecture for the DeskJet printer. After Ulrich & Eppinger (2000)



One method to manage such complexity is for each element to be assigned toits own module and then to cluster elements when advantageous. Some factorsworth considering when clustering elements are:

• Geometric integration and precision. In some cases, it is convenientto cluster elements that control certain functions that are related betweenthemselves. Elements requiring precise location or close geometric inte-gration can often be best designed if they are part of the same module.In the case of the DeskJet, this principle would suggest clustering theelements associated with positioning the cartridge and the paper.

• Function sharing. When a single device can implement several differ-ent functions, it is best to cluster the related components together. Forthe DeskJet it was believed that that the status display and the usercontrols could be incorporated into the same component.

• Capability of vendors. If a specific vendor is know for its capacityin developing and manufacturing certain components, it is best if thosecomponents are cluster together. This will help the vendor to integratemore efficiently the said components.

• Similarity of design or production technology. When two or morecomponents are designed or manufactured using the same or similar tech-nology is best to cluster them in order to save costs. An typical exampleof the application of this principle is the clustering of several electronicdevices into a single circuit board.

• Localization and change. If the design team anticipates that a com-ponent will suffer several changes over time, it is best to isolate thatcomponent into its own module. In the case of the DeskJet, the engi-neers decided that the printer would suffer cosmetic shape modificationsand decided to isolate the enclosure into its own module.

• Accommodate variety. If some components of the product will bechanged to satisfy different market or operative conditions, it is best toisolate these components in a module that can be easily replaced. Forthe Deskjet, the engineers decided to isolate the components associatedwith the DC power supply as the printer was going to be sold in differentparts of the world with different power standards.



• Enabling standardization. When a component or components can beused in different products, it is best to isolated them in separate moduleor modules. This allows the higher production of the elements in themodule. An example of this standardization is the printer cartridge inthe DeskJet printer.

• Portability of the interfaces. Some interactions are more easily trans-mitted over large distances than others. For example, it is easier totransmit electric or light signals over a distance than mechanical forcesand motions. It is also true for the transmission of fluid connections.As a result, it is easier to separate elements with electronic and fluidinteractions. In the case of the DeskJet, the flexibility of electrical in-teractions allowed the design team to cluster control and communicationfunctions into the same chunk. On the other hand, the design team wasconstrained by the geometric and mechanical interactions of the paperhandling mechanism.

Step 3: Create a rough geometric layout.

The next step once the general components have been arranged in modules, isto generate a general geometric layout to analyze if the proposed distribution isphysically possible. This rough sketch can be made out of foam or cardboard,or even as a rough 3-D computer model, as it is not necessary to includegreat detail. Nevertheless, it should be sufficient to decide whether componentand interface distributions are possible. An example of a geometric layout forthe DeskJet printer is shown in figure 6.6. In this example, the design teamrealized that there was a trade off between the height of the machine and howmuch paper could be stored in the paper tray.

In many cases, the design team may decide that the geometric layout or theclustering chosen are not feasible. In this cases, components may be assignedto different modules.

Step 4: Identify the fundamental and incidental interactions.

It is common practice to divide the design so each module can be assignedto an specific person or team. Since the different modules interact in oneway or another, different persons or teams have to constantly coordinate theiractivities and exchange information. To facilitate this interaction, interactionsbetween modules have to be identified.



printcartridge

Printmechanism

Logicboard

Log

ic b

oard

Enclosure

Print cartridge

Roller/guide

PaperPaper tray

Chassis

User interface board

Paper tray

Chassis

Figure 6.6: Geometric layout of the printer. After Ulrich & Eppinger (2000)

According to Ulrich & Eppinger (2000) there are two types of interactionsbetween modules. First, fundamental interactions are those correspondingto the lines on the schematic that connect the chunks to one another (seeFigure 6.4). This interaction is planned and is fundamental to the operationof the system. Second, incidental interactions, are those that arise because ofparticular geometric or physical implementation of modules. In the Deskjetexample, the vibration from the actuator in the paper tray could interfere withthe precise location of the print cartridge in the x−axis.

Even when the principal interaction between modules were described in theschematic, incidental ones should be documented apart. When the systemincludes a reasonable small number of incidental interactions (less than 10),



Enclosure

Paper Tray

Chassis

PrintMechanism

ThermalDistortion

Logic Board

User InterfaceBoard

Power Cordand "Brick"

Host DriverSoftware

Thermal DistortionRF Shielding

Styling

Vibration

RF Interference

Figure 6.7: Incidental interaction graph. After Ulrich & Eppinger (2000)

an incidental interaction graph is convenient. Figure 6.7 shows an example ofan interaction graph regarding the DeskJet example. This graph shows thatvibration and thermal distortion are two incidental interactions that may affectthe performance of the print mechanism. The design team should be carefulto address these issues.

To define the interactions between modules, flows in material, energy andsignals must be investigated and refined at each module boundary. Theseflows usually define the interactions and the boundaries define the interfaces.According to Cutherell (1996), four types of interactions are typically investi-gated:

1. Material interactions: solid, liquids, or gases that flow from one moduleto the next.

2. Energy interactions: energies that must be transmitted or shielded be-tween modules.

3. Information interactions: signals (tactile, acoustic, electrical, visual, etc.)that must be processed from one module to the next, and

4. Spatial interactions: geometrical dimensions, degrees-of-freedom, toler-ances, and constraints that must be maintained between modules.


6.2 Geometry and layout refinement 132

6.2 Geometry and layout refinement

In the quest of creating a robust product, two main activities take place oncerough concepts have been generated and selected:

1. refining the geometry and architecture of the product,

2. systems modeling toward detail design.

Take for example the electric wok presented by Otto & Wood (2001) shown infigure 6.8. In this case the design team was faced with the task of improving anexisting product. As shown, the original concept of the wok evolved to a newone that included more advanced controls and configurations accommodatingimproved product cleaning and storage. As a result of the embodiment designphase, components, parts, assemblies and interfaces were clearly defined, fromboth geometrical, and functional points of view. Another example of the resultof embodiment design is shown in figure 6.9, where an exploded view of thePrestoTM hot air popcorn popper is shown.

In any case, the embodiment process include the following tangible documen-tation:

• detailed drawings,

• exploded views,

• assembly diagrams,

• tool designing,

• manufacturing process plans,

• tolerance design,

• packaging,

• maintenance and warranty information and

• user’s manual.

In order to generate the above documentation, some guidelines described nextmay be followed. The main objective of these guidelines is to transform aconcept sketch into refined geometry and material choices focusing on thefunctional performance of the product, including all relevant engineering spec-ifications.


133 6.2 Geometry and layout refinement

Figure 6.8: Embodiment example of a new electric wok concept. (a) Originalwok concept. (b) Original product realization. (c) Evolved wok concept. (d)Realization of new product concept. After Otto & Wood (2001)

In the embodiment design phase, specific layouts and parameters are generatedin order to logically chose a given concept from a number of solution alterna-tives that have been developed. Ideally, the result of this phase is a singledeveloped concept, in its definitive form, for the product or each subsystemdefined including:

• geometric layout

• material composition

• quality and manufacturability issues

• economics

In practice, the design team may be faced with the situation that furtherrefinement of the selected concepts is needed before commitment for a singlesolution occurs.



Figure 6.9: Embodiment example of the PrestoTM hot air popcorn popper.After Otto & Wood (2001)



In this situation, parallel development of the concepts should be carried out.It is important to try to select one final concept as soon as possible in orderto direct more resources for the development of the final product.

Without doubt, the main challenge of embodiment design is that when param-eters in the different subsystems or modules change, they usually affect othersubsystems or modules, they propagate. This behavior is the result of hav-ing parameters that are highly coupled between product subsystems/modules.This scenario means that embodiment design activities of the different sub-systems/modules must be carried out simultaneously and iteratively. As onechange is made, its effects in the other subsystems/modules are studied and, ifacceptable, the change is approved and its effects mitigated. The process stopwhen the performance of the product becomes acceptable.

In order to deal with these complex characteristics of embodiment design, Pahl& Beitz (1996) suggest a general process to iteratively refine the geometry andlayout of a product from an abstract form to a well-defined one. Figure 6.10illustrates this general process.

The process begins defining customer needs and the engineering specificationsthat fulfill them. Critical specifications/requirements that will drive the em-bodiment process are identified. Some examples of critical specifications are:

• size and geometric specifications

• material specifications

• arrangement specifications

After critical specifications have been selected, the next step is to draw ascale sketches of the product. These sketches should have enough detail toincorporate all critical aspects of the alternatives but care should be takento avoid over-constraining the models. These drawings includes the followingitems:

• maximum/minimum dimensions of the product

• clearance between relative subsystems/modules

• installation paths

• general arrangement of components relative to one another.



Usi

ng th

e en

gine

erin

g sp

ecif

icat

ions

,id

entif

y cr

ucia

l req

uire

men

ts

Dev

elop

pre

limin

ary

layo

uts

poss

ibili

ties

for

the

mod

ule

plac

emen

t

part

dra

win

gs, p

rodu

ct p

roce

ss p

lans

Prep

are

part

s lis

ts (

BO

M),

ass

embl

y dr

awin

gs,

Eva

luat

e ag

ains

t cus

tom

er, t

echn

ical

, rob

ust,

safe

ty, a

nd b

usin

ess

case

cri

teri

a

Com

plet

e fo

rm d

esig

n w

ith d

etai

led

draw

ings

Eng

inee

r th

e di

men

sion

s of

com

pone

nts

Che

ck f

or r

obus

tnes

s an

d er

rors

3. P

relim

inar

y L

ayou

t

Def

ine

the

spat

ial f

orm

with

pre

limin

ary

scal

e dr

awin

gs

Sele

ct a

nd e

ffec

tive

layo

ut f

or m

odul

e pl

acem

ent

Fill

in th

e sp

atia

l for

m w

ith s

cale

dra

win

gs

Inve

ntor

y re

quir

ed s

uppo

rtin

g fu

nctio

nalit

y

Dev

elop

pos

sibl

e la

yout

s fo

r su

ppor

ting

func

tions

Sele

ct a

n ef

fect

ive

layo

ut

Iden

tify

the

mai

n fu

nctio

nal m

odul

es

2. C

once

pt L

ayou

t

1. C

once

pt

4. D

efin

itiv

e L

ayou

t

Figure 6.10: A general process for concept embodiment. After Pahl & Beitz(1996)



Once the sketches have been completed, it should be verify that each productsubsystem, module, part or assembly fulfills its intended function completely.Also, the different sketches should be checked for possible geometry simplifi-cations and function sharing.

Based on the results, alternative sketches should be generated if needed. Next,ranges of geometric, materials, and other variables should be established andlisted for each subsystem/module. Also, decisions regarding the possible use ofstandard components in each subsystem/module should be made. This processstage is finished by choosing between alternative layouts using the productspecifications. The scale drawings are updated with the choices made.

As the next stage, additional functions that may be needed to carry out sup-port and auxiliary requirements should be identified. Then, rough layouts forthese additional functions should be developed ensuring the compatibility ofall subassembly interfaces. This task usually requires the use of standards,mathematical models, design guidelines and experimentation in order to de-termine all appropriate parameters. At the same time, the product should beevaluated against customer, technical, robust, safety an economic criteria andthe layout should be checked to estimate potential faults.

The embodiment process concludes with the testing of physical prototypes andthe design of appropriate tooling.

6.2.1. Embodiment

checklist

A second method to supplement the general embod-iment process, is the application of the embodiment check-list developed by Pahl & Beitz shown in table 6.1. This

table provides a systematic approach to apply proven design principles dur-ing the embodiment phase. The objective of the list is to ensure robustness,clarity, simplicity and safety in a product.

The checklist involves categories of possible engineering specifications, whereeach category has a set of basic questions that should be exhaustively applied tothe product and the different subsystems/modules as they are being detailed.It should be noted that mathematical models or physical prototypes may beneeded to effectively answer each of the questions.



Embodiment Checklist issue

Function Are the customer needs satisfied, as measured by the target values?

Is the stipulated product architecture and function(s) fulfilled?

What auxiliary or supporting functions are needed?

Working principles Do the chosen form solutions (architecture and components per function) produce the desired effects

and advantages?

What disturbing noise factors may be expected?

What byproducts may be expected?

Layout, geometry Do the chosen layout, component shapes, materials, and dimensions provide

and materials minimal performance variance to noise (robustness),

adequate durability (strength),

efficient material usage (strength-to-mass ratio),

suitable life (fatigue),

permissible deformation (stiffness),

adequate force flows (interfaces and strength concentrations),

adequate stability,

impact resistance,

freedom from resonance,

unimpeded expansion and heat transfer, and

acceptable corrosion and wear with the stipulated service life and loads?

Energy and Do the chosen layout and components provide

kinematics efficient transfer of energy (efficiency),

adequate transient and steady state behavior (dynamics and control across energy domains), and

appropriate motion, velocity and acceleration profiles?

Safety Have all of the factors affecting the safety of the user, components, function, operation, and the

environment been taken into account?

Ergonomics Have the human-machine relationships been fully considered?

Have unnecessary human stress or injurious factors been predicted and avoided?

Has attention been paid to aesthetics and economic analysis of the production process, capability, and

suppliers?

Quality control Have standard product tolerances been chosen (not too tight)?

Have the necessary quality checks been chosen (type, measurements, and time)?

Assembly Can all internal and external assembly operations be performed simply, repeatedly, an in the

correct order (without ambiguity)?

Can components be combined (minimize part count) without affecting modular architectures

and functional independence of the product?

Transport Have the internal and external transport conditions and risks been identified and solved?

Have the required packaging and dunnage been designed?

Operation Have all of the factors influencing the product’s operation, such as noise,

vibration, and handling been considered?

Life cycle Can the product, its components, its packaging be reused or recycled?

Have the materials been chosen and clumped to aid recycling?

Is the product easily disassembled?

Maintenance Can maintenance, inspection, repair, and overhaul be easily performed and checked?

What features have been added to the product to aid in maintenance?

Costs Have the stipulated cost limits been observed?

Will additional operational or subsidiary costs arise?

Schedules Can the delivery dates be met, including tooling?

What design modifications might reduce cycle time and improve delivery?

Table 6.1: Checklist for embodying a product concept. Adapted from Pahl &Beitz (1996).


139 6.3 Trends for the design process

6.3 Trends for the design process

In the last decades technology has brought profound changes in the way en-gineers design. Computational tools together with methods to increase thecommunication with all parts involved in the life cycle of a product have short-ened significantly the amount of time needed to put a initial concept or ideainto the market as a final product. From all the techniques that are or havebeen applied, concurrent engineering and computer aided tools have been mostsignificant.

6.3.1. Concurrent

engineering

Traditionally, design, manufacturing and marketing ac-tivities have taken place sequentially rather than concur-rently or simultaneously. The designer team would spend

large amounts of time analyzing components and preparing detail drawings fora new product. After the design was considered satisfactory, the design teamforwarded the information to the manufacturing team who would, once more,spend large amounts of time figuring out how to manufacture the product ac-cording to design specifications. After facilities and manufacturing processeswere ready for production, the marketing team began to prepare a marketingstrategy based on the product features.

Although may seem logical at first, this linear scheme proved to be inadequate.In many occasions the design team ended up with a product that was difficultto manufacture and even difficult to sell. Great efforts were wasted (and some-times still are) doing re-designs to improve manufacturing and to add featuresto improve the marketing of the product.

To avoid the previous problems it is best to include members of the manu-facturing and marketing divisions into the design team from the conceptualstages of the design process. The inclusion of the members will help to achievea better decisions to avoid design features that are difficult to manufactureor no desirable from the marketing point of view. This approach, that mayalso include members from other areas like distribution and disposal, is calledConcurrent Engineering.

Kalpakjian & Schmidt (2001) define Concurrent Engineering as a systematicapproach integrating the design and manufacture of products, with a viewtowards optimizing all elements involved in the life cycle of the product.


6.3 Trends for the design process 140

Iterations

Market Specifica-tions

ConceptDesign

DetailDesign Manufacture Sell

Figure 6.11: Depiction of concurrent engineering in the design process. AfterPugh (1991).

Life cycle means that all aspect of the product, such as design, development,production, distribution, use, disposal, and recycling, are considered simulta-neously. The basic goals of concurrent engineering are to reduce changes in aproduct’s design and engineering to reduce the time and costs involved in tak-ing the product from its design concept to its production and its introductioninto the marketplace. Figure 6.11 shows a simple design process models thatmakes emphasis in the interaction between phases due to the use of concurrentengineering principles.

6.3.2. Design for

manufacture and assembly

As discussed above, while designing a product,several disciplines must be taken into account.One of those disciplines that is specially bounded

to the design process is manufacturing. Many times new products have beendesigned only to find out that the technology needed for its manufacturing wasnot readily available.

Hence, each component of the product must be designed not only to fulfillengineering requirements but also to be easily and cheaply manufactured. Thisemphasis is called design for manufacture, and it groups selection of materials,manufacturing methods, planning, assembly, testing and quality assurance.

The design team must be capable of evaluating the impact that design changeshave in manufacturing processes.

After the individual parts have been manufactured, they usually have to beassembled to make the final product. The importance of the assemblies can-not be understated, in many products assembly takes the largest time of themanufacturing process. Much can be done during the design phase to makethe assembly as simple and fast as possible. Figure 6.12 shows some good andbad design practices regarding assemblies.



� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � �

� � � � � ��

� � � � � � �

� � � � � � �

� � � � � � �

� � � � � � �

� � � � � � �

� � � � � � �

� � � � � � �

� � � � � � �

� � � � � � �

� � � � � � �

� � � � � � �

� ��

� ��

� ��

� ��

� ��

� ��

� � �

� � �

� � �

� � �

� � �

� � �

� � �

� � �

� � �

� � �

� � �

� � �

� ��

� ��

� ��

� � �

� � �

� � �

� � �

� � �

� � ��

� ��

� ��

� �

� � �

� � �

� � �

� � �

� � �

� � �

� � �

� � �

� � �

� � �

� � �

� � �

Bad Good Bad Good

Bad

Good

Figure 6.12: Design considerations for assembly.

6.3.3. Design for the

environment

During the centuries the impact that humans have inthe environment has grown steadily as both, popula-tions and its needs, increase. In the last decades, the

awareness about the consequences of extracting resources and dumping wastewithout control has modified the way engineers design.

If populations is to keep enjoying the advantages of technological advances andhigher standards of living for the centuries to come, products must have littleor no impact in the environment. Design for environment (DFE) is a productdesign approach for reducing the impact of products on the environment.

Most of the times, the impact of products into the environment is thought ofin terms of their disposal. Nevertheless, products have an impact during all ofits life cycle from the extraction of the materials it is made from up to theirdisposal (see figure 6.13).

Products can have adverse impact on the environment during their manufac-ture through the use of polluting processes, the use of high amounts of rawmaterials, or the need of high quantities of energy. They can also have dif-ferent levels of impact on their disposal due to large half-lives or the need oflarge amounts of energy for their destruction. As shown in figure 6.13 thereare many opportunities for recycling, remanufacturing and reuse to reduce en-vironmental impact. Unfortunately, products that are designed without thisvision in mind are difficult to remanufacture, reuse or even recycle. Design-ers must use all their knowledge and creativity to create products that areenvironmentally friendly products throughout their manufacture, packaging,transportation, use and disposal.


6.3 Trends for the design process 142

Materialextraction

MaterialProcessing managment

WasteManufacturing Use

Recycle Remanufacture Reuse

Figure 6.13: Stages of a product life cycle. After Otto & Wood (2001).

6.3.4. CAD, CAM, CAE

and CIM

Nowadays, computers are an integral part of theconceptual, refinement, evaluation and productionphases of the design process either as engineering

or management tools. The use of computers has greatly simplified the rep-resentation, study and construction of analytical models through ComputerAided Design (CAD), Computer Aided Engineering (CAE) and ComputerAided Manufacturing (CAM).

These tools use computer software to assist in the creation and revision ofengineering drawings and models (CAD), manufacturing (CAM), and analysis(CAE) of new products.

The use of CAD/CAM/CAE tools avoids the need of making costly illus-trations, models and prototypes, shortening the time needed to bring a newproduct from concept to production. Although these tools may be applied indifferent parts of the design process, they are better suited for certain parts ofthe process (see figure 6.14).

Regarding Computer Integrated Manufacturing (CIM), Egan and Greene (1989)state that the appearance of CIM is based on the recognition that steps in thedevelopment of a manufactured product are interrelated and can be accom-plished effectively and efficiently by using computers.

CIM provides a mean to integrate all the steps in the manufacturing processtaking into account processes, specifications, instructions and data that needto be controlled and organized.



Definition of product need;marketing information

Conceptual design and evaluation;feasibility study

Design analysis; codes/standards review;physical and analytical models

Prototype production; testingand evaluation

Production drawings;instruction manuals

Material specification; processand equipment selection; safety review

Pilot production

Production

Inspection and quality assurance

Packaging; marketing andsales literature

Product

Computer AidedDesign(CAD)

Computer AidedManufacturing

(CAM)

Computer AidedProcess Planning

(CAPP)

Computer IntegratedManufacturing

(CIM)

Computer AidedEngineering

(CAE)

Figure 6.14: The use of computer aided tools in the different steps of the designprocess (after Kalpakjian & Schmid, 2001).


6.4 Failure Mode and Effect Analysis 144

6.4 Failure Mode and Effect Analysis

The notion of a reliable product comes from two different parts. First, there isthe minimization of performance variation across different environments anduser conditions. Second, is the assurance that the product will work as in-tended, without falling short of a given set of customer expectations. The firstpart is achieved through customer quality. The second part is achieved throughthe more fundamental engineering quality. With the latter, it is ensured thatthe product has adequate strength, reliability and failure prevention. Tradi-tionally, reliability has been achieved through extensive testing at the end ofthe design process. A better idea is to design from the early design stagesincorporating the concepts of quality and reliability.

Historically, engineers have not been very good at designing with reliability andquality. In most occasions, engineers use a safety factor as a way of makingup for all the possible failure modes that were not considered in the design.As the engineer had less idea of what could go wrong with the product, thelarger the safety factor that the engineer would use.

Unfortunately, as stated in the Mechanical Engineering magazine:

A large safety factor does not necessarily translate

into a reliable product. Instead, it often leads to an

overdesigned product with reliability problems.

Failure Mode and Effect Analysis (FMEA) is an analytical methodology usedas means for analyzing potential reliability problems early in the product de-sign process, where it is easier and cheaper to take corrective actions. FMEAis used to identify potential failure modes, determine their effect on the useof the product, and identify counter-actions to correct them. FMEA focuseson the entire product and not just in the different components and interfaces,although failure modes may be related to specific components or interfaces.

6.4.1. Types of FMEA There are several types of FMEAs, each one with itsown focus and objectives. Independently of the task

at hand, FMEA should always be used whenever failures would mean potentialharm or injury to the user. The types of FMEA are:

• System FMEA: focuses on global system functions


145 6.4 Failure Mode and Effect Analysis

• Design FMEA: focuses on components and subsystems

• Process FMEA: focuses on manufacturing and assembly processes

• Service FMEA: focuses on service functions

• Software FMEA: focuses on software functions

6.4.2. FMEA Usage According to the Society of Automotive Engineers (2002),FMEA supports the product development process in re-

ducing the risk of failure by:

• aiding in the objective evaluation of design requirements and design al-ternatives

• aiding in the initial design for manufacturing and assembly requirements

• increasing the probability that potential failure modes and their effectson system operation have been considered in the design/developmentprocess

• providing additional information to aid in the planning of through andefficient design improvements and development testing

• providing an open issue format for recommending and tracking risk re-ducing action

• providing future references to aid in analyzing field concerns, evaluatingdesign changes, and developing advanced designs

Properly used, FMEA provides the engineer with several benefits that include(Crow, 2002):

• Improve product/process reliability and quality

• Increase customer satisfaction

• Early identification and elimination of potential product/process failuremodes

• Prioritize product/process deficiencies



• Capture engineering/organization knowledge

• Emphasizes problem prevention

• Documents risk and actions taken to reduce risk

• Provide focus for improved testing and development

• Minimizes late changes and associated cost

• Catalyst for teamwork and idea exchange between functions

In order to effectively apply FMEA, the greatest challenge that the design teamfaces is to anticipate what might go wrong with a product. While anticipatingevery possible failure mode is almost always impossible, the development teamshould generate a detailed list of potential failures. Some questions that mayhelp in this task are (Stamatis, 1995):

1. What does the product do and what are its intended uses?

2. How does the product perform its function?

3. What raw materials and components are used to build the product?

4. How, and under what conditions does the product interface with otherproducts?

5. What by-products are created by the product or by the use of the prod-uct?

6. How is the product used, maintained, repaired, and disposed of at theend of its useful life?

7. What are the manufacturing steps in the production of the product?

8. What energy sources are involved and how?

9. Who will use or be in the vicinity of the product, and what are thecapabilities and limitations of these individuals?

The above questions should be aimed to gather information in order to addresssix basic questions:



1. What could fail or go wrong with each component of the product?

2. How or why can the part fail to meet its engineering specifications?

3. What circumstances could cause the failure?

4. To what extent might it fail?

5. What are the potential hazards produced by the failure?

6. What steps should be implemented to prevent the failure?

6.4.3. Step by step Design

FMEA Analysis

The use of FMEA is straightforward consistingof a series of steps. Following the procedure sug-gested by Otto & Wood (2000), the 10 steps pro-

cedure is explained in what follows.

Step 1: List each subassembly and component number, along with the basicfunctions or function chains of the component. The component numbers maybe referenced from a product’s bill of materials. Likewise, the componentfunctions should be consistent with the functional models and architecturedeveloped for a product. Any functions listed for a component should conciselyrepresent the design intent. Environmental and operational parameters, suchas temperature, humidity, and pressure ranges, should be listed to clarify thisintent.

Step 2: Identify and list the potential failures for each product component.Simple prototype models and brainstorming techniques can aid in identifyingpotential failure modes. Likewise, sketches, storyboards, free-body diagrams,force-flow diagrams, and process-flow diagrams can help in understanding thephysics of a failure mode. Tables 6.1 and 6.2 should be used to check for typicalproblems with components and product systems. For any listed failure mode,the idea is that the failure could occur, but not that will necessarily occur forthe product under consideration.

Step 3: List possible potential causes or mechanisms of the failure modes. Ex-ample causes include tolerance stack-up, assembly errors, poor maintenance,impact loading, overstressing, and so forth. These causes will provide insightsinto modeling of the failure mode. They will also indicate appropriate preven-tive measures that might be adopted.



List of example failure modes

Corrosion Ingress Delamination

Fracture Vibrations Erosion

Material Yield Whirl Thermal shock

Electrical short Sagging Thermal relaxation

Open Circuit Cracking Bonding failure

Buckling Stall Starved for lubrication

Resonance Creep Staining

Fatigue Thermal expansion Inefficient

Deflections or deformations Oxidation Fretting

Seizure UV deterioration Thermal fatigue

Burning Acoustic noise Sticking

Misalignment Scratching and hardness Intermittent system

operation

Stripping Unstable Egress

Wear Loose fittings Surge

Binding Unbalanced

Overshooting Enbrittlement

Ringing Loosening

Loose Scoring

Leaking Radiation damage

Table 6.2: Abbreviated list of example failure modes. After Otto & Wood(2000).



Step 4: List the potential effects of the failure, including impact on the envi-ronment, property, or hazards to human users. Example effects include noise,poor appearance, flying debris, unpleasant odor, erratic operation and so forth.

Step 5: Rate the likelihood of occurrence (O) of the failure. The ratingsshould be on a scale of 1-10 as given by:

1 No effect

2/3 Low (relatively few failures)

4/5/6 Moderate (occasional failures)

7/8 High (repeated failures)

9/10 Very high (failure is almost inevitable)

Step 6: Estimate the potential severity (S) of the failure and its effect. Again,a 1-10 scale should be used. The following meanings are associated with thisscale:

1 No effect

2 Very minor (only noticed by discriminating customer)

3 Minor (affects very little of the system; notice by average customer)

4/5/6 Moderate (most customers are annoyed)

7/8 High (causes a loss of a primary function; customers are dissatisfied)

9/10 Very high and hazardous (product becomes inoperative; customers are

angered; the failure may result in unsafe operation and possible injury)

Step 7: List current or expected design controls/test for detecting (D) thefailure before the product is released for production. A 1-10 scale is used toassess detection:



1 Almost certain

2 High

3 Moderate

4/5/6 Moderate – most customers are annoyed

7/8 Low

9/10 Very remote to absolute uncertainty

Step 8: Calculate the Risk Priority Number (RPN). An RPN prioritizes therelative importance of each failure mode and effect on a scale of 1-1000. It canbe calculated with the following relation:

RPN = (S) × (O) × (D)

A “1000” rating implies a certain failure that is hazardous and harmful and willoccur, whereas a “1” rating is a failure that is highly unlikely and unimportant.Rating above “100” will occur, whereas ratings below “30” become reasonablefor typical applications. It is important to notice that the RPN scale is non-linear in risk.

Step 9: Develop recommended actions for the failure modes, assign respon-sibilities to appropriate parties and team members, and set a schedule forimplementing the actions. Corrective actions should be first developed forthe highest ranked failure modes based on the RPN. Example actions includerevised component or subassembly design, revised test plan or material spec-ification, design of experiments and prototypes, etc. These actions should bespecific.

Step 10: Implement the corrective actions, update the S-O-D ratings, andrecalculate the RPN for the updated design.

The process and results of the FMEA should be documented, perhaps withthe help of a template like the one shown in figure 6.15.



App

lied

step

s

Fai

lure

Typ

eF

ailu

reC

onse

quen

ceF

ailu

reC

ause

Sugg

este

d re

med

ial

mea

sure

sPr

opos

ed te

stst

eps

OS

D

Cur

rent

sit

uati

on

RPN

OS

DR

PNF

ailu

re L

ocat

ion

Impr

oved

sit

uati

on

Des

ign

FM

EA

Pro

cess

FM

EA

FA

ILU

RE

MO

DE

AN

D E

FF

EC

T A

NA

LY

SIS

Com

pone

nt n

ame

Nam

e/D

epar

tmen

t/Sup

plie

r/T

elep

hone

By

(Nam

e/D

epar

tmen

t/Tel

epho

ne)

Figure 6.15: FMEA Template.



References

1. Cross, N. (1994) Engineering Design Methods, John Wiley & Sons.2. Crow, K. (2002) Failure Modes and Effects Analysis (FMEA), DRM Asso-ciates, www.npd-solutions.com. 3. Cutherell, D. (1996) Product Architecture.Chap 16. in “The PDMA handbook of new product development”, edited byM. Rosenau, Jr. et al. New York: Wiley. 4. Kalpakjian, S. & Schmid, S.(2001) Manufacturing Engineering and Technology, fourth ed., Prentice-Hall.5. Otto, K. & Wood, K. (2001) Product Design - Techniques in Reverse Engi-neering and New Product Development, Prentice-Hall.6. Pahl, G. and Beitz W. (2001) Engineering Design - A systematic Approach.Second Ed. Springer.7. Pugh, S. (1990) Total Design, Addison Wesley.8. SAE (2002) Potential Failure Mode and Effects Analysis in Design (DesignFMEA) and Potential Failure Mode and Effects Analysis in Manufacturingand Assembly Processes (Process FMEA) and Effects Analysis for Machinery(Machinery FMEA). SAE Standard J1739.9. Stamatis, D. H. (1995) Failure Mode and Effect Analysis - FMEA fromTheory to Execution. ASQ Quality Press.10. Ullman, D. (2003) The Mechanical Design Process, Third Edition. McGraw-Hill.11. Ulrich, K. (1995) “The role of product architecture in the manufacturingfirm”. Research Policy, 24, 419-440.12. Ulrich, K. & Eppinger, S. (2000) Product Design and Development. IrwinMcGraw-Hill.


Product Design Techniques for Robustness, Reliability and Optimization

Documents

jos carlos

john wiley

key business

good design

transparent

tea brewing

quality function

gather raw