Gerhard Pahl, em. Pro f. Dr. h.c. mult . Dr.-Ing. E.h. Dr.-Ing. Fachbe reich Maschine nbau Tcchnischc Univcrsitiit Da rmstadt Magdale ns trasse 4 64298 Darmstadt Germa ny
1" Wolfgang Beitz, Prof. Dr.-Ing.
)org Feldhusen, Prof. Dr.-lng. Institut für Allgeme ine
Karl-Heinrich Gro te, Prof. Dr.-Ing. [nstitut für Maschinenko nstruktio n Otto-von-Guer icke-Universitiit Magde burg Universitatsplatz 2 39106 Magdeburg Germa ny
Konstruktionslehre
d es Maschinenbaus Rhe inisch Westfalische Technische Hochschule Aachen Steinbachstrasse 548 52074 Aa che n Germa ny
E.h. Dr.-Ing.
1935-1998
British Library Cataloguing in Publication Data Engineering design :a systematic approach. - 3rd ed. l. Engineering de.sign l. Pahl, G. (Gerhard), 1925- 11. Wallace, Ken 620'.0042 ISBN- JO: 1846283 183 Library of Con greS> Control Number: 2006938893 ISBN 978- 1-84628-3 18·5 3rd cdition
e-ISBN 978-1-84628-319-2 3rd cdition
Printed o n acid-free paper
ISBN 3-540-19917-9 2nd edition
e Springer-Verlag London Limitcd 2007 Translation from theGerman Languageedition: Komtrukt ionslel~re by Gerhard Pahl et al. Copyright @Springer·Yerlag Berlín Heidelberg 2003. All rights reserved. 3rd English edition, Springer 2007 2nd English edition, Springer 1996 1st English edition publishcd by The Design Council, London, UK (ISBN 085072239X) Aparl from any f'air dealing for the purposcs of rescarch or private stud)'t or criticism or review, as permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced, stored or transmitted, in any form or by any means. with the prior permission in writing ofthe publishers,or in thecase of reprographlc reproducti.on in accordance with the terms of licenc.es is.sued by the Copyright Licensing Agency. Enquiries coatcerniatg reproductioo outsidc those tcnns should be sent to the publishers.
The use oí registered names, trademarh, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the re]evant la:ws and regulations and therefore free for general use. The ptablisher makes no representation, expres.s or implied, with regard to the accuracy of the information contained in this book and cannot accept an)' legal respon::;.ibility or liability for any errors or omis.sions that may be made. 987654321
Spriatger Science+Busiatcss Media springer.com
Contents
1
lntroduction .. ..... ... ... ... ... .. . . .. ... . • .. .. .. . . .. . . 1.1 The Engineering Designer ... . ....... ... . ... . ..... . . 1.1. 1 Tasks a nd Activities . ... ... ... . .. .... . . .... . . J. 1.2 Position of the Design Process within a Company . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.1.3 Trends . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . 6 1.2 Necessity fo r Systematic Desig n . . . . . . . . . . • . . . . • . . . . . 9 1.2.1 Requirements and the Need fo r Systematic Design . . . . . . . . . . . . . . . . . . . . . . . 9 1.2.2 Historical Background . . . . . . . . . . . . . . • . . . . . . . 10 1.2.3 Cu rren! Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 1.2.4 Aims and Objectives of this Book . . . . . . . • . . . . . 19
2
l'undame ntals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Fundament.als ofTech nical Systems ... . . ..... .. ... . . . 2.1.1 Systems, Plant, Equipme nt, Machines, Assemblies and Componcnts . . . . . . . . . . . . . . . . . 2.1 .2 Conversion of Ene rgy. Materia l and Signals . . . . 2.1.3 Functional lnterrelationship . . . . . . . . . . . . . . . . . 2.1.4 Working Jnterrelatio nship . . . . . . . . . . . • . . . . . . . 2.1.5 Co nstructionalln terrelationship .. ... ... .... .. 2. 1.6 System lnterrelationship.. . . . . . . . . . . . . . . . . . . . 2. 1.7 Systematic Guideline . . . . . . . . . . . . . . . .. . . . . . . . 2.2 Fund amentals of the Systematic Approach . . ..... . . . . . 2.2.1 Problem Solving Process ... . ..... .. ..... . . . . . 2.2.2 Characteristics o f Good Problem Solvers. . . . . . . 2.2.3 Prob le m So lvi ng as In formatio n Processing . . . . 2.2.4 General Wo rking Mcthodo logy . . . . . . . . . . . . . . . 2.2.5 Generally App licable Methods . . . . . . . . • . . . . . . . 2.2.6 Role of Comp uter Support . . . . . . . . . . . . . . . . . . .
3
27 27 27 29 31 38 42 42 43 45 45 49 51 53 58 62
Product Planni.ng, Solut ion Fi.nding and Evaluation . . . . . . . . 63 3.1 Product Planning... . ..... . .... .. ..... .. .... . ...... 63 3. 1.1 Degree o f Novelty of a Product . . . . • . .. .. . . . . . 64
xviii
Contents
3.1.2 Product Life Cycle . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.3 Com pany Goals a nd Their Effec t . .. .... . ..... . 3. 1.4 Product Plan n ing . . . . . . . . . . . . . . . . . . . . . . . . • . . So lution Fi nd ing Methods . . . . . . . . . . . • . . . . . . • . . . . . . . 3.2.1 Convcntio nal Mcthod s . . . . . . . . . . . . . . . . . . . . . . 3.2.2 lntuitive Method s . . . . . . . . . . . . . . . . . . . • . . . . . . . 3.2.3 Discursive Methods . . . . . . . . . . . . . . . . . • . . . . . . . 3.2.4 Method s for Com bin ing So lutions .... . . ..... . Selection and Evaluation Methods . . .. _. . . .. _..•.. _ . . 3.3.1 Selecting Solu tion Variants .... . .... . . • . . ... . . 3.3.2 Evaluating So lution Variants . ... .... . . ... . . . .
64 65 66 77 78 82 89 103 106 106 109
4
Product Development Process .... .. ... .... ... ... ... ..• .. 4.1 General Problem So lving Process .. . .......... . ..... . 4.2 Flow o fWork Duri ng the Process ofDesig ning ... ... .. 4.2.1 Activity Plan ning . ... .. .... ... .. ..... .. .... . 4.2.2 Timing and Schedu ling .. .. . . .. .. .... .. ..... . 4.2.3 Plan ning Project and Produc t Costs ... .. ..... . 4.3 Effcctivc Organ isation Structures . .. ..... .. .... . . ... . 4.3.1 lnterd isciplinary Cooperation ... .. . . ......... 4.3.2 Leadersh ip and Team Behaviour . ... • . . .. . .. . .
125 125 128 128 134 136 138 138 141
5
Task Clarificat ion . .. _. . . .. .. . . .. .. . .. . .. • . . . .. . • . .. . .. . 5.1 lmportance o f Task Clarification . ... ... . . .... .. ... .. 5.2 Setting Up a Requirements List (Design Specification) .. . ..... . . .... • . ..... • . ... .. . 5.2.1 Contents .. .. .... .. ..... . . . . . . . . . ... . . .... .. 5.2.2 Format .... . .... . . ..... . ..... . . .... .. .... . . 5.2.3 lde ntifying the Requ ire ments ... .. ..... . ..... 5.2.4 Rcfin ing and Extcnding thc Rcquircmcnts . .. .. 5.2.5 Compiling the Requ irements List . .... .. .. . .. . 5.2.6 Examp les . .. .. _. . . ..... . ..... . . .... • . ... _.. 5.3 Using Requiremen ts Lists .. .. .......... • . ..... • ..... 5.3.1 Updating .. . .... . . ..... . ... ... • . . .. •.. .. _. . 5.3.2 Partial Req uirements Lists .. ..... .. .... • . .... 5.3.3 Further Uses .. .... .. .... ... ... ... ... . . .... . 5.4 Practica! Application of Requirements Lists . ... . • . .. ..
145 145
Conceptual Design .. . .. . ..... . .. .. . .. .. . .. . .. . . . . .. . .. . 6.1 Steps of Conceptual Design . .. . . .. .. . . .. ... ... ... ... 6.2 Abstracting to Identify the Essential Problems . ... ... . 6.2.1 Aim of Abstraction .. ..... . .... . . ..... . .. _ .. 6.2.2 Broaden ing the Problem For mulat ion . .. ..... . 6.2.3 Idcntifying the Esscn tial Problcms from the Requ irements List . .... .. ... . . . . . .. . 6.3 Establishing Fu nction Structures .. . .. . . • . ..... • .. . ..
159 159 161 161 162
3.2
3.3
6
146 146 147 149 151 152 153 153 153 156 157 157
164 169
CoJHeJHS
6.3.1 Overall Function 6.3.2 Breaking a Functio n Down into Subfu nctions .. 6.3.3 Practica! Applications of Function Structures .. Developing Working Structures .... . ..... . .......... 6.4.1 Scarcbing for Working Principies . ... .... .. ... 6.4.2 Combining Working Principies ............... 6.4.3 Selecting Working Str uctures . . . ...... . .... .. 6.4.4 Practica! Application ofWorking Structures .... Developing Concepts .... . . ... . . ............ . ..... . 6.5.1 Firming Up into Pri ncipie Solution Variants .... 6.5.2 Evaluating Principie Solution Variants . . ..... .. 6.5.3 Practica! Application of Developing Concepts .. Examples o f Co nceptual Design .... . ..... . ..... . .... 6.6.1 One-Handed Househ old Water Mixing Tap . . ... 6.6.2 Jmpu lse-Loading Test Rig . .. ... ... ... .... .. ..
169 170 178 181 181 184 186 186 190 190 192 198 199 199 210
Embodimcnt Dcsign ... .. ..... . ..... . ..... . . .... . • .... . 7.1 Steps o fEmbodiment Design ... . .... . . . .... . . .... .. 7.2 Checklist for Embod iment Dcsig n . . ..... .. ... . . . ... . 7.3 Basic Rules of Embodiment Design ... . .. . .. . • .. ... .. 7.3.1 Clarity ............. . ........ .. .. . . .. .. . ... 7.3.2 Simplicity . ... .... .. .... .. .... .. . .... .. ... .. 7.3.3 Safety ............................. • ....... 7.4 Princip ies of Embodiment Desig n . .... .. .... . . .... .. 7.4.1 Principies of Force Transm ission . .. .... . .... . 7.4.2 Principie of the Division ofTasks .. . ... .•.. .. . 7.4.3 Principie o f Self-Help ......... .. .... .. .... .. 7.4.4 Principies o f Stability and Bi-Stability . .. .... .. 7.4.5 Principies for Fault-Free Design ........ .... .. 7.5 Guidclincs for Embodimcnt Dcsig n . ... ... .... .. ... .. 7.5.1 General Considerations ........ . ..... . . . ..... 7.5.2 Design to Allow for Expansion ... .. ...... . ... 7.5.3 Design to Allow for Creep and Relaxation . ... .. 7.5.4 Design Against Corrosio n ..... . . .... . . ..... . 7.5.5 Design to Min imise Wear . . ..... • . .... . ..... . 7.5.6 Design for Ergonomics . ... .... . . .... . . .... .. 7.5.7 Design for Aesthetics .. ... . . . .... .. .... .. .... 7.5.8 Design for Productio n ....... ... .... . . .. .... 7.5.9 Dcsign for Assembly ... ... .... . . .... . . .... .. 7.5.10 Design for Maintenance .... .. . .. ... . .. . .. ... 7.5.11 Design for Recycling . ... .. ..... . . .... .. .... . 7.5.12 Design for Mínimum Risk . ..... • . ... . . ..... . 7.5.13 Design to Standards . ... . . .... • . . .... • .. ... .. 7.6 Evaluating Embodiment Desig ns ......... • .... . ..... 7.7 Example of Embodiment Design .. . . .. . • . . . . • .... . .. 7.8 Detail Design .. . ..... . .................... • .......
227 227 233 234 235 242 247 268 269 281 290 301 305 308 308 309 321 328 340 341 348 355 375 385 388 402 410 416 417 436
6.4
6.5
6.6
7
xix
xx
8
Contents
Mechanical Connections, Mechatronics and Adaptronics . . ..... ....... ..... ...... . .......... .. . 8.1 Mcchan ical Con ncctions .. .... .. .... ... .... .. .... .. 8.1.1 Ge neric Func tions and General Behavio ur ..... 8.1.2 Material Con ncctions .. .. ... .... ... ... .... .. 8.1.3 Form Connections . . .... . ...... • .... • ..... . . 8.1.4 Force Con nectio ns . . ..... . . .... • . .... .. .... . 8.1.5 Applicatio ns .. .... .. .... .. .... .. .... .. .... . 8.2 Mechatron ics . ..... . . .... . ..... . .. ... ... ... . . . . ... 8.2. 1 Ge neral Architecture a nd Termino logy ..... . .. 8.2.2 Go als and Limitations .... .. .... .. .... .. .... . 8.2.3 Development of Mechatronic Solutions . .... ... 8.2.4 Examp les .. . .. . ..... . ..... . ..... . .......... 8.3 Adaptronics . . .. . ..... . ................... • . .. .... 8.3.1 Fu ndame ntals a nd Termino logy .. .. ..•.. .. ... 8.3.2 Go als a nd Limitations . ... .. .... .. .... .. .... . 8.3.3 Development of Adap tron ic Solutions .... ... .. 8.3.4 Examples ........ . ..... . ..... . .. ... .... .. . .
439 439 440 440 441 443 447 448 448 450 450 451 458 458 459 460 461
Size Ranges and Modular Products ... ..... ...... . • ... .. .. 9.1 Size Ranges ..... . ..... . ........................... 9.1. 1 Similarity Laws .... ... ... ... ... ... ... ... ... . 9.1.2 Decimal-Geometric Preferred Number Series . . 9. 1.3 Representation and Selection of Step Sizes ... .. 9.1.4 Geometrically Simila r Size Ranges ... ... ... ... 9.1.5 Semi-Similar Size Ranges ........ . ... ..•.. ... 9. 1.6 Developm ent of Size Ranges ........ . . • .... . .. 9.2 Modular Products .... ... ... .... .. .... .. .... . . .... . 9.2.1 Modular Product Systematics ............. . .. 9.2.2 Modular Product Dcvclopmcnt .. ... .... ... . .. 9.2.3 Advantages and Limitations of Modular Systems 9.2.4 Examp les . .. .... .. .... .. ...... . ..... . .... .. 9.3 Recent Rationalisatio n Appro aches .. .... .. .... .. .... 9.3.1 Mod ularisation and Product Architecture . . .... 9.3.2 Platform Construction ... ... ... ... ... ... ... .
465 465 466 469 472 476 481 493 495 496 499 508 510 514 514 SI 5
l O Design for Quality .. ... ... ... ... ... .... . .... . .. ... . .. .. 10. 1 Applying a Systcmatic Approach . ... .... . . ... . . . .... 10.2 Faults and Disturbing Facto rs . . . .. ........ .. . .. . .... 10.3 Fault-Tree An alysis ... ... .... .. .... ... ... . ...... .. . 10.4 Failu re Mo de and Effect Analysis (FMEA) .... • .... . . . 10.5 Quality Function Deployment (QFD) ........ • .... . ..
517 517 521 522 529 531
9
11 Design for Mínimum Cost . ..... . . ... . . .. .. . • .. ... .. ..... 1 J. 1 Cost Factors . . ..... . ..... .. .... .. .... . ..... . • .... . 11.2 Fundamcntals o f Cost Calculations .. .... .. ... . • . ... . 11.3 Methods for Esti mati ng Costs .. . .. . .. . ..... . ........ 11.3.1 Comparing with Rclativc Costs ... ... ... .... .. 11.3.2 Estimating Using Share of Material Costs ...... 11.3.3 Estimating Using Regressio n Analysis . .. .... .. 11.3.4 Extrapolating Using Si milarity Relations . . ... . . 11.3.5 Cost Structures .. .. . . .. .. . . . .... .. .... .. .... 11.4 Target Costi ng . . . . . . . . . . . . . . . . • . . . . . . • . . . . • . . . . . . . 11.5 Rules for Minimising Costs . ... .. . ..... . . .... .. .... .
535 535 537 539 539 544 545 54 7
558 560 561
12 Summary ... .... ... ... ... ... ... ... . .. ... . . . .... .. .... . 563 12.1 The Systematic Approach . ............... • .. . . • ... . . 563 12.2 Exper ie nces of Applying the Systematic Approach in Practice . .... .. .... . . .... 567
References .... .. .... .. ..... . ..... . ..... . . .... . . .... . . .... . 571 English Bibliography . ..... . . ........... . ..... • ..... . ..... . 603 1ndex ..... . .. . .... ....... • .... . .... . . • ...... • .. . . • ....... 609
1
lntroduction
1.1 The Engineering Designer 1.1.1 Tasks and Activities The main task of engineers is to apply their scien tific and engineering knowledge to the so lution of technical problems, and then to optimise those solutions within the requircments and constraints set by material, tcchnological, economic, legal, e nvironmen tal and human-related consid erations. Problems become con crete tasks aftcr thc problems that cng inccrs havc to solvc to crcatc ncw tcchnical products (artefacts) are clari fied and defined . This happ ens in individual work as well as in teams in order to realise interdiscipli nary product d eve lopme nt. The mental creation o f a new product is the task o f d esign and development eng ineers, whereas its physical realisation is the responsibility o f production eogineers. In this book, designer is used synony mo usly to mean d esign and development engineers. Designers contribute to finding solu tions and developing products in a very specific way. They carry a heavy b urden of responsibility, since their id eas, k nowledge a nd skills determine the tech nical, economic a nd ecological properties of thc product in a dccisivc way. Desig n is an interesti ng engineeri ng activi ty that: • affects almost all a reas of hu ma n life • uses the laws and insights of science • b uilds u pon special experience • provid es the prerequisites for the physical realisation of solution i deas • requires professional integrity and responsibility. Dixon [1.39] and later Pen ny [l.l 44] placed thc work of engineering d csig ners al the centre of two intersecting cultural and tech nical streams (se e Fig ure l.l ). However, other models are also available. In psychological respec: ts, designing is a creative activity that calls for a sound g rounding in mathem atics, physics, chemistry, mechanics, thermodynamics, hydrodynam ics, electrical engineering, production eng ineering, materials techno logy, m achine elements and design theory, as well as knowledge and exp erience of the d omain o f interest. Initiative,
2
1 Jntroduction
Politks 1
Sociotogy. psychology 1
Ecanomi
Science 1
_
Enginee
~nglneerlng_ f- ~ngineering
deslgn
te
p d . 1 ro ucuon
lndulcrial design 1 Archit~clure
1
Art
Figure 1.1. The central iKtivity of engineering design. After (1.39, 1.144)
reso lu tion, econo mic insight, tenacity, optimism a nd teamwork a re qualities that stand all designers in good stead and are indispensable to those in responsible positions [1.130] (see Section 2.2.2). In systematic respects, desig ning is the optimisation o f given objectives within partly confticting constraints. Requirements change with time, so that a particular solution can only be optimised for a particular set of circumstances. In organisational respects, dcsign is an essential part of thc product lifc cyclc. This cycle is triggered by a market need or a new id ea. 11 star ts with product planning and e nds-when the product's usefullife is over-with recycling orenvironmentally safe disposal (see Fig ure 1.2). This cycle represents a process of converting raw materials in to economic products of high added value. Designers must u nder take their tasks in close cooperation with specialists in a wide range of d isciplines and with different skills (see Section 1.1.2). The tasks and activities of designers are influenced by severa! characteristics.
Origin ofthe task: Projects related to mass production and batch production are usually started by a produce plann ing g roup after ca rrying out a thoroug h analysis of the ma rket (see Section 3.1). The requirements established by the product planning gro up usually leave a large solution space for designers. In the case o f a customer order fo r a speci fic one-off o r small batch product, howcvcr, thcre are usually tightcr rcquirements to fulfil. In thcse cases it is wise for designers to base their solutions on the existing company know-how that has bccn built up from prcvious d cvclopmcnts and ordc rs. Such d cvclopments usually take place in small incremental steps in order to limit the risks involved . If the d evelopment involves o nly part of a product (assembly or module), the requirements and the d esign space are even tighter and the need to interact with o ther desig n groups is very high. When it co mes to the production of a product,
J. J The Engi1teering Desig1ter
MarkeVNeed/Problem
3
l'<>let>liai!Goalsof
company ProduCI plannlng/
Task senlng
- - ---¡ 1 1 1 1 1
DispOSdl/Environment
Figure 1.2. Life cyde of a product
there are design tasks related to production machines, jigs and 6xtures, and inspection equipment. For these tasks, fulfilling the functional requiremen ts and technological constraints is especially important.
Organisation: The organisation o f the d esign and development process depends in the first instance on the overall organisatio n ofthe compa ny. In product-orien ted companies, responsibility fo r product development and subsequen t production is spli t between separate divisions of the company based on specific product types (e.g. rotary compressor d ivision, piston com pressor division, accessory equipmcnt d ivision). Problem-oriented companies split the responsibility according to the way the overall task is b roken down into partial tasks (e.g. mecha nical engineering, co ntro l systems, materials selection, stress analysis). In this arrangement the project manager must pay particula r attention to the coo rd ination of th e work as it passes fro m group to gro up. In sorne cases the project manager leads independe nt temporary project teams recruited from the vario us g roups. These teams report d irectly to the head o f development o r senior manageme nt (see Section 4.3).
4
1 Jntroduction
Other organ isational structures are possible, fo r example based o n the particular phase o f the design process (co nceptual desig n, embod ime nt d esig n, detail dcsign), the domain (mcchan ical engineering, clcctrical engincering, so ftware development), or th e stage of the pro duct d evelopment process (research, design, development, prc-production) (sec Section 4.2). In largc projccts with clcarly delineated domains, it is o ften necessary to develop ind ividual modules for the product in parallel.
Novelty: New tasks a nd problems that are realised by original designs incor porare new solution principies. These can be realised either by selecting and co mbining known principies and technology, or by inventing completely new tech nology. The term origi nal d esign is also used when existing or slightly changed tasks are solved using new solution principies. Orig inal designs usually proceed thro ugh all design phases, depend on physical and process fu ndamentals and require a careful tech nical and economic a nalysis ofthe task. Orig inal designs can involve the whole productor just assemblies or components. In adaptive design, one keeps to kn own and established so lution p rincipies a nd adapts tbe embodiment to changed requireme nts. It may be necessary to under take o riginal desig ns of ind ividual assemblies or components. In this type of d esign the emphasis is on geometrical (strength, stiffness, etc.), production and material issues.
In variant design, the sizes and ar rangements of parts a nd assemblies are varied within the limits set by previo usly designed product structures (e.g. size ranges and modular products, scc Chaptcr 9). Variant desig n requircs original dcsign effort only once and do es not present significan! d esign problems for a particular o rder. 1t includes designs in which o nly the d imensio ns of ind ividual parts a re changed to meet a specific task. In (1.124, 1.167( this type of design is referred to as principie design or design with fixed principie. In practice it is ofte n not possible to d efin e precisely the bo undaries between the thrcc typcs of design, and this must be considered to be on ly a broad classification.
Batch size: The desig n of one-off and small batch products requires particularly ca reful design of all physical processes and embodiment details to m inimise risk. In these cases it is usually not econom ic to produce dcvelopment prototypes. Often functionality and reliability have a higher priority than economic optimisation. Products to be made in la rge qua ntities (large batch o r mass productio n) must have their technical and econo mic characteristics fully cbecked prior to full-scale production. This is achieved using models and prototypes and often requires several develo pmen t steps (see Figure 1.3). Branc/1: Mechan ical engineeri ng covers a wide range oftasks. As a co nsequence the requirements and the type of solutions are exceptionally d iverse and always require the application of the methods and tools used to be adapted ·to the specific task in hand. Domain-specific e mbodiments are also co mmon. For example, food proccssing machines havc to fulfil spccific rcquiremcnts rcgarding hygicne; machine tools have to fulfil specific requirements regarding precision and operating spccd; prime movcrs have to fulfi l spccific rcqu iremcnts rcgarding powcr-to-wcight ratio and efficiency; agricu ltu ra] machines have to fulfil specific requirements re-
J.1
e
~
E ~
¡¡ Q.
-~
t; ~
~
The Engi1teering Desig1ter
5
e
~
E
~
¡;
"' §
t;
-g .
.;:
Model
One-off product
6atdl·produced product
Mass·produced
product
- - - - - + Product optimisation Figure 1.3. Stepwlse de~~elopment of a mass-produced product. Alter (1.191)
garding func tionality and robustness; and office mach ines have to fulfil specific rcquircmcnts rcgarding crgonomics and noisc Jcvcls.
Goals: Design tasks must be directed towards meeting the goals to be optimised, taking into account the given restrictions. New functions, longer life, lower costs, produc tion problems, and changed ergono mic requ iremen ts are al! examples of possiblc rcasons for cstablishing ncw dcsign goals. Moreover, an increased awareness of environmental issues frequen tly requires complctcly ncw products and proccsscs for which thc task and thc solution pr incipie have to be revisited. This requ ires a ho listic view on the part of designers and collaboration with specialists from o ther discipli nes. To cope with this wide variety of tasks, des ig ners have to adopt di fferent approaches, use a wide range of skills and tools, have broad desig n knowledge and consult specialists on specific problems. Th is becomes easier if desig ners master a general working proccdurc (sec Section 2.2.4), understand gencration and cvaluation methods (see Chapter 3) and are familiar with well-known solutions to existing problems (see Chap ters 7 and 8). The activities of designers can be roughly classified into: • Conceptualising, i.e. searching for solution principies (see Chapter 6). Generally applicable methods can be used alo ng with the special methods described in Chapter 3. • Embodying, i.e. engineering a solution principie by determining the general arrangement and prelirn inary shapes and materials of all components. The methods described in Chapters 7 and 9 are use fui.
6
1 Jntroduction
• Detailing, i.e. finalising prod uction and operating details. • Com puting, representing and information co llecting. These occur d ur ing all phascs of thc d csig n proccss. Ano thcr common classification is thc distinction bctween direct dcsign activities (e.g. conceptualising, embodying, d etailing, computing), and indirect design activities (e.g. collecting and processing infor mation, attend ing meetings, coordinating staff). One should aim to keep the propo rtion of the indirect activüties as low as possible. In the d esign process, the required d esig n activities have to be str uctured in a purposcful way that forms a d car sequen ce of main phascs and ind ividual working steps, so that the flow of work can be plan ned and controlled (se e Chapter 4) .
1.1.2 Position of the Design Process within a Company The design and development departmen t is of central impor tance in any co mpany. Dcsig ncrs determine thc propcrtics of every product in ter m s o f function, safety, ergonom ics, production, transport, operation, maintenance, recycling and disposal. In addition, desig ners have a large influence o n produc tion and operating costs, on quality and on production lead times. Because of this weight of responsibility, d esigners must continuously reappraise the general goals of the task in hand (see Section 2.1.7) . A further reason for the central role o f d esigners in the company is the position of design a nd develop ment in the overall product development process. The links and information fl ows between d epar tments are shown in Fig ure 1.4, fro m which it can be sccn that production and asscmbly dcpend fu ndamentally on information from product pla nning, design a nd development. However, d esign and development a re strongly influenced by knowledge and experience from production and assembly. Because of c urrent market pressures to increase product performance, lower prices a nd reduce the t ime-to -market, product planning, sales and marketi ng must draw incrcasingly u pon spccialiscd engineering k nowledgc. Because of thcir key position in the product development process, it is the refo re particularly importan! to makc full use o f thc theorctical knowlcdgc and product cxpericncc of designers (see Section 3.1 and Chapter 5). Cur rent product liabi li ty legislation [1.1 2j demands not on ly professional a nd respo nsible product development using the best techno logy b ut also the hig hest possible production quality.
1.1.3 Trends The most importan! impact in recent years on the design process, and on the activities of designers, has come from co mp uter-based data processing. Computeraided d esign (CAD) is influencing design methods, o rganisational str uctures, the division of work, e.g. between conceptual designers and detail designers,
J. J The Engi1teering Desig1ter
7
-...,
1
o
/
/
1 1
.--í
1 1 1
1
~- -1 1 1
:---1 1 1
1- -~ 1 1
:---1 1 1
1 1---
l
~----,------"
Pto~ucllon
1
---,
1
/
t>lanning Product\lon contml
1 1 1
- ---'1
n Production ol pans
D
__
1
1
Assembly
D Test.acceptance
D Distribution
D
1 J 1 1 1 1 1 1
...)
Qualilycontrol /
/
( Slockor finished products
e:::=::> Product evolution - - - lnformatlon
Figure 1.4. lnformatíon flows between departments
as well as lhe c reativity and tho ught processes of individual designers (see Section 2.2) . New staff, e.g. systcm managers, CAD specialists, cte., are being introduced into the design process. In the future, routine tasks such as variant d esigns w ill be largely unde rtaken by the co mp uter, leaving desig ners free to concentrate on new designs and custo mer-specific o ne-off p roducts. These tasks will be supported by computer tools that en hance the creativity, engineering knowledge and experience of designers. The d evelopme nt
8
1 Jntroduction
of knowledge-based systems (expert systems) [1.72, 1.108, 1.178, 1.183) and electronic componen! catalogues [1.19, 1.20, 1.53, 1.1 51, 1.1 83 ] wi ll increase the ease with which information can be rctricvcd, includ ing spccific d csign data, dctails of standard components, informatio n about existing products as well as thcir d csign proccsscs and o thcr dcsig n knowlcdgc. Thcse systcms will also aid the analysis, op timisation and combination of solutions, but they will no t replace designers. On the contrary, the decision-m aking abilities of designers will be even more crucial because o f the very large number of so lutions it will be possible to generate, and also because of the need to coordinate the inputs from the many specialists now required in mod ero multidisci plinary projccts. A fu rther strong trend is for com panies to concentrate their desig n and develo pme nt activities on so-called co re co mpetences, and thus actüng as system integrators, buying in assemblies and components as required from other companies (outsourcing). Designers therefore need the ability to assess and evaluate these outsourced items, even though they have not created these themselves. This critica! assessment process is en hanced through broad technical knowledge, accumu lated experience and a systematic use of evaluation procedures (see Section 3.3). Com puter-integrated man ufacturing (CIM) has consequences for designers in terms of co mpany organisation and information exchange. The system within a CIM structure makes better planning and control of the desig n process necessary and possible. The same holds true for simultaneous engineering (see Section 4.3 (1.13, 1.40, 1.188]), where development times are reduced by focusing o n the flexible and partially parallel activities of product opt imisation, production optimisation and quality optimisation. The trend is to bring production planning forward into the design process through the application of co mputers. Apart from thcsc dcvclopmcnts that influcncc thcworking mcthods of dcsigncrs, designers must increasingly take into accou nt rapid technological developments (e.g. new productio n and assemblyprocedures, microelectron icsand software) a nd new materials (e.g. composites, ceramics and recyclable materials) . Tbe integration of mecha nical, electronic and software e ngineering (mechatronics) has led to many exciting product developments. Desig ners now have to give equal weight to these three aspects of modero products. In summary, it ca n be concluded that there is airead y much pressure on designers and this prcssure will increase furthcr. This requircs cont inuous further education for existing designers. However, the initial education of designers m ust take into account the ma ny changes taking place [1.1 27, 1.187). lt is essenti al that future designers not only understand traditional science and engineering fundamentals (physics, chemistry, mathematics, mechanics, thermodynamics, fluid mechan ics, electron ics, electrical engineering, materials science, machin e ele m ents) but also specific d omain knowledge (instrumentation, contro l, transmission technology, production technology, electrical drives, electronic controls). The ed ucation of future designers shou ld include co urses where they actually app ly their d esign knowlcdgc in ordcr to salve dcsign tasks. Thcy also nccd spccialist courscs in design method ology, including CAD and CAE.
J.2
Necessity for Systematk Oesign
9
1.2 Necessity for Systematic Design 1.2.1 Requirements and the Need for Systematic Design In vicw of thc central rcsponsibi lity of dcsig ncrs for thc tcchnical a nd cconom ic properties of a product, and the commercial importance of timely a nd efficient product development, it is importan! to have a defined design proced u re that finds good solutio ns. Th is procedure must be flexible a nd at the same ti me be capab le of being planned, optimised and verified . Su eh a procedure, however, cannot be realised if the desig ners do not have the necessary do main knowledge a nd can not work in a systematic way. Furthcr mo rc, thc use of such a proccdurc should be encouraged and suppor ted by the o rganisation. Nowadays o ne d isti nguishes between design science a nd design methodology (1.90]. Design science uses scien tific methods to analyse the structures of technical systems and their relationships with the enviro nment. The aim is to derive r ules for the d evelopme nt of these systems from the system elements and their relationships. Design methodology, however, is a concrete cou rse of action for the design of technical systems that derives its knowledge from desig n science and cognitive psycho logy, and from practical experience in different domains. lt in eludes pla ns of action that link wo rking steps a nd design phases accordi ng to content and organisation. These plans must be adap ted in a flexible manner to the specific task at hand (see Chapter 4). It also in eludes strategies, rules and principies to achieve ge neral a nd specific goals (see Chapter 7 a nd Chapters 9-11) as well as methods to solve individual d esign problems o r partial tasks (see Chapters 3 and 6). This is not meant to detrae! from the importance of intuition or experience; qu ite the co ntrary- the additio nal use of systematic procedu res can on ly serve to incrcasc thc outp ut and invcntivcncss o f talcntcd dcsig ncrs. Any logical and systematic approach, however exacting, involves a measure of intuition; that is, an inkli ng ofthe overall solutio n. No real success is likely without intu ition. Desig n methodo logy should therefore foster and guide the abilities of designers, encou rage creativity, and at the same time d rive ho me the need for objective evaluatio n of the results. On ly in this way is it possible to raise the general sta nd ing of designers and the rcgard in which their work is held. Systemati.c procedures help to render designing comprehensible and also enable the subject to be taught. Howcver, what is learned and recognised about dcsig n method ology should not be taken as dogma . Such procedu res merely try to steer the effor ts o f desig ners fro m unconscio us into conscio us and more pu rposefu l paths. As a result, when they collaborate with other eng ineers, designers will not merely be .ho lding their own, b ut will be able to take the lead (1.130(. Systematic design provides an effective way to rationalise the d esig n and production processes. In original design, an ordered and stepwise ap proach- even if this is on a partially abstractlevel- wiU provide solutions that can be used again. Structuring the problem and task makes it easier to recogn ise app lica tion possibi lit ics for established so lutions from prcvious projccts and to use d esig n catalogues. The stepwise concretisation of established solution principies makes it possible to
JO
J lntroduction
select and optimise them at an early stage with a smaller amou nt of effort. The approach of develop ing size ranges and modular produc ts is an importan! start to rationalisation in the dcsign arca, but is cspccially importan! for thc production process (see Chapter 9). A desig n methodology is also a prerequisitc for flexible and contin uous computer support ofthe design process usingproduct models stored in the comp u ter. Without this methodo logy it is not possible to: develop knowledge-based systems; use sto red data a nd methods; link separate programs, especially geometric m od ellers with analysis programs; ensure the continuity of data fiow; and link data from different co mpany divisions (CIM, PDM). Systematic procedu res also make it easier to divide the work bctwcen designers and computcrs in a mcaningful way. A rational approach must also cover the cost of comp utation a nd guality considerations. More accurate and speedy preliminary calcu lations with the help o fbetter data are a necessity in the desig n field, as is the early recognition of weak points in a solutio n. All this calls for systematic processing ofthe desig n docume ntation. A design metiJOdology, therefo re, must: • allow a problem -directed approach; i.e. it must be applicable to every type o f design activity, no matter which specialist field it involves • foster inve ntiveness and u ndersta nd ing; i.e. facilitate the search fo r optimum solutions • be compatible with thc conccpts, methods and findings of o thcr disciplines • not rely on finding so lutions by chance • facilitate the application of known solutions to related tasks • be com patible with electronic data pro cess ing • be easily taug ht and learned • refiect the findings of cognitive psychology and modern management science; i.e. reduce workload, save time, prevent human er ro r, and help to maintain active interest • ease the plan ning and management of teamwork in an integrated and in terdisciplinary product development process • provide guidance for leaders of product development teams.
1.2.2 Historical Background It is d ifficult to determine the origins of system atic d esign. Can we trace it back to Leonardo da Vinci? Anyone looking at the sketches o f this early master must be surprised to see - and the modern systematist delights in discovering- the great extent to which Leo nardo used syste matic variation of possible solu.tions [1.118). Right up to thc industrial era, dcsigning was doscly associated with arts and crafts.
1.2 Nece-ssity for Systernatic De.sig11
ll
With the rise of mechan isation in the nineteenth century,as Redtenbacher [ 1.150] poi nted o ut early on in his Prinzípíen der Mechaník und des Masch íne n baus (Pri ncipies of Mcchanics a nd of Machine Constructio n), attcntion bccamc incrcasingly foc used o n a number of characteristics and pri ncip ies that con tinue to be of great importancc, namcly: sufficicnt strcngth, sufficicnt stiffncss, low wear, low friction, mínimum use of materials, easy hand ling, easy assembly and maxim um rationalisation. Redten bacher's pup il Reu leaux [ 1.152] developed these ideas but, in view of their often conflicting requirements, suggested that the assessment of their relative importance must be left to the intell igence and discretion of individual designers. They can not be trcatcd in a general way orbe taug ht. Important con tributions to the development of engineering design were also made by Bach [ 1. 11] and Riedler [ 1.1 53], who realised that the selectio n of mate rials, the choice of production methods and the provision o f adeq uate strength are of equal impor ta nce a nd that they influence one ano th er. Rotscher [ 1. 164] mentions the following essential c haracteristics of design: specified pu rpose, effective load paths, and efficient production and assembly. Loads should be conducted along the shortest paths, and if possible by axial forces rather than by bend ing moments. Longer load paths no t o nly waste materials and increase costs b ut also require consid erable changes in shape. Calculation and laying o ut must go hand-in-ha nd. Designers start with what they a re given a nd wi th rcady-made asscmblics. As soon as possib lc, thcy should make scale drawings to ensure the co rree! spatial layout. Calculation can be used to obtain either rough estimates for tbe preliminary layo ut or precise values that are used to check the d etail design. Laudien [I.I07J, upon exam in ing the load paths in machine parts, gave the followi ng advice: for a rigid con nection, jo in the parts in the di rection of the load; if flcxibility is requircd, join thc parts along indircct load paths; d o not makc unnecessary provisions; do no t over-specify; do not fulfil more demands than are required; save by simplification and econo mical co nstruction. Modern systematic id eas were pioneered by Erkens [1.46 ] in the 1920s. He insisted on a step-by-step approach based o n constant testíng and evaluatíon, a nd also on the balancing of conflicting demands, a process that must be conti nued until a network of ideas-the desig n-emerges. A more comprehensive acco unt ofthe "techniq ue of design" has been presented by Wogcrbauer ( 1.206 ], whose contribution wc consider to be thc orig in of systematic design. He divides the overall task into subsidíary tasks, and these into operational a nd imple mentatio nal tasks. He also examines (but fails to present in systematic form) the numerous interrelationships between the identifiable constraints desig ners must take in to accou nt. Wogerbauer himself do es not proceed to a systematic elabo ration of solutio ns. llis systematic search starts with a so lution d iscovered more or less in tuitively and varied as comprehensively as possible in respect to the basic form, materials and method of production. The resulting profu sion of poss ible solu tions is th en reduced by tests a nd evaluatio ns, w ith cost be ing a cr ucial critcrion. Wogcrbaucr's ver y comprchcnsivc list of characteristics hclps in thesearch for a n optimu m solution and also when testingand evaluating the results.
12
1 Jntroduction
Franke [1.54] d iscovered a comprehensive structure for transm ission systems usi nga logical- fu nctio nal a nalogy based on elements with d ifferent ph ysical effects (clcctrical, mcchan ical, hyd raulic cffccts for idcntical logical functio ns guiding, cou pling and separati ng). For this reason he is regarded as a representative of those working on thc functional com parison of physically diffcrcnt solu tion clcmcnts. Rodenacker in particular used th is a nalogical approach [1.155 ]. T hough so me need to imp rove and rationalise the design pro ces& was felt even befo re World War 11, progress was impeded by the absence of a reliable means of representing abstract ideas and the widespread view that designing is a form of art, nota technical activity like anyo ther. A period of staff shortages in the 1960s [1.1 90) creatcd a st rong ímpetus to ad opt systcmatic th inking more widdy. Impor ta n! pioneers were Kesselring, Tschochner, Niemann, Matousek and Leyer. Their work continues to p rovide most useful suggestions fo r hand ling the individual phases and steps of systematic d esign. Kesselring [1.98) first explained rhe basis ofhis meth od of successive approx.ima tions in 1942 (for a su mmary see (1.96, 1.97] and VD! Guideline 222 5 (1.195)) . Its salien t feature is the evaluation of form variants according to technical a nd economic criteria. In his theory, he mentions five overlying pr incipies: • the princip ie of mínimum production costs • thc princip ie of mínim um space rcqu ircmcnt • the princip ie of mínimu m weigh t • the principie of mín im um losses • the pri ncip ie of op tim um handli ng. The design a nd opti misation of ind ividual parts a nd sim ple tech n ical artefac ts is thc aim o f thc thco ry of for m dcsign. lt is ch aractcriscd by thc sim ultaneo us application of physical and economic laws, and leads to a determlination of the shape and di mensions of com ponents and an appropriate choice of materia ls, production methods, e tc. If selected op timisation characteristics are taken into account, the best solution can be fou nd with the help of mathematical methods. Tschoch ner ( 1.179) mentions four fundamental design factors, nannely the working principie, the material, the form and the size. They are interconnected and dependen! on the requirements, the n umber of u n its, costs, etc. Designers start from the so lution principie, d eter mine the other fund amental factors-material and form-and match them with the help of the chosen dimensions.. Nieman n [1.1 21) starts out with a scale layout of the overall des ig n, showi ng the main dimensio ns a nd the general arrangement. Next he divid es the overall design into parts that can be developed in parallel. He proceeds from a definition of the task to a systematic variation of possible solutions and fi nally to a critica/ and formal selection of the optimum solution. These steps are in general agreement with those used in more recent methods. Niemann also draws attenti·o n to the then lack of meth ods for a rriving at new so lutions. He must be conside:r ed a pionee r o f systcmatic dcsign inasmuch as he consistently dcmandcd and cncouragcd its development.
J.2
Necessity for Systematk Oesign
J3
Matousek (l.ll2 ] lists four essentia l factors: working principie, material, production and form design, and then, following Wiigerbauer [1.206], elaborates a n ovcrall working plan based on thesc four factors considcred in thc ordcr givcn. He adds that, if th e cost aspect is unsatisfactory, th ese factors have to be reexam ined in a n itcrativc manner. Leyer [1.109 ] is mainly concerned with form desig n, for which he d evelops fu ndamental guidelines a nd principies. He distingu ishes three ma in design phases. In the first, the working principie isla id down with the help of an idea, an inventio n, or established facts; the second phase is that of actual design; the third phase is that of implementation. His second phase is essentially that of embodiment; that is, layout and fo rm design supported by calculations. During this phasc, principies or r ules have to be taken into accou nt-for instance, the principie of constan! wall thickness, th e principie of lig htweight constr uc tion, the principie of shortest load paths, and the principie of homogeneity. Leyer's r ules of form design are so valuable because, in practice, failure is still far less frequently the result of bad working pri ncip ies than o f poor detail desig n. These prelim inaryattempts m ade way for the intensive development of methods, mainly by university professors who had learnt the fundamentals of design by desig ning tech nical products of increasing complexity in industrybefore becoming professors. They realised that a greater reliance on physics, mathematics and information theory, and the use of systematic methods, were not o nly possible but, with the growing division of labour, quite indispensable. Need less to say, these developments were strongly affected by the requirements of the particular industries in which theyor iginated. Most carne from precision, power transm ission and electromechanical engineering, in which systematic relationships are more obvious than in heavy engineering. Hansen and o ther members of the Tlmenau School (Bischoff, Bo ck) first put forward their systcmatic dcsign proposals in thc carly 1950s [1.2 1, 1.25, 1.78]. Hansen presented a more com prehensive desig n system in the second ed ition of h is sta ndard work pub lished in 1965 [1.77]. Hansen's approach is d etined in a so-called basic system. The four working steps in this approach are applied in the same way in conceptual, embod iment and detail desig n. Ilansen begins with the a nalysis, critique, and specificatio n of the task, which leads to the basic principie of the development (the crux of the task). The basic principie encompasses the overall function tihat has been derived fro m the task, the prevailing conditions, as well as the required measures. The overall fu nction (the goal and the constraints) and the context (elements and properties) constitute the crux of the task together with the given constraints. The second working step is a systematic search for solution elements and their co mbination into working means and working principies. Hansen attaches g reat importan ce to the th ird step, in which any shortcomings of the developed working means are analysed with respect to their pr operties and quality characteristics, and then, if necessary, improved. In thc fo urth and last stcp, these improvcd working mcans are cvaluatcd to determine the optimu m working means for the task.
14
1 Jntroduction
In 1974 Hansen p ublished another work, entitled Konstruktionswissensclwft (Science of Design) [1.76]. The book is mo re conce rned with theoretical fundamcntals than with rules of practica! dcsign. Simila rly, Mülle r (1.1 16] in his Grundlagen der systematiscl?en Heuristik (Fundamcntals of Systcmatic Hcuristics) prcscnts a thcorctical and abstrae! pic turc o f the design process. This book offers essential fou ndations of design science. Furthe r im portant publi cations are [1.114, 1.115,1.117]. After Han sen, it is Rod enacker (1.1 55- 1.157] who became preeminent by developing an original design method . His approach is characterised by developing the required overall working interrelationship by d efin ing in seque nce the logical, physical and embodiment relationships. He cm ph asiscs thc rccognition and suppression of disturbing inftuences and failures as early as possible during formulation o f the physical p rocess; the adoptio n of a general selection strategy fro m simple to complex; and the evaluation of all parameters of the tech n ical system against the cr iteria quantity, quality and cost. Othe r characteristics of his method are the emphasis o n logical function struc tures based on binary logic (con necting and separating), and on a conceptual design stage based on the recognition that product optimisation can only take place once a suitable solution principie has been fo und. T he most importan! aspect of Rod enacker's systematic design approach is undoubted ly h is emphasis on establishing the physical process. Based o n this, he not o nly deals with the systematic p rocessi ng of concrete design tasks, b ut also with a methodology for inventing new technical syste ms. For the latter he starts with the q uestion: For what new app lication can a known physical effect be used? He then searches systematically to discover complete! y new solutions. In addition to the methods we have been describing, there is a view that a onesided emp hasis on discursive methods do es not present the com pie te picture. Thus Wachtler ( 1.199, 1.200] argues, by a nalogy with cybernetic concepts su eh as control and lcarning, that crcativc dcsign is thc most complcx fo rm of thc "lcarn ing process". Learn ing represents a h igher for m of control, one that involves not o nly q uantitative cha nges at constant quality (rules), but also cha nges in the quality itself. What matters is that, for the purpose of optimisation, the design process should be treated, no t statically, b ut dynamically as a control process in which the information feedback must be repeated until the info rmation content h rus reached the leve! at which the optimum solution can be found. The learn ing process thus keeps increasing the lcvcl of information and hence facilitates the scarch for a solution. The systematic design methods of Leyer, Hansen, Rodenacker and Wachtler are still bei ng app lied today, having be en integrated into the more recent develo pme nts in desig n methodo logy.
1.2.3 Current Methods 1. Systems Theory
In socio- cconomic-tcchnical proccsscs, proccdurcs and mcthods of systems theory are becoming increasing ly importan!. The interdisciplinary science of systems
J.2
Necessity for Systematk Oesign
JS
theory uses special methods, procedures and aids for the ana lysis, plann ing, selection and optimu m d esig n of com plex systems (1.1 4-1.1 6, 1.23, 1.29, 1.30, 1.1 43, 1.208 ]. Techn ical artefacts, includi ng the products o f light and heavy e ng ineering industry, are artificial, concrete and mostly dy namic systems consisting of sets of ordered elements, interrelated by virtue of their properties. A system is also characterised by the fact that it has a boundary which c uts across its Ji n ks with the environment (see Figure 1.5) . T hese Jinks determine the externa] behavio ur o f the system, so that it is possible to define a function expressing the relationship between inputs a nd outp uts, a nd hence changes in the magnitudes of the system variables (sec Section 2.1.3). From the idea that technical artefacts can be represen ted as systems, it was a short step to the applicatio n of systems theory to the design process, the more so as the objectives of systems theory correspond very la rgely to the expectations we have of a good design method, as specified at the beginning of this chapter [1.16]. The systems approach reflects the general appreciation that co mplex problems are best tackled in fixed steps, each invo lving analysis and synthesis (see Section 2.2.5). Fig ure 1.6 shows the steps of the systems approach. The first o f these is the gathering of infor mation about the system under consideration by mea ns of market analyses, trend stud ies or known require ments. In general th is step can be called prob lem analysis. The ai m here is the clear for mu la tion of thc problem (or subproblem) to be solved, wh ich is the actual starting point for the development of the system. In the second step, or perhaps even during the first step, a programme is drawn up in order to give formal expression to the goals of the system (problem form ulation). Such goals provide important criteria for the subsequent evaluation of so lution variants a nd hence for the discovery of the opti mu m so lu tion. Severa] solution variants are thcn synthcsised o n the basis of the information acquired du ring the first two steps. Before these varia nts can be evaluated, the pe rfor ma nceof eaeh m ust be analysed for its properties and behaviour. In the evaluation that fo llows, the pe·rformance of ea eh variant is compared with the o riginal goals, and o n the basis of th is a decision is made and the optimum system selec ted. Finally, informatio n is g iven out in the for m o f system implementation plans. As Figure 1.6 shows, the steps d o no t always lead stra ig ht to the final goal, so that iterative procedu res m ay be needed. Bu ilt-in decisio n steps facilitate th is optimisation process, which constitutes a transforma! ion of in formation. In a systems theory process model [1.23, 1.52], the steps repeat themselves in so-called Ji fe cycle p hases of the system in wh ich the chrono logical p rogression of a system goes from abstract to concrete (see Figure 1.7). 2. Va fue Analysis
The m a in aim of Va/ue Atlalysis, as described in DIN 699 1O( 1.37, 1.66, 1. 196- 1.198], is to reduce cost (scc Chap ter 11). To that end a systcmatic ovcraln approach is proposed which is applicable, in par ticula r, to the further developmen t of existing
16
1 Jntroduction
¡ ·- ·- · -·- ·- ·- · -·-
-1
s,
o,
Figure 1.5. Structure of a system. S; system boundary; S1-S5: subsystems of S; s2l-s 2. : subsystems or elements of S2; / 1- /3: inputs; 0 ¡- 01 outputs
1
System studies State analysh, problem analysis
Goal programme Goalsetting, problem formulation
System synthesis Dtvetopment of sofulion varlants
¡
¡
¡
\
System analysis Properties and behaviour ofvariams
<:
.J._
t System evaluation Evaluation of variants against goal fl(ogramme ..l
~-System de
System implementation plans Planning lhe next systemphase
Figurt1 .6. Sleps oflhe systems apfl(oach
•r
J.2
--
Probl
Problem fonnulation System synthesis
SyStem a nal~;
Evaluatlon Oecision
Preliminary System System study dm!lopment production
~
' 1--;:-l D' i
o !
¿ 2J 1
~
\
17
üfe pha>es ofasystem -
Planning
T T
Necessity for Systematk Desig.n
Sy
System
installation
operation
1?y rr ¡9
System replacement
D
1 !
1
T
~
figure1.7. Modelof the ;y;tems oppro><:h. After (1.23, 1.52(
Oevelop solution ideas
Prepare proje
Assemble team
• Define scope ofValue Analysis • Define organlsatlon and proc:edure
• Collect existing ideas •
Search for new ideas
OeteJmine solutions Anatyse actUcll state
•
Recogni.IA! fimctions • Oelermine funetion costS
• Evaluate ideas • Oevetop ideas intosolutions • Evaluate and decide uponsolutions
Detennine target !tate
•
l
Definetarget funclions
Realise s.olutions
•
klentify additional requ iremen~
• Detailchosen S<>lutions
•
Match target msts with target functions
•
Plan theirimplementation
Figure 1.8. Basicworking steps of Value Anatysis. After OIN 69910
products. Figure 1.8 shows the basic worki ng steps of Value Analysis. In general, a start is madc with an cxisting dcsign, which is a nalyscd with roespcct to thc required functions and costs. So lution ideas are then proposed to meet the new targcts. Bccause of its cmphasis on functions and the stepwise scarch fo r better solutions, Value Analysis has much in common with systematic design. Various metho ds are available to estímate costs and assess cost breakd owns (see Chapte r 11). Tea mwork is essential. Good communication between staff in sales, pu rchase, dcsig n, production and cost cstimation (thc Valuc Analysis tcam) en-
18
1 Jntroduction
sures a holistic view ofthe requirements, embod iment design, materials selection, production processes, storage requirements, standards and marketing. A furthcr csscntial aspcct is thc division of thc rcquircd ovcrall function into subfu nctions in the o rder of desce ndi ng com plexi ty along with their allocatio n to function carricrs (asscmblics, ind ividual compo ncnts) . Thc costs ()f fulfilling all o f the functions up to and including the overall function can be estimated from the costs calc ulated for the ind ividua l compone nts. Such "fu nction costs" can then provide the basis for evaluating the concep ts or embodiment var ian.ts. The aim is to minimise these function costs and where possible eliminate functions that are not really necessary. It has becn suggestcd that thc application of thc Value Analysis mcthod sho uld not be left un til after the layout and detail drawings have been fi nalised, b ut should be started du ring concep tual design in orde r to "design in" value [ 1.65]. In this way, Value Analysis approaches the goals of systematic design.
3. Design Methods VDI Guideline 2222 [ 1.192, 1.193 ] defines an approach and individual methods for
the conceptual desig n of technical products and is therefore particularly suitable for the development of new products. The more recent VDI Guideline 2221 [ 1.191) (English translation: [1.186]) pro poses a generic approach to the design of technical systems and products, emphasising the general applicability of the approach in the fields of mechanical, precision, cont rol, software and process eng ineeri ng. Thc approach (scc Fig ure 1.9) ineludes scvcn basic working stcps that accord with the fundamentals of technical systems (see Section 2.1) and comp any strategy (scc Chaptcr 4). Bo th g uid clincs havc bccn dcvclopcd by a VOl Committcc comprising senior desig ners from industry and many of the previously mentioned design methodologists from the former West Germany. Because the aim is for general applicability, the design process has been only roughly str uctured, thus permitting product-specific a nd company-specific variations. Fig u re 1.9 should therefore be regarded as a gu ideline t() wh ic h detailed worki ng procedures can be assig ned. Special cmphasis is placcd on the iterative nature of thc approach and the sequen ce o f the steps must not be considered rig id. Sorne steps might be omitted, and othe rs repeated frequently. Such flexibi lity is in acco rdance with practica] design exper ience and is very importan! for the application of all d esign methods. The design methodo logists and sen ior designers from industry who co llabo rated to produce these VD! Guidelines often represented differe nt schools of thought or had developed their own design methods. Severa! contributions to design methodo logy were made by colleagues in o ther co untries. In th is book, rcfcrcnccs are madc to all of thcsc many inputs when thc individual mcthods and procedures are discussed in detai l. A comprchcnsivc ovcrvicw of thc intcrnat ional dcsign tcaching and rcscarch activities since 1981 can be fou nd in the proceedings of the ICED con ference series (International Co nference o n Engin eering Desig n) [1.148].
J.2
Stages
Necessity for Systematk Desig.n
Results
19
Phases
.l
Phase 1
'(
Spccification Determine functions and rheir suuauu~
P11ase 11
€ ~
~ !I
'S
"5. ~
"' "'e ~
~
e:
Preliminary la)'outs
"5 ~
ll ll Phaselll
P11ase IV
Productdo
~--11
1
Figure 1.9. Generalapproa(h to deslgn. Alter (1.191(
In Table 1.1, the mai n publications on design methodo logy are given in chronological ordcr. This tablc rcplaces and extends in a more compact form t hc individual efforts and achievements that were d escr ibed in Chapter 1 of the second English edition of this book. Fu rther contributio ns from the a uth ors listed in the table can be seen from their en tries in the list of references at the end of the book.
1.2.4 Aims and Objectives of this Book On closer examination the methods we h ave been describing have been strongly intl uenced by their a uthors' specialist fields. Thcy nevertheless resernble one an other far mo re closely than the various concepts a nd terms might suggest. VDJ Gu idelines 2222 and 2221 confirm these resemblances as they were developed in collaboratio n with a wide ra nge of experienced co ntributors. Bascd on our cxpcricncc in thc hcavy machincry industry and railway and automotive engineering and many years spent in engineering design education at
20
Jntroduction Table l.l. ChronoiQ9Ical ovtrview of the devtfopment of dt$ign methodology
Year
Author
Thememtle
Country
19S3 19SS
Bischoff, Hansen Bock
Rationelles Konstruieren Konstruktionssystematik-die Methode
DDR DDR
[1 .211 [1.2SI
19S6 1963
Hansen Pahl
DOR DE
[1 .781 [1.131]
1966
Oixon
USA
[1.391
1967 1968
Harrisberger Roth
USA DE
[1.791 [1 .1631
1969
Glegg
GB
[1.68-1.701
USA DE GB DE DE GB
[1.177] [1.161 [1.711 [1.1291 [1.1S51 [1581
DE
[1 .142]
USSR DE
(1.51 (1.192)
USA
(1.1)
DDR GB USA
[1 .821 [1.49, 1.501 [1 .126]
DE
(1.134]
DE
(1.1921
DDR DDR
[1.16$1 [1.6D-1.62I
DDR
[1 .99, 1.1001
USSR
(1.146, 1.1471
DE
[1 .671
H
(1.1411
USA
(1.1191
Uterature
der ordnenden Gesichtspunkte
1971
Tribus Beitz Gregory Pahl Rodenacker French
1972
Pahi,Beitz
1973
Altschuller VOl
1974
Adams
1976 1977
Hennig Flursheim Ostrofsky
1970
Pafli,Beitz
VOl 1978 1979
Rugenstein Friek
Klose
Polovnikin
1981
Gierse
Kozma, Straub (Pa hi/Be~z)
Nadler
Konstruktionssystem.atik Konstruktionstechnik im thermischen Maschlnenbau DesignEngin~ring; lnventiveness, Analysis and Oe
Wege zur lOsungsfindung Methodisches Konstruieren (4th Ed~ion 19911 Conceptual Design for Engineers, 1st Edition (Jrd Edition 1999) Series of anides .,Fürdie Konstruktionspraxis" (1972- 1974) Erfinden: Anleitung ftir Neuerer tmd Erfinder VOI·Richtlinie 2222, Blatt 1 [Entwurf]: Konzipieren te
J.2
Necessity for Systematk Oesign
21
Tablo1.1. (
1982
Author
Thome/Title
Country
Prcxeedlngs of ICED by Hubka S
WDKSeries blannually flom 1981 to 2001; Design Sodety Series from 2003 Methodenbewu" tes Ploblemiosen
(H
(1.1481
Dietrych, Rugenstein
CH Pl/0
(1.1701 [1.361
Roth
Einführung indie Konstruktions· wissens(haft Konstruieren mit Konstruktions-
DE
VOl
katalogen, 1st Edition (lrd Edition 2001) VOI·Ri
oe
[1.160, 1.1611. (1.1621 [1.1931
OK OOR
(1.81 (1.841
GB USSR
(1.801 (1.41
CH Pl
(1.86, 1.871 (1.139)
GB
(1.140)
Automationin Thinking in Oesign The lmpficationsfor the Study for Oesign Methods ofRecent Development in Neighbouring Disciplines Kostengilnstig Entwidceln und Konstruieren
J GB
(1.207) (1.10)
DE
(1.41, 1.431
Konstruktionsmethodik und Konstruktions· praxis- eine kritische Betrachtung Konstruktionslehre ffir den Maschinenb.lu. Grundlagen, Arbeitsschritte, PrinzipiOsungen.
oe
(1.51)
oe
(1.101, 1.102), (1.103, 1.104)
Nl
(1.185)
USSR USSR J DK
literature
und Anwendung von Konstruk.tio nska tal og~n
1983
Andreasen etal.
Design for Assembly
HOhne
Struktursynthese und Variationstechnik
Hawkes, Abinett
The Engineering DesignProcess Etfinden- Wege z:ur LOsung te<:hniS(her Probleme Theorie te
belmKonstruleren 1984
Alts
Yoshikawa Archer
EMenspiel, Lindemann Franke
Kolltr
Oesign, 1st Edition (lrd Edition 2006)
(3rd Editlon 1994)
1986
1987
vanden
DesignMethodology as a Conditionfor
Kroonenberg
Computer·Aided Oesign
Odrin
Morphologische Synthese vonSystemen
Alts
Theory ol lnwntive ProblemSolving
Taguchi Andreasen, Hein
lntroductionof Quality Engineering lntegrated Product Oevelopment Applicationof Expert Systems in Machine Design
oe
Dn Oesign Differently Analysis ol the Engineering Design Prcxess
Pl GB
[1.1 221 (1.2, 1.31 (1.1751 (1.71 (1.421 (1.631 (l.?l-1.751
DDR OE/G8
(1.1691 (1.1861
Erlenspiel, Figel Gasparski
Hales
inanIndustrial Context. Managing
Englneerlr>g Design
1988
S
Kons.truktionslehre
VOI/Walla
VOl Oesign Handbook 2221: Systematk
Walla
Approachto the DesignofTechnical Systems and Produc:ts. English translation Oetailed Analysis of <~ n Engineering OesignProje
Oixon
On Research Methodology-Towarc!s A Sdentifi< USA Theory of Engineering Oesign
(1.2031 (1.381
22
1 Jntroduction Tablel.l. {continued)
Year
Themerrrtte
Country
F•ench
Form, Structure and Me
GB
(1.$7, 1.58)
CHICA
[1 .88, 1.89)
N
(1.92)
USA USA
(1.1 73.1.174) (1.182)
USA
(1.205)
Hubka, Eder Jakobsen Suh Ullmann,
Stauffer. Oienerich
1989
literature
Author
Based on Empirical Data
Winner,
The Role of Concurrent Engineering in Weapon
~ n nell, e tal.
Acquls.ition
Cross OeBoer Elmaragh,
Engineering OesignMethods Oedsion Methodsand Techniques Oesign Theory and Methodolfl9Y
GB USA
(1.33) (1.35) (1.45)
funktional e Ges taltbildung-Gestalte nde
oe
(1.93, 1.94)
NL
Seering, Ullmann Jung
Konstruktionslehre für Vorri
Instrumente und Maschinen Pahi!Seitz Ulrich, Seering 1990
Birkhofer
Konttinnen (Pahi/ Beitz} Kostefic MOIIer Pighini Pugh
1991
Rinderle Roozenburg, Eekels Andreasen Bj!rnemo
Boothroyd, Oieter Oart, Fujimoto Flemming Hongo, Nakajima
Chinese translationof Pahi/Beitz Engineering PRC Oesign Synthesis of Schematic Description in USA Mechanical Design Vonder Produktidee zumPtodukt- Eine kritische Betrachtung zur Auswahl und Bewertung in der Konstruktion Finnish translation of Pahi/Beitz Engineering FIN Design Design for Quafity YU A•beltsmethoden der TechnlkwlssenS
oe
[1.138) [1.184) [1.17, 1.18)
(1.137) (1.10S) (1.114) (1.145) [1.149) (1.154) (1.158. 1.1S9)
DK S
(1.6) [1.22)
USA
[1.26)
Product Oevelopment Perfa
(1 .31)
Methodical Design Frame by New Procedures Evaluation and OedsionTechniques in the Engineering DesignProcess Assembly Automatlon and P•oduct Oesign
(1.47, 1.48) (1.85)
J.2
Necessity for Systematk Oesign
23
Tablo1.1. (contlnued) Year
Auth0<
Theme/Title
Country
Deslgn Synthelic ReMOnlng: A Methodology for Me
USA
(1.95)
USA USA USA
(1.172(
1992
Kannapan, MaJShek 5tauffer (e
DE
(1.171 )
USA CH
(1.1 80, 1.181) (1.28)
Erfolgrelch Erfinden. Wide.spruchsorientierte lnnovationsstrategie Concurrent Engineering Design VOI-Richttinie 2221: Methodikzum Entwickeln und Konstruieren te<:hnischer Systeme und Produkte Total Quality Oevelopment A Process·Base
DE
(1.110)
USA DE
(1.113) (1.191)
USA GB
(1.32) (1.24)
DE
(1.127)
DE
(1.40)
GB
(1.135(
Designfor Excellence Analysing Design Activity
USA Nl
(1.27) (1.34)
Systems Engineering: An Appro>
(1.81)
lntroduclion ofTRIZ in Japan
(1.59)
5eeger
1993
Ullmann Breilng, Flemming l inde, HIII
Miller VOl 1994
Clausing Blessing
Pahl (Editor) 1995
1996
1997
1998
1999
Ehrlenspiel Pahi/ Beitz WalliKe, Blessing; Bauen (PahVBeit2) Bralla Cross, Christiaans, Dorst Hazehlgg
Waldron, Waldron Frey, Rivin, Hatamura Magrab Frankenberger:. Badke--Schaub, Birkhofer Hyman Pahi/ Beitz Terninko, Zusman, Zlotin, Herb (e
fntegrated Product and Process Designand Development Konstrukteure alswi
(1.204(
(1.123)
(1.136)
(1.202(
USA
(1.111)
DE
(1.55)
Fundamentals of Englnee.lng Oeslgn USA Koreantranslationof Pahi/Beitz Engineering Design KR Systematic lnnovation: An lntroduction to TRIZ USA
Oen.lc· und Handlungsweisen beim Konstruieren lntroduction to Enginééring Design
literature
DE AU
(1.91) (1.133) (1.1 76(
(1.128( (1.168(
24
1 Jntroduction Tablel.l. {continued)
Year
Author
Themerrrtte
Country
VOl
VDI·Richtlinie 2223 {Entwurf): Methodlsches Entweden te
DE
[1.194)
BR
[1.1321
USA
[1.9)
Produktinnovationmit strategischer Pfanung
DE
[1.64)
lnnovatlve Conceptual Deslgn: Parameter Analysís Entwurfsdenken und Oarstellungshandeln, Verfestigung vonGedanken beimKonzipieren Produk'tdatenmanagement·Systeme Konstruierensicherheitsgerechter Produkte Grundlagen der klassis
USA
[1.106)
DE
[1.166]
DE DE DE DE
[1.44) [1 .120) [1.12SJ [1.201)
2000
Pahi/Beitz
2001
Antonsson, cagan Gauserneytr, Ebbesmeyer, Kallmeyer Kroll, Condoor,
Jansson 2002
Sachse Eigner, Stel!er NeudOrfer Orloff Wagner
literature
the u ndergraduate and graduate levels, this book sets out a comprehensive d esign methodology fo r all phases of the product plan ning, design and development processes for tech nica lsystems. Most ofthe argu men ts a re elabo rations of a seminal series of pape rs published by the authors Pahl and Bei tz [1.142] a nd p revious cd itions o f this boo k. lt should be cm phasiscd that betwccn thc publication of the first Germa n ed ition o f the book in 1977 a nd th e la test edi tion, no n e of the statcmcnts had to be dropp cd bccausc thcy wcrc outdatcd. As befare, altho ugh ou r own approach to design does not claim to be the final word on the subject it tri es to: • be useful in design practice and design education • p rovide a "toolbox" of design methods prese nted in a co mpatible way without exprcssing a particular school of thought or including short-livcd trcnds • em phasise the impor tance of design fundamentals, principies a nd g uidelines at a ti me when more and mo re products a re des ig ned wi th the help o f co mp uters and many assemblies and components are outsourced • serve as a guide to help designers and desig n leaders manage successful product developmen t irrespec tive ofthe o rganisatio nal structure (project :management, however, is no t the focus of this book). We hope that th is systemat ic approach to eng ineering d esign may serve as an introduction a nd springboard fo r the learner, as a help and illustration for the teacher, and as a source of information and further lear n ing fo r the practitioncr. lt is impo rtan! to rcalisc that thc mctbods a nd guid clin cs prcsen ted he re u nd erpin successful product d evelopment and prod uct improvemen t.
J.2
Necessity for Systematk Oesign
25
Readers who a re familiar with the application of generally applicable design methods a nd the fu ndamentals of systema tic d esign can jump to Chap ter S a nd start dircctly with thc systcmatic approach to product dcvclopmcn t, rcturning to the fundamentals described in Chapters 2- 4 when necessary. However, it is cxtremcly impo rtan! that stud cnts and noviccs build a so lid fou ndatio n and d o no t igno re these early chapters.
2
Fundamentals
To develop an approach to design that ca n serve as a strategy for the develop· ment of solutions, we m ust first examine lhe funda mentals of lech n ical systems and procedures along with the prerequ isites for computer su pport . Only when that has been don e is it possib le to make d etailed recommen dations for design work.
2.1 Fundamentals of Technical Systems 2.1.1 Systems, Plant, Equipment, Machines, Assemblies and Components Technical tasks are performed with the help of technical artefacts that include plant, equipmen t, machin es, assemb lies and components, listed h er e in approxi· mate ord er ofthe ir co mp lexity. These terms m ay not have identical uses in differen t tields. Thus, a piece o f equipment (reactor, evaporator) is sometimes c onsidered to be more complex tha n a plant, and artefacts described as "plant" in certa in fields m ay be described as "machines" in olhe rs. A m ach ine consists of assemblies and com pon cnts. Control eq uipment is uscd in plant and machines a likc and m ay a lso be made up o f asse mblies a nd com ponents, and perhaps even of small machi nes. The variation in use of these terms reftect h istorical developments a nd application areas. Th ere are a ttempts to define slandards in which energy·transform ing lech nica l artefacts are referred toas mach ines, materia l·lran sform ing arlefacts as apparatus and signal-transforming artefacts as devices. lt is evid ent that a clear d ivision on the basis of these characteristics is no t a lways possible and that the current term ino logy is not ideal. Therc is much to be said for Hubka's suggcstion [2.22 - 2.24] that tech n ical arte· facts should be treated as systems conn ected lo the environment by mean s of inputs and outputs. A system can be divided inlo subsys1ems. What belongs to a particular system is determine
28
2 1:undamentals
Figure 2.1 . System"coupUng": a ... h systemefements; i ...1connecting elements, Soverall system; S1 subsystem "flexible coupling"; S2 subsystem *dutch"; 1 inpuu; O outputs
machines, can be consid ered to be an assembly. This coupling assembly can be treated as two subsystems- a "flexible co upling" and a "clutch". Each subsystem can, in turn, be subd ivided into system elements, in this case components. The system depicted in Figu re 2.1 is based on its mechan ical construction, referred to as the construction structure, see Figure 2.13. lt is, however, equally possiblc to consid cr it in tcrms of its functions (scc Scction 2.1.3). In that case, the to tal system "coupling" can be split up into the subsystems "d!amping" a nd "clutchi ng"; the second subsystem into the further subsystems "changing clutch operating force into nor mal force" and "transferring torq ue". For example, the system element g could be treated as a subsystem whose fu nction is to co nvert the actuating force into a larger normal force acting on the friction surface, and through its llexibility provide sorne equalisation of the wear. Which viewpoint is used to divide the system depends on the in tended p urpose o f the d ivision. Common vicwpoints are: • Function: uscd to idcntify or describe thc functional rclationships • Assembly: used to pla n assembly operations • Production: used to facilitate production and production plannin.g. Depending on their use, any number of such subdivisions m ay be m a de. Designers have to establish par ticular systems for particu lar p urposes, and must specify their
2.1 Fundamentals ofTcchnical ~'ystems
29
various inputs and o utp uts and fix their boundaries. In doing this, ·t hey may use what ter minology they prefer or is c ustomary in their particular field.
2.1.2 Conversion of Energy, Material and Signals One encounters matter in many shapes and forms. lts natu ral fo rm, or the for m imposed upon it, provides infor matio n about its possible uses. Matter without form is inconceivable-form is a pr imary source of information about the state of matter. With the development of physics, the concept offorce became essential. Force was conceived as being the means by wh ich the motion of matter was changcd. Ultimatcly this proccss was cxplaincd in tcrms of encrgy. Thc thcory of relativity postulated the equivalence of energy and matter. Weizsacker (2.6l] lists energy, matter a nd info rmation as basic concep ts. lf change o r flow is involved, time must be introduced as a fundamental quantity. Only by reference to time does the physical eve nt in questio n beco me comp rehe nsible, and can the interp lay of energy, matter and information be adequately d escribed. In the technical sphere the previous terminology is usually linked to concrete physical or technical representations. Energy is often specified by its manifest for m. We speak of, say, mechanical, electrical or optical energy. For matter, it is usual to substitute material with such properties as weight, colour, condition, etc. The general concep t of info rmation is generally given more concrete exp ressio n by means of the term signal-that is, the physical form in which the in formation is conveyed. Jnformation exchanged between people is often called a message [2.20]. The analysis oftech nical systems- plant, equipment, m ach ine, dev·ice, assembly or componen t-makes it clear that all of them involve technical processes in wh ich energy, material and signa ls are channelled and converted. Such conversions of energy, material a nd sig nals have been analysed by Rodenacker [2.46 ]. Energycan be convcrtcd in a varictyofways. An clcctric motor convcrts clcctrical into mechan ical and thermal energy, a combustion engine converts chemical into mechan ical and thermal energy, a nuclear power station converts nuclear into thermal energy, and so on. Materials too can be co n verted in a variety of ways. They can be mixed, separated, dyed, coated, packed, transpor ted, reshaped and have their state changed. Raw materials are turned into par t-finished and finished products. Mech anical par ts are g iven particular shapes and su rface finishes and so me are destroyed for testing pu rposcs. Every plant must process information in the form of signals. Signals are received, prepared, co mpared and comb ined with others, transm itted, d isplayed, recorded, and so on. In technical processes, one type of conversion (of energy, material or signals) m ay prevail over the o thers, depending on the problem or the type of solution. lt is useful to consider these conversions as flows, and the prevailing one as the main flow. lt is usually accompanied by a second type of flow, and quite frequently all three co me into play. There can, for example, be no flow of material or signals witho ut an accompanying ftow of cncrgy, howcver small. Thc provision and conversion of energy in such cases m ay not dominate, but it remains necessary to
30
2 1:undamentals
allow for them . Energy flow al so im•o h•es the transfer of forces, torq ues, c urrents, e tc., w hich are then referred toas fo rce flow, torque flow and c urrent flow. Thc convcrsion of cncrgy to p roduce cl cctrical powcr, for cxam p lc, is associatcd w ith a material convers io n, even tho ug h no conti nuous m aterial Row is v isible in a n uclear powcr s ta tion comp arcd to a coal-fircd onc. Th c associatcd flow of signals constitutes an importan! subsidiary flow for the con trol a nd regulation of the en ti re process. However, num ero us measuring instru ments receive, tran sform and disp lay signals witho ut any flow of material. In many cases energy has to be specially provided for this p urp ose; in o ther cases latent en ergy can be drawn up on d irectly. Every flow o f signals is associatcd with a flow of energy, though not n eccssarily with a flow of material. In what fo llows, we shall be dealing w ith: • Energy: m echan ical, th ermal, electrical, ch emical, optical, n uclear . .. , also force, cu rre nt, heat ... • Material: gas, Ji quid , so lid , dust . .. , also raw materia l, test sam ple, workpiece ... , end-product, componen! ... • Sig nals: magnitude, d isplay, control impulse, d ata, information ... In this boo k tech nical system s whose mai n flow is energy-based a.re referred to as m achines, those whosc main flow is m aterial-bascd as app aratus, a nd thosc whose main flow is signal-based as devices, u nless th ese ter ms are not in line with established termin ology. In every type of proposed conversion, quantity and quality must be taken into consideratio n if rigoro us criteria for the defin ition of the task, fo r th e cho ice of solution s and fo r evaluation are to be established. No statem ent is fu lly de fined un less its q uantita tive as well as its qualitative aspects a re taken into account. Thus, the statement "lOO kgfs of steam at 80 bar and 500 •e" is nota sufficient definition o f thc inp ut of a stcam turbine un less thcre is the furth er specification that thcse figures refer toa nom inal qua ntity of stea m a nd n ot, for instan ce, to the maximum flow capacity of the turbi ne, and th e ad missible fluc tuations in the s tate of the s team are fixed at, say, 80 bar ± S bar and 500 •e ± JO •e, that is, exte nded by a qu alitative aspect. In many application s, it is also essential to stipulate th e cose or value of the inputs and the m axim um perm issible cost of the outp uts (see [2.46 ], Catego ries: Quan tity-Quality-Cost).
E nergy M aterial
-~~~~~~~= =
E nergy' Material"
Slgn¡¡a ls
----
Signals'
-
-
Figure 2.2. Theconversion of energy, material and signals. Solvtion not yetknown; taskor fvn
2. 1 Fundamentals ofTcchnical ~'ys tems
3J
All technical systems, therefore, involve the com•ersion of energy, material and signals, which m ust be defi ned in quantitative, quali tative and economic ter ms (sce Figure 2.2) .
2.1.3 Functionallnterrelationship 1. Task-Specific Descriprion
In order to solve a technical problem, we need a system with a dear and easily reproduced rela tionship between inp uts and outputs. In the case of material convcrsions, for instance, we rcquire ident ical outputs for identical inp uts. Also, between the beginning and the end of a process, for instance filling a tank, there m ust be a clear and reproducible relationship. Such relatio nsh ips m ust always be plan ned- that is, designed to meet a specification. For the pu rpose of describi ng and solving design prob le ms, it is usefu l to apply the ter m function to the in tended inp ut/o utput relationship of a system whose purpose is to perform a task. For static processes it is eno ugh to determ ine the inputs and outp uts; for pro cesses that change with ti me (dyna mic processes), the task must be defi ned fu rther by a description of the initial and final magn itudes. At this stage there is no need to sti pulate what so lution will satisfy this ki nd offunc tion. The fu nctio n thus beco mes an abstract for mulation of the task, independent of any par ticu lar so lution. If the overall task has been adequately defined- that is, if the inputs a nd o u tputs of all the q uantities involved and their actual or required properties are k nown - then it is possible to specify the overall f rmction. An overall function can often be divided directly into identifiable subfunctions correspond ing to subtasks. The relationship between subfunctio ns a nd the overall fu nction is vcry oftcn govcrncd by certain constraints, inasm uch as so rne subfunctions have to be satisfied befo re others. On the other hand, it is usually possible to li nk subfunctions in various ways and hence to ereate variants. In all such cases, the lin ks m ust be compatible. The mean ingful and compatible combination of subfunctions into a n overall fu nction produces a so-called f unction structure, wh ic h may be var ied to satisfy the overall function. To that end it is useful to make a block d iagram in which the processes and subsystems inside a given block (black box) are initially igno red, as shown in Figure 2.3 (see also Fig ure 2.2). The symbols used to represen! subfunctions in a function structure are summarised in Figu re 2.4. Functions a re usua lly defined by statements co nsisting of a verb and a no un, for example "increase pressure", "transfer torque" a nd "reduce speed". They are derived for each task from the conversions of energy, material and signals d iscussed in Section 2.1.2. So far as is possib le, all of these d ata should be accompan ied by specifications of the physical quantities. In most mech anical en gineering applications, a combination of all three types of conversion is usually in volved, with the conversion either of material or of energy influe ncing the func tion structurc d ccisivcly. An analysis of all thc functions invo lved is always uscful (sec also [2.59 )) .
32
2 1:undamentals
Subfunction •
' 1
-Figure 2.3. E>tabllshing a func:tlon structu•e by bleaklng down an overall functlon lnto subfunctions
Types of flow: E
F!owof energy anddirection M
Rmv of material and dire-ction
S
----- ~
F1ow of signalsand direction
System: System boundary
Function:
D
Main function
,.------, 1
1
1 _ _ _ _J1 ¡_
Auxiliary function
Figure 2.4. Symbols used to represent subfunctions ina functionstructure
Jt is useful to d isting uish between main and auxiliary functions. While main
funcrions are those subfunctions that serve the overall function directly, auxiliary functions are thosc that con tributc to it indircctly. Thcy have a supportivc or complementary character and are often determined by the nature of the solutions for the main functions. These defi nitions are derived from Value Analysis (2.7,2.58, 2.60 ]. Although it may not always be possib le to make a clear distinction between
2. 1 Fundamentals ofTcchnical ~'ys tems
33
main and auxiliary functions, the terms are nevertheless useful. The division between them should be managed in a flexible manner. For example, a change in the systcm bo undary resulting from a changc of focus can transform an a uxiliary fu nction into a mai n functio n and vice versa. It is also neccssary to exam ine thc rclationship betwccn the various subfunctions, and to pay particular attention to their logical sequen ce o r requ ired arrangement. As an example, co nsider the packing of carpet tiles sta mped out of a length of carpet. The first task is to introduce a method of contro l so that perfect tiles can be selected, counted a nd packed in specified lots. The main flow here is that of material, as shown in thc form o f a block d iagram in Figure 2.5, whicb, in this case, is the only possible sequence. On closer exam ination we d iscover that this chain of subfunctions requires the introduction of auxiliary fu nctions because: • the stam ping-out process creatcs offcuts that m ust be removed • rejects must be removed separately a nd reprocessed • packi ng mate rial must be brought in. The result is the fu nction struc ture shown in Figure 2.6. lt wi ll be seen that the subfunction "cou nt tiles" can also give the signa! to pack the tiles into lo ts of a specified size, so it seems useful to introd uce a signa! flow with the subfunction "send signa! to com bine n tiles into one lot" into the function structure. The functions in this case are task-specific functions, whose defin itions are derived fro m the ter mino logy appropriate for the task being considered. Outside the design do main, the ter m function is sometimes used in a broadcr sense, and sometimes in a narrower sen se, depend ing on the con text. Brockh aus [2.40 J has defined functions in general as activities, effects, goals and constraints. In mathematics, a function is the association of a magnitude y with a magnitude x such that a unique va lue (single-valued function) or more than one valu e (multi-valued func tion) ofy is ass ig ned for every value of x. According to the valuc analysis definition g ivcn in [2.7], functions define thc behaviour of artefacts (tasks, activities, characteristics).
loose
Carpet tiles packed in lots
(omblnt lnlo
l
...,. Materialftow Cl Maln functlon
---System boundary
f igure 2.5. Function structure for the packing ofcarpet tiles
34
2 1:undamentals
_,. I.IJ~IIoN
- 51o:;r;IIIOil
r.-'! M~ fJn(\iOII'I
t::iAullliuy(unnkn
- - - 'V¡tMI hlhfiiiJII'F
Figure 2.6. Function SUlKture for the packing of
2. Generally Va/id Description Various desig n methodologists (see Section 1.2.3) have put forward wider or stricter defi nitions of general/y va/id fmrctions. In theory, it is possiible to classify functions so that thc lowcst lcvcl of the function structurc consists cxclusivcly of functions that cannot be subdivided further wh ile remain ing generaaly applicable. They therefo re rep resen! a high leve! of abstraction. Rodenacker [2.46 ) has defined generally va lid functions in ternlS of binary logic, Roth [2.47,2.49 ] in terms of their general applicabi lity, and Koller [2.28,2.29] in ternlS of the required physical effects. Krumhauer [2.31) has exam ined general functions in the light of possible com p uter applications d uring the co nceptual design phase, paying s pecial attention to the relationship between inputs and o utputs after c hanges in type, magnitude, nu mber, place a nd time. By and large, he arrives at the same functions as Roth, except that by "change" he refe rs exclusive! y to c hanges in the type o f inp ut and o utput, wh ile by " increase or d ec rease" he rcfers cxclusivcly to changes in magnitudc. In the context of the design methodology presented here, the generally valid functions ofKr umbauer will be used (see Figure 2.7). T he func tion chain shown in Figure 2.5 can be represented using gene rally valid functions, as shown in Figure 2.8. A comparison between th e fu nctional representations in Figures 2.5 and 2.8 shows that thc dcscription tbat uses gcncrally valid functions has a big hcr level of abstraction. For th is reason, it leaves open all possible solutions and makes a systematic approach easier. Howeve r, usi ng generally valid functions can represen! a problem because such an abstrae! leve! can sometimes h inder the di rect sea rch for solutio ns. For more about the app li cation of taskspecific a nd generally valid func tions, alo ng with further examples, see Section 6.3.
3. Logical Description Thc logical analysis of functio nal rclationships starts with thc scarch for thc cssential ones that must necessarily appear in a system if the overall problem is to
2. 1 Fundamentals ofTcchnical ~'ys tems
Charaderistic Input (1)/0utput (0)
Generally valid functions
Symbols
Explanations
Change
- (.¿J -
ourwOcJrd formof
35
Type ar.d Type
1andO differ M>gnitude
Vary
Number
Conne
Plate
Channel
- ...g:j--
'
::123--
Numberof i > O Numberof i
----g:¡::
-o-
Place ofl ;tO Placeofl = O
--m
.. ···-
-ITII-
Store
Time
1< 0 1> 0
--c:23-
Timeofl ,¡ o
Fig ure 2.7. Generally valid fun
r
-
,...-· -·- · - ·- · -·- ····
·-
·-
....,
1 Signal -Material
1
Energy
L.-
·-
·-
·- · - -· --
·-
·-
·- · - ·
Figure 2.8. Same function structure a.s shown inFigure l .S but represented using generallyvalidfunctions. as defined in
Figure 2.7
be solved. lt m ay equally well be the relationships between subfunctions as those between inp uts a nd outputs of particular subfunctions. Let us first of ali Jo o k at thc relationships between subfunctions. As wc have pointed o ut, certain subfunctions must be satisfied befo re another subfunction can be mea ningfully introduced. The so -called "if- then" relatio nsh ip helps to clar ify this point: if subfunction A is present, then subfunction B can come into effect, and so on. Often severa] subfunctions must all be satisfied simultaneously before ano ther subfunction can be put into effect. The arrangement of subfu nctions thus determines the structure of the energy, material and signa] conversions under consid eration . Thus, d uring a test of tensile strength, the first subfunction - "load specimen" - must be satisfied before the o ther subfunctions-"measure force" and "mcasu rc dcformation"- can be dcp loycd. Thc Jast two subfunctions, morcover, must be satisfied simultaneo usly. Attentio n must be paid to consistency a nd ordcr within thc flow undcr consid cration, and this is do ne by thc u nambiguo us combination of the subfunctions.
36
2 1:undamentals
AND-function (Conjunction)
Designation
x,=EJ-&
Symbol
X¡
Xz
y
Boolean a1gebra {Functíon)
NOT-fl!lnctlon (Negation)
x'E-r x-@-r
y
X¡
¡
1!-. o
Truth table
OR-functlon (Disjunction)
1
oo oo
~
o
1 1 1
x~
Xz y
G 1
o! 1 o 1 t o1 o í 1 1 o' 1 t i 1
$ y,x
y.r., vx,
Y•X1 f,X·
o
Figure 2.9. logical functions.X independent statement (signal); Y dependent statement; "0", "1" value ofstatement. e.g. "off", "'on"'
INHIBITION
AND
X¡
Xz
r
1Signal suppEed) (C1utch engaged) (loH¡ue transmitted}
o 1 o oo 1 3 o o
.,
x, o X¡
1
y
f igurt 2.10. logical functionoftwo dutches
1 1
1 o 1 oo 1 ooo
2. 1 Fundamentals ofTcchnical ~'ys tems
37
Logical relationsh ips, moreover, m ust also be established between the inpu ts and outputs of a partic ular subfunction. Tn m ost cases there are severa! inputs and outputs whosc rclationships can be trcatcd likc propositions in binary logic. Elernentary logicallinks of the inp ut a nd output magnitudes exist fo r th is pu rposc. In binary logic thcsc are statcmcnts sucb as truc/falsc, ycs/no, in /out, fullilled/un fulti lled, present/not p resent, wbich can be com pu ted using Boolean algebra. We distinguish between AND func tions, OR fu nction s a nd NOT func tions, and also between their combination into more complex NOR functions (OR with NOT), NAN D functio ns (AND w ith NOT) and storage func tions w ith the help of ftip-ftops [2.4, 2.45,2.46 ]. Groupcd together, thcsc are callcd logical f unctions. In the case of ANO function s, all s ignals on the inpu t s ide must have the same validity if a valid s ig na! is to a ppear on the outp ut side. In the case of OR functions, only one s igna[ n eeds to be valid on the input side if a valid signa! is to appear on the o utput side. In the case of NOT fun ctions, the signa! on the inp ut side is negated so that the negated s ig na! appears on the outp ut side. All of these logical functions can be expressed by standard s ymb<>ls, which can be fo und in [2.4 ]. The logical validi ty of any s ig na! can be read from the t ruth table shown in Fig ure 2.9, in wh ich all of tbe inputs a re combined systematkally lo yield the releva n! oulp uts. The Boolean equations have been added for th e sake of co mpleleness. Us ing logical fu nctions it is possib le lo con stru ct complex swi tcbess an d thus to increase lhe safety and reliability of control and commu nicalion systems.
Bearing 1
&
&
Bearing 2
&
Bearing 3
&
Bearing 4
&
&
~
1
&
Fig ure 2.11 . Logic.al functionsfor monitoringa bearinglubñcationsystem. A positive signal for every bearing (oil present) permits operation. Monitor pressure p; monitor oil ftow V
38
2 1:undamentals
Figure 2.10 shows two mechan ical clutches with their characteristic logical functions. The workings of the clutch o n the left can be represented by a simple AND function (thc signa! must be scnt and thc clutch cngagcd befo re thc torquc can be transm itted). The clutch on the right has been constructed such that, when thc opcrating sig na! is givcn, thc clutch is discngagcd, mcan ing that X1 must be negative ifthe torque is to be transmitted. ln other wo rds, only X2 m ust be presen t or positive if the desired effect is to be produced . Figure 2.11 shows a logical system for mon itoring the bearing lubrication system of a m ulti-bearing m achine shaft involving AND and OR functions. Every bearing posi tion is mon itored for oil pressure and oi l Row by comparing a specified or target valu e with thc actual val u c. Howcvcr, on ly o nc positivc value for cach bcaring position is needed to allow the system to operate.
2.1.4 Working lnterrelationship Establisb ing a function st ructurc faci lita tcs thc discovcry of so lutions bccausc it simplifies the general search for them, and also because solu tions to subfunctions ca n be elabo rated separately. Ind ividual subfunctions, originally represented by "black boxes", must now be replaced with more concrete statemen ts. Subfunctions are usually fu lfilled by phys ical, c hemical or biological processes-mechan ical eng ineer ing so lu tions are based main ly on physical processes whercas proccss cng incering solutions are based mainly on chemical and biological processes. lf, in what follows, we refer to physical processes, wc tacitly includc thc cffccts of possiblc ch cm ical and biological processes. A phys ical process realised by the selected physical effects and the deter mined geometric and material characteristics results in a working interrelationship that ensures the function is fulfilled in accordance with the task. Hen ce a working interrelationship comes into existence thro ugh p hysical effects in com bination with the chosen geometric and material characteristics. 1. Physical Effects
Physical effects can be described quanti tatively by means of the p hysical laws govcrn ing the physical quantities involvcd. Thus, thc friction cffcct is dcscribcd by Coulomb's law, Fp = pFN; the lever effect by the lever law FA · a = Fs · b; and thc cxpansion cffcct by tbc cxpansion law !l./ = a· f. fl.{} (scc Fig ure 2.12). Rodenacker [2.46] and Koller [2.28], in particular, have collated such effects. Severa! physical effects m ay have to be combined in order to fuUi l a subfunction. Thus the operation of a bimetallic strip is the result of a combination of two effects, namely thermal expansion and elasticity. A subfu nction can often be fulfilled by one of a n umber of physical effects. Thus a force can be amp lified by tbe lever effect, tbe wedge effect, tbe ele·ctromagnetic effect, the hydrau lic effect, etc. T he physical effect chosen for a particular subfunction must, however, be com patible with the physical effects of other related
2.1 Fundamentals ofTcchnical ~'ystems Subfunctlon
Physical errect {lndependent of solutionl
39
WO
Friction
Expansion 1
¡ -:
ñ TI' 'rJ -· tll=a·f·l}
Fig ure 2.12. Fuffilling subfunctions by worting principies built up from physi
characteristics
subfunctions. A hydraulic a mplificr, for instancc, cannot be powcrcd directly by an electric battery. Moreover, a particular physical effect can only fulfil a subfunction optimally u nder certain conditions. Thus a pneumatic control system will be superior to a mechanical or electrical control system only in particular ci.rcumstances. As a ru le, compatibility and optimal fulfilment can only be realistically assessed in relatio n to the overall func tion once the geometric and material characteristics have beco cstablishcd more concrctcly.
2. Geometric and Material Characteristics The place where the physical process actually takes effect is the working location, i.e. the specific active location that is the focus of interest at the time. A function is fu lfi lled by the physical effect, which is realised by the working geometry, i.e. the arrangement of working surfaces (or working spaces), and by the choice of working motions [2.33) . The working surfaces are varied with respect to and determined by: • 'I'ype • Shape • Positio n • Size • Numbcr [2.46 ].
40
2 1:undamentals
Similarly, the required working motions are determ ined by: • Type: translation-rotation • Nature: regu la r-irregu lar • Dircction: in x-, y-, z-dircctions and/or about x -, y -, z-axcs • Magn itud e: velocity, etc. • Number: one, severa!, etc. In addition, we need a general idea of the type of material with which the working surfaces are to be produced, for example, whether it is solid, liquid or gaseous; rigid or flexib le; elastic o r plastic; stiff, hard or tough; or corrosio n-resistant. A general idea of the final embodiment is often insufficient; the main material properties must be specified before a worki ng interrelationship ca n be formu lated adequately (see Fig ure 3.18). On ly the combination ofthe physical effect with the geometric and m aterial characteristics (working surfaces, work ing motions and materials) allows the principie o f the solu tion to emerge. Th is interrelationshi p is called the working principie (Hansen [2.19 ] refers to this as the working means), and it is the first concrete step in the implementation of the solution. Figure 2.12 shows so rne cxam plcs: • Transferring the torque through friction against a cylindrical working surface in accordance with Coulomb's law will, depending on the way in which the normal force is applied, lead to the selection of a shrink fit ora clamp con nection as the wo rking princip ie. • Amplifying muscular force with the help of a lever in accordance with the lever law after d eter mining the pivot a nd force application poi nts (work ing geometry) and considering thc necessary working motion willlcad to a description o f th c wo rking principie {lever solution, eccen tric solution, etc.) . • Making elec trical contact by br idging a gap usi ng the expansion effect, applied in accordance with the li near expansion law, only leads to an ov-erall work ing principie after d etermination of the sizes (e.g. the d iameter and le ngth) and the positions ofthe workingsurfacesneeded for the workingmotion ofthe expanding med ium : a material. For example, either mercury expand ing by a fixed amou nt ora bimetallic strip serving as a sw itch. To satisfy the overall fu nction, the working principies of the ''arious subfunctions have to be co mbi ned (see Sectio n 3.2.4). T here a re obviously severa! ways in which this can be done. Guid eline VDI 2222 [2 .55] calls each co mbination a combination
ofprincipies. T he combination of severa! working princip ies results in the workitrg structure of a solution. 1t is through this com bination of working principies that the solution principie for fulfilling the overall task can be recognised . The working structu re derived from the functio n structure thus represents how the solution wi ll work a t thc fundamental principie leve!. Hubka rcfcrs to the working str ucturc as thc organ structu re (2.22-2.24].
2.1 Fundamentals ofTcchnical ~'ystems
41
For known elements, a circuit d iagram or a flow ch art is sufficient as a means of representing a work ing structure. Mechanical artefacts can be effectively reprcscntcd using cnginccring drawings, though ncw or u ncommon c lcmcnts may require add itional explanatory sketches (see Figures 2.12 and 2.13). Oftcn thc working s tructurc alonc will no t be concrete cnough to cvaluatc the solution principie. 1t may need to be q uantified, for example by preliminary calculatio ns and rough scale drawings, befo re the solution principie ca n be fixed. Th e result is called a principie solution.
lnterrelationships
Elements
Examples
suuctures
r. Clutcll F,~ 11 ' torque F.
Functional
imerrelatioruhip
Functions structure
Functions
~--·-Chang~dutc~----~ operating fon:efs into normal force r,
1
~1 1
.
Input r,
L
Artefact to be developedi
·- -·- -- - - - -·- - ·----'
Lewreffect
F, ~
Physi
geometric and
material
Working structure
chcuacteristics
f Working prindples Constructional interrelationship
Components
Joints Assemblies
j
i
b
¡ro
fo·o • fo· b F, =f, J 0 • Fs
Friction efte
r.~ .H
>;~;;:;
R·Fr •f.J. ·FN F,
1 ~ :j11 1 1 AAI I III bFs r,· ·-
rmiT1ilili1l
r,
Construction structure
r -·- - - - l
i
System interrelatlonship
Artefacts Human beings Environmem
System structure
rr ni1
~o_~ ~
Figurt 2.13. lnterrelationships in technical systems
T;
42
2 1:undamentals
2.1.5 Constructionallnterrelationship The working interrelationship established in the working structure is the starting point for further concretisation leading to the construction structure. This ínter· relatio nsh ip represents the conc rete technical artefact o r system by defining the components, assemblies and machines and their intercon nections. The construc· tion structure takes into account the needs of production, assembly, transport, etc. Figure 2.13 shows the fundamental interrelationships for the clutch shown in Figure 2.1. The increasing levels of concretisation can be seen clearly. The concrete elements of a construction structure must satisfy the requirements o f the selected working structure plus any other requirements necessary for the technical system to opera te as in tended . To identify these requirements fully, it is usually necessary to consider the syste m interrelationship.
2.1.6 System lnterrelationship Technical artefacts and systems d o not operate in isolation and are, in general, part of a larger system. To fulfi l its overall function, such a system often involves human beings who inft uence it through itrput effects (operating, controlling, etc.) . The system returns feedback effects or sig nals that lead to further actions (see Figure 2.14). ln this way, human beings support or enable the intended effect of the tech nical system. Apart from d esired inputs, undesired ones from the environ me nt and from neighbour ing systems can affect a technical system. Such disturbing effects (e.g. excess temperatures) can cause undesired side-effects (e.g. deviations from shape o r shifts in position). Also, it is possible that in addition to the desired working interrelationship (intended effects), unwanted phenomena can occm (e.g. vibrations) as side-effects from individual co mponents within the system or from the overall system itself. These side -effects can have an adverse effect on humans or the envi ron me nt.
E M
S
,-----·-¡-1
-- ' - --
Technical
artifa
f
'
Oisturbing effe
j
rrom env1ronmenc1
1
.r j '
lnJut eflects Feedback effects and slde effects
} lntended effe
·-1-·1Sideeffectson h be d 1 u~an ings an env1ronment
1
Figure 2.1 4. lnterrefationships in te
2.1 Fundamentals ofTcchnical ~'ystems
43
In accordance with Fig ure 2.14 it is usefulto make the following distinctions (after [2.56]): In tended cffcct: Input effect:
Functionally dcsircd cffcct in thc sen se of systcm opcration . Functional relationship due to human action on a technical system. Feedback effect: Functional relationship dueto the action of a tech nical system on a human or another technical system. Disturbing effect: Fu nctio nally undesired infi uence from outside on a technical system or human that makes it difficult for a system lo fulfil its function. Side effect: Func tionally u ndesired and unintended effect of a technical systcm on a human or on thc cnvironmcnt. The overall interrelationship of allthese effects mus! be carefully considered during the development of technical systems. To help recognise the m in time, so that des ired effec ts can be used and undesi red ones avoided, il is helpful to fo llow a systematic guideline that adheres to the general objectives and constraints in Section 2.1.7.
2.1.7 SystematicGuideline The so lution of techn icaltasks is determined by the general objectives and con straints. The fulfilment of the tec/mica/ function, the attaimnent of economic feasibility and the observartce of safety requirements for hum ans and the env iro nm ent can be considered as general objectives. The fulfilment of the techn ical functio n alonc do es not complete the task of designers; it would simply be an cnd unto itsclf. Econom ic feasibility is another essential requirement, and concern with human and environmcntal safcty must imposc itsclf for cthical rcasons. Evcry onc of these objeclives has direct repercussions on the res t. In addition, the so lution o f technicaltasks imposes cer tain cons traints or require ments resulting from ergono mics, production methods, transport facilities, the in tended operation, etc., no matter whe ther these co nstraints are the result of the particular task o r the general state of technology. In the first case we speak of task-specific constraints, in the second of general constraints that, altho ugh often not specified explicitly, must nevertheless be taken into account. Hubka (2 .22- 2.24) separates the proper ties affected by the constraints into catcgorics bascd variously on industrial, crgonom ic, acsthctic, distrib ution, dclivcry, planning, design, production and economic factors. Besidcs satisfying thc functional and working interrelationships, a solution must also satisfy certain general or task-specific constraints. These can be classified under the following headings: • Safety • Ergo nom ics
also in the wider sen se of reliability and availability
• Productio n
production facilities and type of production
human -machine co ntext, also aesthetics
44
2 1:undamentals
• Quality control througho ut the design and production process • Assembly
during and after the production of componen ts inside and o utside of the factory
• Transport • Maintenance
in tended use, handling upkeep, inspection and repair
• Expend iture
costs a nd schedules
• Rccycling
reuse, rcconstitution, disposal, final storagc.
• Operation
The characteristics that can be derived from these constraints, whicn are generally for mulated as requirements (see Sectio n 5.2), affect the function, work ing a nd construction str uctures, a nd also inll uence one a nother. llence they should be treated as guidelines throughout the design process, and ad apted to each leve! of embodiment (see Figs. 2.15 and 12.3). In ad dition thcre are inlluenccs from thc designer, thc developmcnt team a nd the su ppliers as well as the customer, the specific context a nd the environmen t. lt is advisable to consider these gu ideli nes even during the conceptual phase, at least in essence. During the embodiment phase, when the layout and form d esig n of the mo re or less q ualitatively elaborated working structure is first q uantified, both the objectives of the task and al so the general and task-specific constraints must be considered in concrete d etail. This involves severa! steps- the collect ion of fur ther information,layout ancl form design, and the elimination of weak spots, together with a fresh, if limited, search for solutions fo r a variety of subtasks, un til fi nally,
Oesigner Te,¡ m
Suppfier
Guideline Figure 2.1S. lnfluences and constraints duringdesign and dtovelopment. These
2.2 f:undamentals of the Systematic Approach
45
in the detail pllase, the elaboration of detailed production instructions brings the design process toa co nclusio n (see Chap ters 5 to 7) .
2.2 Fundamentals ofthe Systematic Approach Befo re we deal wi th th e specific steps a nd rules of systematic desig n, we m ust first d iscuss cognitive psycho logical relationships and general methodical principies. These help to structu re the proposed procedures and ind ividual me thods so that they can be applied to the so lution o f desig n tasks in a purposeful way. The ideas come from a host of differe nt disci plines, main ly non-tech nical o nes, and a re usually b uilt on interdisciplinary fu ndamentals. Work science, psychology and p hilosophy a re among the mai n inspirations, which is not su rpr is ing when we co nsid er that methods desig ned to improve working procedures im pinge on tbe qualities, capacities and limitations o f human thought [2.411.
2.2.1 Problem Solving Process Designers are often confro nted with tasks co ntaining problems tbey cannot solve imm ediately. Prob le m so lvi ng in d ifferent areas of application and at different levels of concretisation is a charac teristic of tbeir work. Researching the essence of hu man thin king is the focus of cognitive psycho logy. The resu lts of this research m ust be taken into account in engineering design. Tbe following sections are based largely on the wo rk of Di:irner 12.8,2.10]. A problem h as three components: • a n undesirable initial state, i.e. the existen ce of a n unsatisfactory s ituation • a desirable goal state, i.e. the realisation of a satisfacto ry situation • obstaclcs that preven! a transformation fro m thc undcsirablc initial statc to thc desirable goal state ata particular point in time. An obstacle that prevents a transforma! ion can arise from the following: • The means to overcome tbe obstacle are un known and have to be found (synthesis o r operator proble m). • The means are known, b ut they a re so nu merous or involve so many co mbinations that a systematic investigation is impossible (interpolation problem, combination a nd selection p roble m). • The goals are only k nown vague!y o r are no t formu la ted clearly. Find ing a solution involves continuous d elibera! io n and the removal of con tlicts u ntil a satisfactory situation is reached (dialectic problem, search and application problem ). A problem has the following typical characteristics:
• Complexity: many com ponents are involved and these components, tbro ug h lin ks of different strength, influence each other.
46
2 1:undamentals
• Uncertainty: not al! req uirements are known; not al! criteria are established; the effect of a partial solu tion on the overall so lution or on o ther partial solutio ns is no t fully understood, or only emerges gradually. The d ifficulties beco me more prono unced if the charac teristics of the problem area change with ti me. A task is distinct fro m a problem because: • A task imposes mental requirements fo r which various mea ns and methods are available to assist. An example is the desig n of a shaft with g iven loads, connecting dimensions and production methods. Tasks and problems occur in design in a nu mber of ways, often combined and not clearly separable initially. A specific desig n task can, for example, turn o ut to be a problcm when looked at mo re doscly. Many large tasks can be divid cd into subtasks, sorne of which can reveal difficult subproblems. On the other hand, it is sometimes possible for a problem to be solved by fulfilling severa] su btasks in a previously un known combination. Thinking processes take place in the brain and invo lve changes in memory con ten t. When thi nking, the contents o f the memory, and the way in which they are lin ked, play an importan! ro le. In simple ter ms, o ne can say that in order to star t solving a problem humans need a certain leve] of factual knowledge abo ut the domain of thc problcm. In cognitive psychology, when this knowledge has been transfer red into memo ry it rep resents the epistemic structure. Hu ma ns also need cer tain procedures (methods) to fi nd so lutions and to fi nd these effectively. This aspect involves the heu ristic structure of human thought. lt is possib le to d istingu ish between short-ter m and long-term me mory. Shor tterm memor y is a kind o f working storage. lt has limited capacity and can only retain about seven arguments or facts at the same time. Long-term memory probably has u nli mited capacity a nd contai ns fac tual and heu ristic knowledge that appears to be stored in a structured way. In this way, humans are able to recog nise specific relationsh ips in many possib le ways, to use these relationsh ips and to create new o nes. Such relationships are very importan! in the tech nical domain, for example: • co ncrete- abstrae! relationship e.g. angula r contact bearing- ball bearing- rolling element bearing- bearing-gu ide-transfer fo rce a nd lo cate co mponent. • whole- part relationship (hierarchy) e.g. plant- machine-assembly-component. • space and time relationships e.g. arrangement: front- back, below- above, e.g. sequence: th is first-that next. The memo ry ca n be thought of as a seman tic netwo rk with no des (knowledge) and connections (relat ionships) wh ich can be mod ified and extended. Figu re 2.16 shows a possible, though not necessarily comp le te, semantic network related to
2.2 f:undamentals of the Systematic Approach
47
Figure 2.16. Extra
the term "bearing". In this network it is possible to recognise the relationships mentio ned above as well as others, such as property relationships a nd ones in dicating opposites (polar relationships). Thinking involves buildi ng and restr ucturing such semantic netwo rks, and th e thinking process itself can proceed intuitively or d iscursively. Intuitive thinking is strongly associated with flash es of inspiration. The actual thinking process takes place toa large extent unconscio usly. lnsights appear in the conscious mind sudd enly, caused by sorne trigger or association. This is referred toas prima ry creativity [2 .2,2.301 and involves processing quite com plex relations. In this co ntext, Müller [2.36 ) refers to "si lent knowledge", which in eludes co mmon and background k nowlcdgc. This is also thc knowlcdgc that is availablc whcn one deals with episodic memories, vague concep ts and imprecise d efinitions. lt is activated by both co nscious and un conscious thinking activities. Generally time is needed for undisturbed and unconscious "thinking" before sudden insig hts appear. The length of this incubation period can not be predetermined. lnsights can be tr iggered, for example, by producing freehand sketches or
48
2 1:undamentals
e ngineering drawings of solu tion ideas. According to [2.14], these manual ac tivities foc us concentration on the subject, b ut sti llleave space in the mind that can by used by unconscious th in king proccsscs, wh ich can also be stim ulatcd by su eh activitics. Discursive thinking is a conscio us process that can be co mm u nicated and in flucnccd. Facts and rclationships are conscio usly analyscd, varicd, combincd in new ways, checked , rejected , and consid ered further. ln [2.2,2.30] th is is referred to as secondary creativ ity. This type of th inking involves checkin g exact and scien tific knowledge a nd buildi ng th is into a knowledge structure. In co ntras! to intuitive thinking, this process is slow and involves many small conscio us steps. In the memory structure, explicit and consciously acqu ired knowle-dge cann ot be scparatcd prcciscly from the vagucr com mon or backgrou nd knowlcdgc. Bcsid cs, the two types of knowledge in fluence each o ther. For k nowledge to be easily re trieved a nd combined, it is tho ug ht that an orde red and logicalstr uc ture offactual k nowledge in the m ind of the problem solver (epistemic structure) is decisive, and that th is is true whe the r the thinki ng process is intuitive or discu rsive. The heuristic structure includ es explicit knowledge (i.e. knowledge that can be explained ) as well as implicit knowledge. This is necessary in order to organise the sequence of thinking o perations, including modifying operations (searching and fin ding) and testing opera tions (c hecking a nd assessing). lt appears that problem solvers often start without a fixed plan in the hope ofimmediately finding a so lution from their k nowledge bases witho ut much effo rt. On lywhen this approach fai ls, or when co ntradictions e merge, do they adop t a more d early planned or systcmatic sequence of thin king operations. T he so-called TOTE m o d el [2.33 ] represents a n important fundamental sequen ce for thin king processes (se e Fig ure 2.17).It consists of two pro cesses: a modification process and a testing process. The TOTE model shows that befo re an operation o f change takes place, an operatio n of testing (Test) is invoked to analyse the in itial statc. On ly thcn is thc choscn opcration o f changc (Opcration) cxe-cutcd. This is followed by ano ther operation of testing (Test), during which the resulting state is checked. Ifthe result is satisfactory, the process is exited (Exit); if no t, the operatio n is adapted and repeated. In mo re co mplex thin king processes, the ron; sequences are li nked in a chain o r severa! modification processes are executed before a test ing pro cess takes place. Thus, when linking mental processes, many combinations and seque nces are possible, but al! of them can be mapped onto the basic TOTE model.
E
T=Test O = Opetation
E= EJcit
Figure 2.17. Basic TOTE mod ~ for thinking processes [2.8, 2Jll
2.2 f:undamentals of the Systematic Approach
49
2.2.2 Characteristics of Good Problem Solvers The following statements are the result of tbe wo rk of Dorner [2.9 J an d of research which has been un der taken with him by Ehrlenspiel a nd Pahl. The results of the rescarch led by Ehrlenspiel and Pahl can be found in thc publications ofRutz [2.50 ], Dylla (2.11,2.12) a nd Fricke [2.1 S, 2. 16) . Th is sect io n provid es a su m mary of their find ings [2.42] .
1. lntelligence and Creativity In general, intelligence is thoug ht to involve a certain cleverness, C()mbined with the ability to understand and judge. Analytical app roaches are often em phasised. Creativity is a n inspirational force that generales new ideas o r produces novel com binations of existing id eas, leading to further solutions or deeper understanding. Creativity is often associated wi th a n intu itive, syn thesisi ng approach. Intelligence and creativity are personal characte.ristics. Up until now it has not been possible to come u p with precise scientific definitions of ora clear distinction between intelligence a nd creativity. Attempts have been made to meas u re the leve! of intclligcnce of individuals using intclligcncc tests. The rcsulting Intelligence Quotien ts provide measures compared to the average of a large sample. Because of the different for ms in which intelligence appears, various tests are needed to get a com plete picture and d raw tentative conclusions. The same is true for creativity tests. Fo r problem so lvi ng, a mínimu m leve! of intelligence is req uired a.nd it appears that pcople with hig h Intelligencc Quoticnts are more likcly to be good problcm solvers. However, acco rding to [2.8, 2.9], intelligence tests on their own do not g ive m u eh insight into which combination of factors makcs a particular individual a good problem solver. The reason, according to Dorner [2.8), is that intelligence tests use tasks or problems that on1y require a few thin king steps to find a so1ution, so the sequence of steps seldom becomes consc io us. Few intelligence tests require a large number of steps to be organised into a s pecific problem solving procedure. Such organisation requires switchi ng between the d ifferent levels and possibilities of a general problcm solving procedurc, and is essential for the execution o f longterm thin king activities. Creativity tests too are often at such a low leve! that th ey do no t add ress complex problem solving which involves plan n ing a nd gu iding one's own approach. Furthermore, in engineering design, creativity is always focused on a specific goal. Purely unfocused generation of ideas and variants can in fact h indcc the problem solving process [2.2) or at best suppor t a specific phase of the process.
2. Decision Making Behaviour Apart from having well-structured factual knowledge, applying a systematic approach, a nd using focused creativity, des ig ners have to maste r dedsion mak ing proccsscs. For dccision making, thc following mental ac tivitics and skills are essentía!:
50
2 1:undamentals
• Recognising Dependencies In com plex systems the dependencies between the individual elements can vary in strcngth. Rccogn ising thc typcs and st rcngths of such dcpcndcncics is an essential prerequisite for dividing the problem into more manageable, less comp lcx subproblcms or subgoals so that thcsc can be addrcsscd scparatcly. However, those working on each separate subproblem must check to see how the short- and long-term effects of their own decisions infl uence the overall desig n.
• Estimating Importance and Urgency Good problem so lvers know how to recognise importance (factual sig nificance) and urgency (tempo ral sign ificancc}, and how to use this information to modify the ir approach to problems. T hey try to resolve the most important things first and then tackle the dependent subproblems. They have the co urage to be satisfied with suboptimal solutions for less significan! problems if they have good or acceptable solutions for the most sig nifican! ones. By doing th is they avo id immersi ng themselves in less relevan! issucs and thcrcby losing valuable time. The same is true when estimating the urgency. Good problem solvers estímate the time they need accurately. They prepare a demanding-b ut not impossibletime plan. )an is and Mann [2.25) have concluded that mild (i.e. bearable) st ress is importa n! fo r creativity. Therefo re, realistic ti me planning has a positive effect on thinki ng processes, and new developments should take place under reasonable time pressure. But, of course, individuals react differently to time pressure.
• Continuity and Flexibility Continuity means an appropriate and continuous focus on achieving the goals, but there is a danger that excessive focus leads toa r igid approaoeh . Flexibility means a ready ability to adap t to changing requirements. However, this sho uld not lead to purposeless jumping from one approach to another. Good problem solvers fi nd a suitable balance between continuity a nd flexibili ty. They demonstratc continuous and consisten!, but at the same time flexible, beha,•iour. They stick to the given goa ls despite any hold-u ps and difficulties they encou nter. On the other hand, theyadap t their ap proach immediatelywhen the situation changes and when new problems occur. They consider heuristics, procedures and instructions first of all as g uidelines and not as rigid prescriptions. Dor ner states [2 .8[: "Heu ristics or heuristic plans should not degenerate into automatic procedures. lndividuals should learn to develop what they have learnt. Heuristics shou ld not be misi nterpreted as prescriptions, but sho uld be treated as g uidelines that can, and often sho uld, be dcvclopcd."
• Failures Cannot be Avoided In complex systems with strong interna! dependencies, at least p ar tial failures are difficult to avoid because it is not possible to recognise all the potential
2.2 f:undamentals of the Systematic Approach
S1
effects simultaneo usly. When recognising such failures, the mo·st important thing is the way o ne reac ts. Being flexible is c rucial, supported by the ability to analysc onc's approach and thc ability to make decisions that lead to correctivc actions. The results of cogn itive psychology rcscarch are summarised bclow. Good problem solvers: • have a sound and structured technical knowledge, i.e. they have a well-structured m o del in their minds • find an appropriate bala nce between concreteness a nd abstraction, depending o n the situation • can deal with uncer tainty and fuzzy data • continuously focus on thc goals whilc ad opting a flexible decision making behaviour. Su eh heu ristic compctence depcnds largely on personal characteristics, but can be developed considerably through training on different types of proble m. The research mentioned earlier reveals that good designe rs demonstrate the following behaviour (2.42 ]: • They thoroughly analyse the goals at the begin ning of a task and contin ue todo so throug hout the design process when formulating partía! goals, in particular whe n the orig inal problem for mulation is vague. • They first generate o r identify the most suitable solution principies in a conceptual phase befo re develop ing concrete embodiments. • They initially adopt a divcrging scarch without gencrating too many variants and then q uickly converge onto a small number of solutions; they choose the appropri ate leve! of concretisation and switch easily between perspectives, e.g. abstract/concrete, overall problem/subproblem, working interrelationship/constructional in terrelationship. • They regularly assess their solutions using a comprehensive set of criteria, avoiding emp hasising personal preferences. • They con tinuously rcflect on thcir approach and adap t it to the situation at hand. These characteristics are in line with the ai ms and proposals for the desig n approach in this book.
2.2.3 Problem Solving as lnformation Processing When we disc ussed the basic ideas o f the systems approach (see Sectio n 1.2.3), we found that problem solving d emands a large and constan\ flow of information. Dürner (2.8] also views problem solving as information processing. The most impo rtant terms used in the theory of infor mation processing are describcd in (2.5, 2.6 ]. Information is received, processed and transmitted (scc Figu re 2.18).
52
2 1:undamentals
--
Reception of information
-
Processing of information
r--
Transmission
ofinfonnation
- r-
lmproveand refine
figure 2.18. The
Information is received fro m market analyses, trend studies, patents, technical journals, research results, licenses, inquiries from c ustomers, concrete assignments, design catalogues, a nalyses of natural a nd artificial systems, calc ulations, cxpcrimcnts, a nalogics, general and in-housc standards and rcgu.lations, stock sheets, delivery instructions, comp uter data, test repor ts, accident reports, and also by "asking questions". Data collectio n is a n essential element of problem solving [2.3 ]. Information is processed by analysis and sy nthesis, the development of so lution concep ts, calculation, experiment, the elaboration of layout d rawings a nd also the evaluation of solutions. l nformation is transmitted by means o f sketches, drawings, reports, tables, production d ocu ments, assemb ly ma nuals, user manuals, etc. T hese can be both in hard copy and electronic forms. Quite often provision must also be made for infor matio n to be stored. In [2.32] sorne criteria for characterisi ng information are given, a nd these can be used for formulating user information requ irements. They include: • Reliability: the probability of the information being available, trustworthy and correct. • Shar pness: the precisio n and clarity of the infor matio n content. • Yolume and density: an ind ication of the n umber of words and pictures needed fo r the description of a system o r process. • Value: the importance of the information to the recip ient. • Actua li ty: a n indication of the po int in ti me when the information can be used. • Form: the d istinction between graphic and alphanumeric data . • Originality: an ind ication of whether or not the orig inal character of the information must be preserved. • Complexity: the structure of, or con nectivity between, information symbols and info rmation elements, u nits or comp lexes. • Degree o f refinement: the quantity of detail in the information . Information conversio n is usually a very comp licated process. Solving problems rcquircs information of diffcrcnt typcs, contcnt and rangc. In ad dition, to raisc tbc level of info rmation and improve it, it m ay be necessary to reiterate certain steps.
2.2 f:undamentals of the Systematic Approach
53
Iteratiol! is the process by which a so lution is approached step -by-step. In this process, one o r mo re steps are repeated, each time ata higher leve! o f information based on the rcsults of the prcvious loop. Only in th is way is it possible to obtain the info rmation to refine a solutio n a nd e nsure co ntin uous improvement (see Figure 2.18). Sueh itcrations occur frequently at all stagcs of tbe problem-solving process.
2.2.4 General Working Methodology A general working method ology should be widely applicable, independent of d iscipline and sho uld not req uire specific technical knowledge from the user. lt sho uld suppo rt a structured and effective thinking pro cess. The following general id eas appcar t ime and time again in specific approachcs, cither d ircctly or sligbtly amended to adapt tbem to tbe special req uirements of developing tecbnical systems. T he purpose o f this sect io n is to provide a general introduct io n to systematic procedures. The following pro cedu res are based no t only on our own professional exper ience a nd on the find ings of cognitive psychology mentioned in Section 2.2.1, bu t also o n the wo rk of Holliger (2.20,2.21], Nad le r (2 .38, 2.39], Müller (2 .35,2.36] and Schmid t [2.51]. They are also k nown as " heuristic principies" (a heuristic is a me thod fo r generating ideas and finding solutions) or "creativity techniques". Tbe following conditions must be satisficd by anyone using a systcmatic approacb:
• Defi ne goals by formulating the overall goal, the individual subgo als a nd their importan ce. This ensures the motivation to so lve the task a nd supports insight into the problem.
• Clarify col!ditiolls by defining the initial and bounda ry constraints . • Dispel prejudice to ensure the most wide-ranging search for so lut ions possible a nd to avoid logical errors.
• Search fo r varia11ts to find a nu mber of possible so lutions or combinations of so lutions from which the best can be selected.
• Evaluate based on the goals and conditions. • Make decisiol!s. This is facilitated by objective evaluations. Withou t d ecisions and experie ncing their consequences there ca n be no progress. To make these general methods work, the following thi11king and acting operatio11s must be considered. 1. Purposefu/ Thinking
As d escribed in Section 2.2.1, intuitive and discursive thinking are p ossible. The fo rm er tends to be mo re unconscio us, the latter more co nscious. llltuitioll has lcd to a largc n umber o f good and cvcn cxccllcnt solutions. Thc prerequisite is, however, always a ver y conscious and intensive invo lvement with
54
2 1:undamentals
the given problem. Nevertheless, a p urely intuitive approach has the following disadvantages: • thc right idea rarcly comes at thc right mo mcnt, sincc it can not be clidtcd a nd elaborated at will • the result depends strongly on individual talent and experience • there is a danger that solutions will be circumscribed by precoJlceived ideas based on one's special t rain ing a nd experience. lt is therefore advisable to use more deliberate procedures that tackle problems step-by·step, ancl such procedures are clenoted discursive. Here the steps a re cho· sen intentio nally; they can be intl uenced and co mmun icated. Usua.lly ind ividual ideas or so lution attcmp ts are consciously analyscd, varicd and com bined. lt is a n importan! aspect of this procedure that a problem is ra rely tackled as a wh ole, bu t is first div ided into manageable parts a nd then a nalysed. lt m ust, however, be stressed that intuitive and discursive methods are no t opposites. Experience has shown that in tuition is stim ulated by discursive thoug ht. Thus, while complex assign ments must always be tackled one step a t a t ime, the subsid iary problems invo lved may, and often shou ld, be solvcd in in tuitive ways. In addition, it should be realisecl that creativity can be inhibited or encouraged by different in tlue nces [2.2). lt is, for example, often necessary to e ncourage in tuitive thinking by interrupting the activity lo provide sorne periods of incubation (see Sectio n 2.2.1) . On the other hand, too ma ny interrup tions can be disturbi ng and thereby inh ibit creativity. A systematic approach including d iscursive elements and adopting d ifferent viewpoints encourages creativity. Examp les in elude using d ifferent solution methods; moving between abstract and concre te ideas; collecting information using solution catalogues; and divid ingwork between team members. Furthermore, according to [2.25), realistic plan ning encourages rather than inhibits motivation a nd creativity.
2. Individual Workíng Styles Designers sh ould be given sorne freedom of action in their work to enable them to realise their own optim ised work ing style. They sho uld be free to select thei r preferred me thods, the sequence in wh ich they u ndertake individual working steps, a nd the sou rces of information they wish to co nsu lt. They sho uld therefore be allowed to ma ke their own plans for their area o f responsibility and for them to h ave control over these plans. Obviously the ind ividual working plans have to be compatible with the o'•e rall approach and make a useful contributio n. In general it is nccessary to co nsid er severa! subfunctions (subproblems) when developing new prod ucts. These functions, or combinations of them, lead to par· tia! so lutions. In such situatio ns desig ners can proceed in d ifferent ways. One possibility is to scarch for working princip ies (so lution p rincipies) for cvery subfunction (o r gro up of subfunctions), to ro ug hly check their compatib.ility, anclthen to combine them into an overall wo rking structure (solution concept). Finally the componen ts are embodied , ma king sure their overall combination lis com patible.
2.2 f:undamentals of the Systematic Approach
55
From a methodical point of view, this approach is systematic, stepwise and processorie nted; that is, the designer develops the d iffere nt functional a reas in parallel, fro m abstrae! (idea gencration) to concrete (fi nal cmbod imcnt) (sec Figure 2.19a). Ano th er possib ility is to proceed fro m idea generation to fi nal embodim ent fo r evcry problcm o r functio nal arca, onc aftcr thc o thcr, and finally co mbine and modify these to make them all fit together. From a methodical point of view, this approach is proble m-o riented; that is, the desig ner develops the d ifferent fu nctional areas in sequence (see Fig ure 2.1 9b). The investigations ofDylla [2.11,2 .12) and Fricke [2.15,2 .16) showthat novices educated in systematic desig n ten d to fo llow the process-or iented approach, whcrcas expcrienccd dcsigncrs tcnd to follow thc problcm-orientcd approach. Experienced designers apply their wealth of experience, know a wide range of possible subsolu tions, and are able to represent these so lutions quickly. Hence they arrive relatively qu ickly at a concrete result. Then, using a corrective approach, they bring this togethe r into an overall solutio n. This type of app roach is successful in those cases where the ind ividual components do no t inll uence each other strongly and their properties are apparent. If these conditions are not met, this ap proach can lead to a relatively late recognition of a possible lack of compatibility between the func tional areas. This a pproach can also result in d ifferen t subsolutions being selected for identical, or similar, subfunctions, which is often not eco nom ic. In such cases further ite ratio ns are required to fi nd other solutions. The process-oriented approach largely avoids the potential disadvantages of the problem-oriented approach. However, more time is req uired because ofthe wider, more systematic perspcctivc. Thiscarries the dangcr o f generatingan unnecessarily large solution space. The process-oriented approach therefore requires designers to achieve an appropriate balance between abstract and concrete; that is, to know whcn a sufficicntly Jarge, b ut not too largc, n umbcr of solu tion ideas has becn generated (divergence}, and tbe time has cometo combine tbese into a concrete co ncept (co nvergence). In practice, these two approaches (pro cess-oriented and problem-oriented ) are often not found in their pure fo rm. They usuall y appear in various combinations depend ing o n the proble m situation. Ilowever, ind ividual designers naturally tend to adopt onc approach in prcfercncc to thc otber. Process-oriented approachcs are recommended when subproblems are strongly interrelated and w hen breaking new g round. A problem-orientcd approach is useful wben the conncctivity bctween fu nctional areas is low and when subsolutions are known to exist in tbe area of application . Similarly in dividual differences in approach can be observed d uring tbe searcb for so lutions. If desig ners develop and investigate d ifferent solution principies or embod iment variants in parallel while searching for solutions for the individual subfunctions, and then compare tbese with one another to find the most suitable, this approach is called a generative search for solutions (see Fig ure 2.20a). 1f, o n the o ther hand, a particu lar id ea or exa mple is used as a starting poi nt and is thcn improvcd and adapted in a stcpwisc approach until a satisfactory solution emerges, this is called a corrective search for solutions (se e Figure 2.20b). Adop ting
56
2 1:undamentals
.. ........
·- ·. ·... ..
........
• MainwoRing sttps •
~Cip
«rbod1mMI .
G?nera~
soMion (()n(q)f
9 c:J y
-
.'
'' '
•
c:J
y
SearchXIr
~;nlleing
'
''
'' '
''
F111xtion8
Ftmroon (
9
.'
...
¡mnc"le
b
function A
Figurt 2.19. Oiffetent individual approaches during the development of solution:s for a tea·making machinewith several
linke
representation aher fn(ke (2.15,2.16))
this latter approach will also result in a range of solution variants, if ind ividual variants a re not rejec ted . A generative search for solutions increases the chances o f finding new and unconventional ideas and considers many differen t principies, and thus may result in a larger solution space. The challenge, however, is a tim ely and goal-o riented
2.2 f:undamentals of the Systematic Approach
57
¡~ · -== E~ (1 -· (,-:-=. ~ :.¿ -
·
a
b
Fig ure 2.20. Different individual approaches duñng the search for solutions for an elastic supporlt. a Generative, Le. generation or various solutions and goal-oriented selection. b Corrective, i.e. search for solutions by improvement and adaptationof one idea
sclcction to avoid wasting time on u nfcasiblc solutions. Th is typc of scarch is typical for novices who have been taught systematic design and for des ig ners who have adoptcd thc systcmatic approach. A corrective search for so lutions is often used by inexp erienced desig ners, in partic ular when they can think of a sim ilar k nown solution in the application area. The advantage is that it is possible to concretise the solutions relat;vely quickly, even if these in itia l solutions are not really satisfactory. When adop ting this type of search, desig ners tend to remain in their area o f expertise a nd only expan d this slowly. Possible dangcrs in elude fixating on solu tion ideas that are less suitablc in principie and failing to recognise other better solution principies . In practice, designers tend to adopta mixture of search types wi th the main aim of minimising their work effort. llowever, d esigners clearly favour one or the other search type because of thei r individual tale nts and experience, usually without being aware of the advantages or dangers of their par ticular styles. The consciously or u nconsciously app lied approaches depend on education and experience and can be in fluenced . Designers shou ld not be fo rced into adopting
58
2 1:undamentals
a particular approach. On the contrary, it is better to make them aware of the advantages and dangers of the various ap proaches a nd leave the final decision up to tbem. lt is, however, uscful thro ug h train ing and further education, along with appropriate management during the project, to identify the most suitab le overall approacb and to agree on tbis.
2.2.5 Generally Applicable Methods The following general meth ods provide further support fo r systematic wo rk, a nd are widely used (2.21). Often so -called "new" methods on ly involve repackaging one of the general methods described below.
J.Analysis Analysis is the resolution of anything comp lex into its elements a nd the study o f tbese clements and their interrclationships. It calls for id entification, definition, structuri ng and arrangement. The acquired information is transfor med into knowledge. If crrors are to be min imised, the n problems m ust be fo rm ulated clcarly and unambig uously. To tbat end, tbey bave to be analysed. Problem analysis means separati ng the essential fro m the nonessential and, in the caseof complex problems, p reparing a discursive solution by resolution into ind ividual, more transparent, subproblems. If tbe searcb for the solution proves difficult, a new fo rm ulation of the problem may provide a better star ting point. The reformulation of statements is often an effective means of finding new ideas and insights. Experience has shown that careful analysis and formulation of problems are among the most importan! steps of th e systematic approach. T hc solution of a problem can also be bro ug ht nearcr by structure analj,sis, that is, the search for h ierarchical structures or logical con nections. In general, this type of analysis can be said to aim at the demonstration of similarities or repetitive features in d ifferent systems, for exam ple by means of analogical reason ing (see Section 3.2.1). Another helpful approach is weak spot analysis. lt is based on the fact that every system has weaknesses cau sed by ignoran ce, mistakcn ideas, externa! disturbanccs, physicallimitations and production errors. Dur ing the development of a system it is therefore importan! to analysc tbe design concept or design cmbodiment for tbe express pu rpose of discovering possible weak spots and prescribing remedies. To that end, special selectio n and evaluation p rocedu res (see Sectio n 3.3) and weak spo t identification methods (see Section 10.2) have been developed. Experience has shown that this type of analysis m ay not only lead to specific improvements ofthe chosen so lution pr incip ie, b ut m ay also trigger offnew solution principies.
2. Abstraction Th rough abstraction it is possible to fi nd a higher leve! interrelatio.n ship, that is, one which is more gcneric and comprchensive. Such a proccdurc xcduces complexity a nd em phasises the essential characteristics of the problem and tbereby
2.2 f:undamentals of the Systematic Approach
59
provides an oppor tunity to search for and find o the r so lutions con taining the id entifi ed characteristics. At the sa me time new st ructures emerge in the minds of desig ncrs a nd thesc assist with thc organisation and rctrieval of the many ideas and representations. So abstrac tion suppor ts bo th c reativity a nd systematic thinki ng. It makes possiblc the d cfin ition of a problem in such a way that a coinciden tal solution path is avo ided and a more generic solu tion is fou nd (see example in Secti on 6.2). 3. Synthesis
Synthesis is the fitting togethe r of parts o r elements to prod uce new effects and to de monstrate that these effec ts create an 0\•erall orde r. lt involves search a nd d iscovcry, and also co mposition and combination. An csscnt ial fcaturc of all desig n work is the combination of ind ividual findings or subso lutions in to an overall wo rking system- in other words, the association of com ponen ts to forma whole. Du ring the process of synthesis the info rmation d iscovered byanalyses is processed as weU. ln general, it is advisable to base synthesis o n a holistic or systems a pproach; in other words, to bear in mind the general task o r course of events wh ile wo rking on subtasks or ind ividual stcps. Unless this is done, there is thc grave risk that, despite the optimisation o f ind ividual assemblies or steps, no suitable overall solution will be reached. Appreciation o fthis fact is the basis ofthe interdisciplinary method known as Value Analysis, which pro ceeds from the analysis of the problem and structu re to a holis tic systems a pproach involving the early collaboration of all departments co ncerned with prod uct development. Such a n approach is also needed in large-scale projects, especially when preparing schedules by such tech niques as critica! path analysis (see Section 4.2.2). The entire systems approach and its metho ds are strongly based on holistic thin king, which is par ticula rly importan! in the selection of eva luation criteria, beca use the value o f a partic ular solu tion ca n only be gauged after overall assessment o f all of the expectatio ns, rcquircmcnts and constraints (see Section 3.3.2). 4. Method of Persistent Questions
When using systematic proced ures it is o ften a good idea to keep asking questions of both oneself a nd o f others as a stimulus to fresh thought and intuition . A stand ard list of questions also fosters the discursive method . In short, asking questions is one of the most impor ta n! method ological too ls. T his explains why many authors havc drawn u p spccial chccklists for various working steps to support this mcthod. S. Method of Negation
The method of d eliberate negat ion starts from a known solut ion, splits it into ind ividual parts or d escribes it by ind ividual statements, and negates these statements one-by-one or in groups. This delibera te inversion often creates new solution possibilities. Thus, when consider ing a "ro tating" m achine eleme nt, one might also examine thc "static" case. Morcovcr, the mere omission of an elemcn t can be tantamount toa negation. This method is also k nown as "systematic d oubting" [2.21 [.
lndex viscoelaslic behavio ur
323
viscometer design 89 volume-surface area relationship
winding machine, mínimum risk design workflow, planning 128, 129, 13 1
3 19
W~ch tl er, R li warning, safety 256 wcak spot analysis .58. Weber- fec:.hner law 470 wclding cost estima tes 554 wélding, dcsign fo r 368, 372
working loc.ation
J2
working motions 32 working principies :1!!. 181,184, 208 combinin¡¡ U\4-186 workingstructurc-s ~ practical use 186, 187, 189, 190
selection
186
wo rking s urfaces/spaces
.12
wheel propulsion, simultaneo\IS 276 white goods, rec)'cUng
399-400
yield point, effect of exceeding
272
Zwicky morphological matrix
104, 184
windlng machine differential construction
357
617 408
