1、raising standards worldwideNO COPYING WITHOUT BSI PERMISSION EXCEPT AS PERMITTED BY COPYRIGHT LAWBSI Standards PublicationBS ISO 3534-3:2013Statistics V o c a b u l a r y a n d s y m b o l sPart 3: Design of experimentsBS ISO 3534-3:2013 BRITISH STANDARDNational forewordThis British Standard is the
2、UK implementation of ISO 3534-3:2013. It supersedes BS ISO 3534-3:1999, which is withdrawn.The UK participation in its preparation was entrusted to T e c h n i c a l C o m m i t t e e S S / 1 , S t a t i s t i c a l t e r m i n o l o g y a n d s y m b o l s .A list of organizations represented on th
3、is committee can be obtained on request to its secretary.This publication does not purport to include all the necessary provisions of a contract. Users are responsible for its correct application. The British Standards Institution 2013. Published by BSI Standards Limited 2013.ISBN 978 0 580 66856 2
4、ICS 01.040.03; 03.120.30 Compliance with a British Standard cannot confer immunityfrom legal obligations.This British Standard was published under the authority of the Standards Policy and Strategy Committee on 30 April 2013.Amendments issued since publicationDate T e x t a f f e c t e dBS ISO 3534-
5、3:2013Reference numberNumro de rfrenceISO 3534-3:2013(E/F)ISO 2013INTERNATIONAL STANDARD NORME INTERNATIONALE ISO3534-3Third editionTroisime dition2013-04-15Statistics Vocabulary and symbols Part 3: Design of experiments Statistique Vocabulaire et symboles Partie 3: Plans dexpriences BS ISO 3534-3:2
6、013ISO 3534-3:2013(E/F) COPYRIGHT PROTECTED DOCUMENT DOCUMENT PROTG PAR COPYRIGHT ISO 2013 The reproduction of the terms and definitions contained in this International Standard is permitted in teaching manuals, instruction booklets, technical publications and journals for strictly educational or im
7、plementation purposes. The conditions for such reproduction are: that no modifications are made to the terms and definitions; that such reproduction is not permitted for dictionaries or similar publications offered for sale; and that this International Standard is referenced as the source document.
8、With the sole exceptions noted above, no part of this publication may be reproduced or utilized otherwise in any form or by any means, electronic or mechanical, including photocopying, or posting on the internet or an intranet, without prior written permission. Permission can be requested from eithe
9、r ISO at the address below or ISOs member body in the country of the requester. La reproduction des termes et des dfinitions contenus dans la prsente Norme internationale est autorise dans les manuels denseignement, les modes demploi, les publications et revues techniques destins exclusivement lense
10、ignement ou la mise en application. Les conditions dune telle reproduction sont les suivantes: aucune modification nest apporte aux termes et dfinitions; la reproduction nest pas autorise dans des dictionnaires ou publications similaires destins la vente; la prsente Norme internationale est cite com
11、me document source. la seule exception mentionne ci-dessus, Sauf indication contraire, aucune partie de cette publication ne peut tre reproduite ni utilise sous quelque forme que ce soit et par aucun procd, lectronique ou mcanique, y compris la photocopie, laffichage sur linternet ou sur un Intranet
12、, sans autorisation crite pralable. Les demandes dautorisation peuvent tre adresses lISO ladresse ci-aprs ou au comit membre de lISO dans le pays du demandeur. ISO copyright office Case postale 56 CH-1211 Geneva 20 Tel. + 41 22 749 01 11 Fax + 41 22 749 09 47 E-mail copyrightiso.org Web www.iso.org
13、Published in Switzerland/Publi en Suisse ii ISO 2013 All rights reserved/Tous droits rservsBS ISO 3534-3:2013ISO 3534-3:2013(E/F) ISO 2013 All rights reserved/Tous droits rservs iiiContents Page Foreword .v Introduction.vii 1 Scope1 2 Normative references1 3 Terms and definitions.2 3.1 General terms
14、.2 3.2 Arrangements of experiments25 3.3 Methods of analysis 54 Annex A (informative) Concept diagrams 66 Annex B (informative) Methodology used to develop the vocabulary 82 Annex C (informative) Experimental design checklists 85 Annex D (informative) Experimental design from the system model perspe
15、ctive.88 Bibliography93 Alphabetical index94 BS ISO 3534-3:2013ISO 3534-3:2013(E/F) iv ISO 2013 All rights reserved/Tous droits rservsSommaire Page Avant-propos. vi Introduction . viii 1 Domaine dapplication . 1 2 Rfrences normatives 1 3 Termes et dfinitions . 2 3.1 Termes gnraux 2 3.2 Dispositifs e
16、xprimentaux . 25 3.3 Mthodes danalyse 54 Annexe A (informative) Schmas conceptuels . 66 Annexe B (informative) Mthodologie utilise pour laborer le vocabulaire. 82 Annexe C (informative) Listes de contrle dun plan exprimental 85 Annexe D (informative) Plan dexpriences du point de vue du modle de syst
17、me. 88 Bibliographie 93 Index alphabtique 96 BS ISO 3534-3:2013ISO 3534-3:2013(E/F) ISO 2013 All rights reserved/Tous droits rservs vForeword ISO (the International Organization for Standardization) is a worldwide federation of national standards bodies (ISO member bodies). The work of preparing Int
18、ernational Standards is normally carried out through ISO technical committees. Each member body interested in a subject for which a technical committee has been established has the right to be represented on that committee. International organizations, governmental and non-governmental, in liaison w
19、ith ISO, also take part in the work. ISO collaborates closely with the International Electrotechnical Commission (IEC) on all matters of electrotechnical standardization. International Standards are drafted in accordance with the rules given in the ISO/IEC Directives, Part 2. The main task of techni
20、cal committees is to prepare International Standards. Draft International Standards adopted by the technical committees are circulated to the member bodies for voting. Publication as an International Standard requires approval by at least 75 % of the member bodies casting a vote. Attention is drawn
21、to the possibility that some of the elements of this document may be the subject of patent rights. ISO shall not be held responsible for identifying any or all such patent rights. ISO 3534-3 was prepared by Technical Committee ISO/TC 69, Applications of statistical methods, Subcommittee SC 1, Termin
22、ology and symbols. This third edition cancels and replaces the second edition (ISO 3534-3:1999), which has been technically revised. ISO 3534 consists of the following parts, under the general title Statistics Vocabulary and symbols: Part 1: General statistical terms and terms used in probability Pa
23、rt 2: Applied statistics Part 3: Design of experiments Part 4: Survey sampling BS ISO 3534-3:2013ISO 3534-3:2013(E/F) vi ISO 2013 All rights reserved/Tous droits rservsAvant-propos LISO (Organisation internationale de normalisation) est une fdration mondiale dorganismes nationaux de normalisation (c
24、omits membres de lISO). Llaboration des Normes internationales est en gnral confie aux comits techniques de lISO. Chaque comit membre intress par une tude a le droit de faire partie du comit technique cr cet effet. Les organisations internationales, gouvernementales et non gouvernementales, en liais
25、on avec lISO participent galement aux travaux. LISO collabore troitement avec la Commission lectrotechnique internationale (CEI) en ce qui concerne la normalisation lectrotechnique. Les Normes internationales sont rdiges conformment aux rgles donnes dans les Directives ISO/CEI, Partie 2. La tche pri
26、ncipale des comits techniques est dlaborer les Normes internationales. Les projets de Normes internationales adopts par les comits techniques sont soumis aux comits membres pour vote. Leur publication comme Normes internationales requiert lapprobation de 75 % au moins des comits membres votants. Lat
27、tention est appele sur le fait que certains des lments du prsent document peuvent faire lobjet de droits de proprit intellectuelle ou de droits analogues. LISO ne saurait tre tenue pour responsable de ne pas avoir identifi de tels droits de proprit et averti de leur existence. LISO 3534-3 a t labore
28、 par le comit technique ISO/TC 69, Application des mthodes statistiques, sous-comit SC 1, Terminologie et symboles. Cette troisime dition annule et remplace la deuxime dition (ISO 3534-3:1999), qui a fait lobjet dune rvision technique. LISO 3534 comprend les parties suivantes, prsentes sous le titre
29、 gnral Statistique Vocabulaire et symboles: Partie 1: Termes statistiques gnraux et termes utiliss en calcul des probabilits Partie 2: Statistique applique Partie 3: Plans dexpriences Partie 4: chantillonnage pour sondages BS ISO 3534-3:2013ISO 3534-3:2013(E/F) ISO 2013 All rights reserved/Tous droi
30、ts rservs viiIntroduction Design of experiments (DOE) catalyses innovation, problem solving and discovery. DOE comprises a strategy and a body of methods that are instrumental in achieving quality improvement in products, services and processes. Although statistical quality control, management resol
31、ve, inspection and other quality tools also serve this goal, experimental design represents the methodology of choice in complex, variable and interactive settings. Historically, design of experiments has evolved and thrived in the agricultural area. Medicine has also enjoyed a long history of caref
32、ul experimental design. Industrial settings particularly benefit from the methodology due to the ease of initiating efforts (user-friendly software packages), improved training, influential advocates, and accumulating successes with experimental design. Design of experiments is fundamental to contin
33、uous improvement and product development. Experimentation often evolves sequentially with improvements taking place following each stage of the learning process. If the objective is to optimize a response, then response surface designs (3.2.19) play a critical role. Multiple levels of factors recogn
34、ized to be important are considered to accommodate neatly curvilinear effects, for example in the vicinity of the optimum settings. Factorial experiments (3.2.1) and fractional factorial experiments (3.2.3) provide a methodology for studying the interrelationships among multiple factors of interest
35、to the experimenter. These types of experiments can be far more resource efficient and effective than intuitive one-factor-at-a-time experiments. Factorial experiments are particularly well-suited for determining that a factor behaves differently (as reflected in the experimental response) at differ
36、ent levels of other factors. Frequently, the “breakthrough” in quality comes from the synergism revealed in a study of “interactions” (3.1.17). If the number of factors under consideration is large, then factorial experiments could exceed resources. However, fractional factorial experiments offer a
37、possible compromise. Actually, if the initial goal is to identify factors warranting further investigation, then screening designs (3.2.8) can be useful. In planning an experiment, it is necessary to limit biases introduced by the experimental conditions or in the assignment of treatments to experim
38、ental units. Topics such as “randomization” (3.1.30) and “blocking” (3.1.26) deal with minimizing the effects of nuisance or extraneous elements. Specific blocking strategies include randomized block designs (3.2.10), Latin square designs (3.2.11) and variants, and balanced incomplete block designs
39、(3.2.14). Designs for experiments with mixtures mixture designs (3.2.20) apply in situations where factors constitute proportions of a total, such as ingredients in an alloy. Nested designs (3.2.21) are particularly useful in inter-laboratory testing and in measurement system analyses. Methods of an
40、alysis of the collected data are straightforward, if the experiment has been carried out according to the plan. Graphical methods (3.3.1) can be particularly effective in revealing overall conclusions. Estimation of parameters from a model is commonly handled using regression analysis (3.3.7). Regre
41、ssion analysis methods can also handle difficulties with missing data, identification of outliers, and other problems. Annex A provides associated Concept Diagrams that relate the various terms. To assist users of this part of ISO 3534, an explanation of Concept Diagrams is provided in Annex B. Desi
42、gn of experiments consists of a complex process to implement experimental plans (3.1.29). Annex C provides checklists that are intended to identify key items to be considered in designing and implementing a designed experiment (3.1.27). Annex D describes experimental design from the systems model pe
43、rspective. BS ISO 3534-3:2013ISO 3534-3:2013(E/F) viii ISO 2013 All rights reserved/Tous droits rservsIntroduction Les plans dexpriences (DOE, design of experiments) catalysent linnovation, la rsolution de problmes et la dcouverte. Les DOE comprennent une stratgie et un corps de mthodes qui sont les
44、 instruments permettant damliorer la qualit des produits, des services et des processus. Bien que la matrise statistique de la qualit, les solutions managriales, les inspections et autres outils de qualit remplissent galement cet objectif, les plans dexpriences reprsentent la mthodologie par excelle
45、nce dans le cas dun environnement de paramtres complexes, variables et interactifs. Dun point de vue historique, les plans dexpriences ont volu et se sont dvelopps dans le secteur de lagriculture. La mdecine a galement bnfici dune longue histoire de plans dexpriences labors avec soin. Les environnem
46、ents industriels tirent particulirement profit de la mthodologie, en raison de la facilit dinitiation des efforts (logiciels dapplication conviviaux), dune meilleure formation, de dfenseurs influents et des nombreux succs obtenus grce aux plans dexpriences. Les plans dexpriences sont indispensables
47、lamlioration continue et au dveloppement de produit. Lexprimentation volue souvent de manire squentielle, les amliorations intervenant aprs chaque tape du processus dapprentissage. Si lobjectif est doptimiser une rponse, alors les plans surface de rponse (3.2.19) jouent un rle critique. De multiples
48、 niveaux de facteurs jugs importants sont pris en compte pour sadapter parfaitement aux effets curvilignes, par exemple proximit des valeurs optimales. Les plans factoriels (voir 3.2.1) et les plans factoriels fractionnaires (3.2.3) fournissent une mthodologie dtude des interrelations entre les multiples facteurs dintrt pour la personne qui ralise lexprience. Ces types de plans dexpriences peuvent tre bien plus efficaces et conomes en ressources que les plans dexpriences intuitifs du type un facteur la fois. Les plans dexpriences factoriels convien
