SPC and CONTINUOUS IMPROVEMENT

Transcription

1 SPC and CONTINUOUS IMPROVEMENT

2 SPC and CONTINUOUS IMPROVEMENT Mal Owen Springer-Verlag Berlin Heidelberg GmbH 1989

3 British Library Cataloguing in Publication Data Owen, M. (Maldwyn), Statistical process control. 1. Industries. Quality control. Statistical methods I. Title 658.5'62 ISBN ISBN (ebook) DOI / Springer-Verlag Berlin Heidelberg Originally published by Springer-Verlag Berlin Heidelberg NewYorkTokyo 1989 Softcover reprint of the hardcover 1st edition 1989 This work is protected by copyright. The rights covered by this'are reserved, in particular those of translating, reprinting, radio broadcasting, reproduction by photo-mechanical or similar means as well as the storage and evaluation in data processing installations even if only extracts are used. Should individual copies for commercial purposes be made with written consent of the publisher then a remittance shall be given to the publisher in accordance with 54, Para 2, of the copyright law. The publisher will provide information on the amount of this remittance. Phototypeset by MFK Typesetting Ltd, Herts.

4 Contents 1 Introduction Introduction What is SPC? SOC: from Shewhart to McArthur The Japanese phenomenon Why go in for SPC? Prevention versus detection Quality as customer satisfaction From SOC to SPC TOC and SPC Quality, productivity and people 15 2 Data collection and representation Introduction Data type Data collection Designing a check sheet Tally charts The histogram The histogram and capability Generating a histogram in practice Simplifying the data Limitations of tolerance-based systems Non-symmetrical patterns Frequency and area 31 3 Problem-solving techniques Introduction The Pareto principle Brainstorming Cause and effect analysis Other techniques 55 4 Measuring location and variability Introduction Accuracy and precision 60

5 4.3 Measures of location Measures of variability Sample and population Measuring improvement Continuous improvement 76 5 The normal distribution and sampling Introduction Measuring risk From theory to practice Histogram and probability distribution The normal distribution Continuous improvement Sample and process Comparing sample and population Sampling from a non-normal distribution 96 6 Control charts for variables Introduction Charts for variables Common and special causes Measuring the process The control chart Calculating and plotting The basis of control limits Setting up the charts Rules for special causes On-going control Use of rules for on-going control Alternative control rules Sample size Capability and control Introduction The histogram The R chart and capability Measuring improvement Continuous improvement Limitations of the histogram Interpreting patterns and setting up charts Introduction Recognising patterns From chart to process The control chart The role of the Quality Control Department Alternative techniques for charting variables Introduction Control chart for medians Moving mean/moving range chart Control chart for X and s Dealing with drifting processes Dealing with small batches Multi-vari analysis 199

6 10 Control charts for attributes Introduction What are attributes? Definition of a fault Size of sample Types of attribute charts The attribute chart Interpreting attribute charts Which chart to use? Capability Continuous improvement Attribute charts for defective units Introduction The p chart The np chart Attribute charts for defects Introduction The u chart The c chart Multiple characteristics chart The Cusum technique Introduction What is the Cusum technique? From Shewhart to Cusum Interpreting the Cusum chart The Cusum technique as a retrospective tool Detecting process changes Machine capability analysis Introduction Process or machine? The machine capability chart Carrying out a capability study Calculating results Other patterns Measuring machine capability Implementing the SPC programme Introduction The implementation programme Obtaining commitment Formulating a policy Appointing a facilitator Developing a training strategy Training managers and supervisors Informing the unions Obtaining union commitment Informing the operat6rs Involving suppliers Collecting data Planning for reaction to out-of-control signals Reviewing processes 326

7 Arranging administration Training the operators Introducing the charts Improving the process Avoiding the pitfalls Introduction Lack of understanding and commitment among top management Lack of a plan SPC is not company-wide Lack of long-term commitment Inadequate training Failure to involve suppliers Emphasis on short-term profits Commitment in only one department Lack of funds Failure to consult the workforce Underestimating the workforce Failure to acquire adequate statistical support Lack of market research Management by fear Lack of middle management support. Lack of quality materials Over-emphasis on computers Moving too quickly Lack of projects Pilot areas not chosen carefully Monitoring products instead of processes Over-emphasis on one technique Failure to respond to chart signals Failure to understand SPC Reluctance to change General lack of knowledge and expertise in SPC Lack of concern for detail APPENDICES A. Deming's 14 points for management B. Table of Pz values C. Control chart constants and formulae D. Control chart constants (British Standards) E. Derivation of US and British control chart constants F. Derivation of constants for individual /moving R chart G. Symbols and definitions H. Various charts I. Further information

8 PREFACE There is no doubt that quality has become a major feature in the survival plan of organisations. With diminishing markets resulting from improved competitive performance and the associated factor of single-sourcing arrangements by the major organisations, it is clear that unless there is a commitment to change, organisations will lose their competitive edge. This will unfortunately mean elimination and the resultant harsh realities that come with it for the employees. It has been said on many platforms that unemployment is not inevitable. Those organisations which recognise the requirements for survival know that quality, and its association with customer satisfaction, is now a key issue. Survival programmes based on quality improvement require an unrelenting commitment to include everyone, from the Managing Director down, in an ongoing, never-ending involvement based on monitoring, and improving, all our activities. These Total Quality Management (TOM) programmes, whatever their specific nature, have a common theme of measuring and then improving. This text describes the philosophy and techniques of one type of involvement programme-statistical Process Control (SPC). The material to follow suggests that SPC is a major element of any programme and, if properly applied, could be a complete programme in itself. Measuring and improving means that data must be collected, used, understood, interpreted and analysed, and thereby lies the difficulty. The West is only too aware of the reputation of the Japanese in terms of quality performance. It may not be aware of the vast training programmes in simple statistical methods which have been ongoing since the early 1950's, and which have proved a key feature in the progress of Japan to world leadership in many areas of activity. This book covers the essential elements of these techniques. There is a constant theme of relating the techniques to actual practical applications, making use, as far as possible, of company experiences in applying the particular methods. Genuine company experiences havethereforebeenheavily leant on in developing the text. The intention has been to provide a book which could be useful to as wide a readership as possible. Senior executives should find the information of value in setting up Quality Improvement programmes and, at the same time, recognise the role they must play in the SPC implementation. Equally, one would like to think that operators would find the material sufficiently well presented, in a basic practical fashion, so that they would be able to glean information and ideas which would assist them in their important task of monitoring, and improving, the processes with which they are involved on a daily basis. At the same time there should be enough material to interest the critical group of middle management/supervision who are caught in the treadmill of production targets, schedules and deliveries.

9 The book has intentionally not been written for the technical specialists- Quality Managers, mathematicians, statisticians and the like. There are sufficient texts available already to satisfy their requirement. The contents should appeal to those undergoing formal courses in Universities, Polytechnics and Colleges, but this remains to be seen. Educational qualifications are traditionally based on syllabuses measured more in terms of theoretical and academic standards rather than practical relevance to industry and commerce. It will take some time for this to change. Much wider issues are involved here, appertaining to the role of educational institutions to the outside world. What can be said is that in the area of statistical education/training the West is some thirty years behind Japan and it will be left to the reader to judge why this is so. An attempt has been made in this book to eliminate, as far as possible, any unnecessary mathematics or statistical theory which interferes with understanding of how the techniques operate in practice. Again, it will be left to the judgement of the reader as to how far the book has succeeded in this respect. Symbols and conventions constantly interfere with the process of technical understanding. No apology is made for the fact that whilst the author is British, and the case studies relate to U.K. companies, there is a heavy emphasis on the use of standard approaches which are N. American in origin. Appropriate references are provided to British Standards, but in view of the current widespread usage, on both sides of the Atlantic, of control chart techniques which stem from major U.S. organisations, it makes sense not to complicate the issue by trying to carry two different approaches. A conscious effort has been made to widen the rather narrow traditional view of SPC as an activity restricted to machine shop operations. Administrative applications have been introduced, directly or indirectly, so that the message of a company-wide activity can be understood as relating to all areas of the organisation. It is a belief that the book has been considerably enriched by the inclusion of real case study material which has been provided by leading U.K. based organisations. The author owes a great deal to the contacts made over the years with colleagues in a variety of positions in many organisations across the U.K. and elsewhere. They have provided willing assistance and close collaboration and have been a source of encouragement and support. In formally acknowledging the organisations involved in the preparation of the book, the support of many colleagues, too numerous to mention in person, is also acknowledged with thanks. I am particularly grateful to John Parsloe, Sauer Sundstrand, who has provided invaluable support over many months in reading and commenting on the scripts as they appeared. SPC programmes seem to bring out missionaries for the cause and John is a leading SPC disciple. His co-operation is much appreciated. I should also like to thank colleagues associated with IFS for their considerable help and encouragement in the writing of this book. Finally my thanks to my family. My wife has provided invaluable support in converting an untidy script to a neat word-processor output. My family have also allowed me precious time to write a book which one hopes will prove useful to the increasing numbers who are looking for material which may help them on the road to never-ending improvement. February 1989 MO

10 ACKNOWLEDGEMENTS Various diagrams appearing in the book are based on similar ones in an SPCoperator manual. The manual forms part of a training package on SPC which was developed with Further Education Unit/(DES (Pickup) funding in the course of FEU project RP320, "The development ofteaching/learning materials in Statistical Process Control:' The co-operation of FEU in allowing use of this material is gratefully acknowledged. (The package is available from the Training for Quality Unit at Bristol Polytechnic) Thanks are due to the following organisations who have provided case-study and other material to support the various SPC techniques. Their co-operation and support is gratefully acknowledged. British Alcan (Rolled Products), Rogerstone, Newport, Gwent British Steel (Tinplate Group), Swansea Century Oils, Stoke-on-Trent Ford of Europe, Statistical Methods Office, Basildon, Essex Ford New Holland, Basildon, Essex Goodyear Tyre & Rubber Company, Wolverhampton Hewlett Packard (Computer Peripherals), Bristol ITT Cannon, Basingstoke Jaguar Cars, Radford, Coventry Lucas Electrical (Starter & Alternator Division), Birmingham PPG Industries, (U.K.) Ltd, Birmingham Sauer Sundstrand, Swindon.