ISO-7966-1993.pdf

INTERNATIONAL STANDARD IS0 7966 First edition 1993-l 2-l 5 Acceptance control charts Cartes de contrUe pour acceptation Reference number IS0 7966.1993(E) IS0 7966:1993(E) Contents Page Scope 1 Normative references . 1 Definitions . 1 Symbols and abbreviations . 1 Description of acceptance control chart practice . 2 Acceptance control of a process 3 Specifications 3 Calculation procedures 4 Examples . 8 10 Factors for acceptance control limits 11 Annexes A Nomographs for acceptance control chart design 12 B Bibliography 21 0 IS0 1993 All rights reserved. No part of this publication may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying and microfilm, without per- mission in writing from the publisher. International Organization for Standardization Case Postale 56 l CH-1211 Geneve 20 l Switzerland Printed in Switzerland ii IS0 7966:1993(E) Foreword IS0 (the International Organization for Standardization) is a worldwide federation of national standards bodies (IS0 member bodies). The work of preparing International Standards is normally carried out through IS0 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 with ISO, also take part in the work. IS0 collaborates closely with the International Electrotechnical Commission (IEC) on all matters of electrotechnical standardization. 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. International Standard IS0 7966 was prepared by Technical Committee lSO/TC 69, Applications of statistical methods, Sub-Committee SC 4, Statistical process control. Annex A forms an integral part of this International Standard. Annex B is for information only. . . . III -,-,- IS0 7966:1993(E) Introduction An acceptance control chart combines consideration of control impli- cations with elements of acceptance sampling. It is an appropriate tool for helping to make decisions with respect to process acceptance. The bases for the decisions may be defined in terms of a) whether or not a designated percentage of units of a product or service derived from that process will satisfy specification requirements; b) whether or not the process has shifted beyond some allowable zone of process level locations. A difference from most acceptance sampling approaches is the emphasis on process acceptability rather than on product disposition decisions. A difference from usual control chart approaches is that the process usu- ally does not need to be in control about some single standard process level, but that as long as the within-subgroup variability remains in control, it can (for the purpose of acceptance) run at any level or levels within some zone of process levels which would be acceptable in terms of tol- erance requirements. Thus, it is assumed that some assignable causes will create shifts in the process levels which are small enough in relation to requirements that it would be uneconomical to attempt to control them too tightly for the purpose of mere acceptance. The use of an acceptance control chart does not, however, rule out the possibility of identifying and removing assignable causes for the purpose of continuing process improvement. A check on the inherent stability of the process is required. Therefore, variables are monitored using Shewhart-type range or sample standard deviation control charts to confirm that the variability inherent within ra- tional subroups remains in a steady state. Supplementary examinations of the distribution of the encountered process levels form an additional source of control information. A preliminary Shewhart control chart study should be conducted to verify the validity of using an acceptance control chart. iv INTERNATIONAL STANDARD IS0 7966:1993(E) Acceptance control charts 1 Scope This International Standard gives guidance on the uses of acceptance control charts and establishes general procedures for determining sample sizes, ac- tion limits and decision criteria. Examples are included to illustrate a variety of circumstances in which this technique has advantages and to provide details of the determination of the sample size, the action limits and the decision criteria. 2 Normative references Indifference zone The following standards contain provisions which, through reference in this text, constitute provisions of this International Standard. At the time of publi- cation, the editions indicated were valid. All standards are subject to revision, and parties to agreements based on this International Standard are encouraged to investigate the possibility of applying the most re- cent editions of the standards indicated below. Members of IEC and IS0 maintain registers of cur- rently valid International Standards. IS0 3534-l :1993, Statistics - Vocabulary and sym- bols - Part I: Probability and general statistical terms. IS0 3534-21993, Statistics - Vocabulary and sym- bols - Part 2: Statistical quality control. IS0 8258:1991, Shewhart control charts. 3 Definitions For the purposes of this International Standard, the definitions given in IS0 3534-l and IS0 3534-2 apply. An acceptable process would be a process which is represented by a Shewhart control chart (see IS0 8258) with a central line within the acceptable Process zone (see figure 1). Ideally the average value X of such a control chart would be at the target value. Rejectable processes A RPL, Indifference zone APL, + Target level Rejectable processes RPLL Figure 1 - Two-sided specification limits: Upper and lower APL and RPL lines in relation to processes of acceptable, rejectable, and indifference (borderline) quality 4 Symbols and abbreviations USL upper specification limit LSL lower specification limit ACL acceptance control limits APL acceptable process level RPL rejectable process level or non-acceptable process zone n subgroup sample size PO acceptable proportion nonconforming items Pl pa T rejectable proportion nonconforming items probability of acceptance target value, i.e. optimum value of the characteristic 1 IS0 7966:1993(E) x average value of the variable X plotted on a control chart Z variable that has a normal distribution with zero mean and unit standard deviation Z! P normal deviate that is exceeded by 100 % of the deviate in a specified di- rection (similarly for z, zs, etc.) o! risk of not accepting a process centred at the APL CT risk of not rejecting a process centred at the RPL process mean standard deviation corresponding to the inherent process variability The acceptance control chart is a useful tool for covering this wide range of approaches in a logical and simple manner. It distinguishes between the in- herent variability components randomly occurring throughout the process and the additional location factors which contribute at less frequent intervals. When shifts appear, the process may then stabilize at a new level until the next such event occurs. Be- tween such disturbances, the process runs in control with respect to inherent variability. An illustration of this situation is a process using large uniform batches of raw material. The within-batch variability could be considered to be the inherent variability. When a new batch of material is intro- duced, its deviation from the target may differ from that of the previous batch. The between-batch vari- ation component enters the system at discrete inter- within-subgroup standard deviation standard deviation of the subgroup aver- age corresponding to the inherent process variability: OF = a/ Jn. 5 Description of acceptance control chart practice In the pursuit of an acceptable product or service, there often is room for some latitude in the ability to centre a process around its target level. The contri- bution to overall variation of such location factors is additional to the inherent random variability of individ- ual elements around a given process level. In most cases, some shifts in process level must be expected and can be tolerated. These shifts usually result from an assignable cause that cannot be eliminated be- cause of engineering or economic considerations. They often enter the system at infrequent or irregular intervals, but can rarely be treated as random com- ponents of variance. There are several seemingly different approaches to treating these location factors contributing variation beyond that of inherent variability. At one extreme is the approach in which all variability that results in de- viations from the target value must be minimized. Supporters of such an approach seek to improve the capability to maintain a process within tighter toler- ance limits so that there is greater potential for pro- cess or product quality improvement. At the other extreme is the approach that if tolerance limits are satisfied, it not only may be uneconomic and wasteful of resources to tightly control the process, but it is very likely to be counterproductive to im- proving the capability of reducing variability. This often is the result of the introduction of pressures which encourage “tampering” with the process (over- control) by people qualified to work on control aspects but not product or process quality improvement pro- grammes. An example of this within- and between-batch vari- ation might very well occur in a situation where a blanking die is blanking a machine part. The purpose of the chart is to determine when the die has worn to a point where it must be repaired or reworked. The rate of wear is dependent upon the hardness of the successive batches of material and is therefore not readily predictable. It will be seen that the use of an acceptance control chart makes it possible to judge the appropriate time to service the blanking die. The acceptance control chart is based on the Shewhart control chart but is set up so that the pro- cess can shift outside of control limits if the specifi- cations are sufficiently wide, or be confined to narrower limits if the inherent variability of the pro- cess is comparatively large or a large fraction of the total tolerance spread. What is required is protection against a process that has shifted so far from the target value that it will yield some predetermined undesirable percentage of items falling outside the specification limits, or exhib- its an excessive degree of process level shift. When a chart of the average value of data sets from a process is plotted, in sequence of the production, one notices a continual variation in average values. In a central zone (acceptable process, figure I), there is product that is indisputably acceptable. Data in the outer zones (figure 1) represent a process that is producing product that is indisputably not acceptable. Between the inner and the outer zones are zones where the product being produced is acceptable but there is an indication that the process should be watched and as the outer zone is approached correc- tive action may be taken. These criteria are the basic concepts for the acceptance control chart, The de- scription in this International Standard is designed to provide practices for the establishment of appropriate 2 -,-,- IS0 7966:1993(E) action lines for one- and two-sided specification situ- ations. Since it is impossible to have a single dividing line that can sharply distinguish a good from an unsatisfactory quality level, one must define a process level that represents a process that should be accepted almost always (1 -a). This is called the acceptable process level (APL), and it marks the outer boundary of the acceptable process zone located about the target value (see figure 1). Simplicity of operation is of critical importance to the use of a procedure such as an acceptance control chart. Only the acceptance control limits and the sampling instructions such as sample size, frequency, or method of selection need be known to the operator who uses the chart, although training him to under- stand the derivation is not difficult and can be helpful. It is thus no more complicated to use than the Shewhart chart, The supervisor, quality expert, or trained operator will derive these limits without much effort from the above considerations and will obtain a more meaningful insight into the process accept- ante procedure, and a better understanding of the control implications. 6 Acceptance control of a process Any process centred closer to the target value than the APL will have a risk smaller than rx of not being accepted. So the closer the process is to target, the smaller the likelihood that a satisfactory process will not be accepted. It is also necessary to define the process level that represents processes that should almost never be accepted (1 -fi). This undesirable process level is labelled the rejectable process level (RPL). Any pro- cess located further away from the target value than the RPL will have a risk of acceptance smaller than 8. The process levels lying between the APL and RPL would yield a product of borderline quality. That is, process levels falling between the APL and RPL would represent quality which is not so good that it would be a waste of time, or represent over-control, if the process were adjusted, and not so bad that the product could not be used if no shift in level were made. This region is often called the “indifference zone”. The width of this zone is a function of the re- quirements for a particular process and the risks one is willing to take in connection with it. The narrower the zone, i.e. the closer the APL and RPL are to each other, the larger will the sample size have to be. This approach will permit a realistic appraisal of the effec- tiveness of any acceptance control system, and will provide a descriptive method for showing just what any given control system is intended to do. As with any acceptance sampling system, four el- ements are required for the definition of an accept- ance control chart. They are the following: a) an acceptable process level (APL) associated with a one-sided a-risk; b) a rejectable process level (RPL) associated with a one-sided b-risk; c) an action criterion or acceptance control limit (ACL); d) the sample size (n). NOTE 1 Generally, the defined risks are one-sided in this International Standard. In the case of two-sided specifi- cations, the risks are a 5 % risk to go above an upper limit or a 5 % risk to go below a lower limit. This results in a 5 % (not 10 %) total risk. 6.1 Plotting the chart The sample average value of the quality characteristic is plotted on acceptance control charts in the