BS EN 9138:2019 BSI Standards Publication Aerospace Series — Quality Management Systems — Statistical Product — Acceptance Requirements bsi.BS EN 9138:2019 EUROPEAN STANDARD EN 9138 NORME EUROPEENNE EUROPAISCHE NORM November 2019 ICS 03.120.10; 49.020 English Version Aerospace Series - Quality Management Systems - Statistical Product - Acceptance Requirements Série aérospatiale - Systemes de management de la Luft und Raumfahrt- Qualititsmanagementsysteme ‘qualité - Exigences d'acceptation statistique des Statistische Produktannahmeanforderungen produits ‘This European Standard was approved by CEN on 10 December 2018. ‘CEN members are bound to comply with the CEN/CENELEC Internal Regulations which stipulate the conditions for giving this European Standard the status ofa national standard without any alteration. Up-to-date lists and bibliographical references ‘concerning such national standards may be obtained on application to the CEN-CENELEC Management Centre or to any CEN member. ‘This European Standard exists in three official versions (English, French, German). A version in any other language made by translation under the responsibility of a CEN member into its own language and notified to the CEN-CENELEC Management Centre has the same status as the official versions, CEN members are the national standards bodies of Austria, Belgium, Bulgaria, Croatia Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia Lithuania, Luxembourg, Malta, Netherlands, Norway, Poland, Portugal, Republic of North Macedonia, Romania, Serbia Slovakia, Slovenia, Spain, Sweden, Switzerland, Turkey and United Kingdom. am! EUROPEAN COMMITTEE FOR STANDARDIZATION COMITE EUROPEEN DE NORMALISATION EUROPAISCHES KOMITEE FOR NORMUNG CEN-CENELEC Management Centre: Rue de la Science 23, B-1040 Brussels ©2019 CEN All ights of exploitation in any form and by any means reserved Ref, No. BN 9136:2019 E ‘worldwide for GEN national Members,BS EN 9138:2019 EN 9138:2019 (E) Contents Page 0.1 Paradigm of quality requirements and product acceptance plans... 0.1.1 Engineering 0.1.2 Operations 0.2 Acceptance method considerations 1 Scope. 1.1 Purpose... 1.2 Application 2 Normative references.. 24 International Aerospace Quality Group publications.. 2.2 — American National Standards Institute publication: 2.3 International Organization for Standardization publications. 24 — Japanese standards 2.5 Published books and periodical: 2.6 SAE International publication: 2.7. Standardization Administration of China publications .. 28 — United States Military standards. 3 Terms and definitions mrmmmn 4 General requirements nnn 4.1 Introduction to general requirements. 4.2 Trained personnel eurmmnnn 4.3 Safety/critical characteristics 44 Quality requirements and parameters 44.1 Quality parameter. 44.2 Quality parameter value. 44.3 Alternative quality plans. 4.4.4 Regulatory requirements... 4.5 Selection of samples.. 4.6 Consideration of product and process characteristics. 4.7 Evaluation systems. 4.8 Non-conformances (within the sample) 4.9 Statistical product acceptance plans and records. 4.10 Alternate statistical techniques for product acceptance... 4.11 Auditing of statistical techniques for product acceptance. 5 Required information to be documente 5.1 Documented procedure requirements 5.2 Initial Reliability Requirement or quality parameter values for product acceptance. 5.3 Assigning responsibilities. 5.4 Assigning where applied 5.5 Establishing training program 5.6 ig non-conformances 5.7 Monitoring performance and effectiveness 5.8 Defining adjustments to inspection frequency 5.9 Customer approval of procedures and plan:61 6.2 6.24 6.2.2 63 7A 72 73 TA 74d TAL 743 Ad 75 151 75.2 75.3 8.1 814 8.1.2 8.13 B14 B15 8.1.6 8.7 8.18 8.1.9 8.1.10 8.1.11 8.1.12 8.1.13 B2 B24 8.2.2 8.2.3 8.24 82.5 8.2.6 8.2.7 82.8 8.2.9 83 84 On O41 91.2 BS EN 9138:2019 EN 9138:2019 (E) Requirements for accepting product by individual lot General requirements for individual lot sampling Quality parameters for individual lot sampling Delivered product conformance probability parameters... Probability of acceptance parameters.. Sample sizes for individual lots. Requirements for accepting product under switching rules. General requirements for switching rules Background information for switching rules. Quality parameter with switching rules... Requirements for accepting product with switching rules. Sampling documentation wv General conditions. ig plan criteria... ication criteria for sampling. Requirements for accepting product by variables sampling plans using switching rules. Variable sampling appli Evaluating lot quality Related requirements... Requirements for accepting product produced under process controls General process control requirements. Data. Reducing inspection with process controls Measurement systems Training program. Process control points. Process control acceptance plans and records. Unbiased sampling. Time or production order sequence.. Data retention plan. Periodic internal audits. Retrievability system. Limiting errors in measurement and recording... Evaluating effects of corrective actions. Statistical Process Control Methods, Location and dispersion. Basis for control limits.. Basis for stability Investigatio Violation of stability Re-evaluation of stability and capability Capability and process control Process monitoring frequency. Variation restrictions. Process parameter controls... Additional requirements for accepting products based on continuous sampli skip-lot sampling, or methods for special cases Continuous sampling. Quality parameters for continuous sampling Inspection sequenceBS EN 9138:2019 EN 9138:2019 (E) 9.1.3 General conditions. 9.1.4 Supply Chain Management Handbook continuous sampling tables. 9.2 Skip-lot sampling... 9.2.1 General conditions for skip-lot samplin; 9.2.2 Quality protection equivalency... 9.2.3 Considerations for subsequent lot: 9.3. Sampling strategies for special case: 9.3.1 Special sampling customer approvals. 9.3.2 Special case statistical validity... 9.3.3. Predicting quality from correlated variables... 9.3.4 Continuous manufacturing process - First and last unit for a lot produced ... 9.3.5 Sequential sampling. 9.4 Other methods for special cases BREESE ESS Be Baad Annex A (informative) Guidelines for assignii levels. AA Introduction nnn A2 Assigning quality levels. A3 Initial Reliability Requirements Annex B (informative) Acronym log. Annex C (informative) Mathematics for accepting product under lot-by-lot inspection. C1 Yield estimation.. C11 Yield estimates. 1.2 Confidence limits on yield estimates €2 Probability of acceptance mathematics.. 2.1 Hypergeometric probability of acceptance .. 2.2 Legacy Pa lot acceptance mathematic: €2.3 Yield rates, discrete units, and conformance probabilities. 2.4 Impact lot sampling methods on delivered yield. €3 Controlling the probability of conformance for delivered product. €3.1 Average Outgoing Quality Limits. €3.2_ Average Outgoing Quality Limit sample sizes.. €.3.3 Outgoing Quality Confidence Limit sample sizes €3.4 Lot sampling with deliveries from the sample of failed lots €3.5 Probability of acceptance sample size: €3.6 Lot sampling with destructive tests C4 — Mathematics to control clustering. 4.1 Formulation for cluster control.. 4.2 Fixed number of maximum escapements.. 4.3 Controlling clusters in increasing lot sizes... SSseqaaa e 238 BRRAaS Annex D (informative) Guidelines and recommendations for successful implementation of statistical product acceptance methods Guidelines for 4.2 — Trained personnel. General training. Training for lot sampling Training for product acceptance based on process controls Training for continuous sampling, lot sampling with switching rules, skip-lot sampling, or process control methods Non-published statistical methods. Guidelines for 4.3 — Safety /critical characteristics. ‘ements and parameters nnn Guidelines for Initial Reliability Requirement assignment.D3.2 D4 D41 DA2 D43 DAs DAS D46 DS D6 D61 D.6.2 D6.3 D7 D714 D.7.2 D.7.3 Ds D81 D.82 D83 D9 D.10 DAL DAZ D3 D4 DAS BS EN 9138:2019 EN 9138:2019 (E) Guidelines for quality parameter measures. Guidelines for 4.5 — Selection of sample: ‘Sampling recommendations.. Simple random sampling. Stratification sampling ‘Systematic sampling Cluster sampling... Representative sampling without randomization.. Guidelines for 4.6 — Identification of product characteristics, Guidelines for 4.7 — Evaluation systems. Measurement devices .. Non-Destructive Testing... Destructive testing. Guidelines for 4.8 — Non-conformances (within the sample).. Non-conformance risks . C= 0 sampliny Retrievability in reduced, continuous, or skip-lot sampling Guidelines for 4.9 — Statistical product acceptance plans and records.. Statistical product acceptance plans. Acceptance records. Training records... Guidelines for 4.10 — Alternate statistical techniques for product acceptance Guidelines for 4.11 — Auditing of statistical techniques for product acceptance.. Guidelines for 5.3 — Assigning responsibilities Guidelines for 5.4 — Assigning where applied. Guidelines for 5.5 — Establishing a training program. Guidelines for 5.6 — Controlling non-conformances Guidelines for 5.7 — Monitoring performance and effectiveness... Annex E (informative) Discussion on the technical rationale and history of the relationship E1 ELd E12 E13 E2 E21 E22 E23 E3 Es of quality parameters to reliability. Technical rationale for quality — Reliability relationship in quality parameters. Minimum probability of conformance... Probability of non-conformance... Reliability... History of quality — Reliability relationship in quality parameters. Acceptable Quality Level. Average Outgoing Quality Limit. Short run sampling. Recommended quality parameter choices.. Consideration of cluster effects. Annex F (informative) Guidelines for choosing a statistical product acceptance technique Fa Fd FA2 F2 21 F.22 F.23 F.24 F.25 F.2.6 and quality parameter... Guidance for choosing a method of statistical product acceptance. General guidance Clause-specific guidance... Directions to match a quality parameter to a statistical acceptance technique.. Guidelines for applying quality parameters. Overview of quality parameter relationships to Initial Reliability Requirement... Quality parameters for individual lot sampling Quality parameters for lot sampling with switching rules. Quality parameters for product acceptance using process controls Quality parameters for continuous sampling and special methods ..BS EN 9138:2019 EN 9138:2019 (E) F.3__ Assigned by feature and tolerance. Ga G.1.1 Process steps for evaluating individual or isolated lots G12 Selecting an appropriate sampling model... G13 Screening lot: G2 Guidelines for 6.1.2 — Deliver G3 Guidelines for 6.2.2 — Probability of acceptance parameters. G4 — Guidelines for relating AQL to a quality parameter ... G5 Guidelines for 6.3 — Sample sizes for individual lots.. Annex H (informative) Guidelines for some special acceptance method: H.1 Continuous sampling extensions. Guidelines for skip-lot sampling plans. Guidelines for variables continuous sampling. Poisson continuous sampling.. Process characterization. Guidelines for 9.3.3 — Predicting quality from correlated variable: Guidelines for 9.3.4 — Lot acceptance by first and last unit inspections... Guidelines for Expected Outgoing Quality. Probability Outgoing Quality Limit Validation of special acceptance methods — Simulation.. Guidance for small production volumes. Annex I (informative) Guidel under switching rules.. 11 Guidelines for Clause 7 — Requirements for accepting product produced under switching rules. 12 Guidelines for 7.4 — Requirements for accepting product produced in repetitive lots by attribute sampling plans .. 13. Guidelines for 7.5 — Requirements for accepting product produced in repeti by variables sampling plans.. Annex J (informative) Guidelines and recommendations for accepting product produced under process controls JA Guidelines for 8.1 — General process control requirements wn J.1.1 Guidelines for 8.1.12 — Limiting errors in measurement and recording. J.2 Guidelines for 8.2 — Statistical Process Control J.2.1 Guidelines for 8.2.1 — Methods. J.2.2 Guidelines for 8.2.3 — Basis for control limits... J.2.3 Guidelines for 8.2.5 — Investigation. J.24 Guidelines for 8.2.6 — Violation of stability. J.3. Guidelines for 8.3 — Variation restrictions.. J4 Guidelines for 8.4 — Process parameter controls. J.5 Step-by-step summary es and recommendations for accepting product produced Annex K (informative) Guidelines and recommendations for accepting product based on continuous sampling. Applicability of continuous sampling. Conditions necessary for continuous sampling plans. Conditions when continuous sampling inspection might be desirable. Conditions when sampling plans require approval ‘Types of continuous sampling plans. Implementing continuous samplingBS EN 9138:2019 EN 9138:2019 (E) Guidelines for 9.1.1 — Quality parameters. Guidelines for 9.1.2 — Inspection sequence. ‘Submission of product.. Sample selection... Corrective action indications 106 Example continuous sampling plan 106 graphy, 109 Figures Figure 1 — Engineering and operations inputs into product acceptance plan. 1 Figure 2 — EN 9138 clause selection guide... 13 Figure 3 — Typical Statistical Process Control chart depicting control limits, warning limits, and zones, 40 Figure C1_— Average Outgoing Qualty Operating Characteristic curve with Average Outgoing Quality Limit and Initial Reliability Requirement limit. 62 Figure C.2— Average Outgoing Quality Operating Characteristic curve where failed lots are scrapped. 63 Figure C3 — Points on Average Outgoing Quality Operating Characteristic curve for meeting Initial Reliability Requirement when salvaging sample units from failed lots... 64 Figure G.1 — Flowchart for identifying lot sampling method.. 88 Figure H.1— Example graphic Confidence Level to match Initial Reliability Requirements. 94 Figure K.1 — Illustration of CSP-1 continuous sampling. 1-107 Tables Table 1 — Acceptable Quality Level to Average Outgoing Quality Limit conversion. ‘Table 2 — True Cx requirements Table A.1 — Minimum probability conformance values. Table C.1 — Table of mathematical variables, Table C.2 — Minimum sample sizes needed for a yield Confidence Level when no non-conformities are detected. ‘Table C.3 — Expected escapements and delivered un Table C4 — Average Outgoing Quality formulas for lot sampling categories. 60 Table C.5 — Average Outgoing Quality formulas for examining lot size limits 61 Table C.6 — Expected escapements delivered units when tests are destructive 64BS EN 9138:2019 EN 9138:2019 (E) Table E.1 — Formulas for converting quality parameter specifications into an Initial Reliability Requirement. ' Table E.2 — Conversion of Acceptable Quality Level specification to Average Outgoing Quality Limit... ‘Table F.1 — Guidelines to determine inspection logistics. 84 Table F.2 — Quality parameter value to Initial Reliability Requirement value comparisons. 96 Table H.1 — Example computations for Expected Outgoing Quality. Table1— Example Acceptable Quality Level values to use to meet a required 3 % Average Outgoing Quality Limit..... ‘Table K.1 — Values "/" for choice of" 108 and Average Outgoing Quality Limit.BS EN 9138:2019 EN 9138:2019 (E) European foreword This document (EN 9138:2019) has been prepared by the Aerospace and Defence Industries Association of Europe - Standardization (ASD-STAN).. After enquiries and votes carried out in accordance with the rules of this Association, this Standard has received the approval of the National Associations and the Official Services of the member countries of ASD, prior to its presentation to CEN. ‘This European Standard shall be given the status of a national standard, either by publication of an identical text or by endorsement, at the latest by May 2020, and conflicting national standards shall be withdrawn at the latest by May 2020. Attention is drawn to the possibility that some of the elements of this document may be the subject of patent rights. CEN shall not be held responsible for identifying any or all such patent rights. According to the CEN-CENELEC Internal Regulations, the national standards organisations of the following countries are bound to implement this European Standard: Austria, Belgium, Bulgaria, Croatia, Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Norway, Poland, Portugal, Republic of North Macedonia, Romania, Serbia, Slovakia, Slovenia, Spain, Sweden, Switzerland, Turkey and the United Kingdom,BS EN 9138:2019 EN 9138:2019 (E) Rationale This standard is an upgrade and replacement for the Aerospace Recommended Practice (ARP) 9013-series of documents (i.e., ARP9013, ARP9013/1, ARP9013/2, ARP9013/3, ARP9013/4) prepared and published by the Americas Aerospace Quality Group (AAQG) in 2005. Technically equivalent standards are published in all International Aerospace Quality Group (IAQG) sectors (ie, Americas, Europe). Reasons for publishing this standard include the following: 1) Quality Engineers and planners within many organizations which utilized sampling standards, prior to the publishing of the ARP9013-series of documents, found previous standards difficult to iterpret and/or to implement correctly. The IAQG recognized this situation and chartered a committee in 2001 to develop a new sampling standard. The assignment to write a new statistical standard was given to the Americas sector of the IAQG with the stated strategy that once the new document was published and tested in the AAQG, it would be brought back to the IAQG for global mplementation. 2) The ARP9013-series of documents was published in 2005 to be simpler than existing legacy standards, It also marked a transition from legacy statistical product acceptance requirement documents that were organized around a measure of producer protection rather than consumer protection. To accomplish this without extreme increases in inspection required the simultaneous introduction of new statistical tools and tables. 3) After experience gained from the initial release of the ARP9013-series of documents, the IAQG believed that the goals of simpler and more effective statistical methods for Quality Engineers and planners had been achieved, but that further improvements were possible, both in the development of new tools and in providing further detail clarifications within the writing. 4) Furthermore, there have been changes in the EN9100/EN 9110/EN 9120 quality management system standards relative to the language associated to statistical product acceptance. This EN 9138 standard incorporates those changes. Foreword To assure customer satisfaction, the aviation, space, and defence industry organizations produce, and continually improve, safe, reliable products that meet or exceed customer and regulatory authority requirements. The globalization of the aerospace industry and the resulting diversity of regional/national requirements and expectations have complicated this objective. End-product organizations face the challenge of integrating and assuring the quality of product purchased from suppliers throughout the world and at all levels within the supply chain, while suppliers and processors face the challenge of delivering product to multiple customers having varying quality expectations and requirements. ‘The aerospace industry established the IAQG for the purpose of achieving significant improvements in quality and safety, and in reduced costs throughout the value stream. This organization includes representation from aerospace companies in the Americas, Asia/Pacific, and Europe. This standard has been prepared by the IAQG. This document standardizes and streamlines, to the greatest extent possible, the requirements and flexible resources on statistical techniques for product acceptance across the diversity of aerospace industry processes. The establishment of common requirements, for use at all levels of the supply-chain by organizations around the world, should result in improved quality and safety, and decreased costs due to the elimination or reduction of organization-unique requirements and the resultant variation inherent in these multiple expectations. 10BS EN 9138:2019 EN 9138:2019 (E) Introduction ‘This standard establishes the general requirements applicable to any method of statistical product acceptance to reduce inspection costs while still assuring acceptable quality. There is no single specific plan that can be considered best suited for all applications or processes. This document applies only to statistical methods used for product acceptance and does not apply to statistical methods that are not related to product acceptance. Many companies use excellent statistical methods solely to monitor and to improve their product quality, and those methods are not subject to the requirements of this document. Products which are eligible for the methods defined in this standard include, but are not limited to: end items, cast, forged, wrought, machined, fabricated, plastic, moulded, powdered metal, or stamped ‘components and raw material; electronic, electrical, and mechanical components. 0.1 Paradigm of quality requirements and product acceptance plans Figure 1 shows how requirements and acceptance strategies come together to develop a product acceptance plan. Review applicable definition (e.g. drawings, Process logistics specifications) to (eg lot sizes) determine IRR value Review Table A.1 to Determine find closest matching acceptance strategy entry or contact design owner for applicable value Type of quality parameter definition has IRR Initial Reliability Requirement (IRR)_[—~L__Wnite plan Figure 1 — Engineering and operations inputs into product acceptance plan 0.1.1 — Engineering Each product is engineered to meet functionality and reliability goals. Engineering provides requirements that are needed to meet the goals. Assessment of the goals involves statistics. To compare the goals with engineering requirements requires a statistical benchmark. This document expresses the engineering requirement as an Initial Reliability Requirement (IRR) or Inspection Reliability Requirement. The IRR defines a minimum acceptable outgoing yield or probability of conformance. Itis recommended by regulatory authorities that, “Engineering and manufacturing organizations should participate in the review, implementation, and maintenance of statistical quality/process control techniques used for product or article acceptance” [reference Federal Aviation Administration (AA) Advisory Circular (AC) 21-43]. If Engineering does not provide a specific protection value (e.g, IRR), then the values in Annex A (see Table A.1) are provided as conventional levels of protection. cerBS EN 9138:2019 EN 9138:2019 (E) 0.1.2 Operations The attainment of product functionality and reliability goals involves several kinds of production logistics. Product may be produced in batches, large or small lots, continuous processes, or single-piece flow manufacturing processes. Assessments of the product may occur in receiving inspection, in- process inspections, final inspection, or in-storage inspection. The statistical product acceptance requirements for each kind of production logistics are documented in separate clauses of this document. Each clause cites one or more quality parameters that are used to verify that produced parts meet the Engineering IRR. Manufacturing and/or inspection should provide the range of lot sizes and frequencies that are expected to be used to produce the subject product. Additional information from these organizations should include measurement accuracy (see 4.7), randomization tools (see 4.5), product retrievability limits (see 4.8), and available resources for training, auditing, and records (see 4.2, 4.11, and 4.9 respectively). Product may be accepted as one or more isolated lots under the instructions of clause 6; sample sizes may be adjusted from lot-to lot based on the history of lot rejections under the instructions of Clause 7; product may be accepted based on process controls under the instructions of Clause 8; and product may be accepted one unit at a time or by other advanced methods under the instructions of Clause 9. Each clause has further explanations and guidance in an associated Annex. There are also guidance and sampling tables relating to these clauses in the Supply Chain Management Handbook (SCMH) published. online by the IAQG. 0.2 Acceptance method considerations Selecting which of EN 9138 clauses 6, 7, 8, or 9 is most appropriate for a statistical product acceptance application depends on the answers to the five decision diamonds (see Figure 2). Each clause covers a family of statistical techniques and their associated quality parameters; Figure 2 presents how the decisions result in the selection of the most appropriate clause. 12a) ») BS EN 9138:2019 EN 9138:2019 (E) Statistical Acceptance Application Custom design? Lot or Batch Inspection? Attribute or Variable? Variable Switching Rules? lo Yes 9138 Clause 6 9138 Clause 7 9138 Clause 8 9138 Clause 9 Figure 2 — EN 9138 clause selection guide In the first decision diamond, the question is whether the specific statistical tool is drawn from published sources (e.g, SCMH) or whether the organization intends to develop their own method; this recognizes that the EN 9138 standard allows for the development of new methods. The skill level for doing this requires competency with the mathematical tools in the definitions clause of this document, support from individuals knowledgeable about the measurement or evaluation processes, and support from individuals who understand the consequences of non-conformances (these should include design authority individuals). In the second decision diamond, the question is whether the product in question is to be accepted one unit at a time versus being accepted in lots or batches. This decision may be affected by “Economic Order Quantity” mathematics or by direct contractual stipulation. In the third decision diamond, the question is associated with the use of switching rules. For many years, most sampling standards required users to keep track of their records of previous accepted and rejected lots from each sampled process, and to increase or to decrease the sample sizes in response to the process history. The rules for increasing or decreasing the amount of inspection were called “switching rules”. Switching rules may be helpful in processes that have long production runs, stable quality, relatively large lots, and good computing support to handle the administrative details. Switching rules are also sometimes required by contractual reference to legacy standards. In other cases, the simplicity of individual lot sampling plans may offset the potential savings of using more complicated switching rules. 13BS EN 9138:2019 EN 9138:2019 (E) 4) The fourth decision diamond asks whether the product quality is evaluated on an attribute or variables basis. This question depends primarily on the level of computing and administrative resources available, because variables data involves collecting the detailed measurements of every accepted feature. The additional information from variables sampling does allow for fewer inspections to be performed, so the tradeoff would usually apply variables inspection to products with few characteristics and very high inspection costs per unit. e) The last decision diamond asks whether process controls alone will control the probability of an escape to the customer. 14BS EN 9138:2019 EN 9138:2019 (E) 1 Scope 1.1 Purpose ‘This European standard establishes requirements when implementing statistical product acceptance methods to meet defined risk requirements. This standard also establishes the minimum content required to be covered in an organization's documented procedures that govern their application of statistical product acceptance methods. ‘These general requirements and documented procedures apply the requirements of the EN 9100/EN 9110/EN 9120 quality management system standards, in addition to establishing requirements for retrievability, safety/critical characteristics, and quality parameters that protect the customer. 1.2 Application This standard is applicable when invoked in a purchasing contract or specification, contractual document, customer agreement, or adopted by the organization. The purchase contract/agreement may or may not identify the appropriate EN 9138 clause(s) to be applied by the organization. All statistical methods of product acceptance require the use of Clause 4 and Clause 5. ‘To accept product produced: — by individual lots, see Clause 6; — under switching rules, see Clause 7; — under process controls, see Clause 8; and — by continuous sampling or special case methods, see Clause 9. 2 Normative references ‘The following referenced documents are valuable for the application of specific tools presented in this document. For dated references, only the edition cited applies. For undated references, the latest edition of the referenced document (including any amendments) applies. When a conflict in requirements between this document and the referenced standards/related writings exists, the requirements of this document shall take precedence; however, nothing in this document supersedes applicable laws, regulations, and contractual requirements. 2.1. International Aerospace Quality Group publications EN 9100, Quality Management Systems — Requirements for Aviation, Space and Defence Organizations EN 9110, Quality Management Systems — Requirements for Aviation Maintenance Organizations! EN 9120, Quality Management Systems — Requirements for Aviation, Space and Defence Distributors! 1 As developed under the auspice of the IAQG and published by various standards bodies [eg, Aerospace and Defence Industries Association ~ Standardization (ASD-STAN), SAE International, European Committee for Standardization (CEN), Japanese Standards Association (JSA)/Society of Japanese Aerospace Companies (SJAC), Brazilian Association for Technical Norms (ABNT)]. 15BS EN 9138:2019 EN 9138:2019 (E) Supply Chain Management Handbook? NOTE References to specific EN 9100/EN 9110/EN 9120 revision dates are included in this document in order to allow for the corresponding standard clause numbers to be identified and to recognize the potential for future changes to these standards; however, the applicable clause(s) from the latest revision have precedence. 2.2 American Na‘ nal Standards Institute publications ANSI/ASQ Z1.4, Sampling procedures and tables for inspection by attributes? ANSI/ASQ/Z1.9, Sampling procedures and tables for inspection by variables for percent nonconforming? ANSI/ASQC B1, Guide for Quality Control Charts ANSI/ASQC B2, Method of Analyzing Data’ ANSI/ASQC B3, Controlling Quality during Production? ANSI/ASQCS1, An Attribute Skip-Lot Sampling Program’ 2.3. International Organization for Standardization publications ISO 2859-1, Sampling procedures for inspection by attributes — Part 1: Sampling schemes indexed by acceptance quality limit (AQL) for lot-by-lot inspection* IS 3951-1, Sampling procedures for inspection by variables — Part 1: Specification for single sampling plans indexed by acceptance quality limit (AQL) for lot-by-lot inspection for a single quality characteristic anda single AQL* 180.9000, Quality management systems — Fundamentals and vocabulary* ISO 11462-1, Guidelines for implementation of statistical process control (SPC) — Part 1: Elements of sPCt ISO 1462-2, Guidelines for implementation of statistical process control (SPC) — Part 2: Catalogue of tools and techniques* ISO/IEC Guide 98-1, Uncertainty of measurement — Part 1: Introduction to the expression of uncertainty in measurement* ISO/IEC Guide 98-3, Uncertainty of measurement — Part 3: Guide to the expression of uncertainty in measurement (GUM:1995)* ISO/IEC Guide 98-4, Uncertainty of measurement —Part4: Role of measurement uncertainty in conformity assessment* 2 See http://www.sae.org/iaag/ 3. Available from American National Standards Institute, 25 West 43rd Street, 4th Floor, New York, NY 10036, ‘Tel: 212-642-4900, www.ansi.org 4 Available from International Organization for Standardization, ISO Central Secretariat, 1, ch. de la Voie-Creuse, CP 56, CH-1211 Geneva 20, Switzerland, Tel: +41 22 749 01 11, www.iso.org 16BS EN 9138:2019 EN 9138:2019 (E) 2.4 Japanese standards J1SZ9003:1979, Single sampling inspection plans having desired operation characteristics by variables (standard deviation known)5 J1S29004:1983, Single sampling inspection plans having desired operating characteristics by variables (standard deviation unknown and single limit specified)® J1S29010:1999, Sequential sampling plans for inspection by variables for percent nonconforming (known standard deviation)S J1S29015-0:1999, Sampling procedures for inspection by attributes — Part 0: Introduction to the JISZ9015 attribute sampling system’ J1SZ9015-1:2006, Sampling procedures for inspection by attributes — Part 1: Sampling schemes indexed by acceptance quality limit (AQL) for lot-by-lot inspection® J1SZ9015-3:1999, Sampling procedures for inspection by attributes —Part 3: Skip-lot sampling procedures J1SZ9020-1:2011, Control charts — Part 1: General guidelines J1S29021:1998, Shewhart control chartsS J1S29041-1:1999, Statistical interpretation of data — Part 1: Statistical presentation of data J1S29041-2:1999, Statistical interpretation of data — Part 2: Techniques of estimation and test relating to ‘means and variancess J1SZ9041-3:1999, Statistical interpretation of data — Part 3: Tests and confidence intervals relating to proportions® J1SZ9041-4:1999, Statistical interpretation of data— Part 4: Power of tests relating to means and variances> J1829041-5:1999, Statistical interpretation of data— Part 5: Median— Estimation and confidence intervalsS 2.5 Published books and periodicals Certified Reliability Engineer Primer, 4th Edition (October 1, 2009); Quality Council of Indiana; page VI-21 The Mathematical Background of QSTAG 340; American-British-Canadian-Australian Armies Standardization Program, Quadripartite Advisory Publication 16; pages VI-2 and VI-3 ‘Zero Acceptance Number Sampling Plans; N. L. Squeglia; ASQ Quality Press 5 _ Available from Japanese Standards Association (JSA), Mita MT Bldg, 3-13-13 Mita, Minato-ku, Tokyo, 108- 0073, Japan, Tel: +#81-3-4231-8503, http://www.jsa.or.jp/ 17BS EN 9138:2019 EN 9138:2019 (E) 2.6 SAE International publications ARP9013, Statistical Product Acceptance Requirements® ARP9013/1, Statistical Product Acceptance Requirements Using Isolated Lot Sampling Methods® ARP9013/2, Statistical Product Acceptance Requirements Using Attribute or Variable Lot Acceptance Sampling Plans® ARP9013/3, Statistical Product Acceptance Requirements Using Process Control Methods® ARP9013/4, Statistical Product Acceptance Requirements Using Continuous Sampling, Skip-Lot Sampling, or Methods for Special Cases® 2.7. Standardization Administration of China publications GB/T 2828.1, Sampling procedures for inspection by attributes — Part 1: Sampling schemes indexed by acceptance quality limit (AQL) for lot-by-lot inspection? GB/T 2828.3, Sampling procedures for inspection by attributes — Part 3: Skip-lot sampling procedures? GB/T 6378.1, Sampling procedures for inspection by variables — Part 1: Specification for single sampling plans indexed by acceptance quality limit (AQL) for lot-by-lot inspection for a single quality characteristic and a single AQL? 2.8 United States Military standards MIL-STD-105, Sampling Procedures and Tables for Inspection by Attributes® MIL-STD-414, Sampling Procedures and Tables for Inspection by Variables for Percent Defective® MIL-STD-1235C, Single and Multilevel Continuous Sampling Procedures and Tables for Inspection by Attributes® NOTE — MIL-STD-1235 has been cancelled, but can be used as a guide for derivation and application of continuous sampling plans. The SCMH contains the essential text and tables of MIL-STD-1235C with some editing, reformatting, and paragraph renumbering. 6 Available from SAE International, 400 Commonwealth Drive, Warrendale, PA 15096-0001, Tel: 877-606-7323 (inside USA and Canada) or +1 724-776-4970 (outside USA), www.sae.org, 7 Available through Standardization Administration of the People's Republic of China (SAC), hitp://www:sacgov.cn/sacen/, controlled by a Technical Committee of SAC (TC21-Application of Statistical Methods), contact Ding Wenxing, e-mail: dingwx@cnis.gov.cn 8 Copies of these documents are available online at http://quicksearch.dla.mil 18BS EN 9138:2019 EN 9138:2019 (E) 3 Terms and definitions For the purpose of this document, the terms and definitions provided in 10 9000 and the following apply. An acronym log for this standard is presented in Annex B. 34 Acceptable Quality Level AQL acceptance quality limit maximum percentage or proportion of non-conforming units in a lot or batch that, for purposes of acceptance sampling, would be considered satisfactory as a process average 32 acceptance sampling inspection in which decisions are made to accept or not accept a lot (or other grouping of product or material), based on the results ofa sample or samples selected from that lot or grouping 33 Analysis of Variance ANOVA arithmetic process developed for the purpose of comparing several sample means by partitioning the total variation of a set of data into components associated with recognized sources of variation 34 attribute data qualitative data that counts the presence or absence of a characteristic evaluated to a specified criterion, rather than data from a continuous measurement scale Note 1 toentry: Attribute data is typically measured on a go/no-go, pass/fail, or yes/no basis. 35 Average Outgoing Quality A0Q average quality of outgoing product, after sampling inspection, for a given steady value of incoming product quality 36 Average Outgoing Quality Limit AOQL maximum AQQ over all possible values of incoming product quality level for a given acceptance sampling plan 37 Average Run Length ARL average number of plotted points on a control chart or the average number of items inspected before a shift in process level can be expected to be signalled (e.g,,a point plotting outside the control limits) 38 batch same as a “lot” or “grouping” 19BS EN 9138:2019 EN 9138:2019 (E) 3.9 Bernoulli trials model model where a trial has exactly two possible outcomes (ie, success and failure) and the probability of success is the same every time the experiment is conducted _ (lf x occurred *ernoull = { 0,if x did not occur a P if xoccurred Pr(saemout) = {1 » “ip eid not occur ® Bernoulli (xp) = p*(1- p)'*, oa @) 3.10 binomial distribution distribution of counts of successes (or failures) in a number of trials when there are only two kinds of outcomes and the probability of a success (or failure) is constant for all trials (ie., the trials are independent) b(b:p.n) = Y”xXpernoun = (7) (P*)G = py where (Kk; p,n)_ is the probability of k non-conforming items in n trials; n is the fixed number of Bernoulli trials; k is the number of non-conforming units in n trials; P isthe probability of a non-conforming unit; 1-p isthe probability ofa conforming unit Note 1 to entry: ‘The binomial distribution results from summing repeated Bernoulli trials. 341 0 (Cis equal to zero) acceptance sampling plan that requires rejection of the lot, batch, or grouping if the sample contains one or more non-conformances Note 1 to entry: The variable “C” stands for the number of non-conformances in a sample that allows acceptance ofthe lot. 312 capability natural limits of a process; typically, the range of six standard deviations of the values in a process having no changes in its mean or any other characteristic of its distribution; inherent variation in a stable process; more precisely, a metric that characterizes the output of a stable process following a statistical distribution Note 1 to entry: For purposes of establishing a natural variation limit, the distribution is of a continuous variable. 20BS EN 9138:2019 EN 9138:2019 (E) a) Ananalysis of a process that compares the output process with some criteria for conformity. USL -LSL si és (8) *) 6) Note2toentry: Continuous variables analyses compare a model distribution with specifications. . z Cox = Min LSL USL 30 3a Note3toentry: G, and Cx distribution. The appropriateness of the normal distribution is to be verified. translations to a probability of conformance commonly depend on a normal b) For the case of attribute data, it is usually defined as the average proportion or rate of defects or defectives (e.g, centreline of an attribute control chart). Note4toentry: The Bayesian estimator for the defect rate when no defects are observed in the current sample and the prior distribution adds no information is Formula (7). 1 Prd) = 5 ” where Pr(d)_ isthe probability of a defect (the defect rate); n is the number of representative observations of the process, no defects observed (see “Probability and Statistics for Engineers” by R.L. Scheaffer and J.T. McClave). 3.13 combinations number of distinct groupings of m distinct objects taken “/" at a time, where order of selection is immaterial (e.g, 123, 321, 213) ' ()= tegen ® where m!=(m) x (m—1) x (m—2)x + x (3) x (2) x (1), 344 conformance when a character requirements tic of a product or process meets its requirements; fulfilment of specified 3.15 consumer risk probability that a lot with a fraction of non-conforming material equal to a stated index value will be accepted by a sampling plan 3.16 continuous sampling inspection intended for application on the continuous flow of individual items of product that involves acceptance or non-acceptance on a unit-by-unit basis for the applicable characteristics using periods of one-hundred percent (100%) inspection and periods of sampling, depending on the quality of the observed product (see the SCMH) 21BS EN 9138:2019 EN 9138:2019 (E) 3.17 critical characteristic safety characteristic safety critical characteristic characteristic designated with the words “critical” or “safety” by the design authority, where the responsibility for its definition is outside the scope of this standard 3.18 customer buying entity that is the issuer of a contract to a supplier Note 1 to entry: All aviation, space, and defence suppliers are also customers to their suppliers. Protection to the customer is the key goal of this standard. 3.19 defect non-fulfilment of a requirement related to an intended or a specified use; departure of a quality characteristic from its intended level or state that occurs with a severity sufficient to cause an associated product or service not to satisfy intended normal or foreseeable usage requirements Note 1 to entry: This is different from a non-conformance. 3.20 determine decide conformity in a way that achieves the level of probability of conformance in accordance with the applicable quality parameter 3.21 Equal Risk Point ERP fraction non-conforming at which the probability of rejecting a lot is fifty percent (also noted as Pp 50) Note 1 to entry: At this fraction non-conforming, the risk to the producer and consumer are the same, When no integer number of non-conforming units produces exactly 50 % probability of acceptance, the ERP for that sample size is computed as the weighted average of the two nearest points on its discrete Operating Characteristic (0C) curve [see Annex ¢ (C.2.2)]. 3.22 escapement escape non-conformities that were produced, not detected and remedied, and subsequently sent to the customer 3.23 Expected Outgoing Quality EOQ probability that a unit accepted by sampling inspection is non-conforming, given both the sampling plan and a statement of what quality patterns have been demonstrated for the process Note 1 to entry: An £0Q can only be computed if there is a distribution, whether based upon parameters or an empirical distribution. 22BS EN 9138:2019 EN 9138:2019 (E) 3.24 Gage Repeatability and Reproducibility Gage R&R study used to determine the degree of variability within a measurement system, and to separate this variability into components associated with the gage and the use of the gage 3.25 grouping same as a “lot” or “batch” 3.26 homogeneity homogeneous when applied to a process, homogeneity is where there is repetitive production running under the same conditions with similar behaviour. Theoretical homogeneity is distinct from practical homogeneity: a) “Practical homogeneity”: where all units of product are made according to the same design requirements under stable consistent conditions of production; a fixed process. Note 1 toentry: Fixed processes will have identifiable streams or groups of product according to type, grade, lass, size, composition, or other conditions of manufacture that could cause inconsistencies. b) “Theoretical homogeneity”: the state of having identical cumulative distribution functions or values. 3.27 hypergeometric distribution discrete probability distribution that describes the number of items of interest foun froma finite population a sample taken Note 1 toentry: The probability describes sampling without replacement. Applies whenever one is computing the probability of accepting a single finite lot and it is especially valuable when the sample is large compared to the population. Pr[X =x] = ons o) where x =0,1,2,..,min. of n or M; = 1,2, «(population size); M_ =0,1,2,3, ...N (occurrences in population); n =1,2,..,N (sample size); (2) = mem 23BS EN 9138:2019 EN 9138:2019 (E) 3.28 Initial Reliability Requirement IRR Inspection Reliability Requirement minimum probability of conformance for each unit of product accepted through the associated sampling plan for design purposes 3.29 inspection process of measuring, examining, testing, gauging, or otherwise comparing product or material with requirements and determining conformance 3.30 isolated lot lot which is accepted solely on the basis of information derived from within itself; separate from the sequence of lots in which it was produced or collected, and not depending on information from any other part ofa current sequence of inspection lots 3.31 Limiting Quality 1Q Lot Tolerance Percent Defective (LTPD) Lot Quality Protection (LQP) Unacceptable Quality Level (UQL) Limiting Quality Level (LQL) quality parameter that is the fraction non-conforming in a lot where there is a small chance that a lot will be accepted by the specified sampling plan Note 1 to entry: For use throughout this standard, the term L@ is established at the percent non-conforming having probability of lot acceptance equal to 10 %. Note2 to entry: For precise computation of the LQ value for finite lots of discrete parts, it is generally the case that no integer number of non-conforming units produces exactly 10% probability of acceptance, so the LQ for that sample size can be computed as the weighted average of the two nearest points on the discrete OC curve [see Annex € (C.2.2)}. 3.32 lot batch grouping defined quantity of some product, material, or service collected and submitted together for acceptance Note 1 to entry: _Itis often expected that a lot be homogeneous in its manufacturing conditions. A lot for product acceptance purposes is not necessarily the same as a lot formed for purposes of a contract, shipment, or anything other than acceptance. 3.33 lot-by-lot inspection inspection of each sequential lot in a series 3.34 lot homogeneity quantity of product, material, or service that shows consis within alot rent characteristics throughout the quantity 24BS EN 9138:2019 EN 9138:2019 (E) 3.35 Lot Tolerance Percent Defective LTPD level of quality in a lot that the lot-sampling plan will accept 10% of the time [see 3.31, Limiting Quality (1@)] process of determining the disposition of non-conforming material 3.37 multi-level continuous sampling plan continuous sampling plan that changes frequency or amount of inspection by a set of rules, based upon the results of past plan inspections 3.38 must step that is inherent in meeting obligations from authorities outside this document, especially those of customers, regulatory authorities, or the laws of mathematics Note 1 toentry: The use of the term “must” in this document is distinct from the use of the term “shall” 3.39 non-conformance parts or materials that do not meet specifications or requirements. The failure of a characteristic to conform to the requirements specified in the contract, drawings, specifications, or other approved product description 3.40 non-conformity specific occurrence of a condition that does not conform to a specification or other inspection standard 3.41 Non-Destructive Testing NDT Non-Destructive Inspection (NDI) technique of testing material properties without impairing their future usefulness 3.42 normal inspection default level of lot-by-lot inspection sampling plans with switching rules Note 1 toentry: Switching rules in sampling standards begin with this level of inspection, and direct the inspection to greater or lesser inspection according to the patterns of quality in the process. 3.43 one-hundred percent (100 %) inspection inspection of every item of product or service 25BS EN 9138:2019 EN 9138:2019 (E) 3.44 Operating Characteristic curve OC curve curve on a graph for a given sampling plan showing the lot fraction defective (x-axis) versus the probability that the sampling plan will accept the lot (y-axi Note 1 to entry: This term is also applied to AOQ, Average Fraction Inspected (AFI), Average Total Inspected (ATI, and ARL curves. 3.45 organization entity or party that is performing the work 3.46 Outgoing Quality Confidence Limit ogc computed confidence limit on the probability of conformance for delivered product after sampling Note 1 toentry: Computations depend on the lot size, sample size, acceptance criteria, screening policy, rectification policy, and Confidence Level (CL), 3.47 Poisson distribution discrete frequency distribution that gives the probability of a number of independent events occurring ina fixed population, where the population can be an interval of time, volume, area, or length aj="em (20) a where x isthe number of occurrences (0, 1,2, ..); m_ isthe population mean (m > 0); e isthe base of natural logarithms (2,71828... 3.48 Probability Outgoing Quality Limit POQL upper confidence limit on the outgoing percent non-conforming in a finite series of lots Note 1 toentry: This is the maximum of the curve that represents the 90% upper confidence limit on the Outgoing Quality (0Q) and it depends on the sample size, acceptance criteria, the length of the series of lots, and any switching rules that apply. 3.49 process combination of activities, methods, and resources, including people, material, machines, tools, and environment, that transform inputs into output products or services 26BS EN 9138:2019 EN 9138:2019 (E) 3.50 process control disciplines of controlling the variables of a process to ensure the resultant characteristic meets requirements Note 1 toentry: The adequacy of process controls is described in Clause 8. 3.51 process control point step (in a process) at which a full or partial control is applied to prevent, eliminate, or reduce the risk of a non-conforming part continuing through the process 3.52 process monitoring planned evaluations of the characteristic being accepted via process control to ensure that the process maintains stability 3.53 process stability stability condition or state of behaviour of a process where, through the use of past experience, limits of the process can be predicted for the near future; condition where there is no indication of a special cause of variation, but where only random common cause variation is present 3.54 quality parameter statistic of a product acceptance method that indicates the probability of conformance for each accepted unit of product to determine what it takes to ensure that the product conforms to its requirements (eg, the IRR) Note 1 toentry: Quality parameters include AOQL, OQCL, PPM, ERP, and Rejectable Quality Level (RQL). 3.55 random sample selection of units collected in such a manner that all units being accepted have an equal chance of being selected for inspection or evaluation 3.56 reduced inspection one of several levels of inspection connected by switching rules, which involves a lower fraction inspected than one or more of the others Note 1 toentry: In such a system the quality parameter is usually the AOQL and that AOQL is maintained, despite the lower fraction inspected, by the fact that the process generally only gets to that level when the quality is as good as or better than the AOQL for some period of time. 27BS EN 9138:2019 EN 9138:2019 (E) 3.57 Rejectable Quality Level RQL minimum percent defective at which a lot will have 5 % or less chance of being accepted when using a sampling plan Note 1 to entry: For precise computation of the RQL value for finite lots of discrete parts, itis generally the case that no integer number of non-conforming units produces exactly 5 % probability of acceptance, so the RQL for that sample size can be computed as the weighted average of the two nearest points on the discrete OC curve [see Annex € (C.2.2)}. ability of a production system to identify or trace units, and make them available for appropriate disposition 3.59 sample one or more units of product drawn from a lot, batch, or grouping to provide information 3.60 sampling plan statement of sample size(s), frequencies, associated acceptance and rejection criteria, and switching rules (ifapplied) 3.61 sequential sampling sampling acceptance where a sequence of samples is taken from a lot or grouping and after each unit a decision is made to accept or reject the lot or grouping, or to continue sampling 3.62 shall indicates a requirement 3.63 should indicates a recommendation Note 1 to entry: If the organization chooses not to follow a “should” or other recommendation, then it is advisable to document the rationale for that decision. 3.64 sigma Greek letter (0) symbolizing standard deviation 3.65 single-level continuous sampling alternating periods of 100 % inspection and sampling inspection wherein the sampling rate is constant Note 1 to entry: The extent of periods of 100 % inspection depends on the pattern of non-conformances. 28BS EN 9138:2019 EN 9138:2019 (E) 3.66 skip-lot inspection sampling inspection in which some lots in a series are accepted without inspection, when the sampling results for preceding lots from the same process meet stated criteria and all retrievability requirements for skipped lots have been met 3.67 standard deviation measure of the spread or dispersion of the process output that has the same unit of measure as the original observations and the process mean (i.e, square root of the variance in a population); usually identified by the Greek letter (0) symbolizing standard deviation of the population 3.68 statistical control process operating with a constant system of causes that result in a stable probability distribution of outcomes Note 1 toentry: Evidence of statistical control is provided by observations or observational statistics (eg. average, range, standard deviation, maximum) appearing within limits calculated from extremes of the estimated stable distribution. Note2toentry: Control charts depict graphically the observations and limits over time. Observational statistics within control limits are equivalent to failing to reject the null hypothesis that the process outputs come from a stable distribution. Note3to entry: Tolerance limits do not enter into calculations of control limits or determination of statistical control Note4 to entry: See clause 8 for more details and the method for determining when to accept a process as being adequately in statistical control. 3.69 statistical principles legitimate conclusion, based on samples taken from a defined population, claiming some probability value that is achieved by following a defined procedure; verifiable mathematics exists that shows this probability value is accurate; refers to the phrase “justified on the basis of recognized statistical principles”, as used in quality management system standards 3.70 Statistical Process Control SPC graphical/statistical method by which a process is monitored for patterns of variation Note 1 toentry: The resulting data is analysed statistically to identify incidents of special causes of variation (see Clause 8 for more details). 371 statistically valid legitimate conclusion, based on samples taken from a defined population, claiming some probability value that is achieved by following a defined procedure; by definition, verifiable mathematics exists that shows this probability value is accurate 29BS EN 9138:2019 EN 9138:2019 (E) 3.72 subgroup segregation of process output such that it is reasonable to expect the same causal factors are present throughout the subgroup Note 1 to entry: A subgroup is represented on a control chart by a single point. 3.73 supplier entity or party that supplies product or services to a customer in accordance with contractual requirements 3.74 switching rules requirements used to adjust the level of inspection, such as between “Normal’, “Tightened”, and “Reduced” inspections 3.75, tightened inspection level of sampling with switching rules that has more inspection than a “Normal” inspection level, but generally less than 100 % inspection 3.76 unit single part or item Note 1 to entry: For continuous sampling application, this term can also apply to a lot with a small quantity of parts froma fixed process, where the sample size equals the lot size. 377 variables data variables measurement quantitative measurements taken on a continuous scale 4 General requirements 4.1 Introduction to general requirements ‘The requirements of this clause are common to the application of methods and techniques defined in Clause 6, Clause 7, Clause 8, and Clause 9 which might be restated therein, Additional requirements specific to each method are defined in its respective clause of the document. Refer to Figure 2 for guidance in determining which method and technique is appropriate for the product being assessed. Regardless of the use of these statistical product acceptance methods, the organization shall be responsible to ensure that each part conforms to the engineering design requirements. 4.2. Trained personnel Personnel involved in the planning, application, and execution of statistical product acceptance do affect product quality and therefore shall be subject to the training requirements of the EN 9100/EN 9110/EN 9120 quality management system standards. NOTE Additional training requirements specific to process control are contained in 8.1.4. 30BS EN 9138:2019 EN 9138:2019 (E) 43. Safety/critical characteristics Any characteristics identified in the design documentation as “safety” or “critical” characteristics shall not be accepted using statistical product acceptance methods, unless prior written authorization is, granted from the customer (eg, a customer-approved procedure) such as when the method for acceptance is defined in the design documentation [see Annex D (D.2) for guidance]. No’ Written authorizations for the use of statistical product acceptance on a “safety” or “critical” characteristic in accordance with the EN 9138 standard do not provide adequate information unless that authorization includes a quality parameter value. 44 Quality requirements and parameters When the organization uses statistical techniques for product acceptance, they shall identify a quality parameter and value for each plan. NOTE This is an essential part of the EN 9138 standard fulfilment of the EN 9100 and EN 9120 requirement that “the sampling plan is justified on the basis of recognized statistical principles and appropriate for use.” The value of the quality parameter represents the risk level of sampling; when the risk level is acceptable for a statistical product acceptance plan, the "appropriate for use" stipulation is satisfied. 44.1 Quality parameter A statistical technique for product acceptance shall have a quality parameter that controls the probability of conformance of delivered product (see Annex D for guidance). ‘When engineering specifications require an AQL, the consumer protection quality parameter shall be an AOQL. NOTE The column headings in many existing sampling tables are indexed by the AQL; however, the AQL. is not a measure of protection for the customer. This standard requires the defining ofa different quality parameter that provides protection for the customer. Along with other statistical acceptance requirements contained in this standard, the AQL figures must not be construed as allowing a non-conforming part to be delivered. 44.2 Quality parameter value ‘The quality parameter value shall be selected or approved by individuals who understand the consequences of non-conformance. a) The preferred method to identify the quality parameter value is for the design authority to specify an IRR via instructions on the purchase order, contract, engineering drawing, and/or related product definition documents. This RR determines a maximum for the quality parameter value for the production logistics according to Annex A. b) The quality parameter(s) chosen shall protect the customer by controlling the probability of conformance for delivered product. ©) When instructions on the purchase order, contract, engineering drawing, and/or related product definition documents from the customer require a specific quality parameter and value, the statistical method shall provide at least that degree of protection. 4) Quality can recommend alternative quality parameters or values to the customer authority using guidance in Annex E (see E.3). e) Ifno other authority identifies an /RR value, then Annex A (see Table A.1) can be used as a default to demonstrate conventional levels of protection. 31BS EN 9138:2019 EN 9138:2019 (E) f) When engineering specifications require an AQL, Table 1 shall be used to convert the AQL value to an AOQL value. Table 1 — Acceptable Quality Level to Average Outgoing Quality Limit conversion AQL | 0,010% 0,015% 0,025% —0,040% 0,065% —0,100% —0,150% _0,25% AOQL | 0,018% —0,033% —0,046% —0,074% 0,113% —0,143% 0,198 % —_0,33% AQL | 040% 065% —1,00% 150% 2.50% —4,00% 650% — 10,00% AOQL | 053% 0.79% —1,22% —1,90% 2.90% 494% 7.12% 11.46% 44.3. Alternative quality plans ‘The customer shall retain the right to approve alternative classification methods. 4.4.4 Regulatory requirements ‘The organization shall be responsible for ensuring that any product acceptance methods comply with regulatory requirements, 4.5 Selection of samples Samples shall be representative of the unchecked production. Articles presented for acceptance by sample inspection shall share similar variability, quality level, and history as the unchecked production (see Annex D for guidance). Samples may be predetermined, if it has been determined by an ANOVA, Process Failure Modes and Effects Analysis (PEMEA), or other valid statistical analysis that the predetermined sample shall be over representative of non-conforming product. 4.6. Consideration of product and process characteristics The statistical product acceptance plan shall identify the product characteristics to be accepted under the plan. The organization shall create a statistical product acceptance plan that covers each applicable characteristic. In addition to the general requirements of Clause 4 and Clause 5, the plan shall meet the requirements. of the supporting Clause 6, Clause 7, Clause 8, or Clause 9. Refer to Figure 2 for guidance in determining which method and technique is appropriate for the product being assessed. See examples of these in Annex F (see Table F.1) 4.7 Evaluation systems Monitoring and measuring devices used to accept product shall be controlled as required by 7.6 in the EN 9100:2009/EN 9110:2012/EN 9120:2009 quality management system standards. When reliability values are specified in the design, measurement capability shall support those reliability requirements (see Annex D, Annex E, and Annex F for guidance). Statistical methods for product acceptance may be used in conjunction with NDT, unless disallowed by product engineering. If design drawings or specifications give instructions on destructive testing, those instructions take precedence over this standard. 32BS EN 9138:2019 EN 9138:2019 (E) 48 Non-conformances (within the sample) No discretion shall be permitted for measurements or conditions outside of the tolerances. Any deviations from design tolerance shall be reported as non-conforming and the affected production represented by the sample shall be controlled per the organization's quality instructions that govern non-conformances. In the event that any characteristic is rejected by the plan when lot or batch sampling is applied, the entire lot or batch shall be rejected and dispositioned in accordance with EN 9100-series standard requirements for non-conforming material NOTE The primary options for disposition are to inspect the remaining product in the lot, scrap the product, or return the product to its source. A material review decision to accept the product as-is with no further inspection is not recommended due to the high probability that other uninspected units contain even more extreme non-conformances than the sample did (see Annex D for guidance). This policy is known as "C=0". When the sampling is continuous or skip-lot (see Clause 9), or when the sampling uses reduced sampling levels that are less than would be required under isolated lot sampling (see Clause 7), then the sampling plan shall accommodate adequate retrievability. Replacement or rectification of product is required. Production rates and frequencies of inspection shall be linked so that if non-conforming product is found, any product skipped since the previous inspection can be retrieved. Refer to the EN 9100:2009/EN 9110:2012/EN 9120:2009 quality management system standards (see 8.2.4) for instructions on monitoring and measuring of product. 4.9 Stat ‘ical product acceptance plans and records ‘The documentation of the sampling plan shall not provide for the acceptance of characteristics that are inadequately inspected in the samples. A sampling plan shall show the number of units of product from each lot or batch that are to be inspected and the criteria for determining the acceptability of the lot or batch. Both the statistical product acceptance plan and associated acceptance records shall show conformance to design requirements, account for all non-conformities in the sample, and shall be maintained in accordance with established procedures. Plans for compiling and retrieving statistical acceptance information and acceptance data shall provide for adequate retrievability and shall provide for the verification that the plan is being executed correctly (see Annex D for guidance). 4.10 Alternate statistical techniques for product acceptance Alternate statistical techniques and unique applications intended for special cases are covered in Clause 9. Techniques, other than those stated in this standard, that are developed and used shall be justified on the basis of recognized statistical principles to meet or exceed the required consumer protection quality parameter value or IRR (see 4.4 and 5.2). Reference Clause 5 requirements for documenting these alternatives in written procedures and plans (see Annex D for guidance). 4.11 Auditing of statistical techniques for product acceptance Any statistical technique for product acceptance shall be reviewed as a part of periodic internal audits conducted in accordance with the EN 9100:2009/EN 9110:2012/EN 9120:2009 quality management system standards (see 8.2.2). A technical review of the statistical application should also be performed regularly. A technical review shall be conducted by individuals who understand the design of the statistical techniques used [see Annex D (D.10) for guidance}. 33BS EN 9138:2019 EN 9138:2019 (E) 5 Required information to be documented 5.1 Documented procedure requirements The organization shall document their procedures, plans, and implementation activities of statistical product acceptance applications. The organization shall perform their planning and implementation of statistical product acceptance activities in accordance with those documented procedures and plans. At a minimum, these documents shall address the requirements of this clause. 5.2 Initial Reliabil ry Requirement or quality parameter values for product acceptance ‘The documentation governing application of statistical techniques for product acceptance shall provide instruction on how to obtain specific IRR values for specific product characteristics (see Annex A and Annex E for guidance), a) Every assigned statistical technique for product acceptance, under the governing document, shall identify its quality parameter and the specific value that is achieved, as well as the basis for selecting that value (e.g, specification /RR). Every sampling plan or procedure shall show its quality parameter{s). Those quality parameter values shall be verifiable mathematically. If the statistical validity is claimed because the plan is copied from a published standard, then the specific standard shall be identified. b) Safety or critical characteristics shall be 100 % inspected, except as directed by the design authority under the authorization of the customer. 5.3 Assigning responsibilities Procedures shall include assigned responsi (see Annex D for guidance). nistering statistical product acceptance plans 5.4 Assigning where applied ‘The organization shall define all current statistical techniques for product acceptance and the locations in the organization where they are applied (see Annex D for guidance) 5.5 Establishing a training program To assure effective implementation of statistical product acceptance applications, procedures shall require and define the training requirements of general and specific statistical requirements, methods, and intended audiences (see Annex D for guidance) 5.6. Controlling non-conformances Procedures shall require methods for handling non-conforming product to assure that they do not escape and to provide for their prevention in the future. Procedures shall require that all discovered non-conformities be reported as non-conforming, Methods used to control non-conformances shall be defined or referenced in the established procedure. 5.7. Monitoring performance and effectiveness Procedures shall define the processes used to ensure proper performance of statistical acceptance applications, including the use of any switching rules or changes to inspection frequency (see Annex D for guidance). Procedures should also define the processes used to monitor other indications of inferior quality. 34BS EN 9138:2019 EN 9138:2019 (E) 5.8 Defining adjustments to inspection frequency As applicable, procedures shall define any rationale used to adjust the level of inspection. The adjustment of inspection levels defined in the procedure shall be based on quality history. 5.9 Customer approval of procedures and plans When statistical acceptance procedures and plans are subject to approval by the customer prior to use, the organization's procedure shall document the process. 6 Requirements for accepting product by individual lots 6.1. General requirements for individual lot sampling Applicable requirements defined in Clause 4 and Clause 5 shall be followed when applying this clause. ‘This clause provides additional requirements for accepting product by individual lots. 6.2 Quality parameters for individual lot sampling Quality parameters that are based on the probability of conformance for delivered product are preferred; however, probability of acceptance parameters are acceptable if the probability of acceptance is 50 % or less. Implementing documentation shall specify which parameter is used as well as its value. The mathematics associated with the parameter shall be available either by reference or inclusion. 6.2.1. Delivered product conformance probability parameters Probability of conformance for delivered product parameters are the AOQL, OQCL, or E0Q. The AOQL may be used when provisions are made to assure that all delivered units from a failed lot (one with one or more detected non-conformances) are confirmed to be conforming. Provisions can include screening operations and non-conformance scrapping or rectification. The OQCL should be used instead of the AOQL, if the entirety of a failed lot is simply scrapped. The EOQ may be used, if the distribution of the production process that produced the lot is known. See Annex G (with emphasis on Figure G.1) for further guidance on A0QL and OQCL, and Annex H (see H.2.3) for OQ guidance. 6.2.2 Probability of acceptance parameters Acceptable parameters for lot sampling via probability of acceptance statistics are the RQL, the LQ, and the indifference point or ERP. These correspond respectively to non-conformance rates with probability of acceptance values on the probability of acceptance OC curve of 5 % or less for the RQL, 10.% or less for the LQ, and 50 % or less for the ERP. See Annex G (G.1.2) for further guidance. 6.3 Sample sizes for individual lots Tables of sample sizes for many individual sampling plans are found in the SCMH (http: //www.sae.org/scmh/). 7 Requirements for accepting product under switching rules 7.1 General requirements for switching rules Applicable requirements defined in Clause 4 and Clause 5 shall be followed when applying this clause. This clause provides additional requirements for accepting product produced under switching rules. 35BS EN 9138:2019 EN 9138:2019 (E) 7.2, Background information for switching rules Switching rules are rules that change the sample size in response to the history of lot rejections and acceptances. Many applications have used three approaches for lot sampling: —_ANSI/ASQZ1.4, which is equivalent to ISO 2859-1, JISZ9015-0 and JISZ9015-1, and GB/T 2828.1; — ANSI/ASQ2Z1.9, which is equivalent to 180 3951-1, GB/T 6378.1, or J1SZ9003 with JISZ9004 and J18Z9010 — “Zero Acceptance Number Sampling Plans” (by N. L. Squeglia). ‘These have been used with switching rules and have a common characteristic of being organized around the AQL parameter. This clause shows how to provide consumer protection under these conditions, which are challenging because of the AQL parameter being a measure of producer's risk. If the design engineering does not give an QL, then the above documents and their switching rules do not apply; use other sampling strategies. NOTE1 If the user intends not to use switching rules, the requirements of Clause 6 apply (see Annex for guidance). NOTE2 Where alternative statistical techniques are required for acceptance of product in attribute or variable lot acceptance sampling plans, other than those contained herein, those techniques are subject to the requirements of Clause 9. 7.3, Quality parameter with switching rules Each sampling plan shall identify a parameter used to control consumer risk and identify the value of that parameter. If the AOQL parameter alone is applied, then the plan switching rules in the applicable standard shall be used. The key requirement for any sampling plan is the level of consumer protection provided; that is the level of assurance that an individual characteristic on outgoing product meets design requirements (see Annex G for guidance). 7.4 Requirements for accepting product with switching rules 7.4.1 Sampling documentation Sampling plans based on other procedures shall explicitly identify the documents on which they are based. 7.4.2. General conditions To achieve a required minimum probability of conformance (Le., an IRR) the use of the tables in the subject standards referenced in the beginning of this clause requires one of two approaches: a) different AQL columns in the standard can be used for different lot sizes in order to achieve the matching A0QL or ERP value; or b) an AQL column can be selected in which every AOQL or ERP entry satisfies the minimum probability of conformance. See Annex I (1.2) for examples of these approaches. 36BS EN 9138:2019 EN 9138:2019 (E) 7.4.3 Sampling plan criteria The following conditions also apply to all subject lot acceptance sampling plans with switching rules. ‘The subject lot sampling plans allow a lot or batch of product to be accepted based on the inspection of a portion of that lot or batch. To maintain confidence in long-term quality of the product accepted by this inspection, the following criteria shall be included in the organization's sampling plan: a) The lot or batch shall be clearly identified and segregated throughout the process of sampling inspection. b) Each unit in the lot or batch shall be manufactured under the same design requirements and ‘manufacturing conditions, ©) The sample size shall be selected with respect to the size of the lot or batch, due to these conventional lot sampling plans providing different amounts of customer protection for different lot sizes. The sample size in these sampling tables is determined by the use of “special” or “general” inspection levels; one of which shall be selected. The default level shall be General Inspection Level II for ANSI/ASQ Z1.4, ISO 2859-1, or equivalent documents. See Annex G for guidance. 7.44 Qualification criteria for sampling Prior to the use of sampling plans for inspection, the organization shall ensure that the processes demonstrate the ability to produce conforming product at the /RR level of quality or better. 7.5 Requirements for accepting product by variables sampling plans using switching rules 7.5.1 Variable sampling application limits For the purposes of this standard, the requirements of variable sampling plans are limited to the application of ANSI/ASQZ1.9, ISO 3951-1, or equivalent (see Annex G for guidance). If other variables sampling plans are used, see 9.4. 7.5.2. Evaluating lot quality Inspection by variables requires that a quality characteristic of each unit of product in a sample be measured. The accept/reject decision for the lot is usually based upon the sample mean and spread of the sample measurements (eg, standard deviation, subgroup ranges). The accept/reject decision shall be made after the sample statistics have been used to obtain an estimate of the lot percent non- conforming, 7.5.3 Related requirements Users of variables sampling plans, in accordance with this standard, shall follow the requirements of ANSI/ASQZ1.9, ISO 3951-1, or equivalent documents. These documents model the underlying distribution of individual measurements with a normal distribution. Therefore, an evaluation shall be performed to verify how the data is distributed [see Annex1(I.3) for guidance]. Non-normal distributions shall be handled with a statistically valid method. 37BS EN 9138:2019 EN 9138:2019 (E) 8 Requirements for accepting product produced under process controls This clause provides additional requirements for accepting products produced under process controls, covering the following: — SPC (see 8.2); — Variation restrictions (see 8.3); and — Process parameter controls (see 8.4). When processes are consistent enough to make the resulting product predictably conforming, then the tools presented in this clause are appropriate. 8.1. General process control requirements ‘The content of 8.1.1 to 8.1.13 shall be in place at the organization's process site for all process control methods. These requirements are in addition to those defined in Clause 4 and Clause 5. All products shall be verified unless the total relative frequency of non-conformances is less than 1-IRR. 8.1.1 Data The requirement is to show linkage between measured variables and outcomes. The C = 0 policy limits the use of SPC for product acceptance to data of a quantitative nature. This data shall be sufficiently correlated with the acceptance criteria. If attribute data is allowed, the data shall show suffici to predict the acceptability of affected characteristics (see Annex I for guidance). 8.1.2 Reducing inspection with process controls All products shall be verified unless the total frequency of special cause and common cause non- conformances is less than the complement of the minimum probability of conformance (1-1RR), 8.1.3. Measurement systems See requirements defined in 4.7. 8.1.4 Training program Personnel training shall be required as defined in 4.2. Additional training for process control techniques shall assure the operator understands: a) The use and interpretation of all relevant process control tools. — For SPC, this includes the use and interpretation of the control charts. — For mistake-proofing, this includes operator knowledge of all process elements that prevent frequent or occasional non-conformances. — For process parameter controls, this includes the necessary elements in controlling the variation of the process parameters. b)_ The appropriate and timely corrective action required when process control is violated. 38BS EN 9138:2019 EN 9138:2019 (E) 8.1.5 Process control points Process control points shall be adequately maintained to ensure the characteristics conform to the specification from the last inspection point until the point of shipping to the customer (i.e, the integrity ofthe characteristic is not violated). 8.1.6 Process control acceptance plans and records Requirements for statistical product acceptance plans and records shall apply (see 4.9); see Annex | for guidance. 8.1.7. Unbiased sampling See requirements defined in 45. 8.1.8 Time or production order sequence Any samples used for process control purposes shall be recorded and maintained in time or production sequence (see Annex J for guidance). 8.1.9 Data retention plan See requirements defined in 4.9. 8.1.10 Periodic internal audits See requirements defined in 4.11. Internal audits should not be replaced by dependence on any external auditing function. 8.1.11 Retrievability system ‘The organization shall have the ability to retrieve all products, including any untested or unmeasured products that are deemed suspect; traceability of the products shall be tied to the process control system. Points that are unstable shall be included in the quality record associated with products manufactured while the process was unstable. 8.1.12 Limiting errors in measurement and recording See EN9138, 4.7 and 7.6 of the EN 9100:2009/EN 9110:2012/EN 9120:2009 quality management system standards for measurement process requirements and expectations. Refer to Annex E (see E.1) and Annex J (see J.1.1) for guidance. 8.1.13 Evaluating effects of corrective actions Corrective action shall be evaluated to determine the effect on the characteristics being accepted by the process control plan. A process control plan shall avoid tampering or changing a process without statistically valid indications of a process shift. 8.2 Statistical Process Control ‘The following paragraphs provide additional requirements and guidance necessary for accepting product. 39BS EN 9138:2019 EN 9138:2019 (E) 8.2.1 Methods Where SPC tools are used to meet these requirements, the methods in one of the following groups of documents shall be applied: — ANSI/ASQC B1, ANSI/ASQC B2, and ANSI/ASQC B3; — 180 11462-1 and ISO 11462-2; or — J1829020-1, J1S29021, J1SZ9041-1, ISZ9041-2, JISZ29041-3, JISZ9041-4, and JISZ9041-5, 8.2.2 Location and dispersion SPC charts shall track changes in both the process mean and the process variation. 8.2.3 Basis for control limits SPC charts shall have established statistical control limits based on 1/(1-IRR) subgroups of data, but not less than 20. Appropriate control limits and rules (see Figure 3) shall be in place to separate common cause from. special cause variation. Such limits shall be based on data from the process and not influenced by product requirements (eg,, engineering tolerances). See Annex J for guidance. Upper Control Limit Warning Limit A Warning Limit B Centreline Warning Limit B Warning Limit A Lower Control Limit Figure 3 — Typical Statistical Process Control chart depicting control limits, warning limits, and zones 8.2.4 Basis for stability Control limits or other process stabi one of the following conditions: criteria shall be based on process measurements taken under a) When the process is operating under stable conditions, including sources of variation inherent in routine production. b) When the process has occasional special causes, the variation uncharacteristic of routine production (e.g,, process set-up, outliers) is removed from the calculations, ©) For cases where the process capability exceeds product requirements, control limits may be set wider than three-sigma; however, when using SPC to accept product it shall be demonstrated that controlling to these wider limits does not violate the minimum probability of conformance. If either the process mean or the process variation appears to have changed, then see the following instructions in 8.2.5 through 8.2.7 before implementing 8.2.8 and 8.2.9. 40BS EN 9138:2019 EN 9138:2019 (E) 8.2.5 Investigation ‘When one point falls beyond Zone A (see Figure 3), special causes of variation shall be investigated. The out-of-control point shall be annotated on the control chart and any corrective action taken recorded. If after thorough investigation of the process no special cause of variation can be found, the process may be treated as having had a false alarm. Out-of-control points shall not be designated as false alarms more frequently than can be explained through random chance. See Annex J (J.2.2) for guidance on the level of investigations. Annex C (see Table C.2) provides minimum acceptable intervals between out-of- control points, 8.2.6 Viola mn of stability If the process provides evidence of violating statistical stability, product acceptance by process control methods shall be suspended and an alternative method of product verification shall be employed until stability is revalidated. 8.2.7. Re-evaluation of stability and capability Stability measures (e.g, control limits, limits on first/last piece) shall be re-evaluated whenever identified changes are made to the process. When such limits are modified, the associated capability measure shall also be recalculated. 8.2.8 Capability and process control 8.2.8.1 Valid process capability measure ‘A measure of process capability shall only be valid for data taken when the process shows adequate statistical control. 8.2.8.2 Adequate capability If not otherwise specified by engineering drawings, specifications, or other design documentation, the data shall provide 90% confidence that the process capability is within the engineering tolerance (see Table 2). If only one specification limit exists, the data shall provide 90.% confidence that the process capability is on the acceptable side of that limit. Statistically valid capability measures (8, Cpe Pp) shall be used. Capability calculations shall include all data used in demonstrating statistical stability (see 8.2.1 and Annex J) Table 2— True Cy, requirements Number of Measurements Taken ‘Sample C,, Required for 90% Confidence that | 1,07| 1,08 | 1,09 | 1,10 | 111 | 112 | 1,13 | 114 | 1,15 | 1,17 True C,y.2 1.00 250 | 200 | 150 | 125 | 100 | 90 | go | 70 | 60 | 50 Number of Measurements Taken | 46 | 42 | 38 | 34 | 30 | 28 | 26 | 24 | 22 | 20 (cont.) ‘Sample C,, Required for 90% Confidence that | 1,18 | 1,19 | 1,20 | 1,21 | 1,23 | 1,24 | 1,25 | 1,26 | 1,28 | 1,30 True Cyj,2 1.00 (cont) 41BS EN 9138:2019 EN 9138:2019 (E) 8.2.8.3 Inadequate capability Ifthe characteristic is not capable, the organization shall identify and control sources of variation in the processes that are correlated with the characteristic and take corrective action. These investigation findings and corrective actions shall be documented. If the process does not meet minimum capability requirements, an approved acceptance sampling plan or 100 % inspection shall be enacted until the cause has been identified and capability revalidated. 8.2.8.4 Confidence limits on capability ‘The values in the body of Table 2 are the calculated C,,, values required to be 90 % confident that the actual Cy is greater than or equal to 1,00 given the sample size at the top of the respective column. The values listed in the column titled "Number of Measurements Taken” are the actual number of measured characteristic values, not necessarily the number of plot points. Table calculations assume ‘measurements are normally distributed (see the SCMH for further guidance on use of this table). Evaluations to show how well the data fits the normal distribution model can use either graphical techniques or a goodness of fit test (e.g, Anderson-Darling). Non-normal data should be appropriately sorted or transformed, prior to using standard control charts. 8.2.9 Process monitoring frequency On ceasing 100 % inspec mn and switching to sampling inspection, the follo\ ig rules shalll apply: a) Select a process monitoring frequency that enables at least one observation to be taken before affected products leave the immediate process area. The "immediate process area” shall be restricted to the property of the subject organization. b)_ The process monitoring frequency shall be able to detect any process shifts or trends in a manner timely enough to recover all untested or unmeasured product produced since the previous in- control point. This frequency shall be documented. If there are rules for changing the frequency, those shall also be documented. 8.3. Variation restrictions Tool stops or other variation restricting mechanisms, control of process settings [eg, bought off Computer Numerically Controlled (CNC) programs], standard processes, and/or mistake-proofing devices may be used to accept product if they satisfy a PFMEA as being adequate and consistent with the applicable IRR (see Annex J for guidance). 8.4 Process parameter controls Process parameter controls may be used to accept product, if they are adequate and consistent with the applicable IRR (see Annex J for guidance). If controls, parameters, or characteristics other than those required in the end inspection are used as a basis for determining conformance, adequate relationships between the controls and requirements shall be established. An adequate relationship shall show that the probability is high (relative to the quality parameter) that the inspected feature will predict when the uninspected feature is non-conforming, Records of these correlation studies shall be maintained for as long as they are used in the acceptance of product. 42BS EN 9138:2019 EN 9138:2019 (E) 9 Additional requirements for accepting products based on continuous sampling, skip-lot sampling, or methods for special cases 9.1. Continuous sampling Continuous sampling can apply whenever product is being accepted one unit at a time. All continuous sampling plans shall include one or more 100% inspection phases, until some sequence of units is inspected without any non-conformances being found. When not conducting 100% inspection, a fraction of the units are sampled and inspected. There are both single level continuous sampling plans and multi-level continuous sampling plans, the latter allowing some decreases in the fraction inspected when a process proves sufficiently high quality. Whenever a sampled unit is found to be non-conforming, the specified sampling plan requirements shall be followed, usually increasing inspection (see Annex K for guidance). 9.1.1 Quality parameters for continuous sampling When writing an acceptance plan that uses continuous sampling, the AOQL parameter shall be shown in. the documentation. The A0QL shall be less than or equal to 1-/RR, if an IRR has been specified. If the quality requirement is stated as an AQL, Table 1 provides an acceptable AOQL replacement. 9.1.2. Inspection sequence Under continuous sampling inspection, units shall be submitted consecutively for inspection. 9.1.3 General conditions When a non-conforming unit is found during a fractional inspection phase, the sampling plan must control suspect product. Any skipped units shall be retrievable at the point in time when the results of the next inspected unit become known. All sample units shall be selected using a verifiable randomization method, or if randomization is not possible, an adequate PFMEA should be used following clause 8 instructions. Annex K (see K.3) offers further guidance on implementation. 9.1.4 Supply Chain Management Handbook continuous sampling tables The tables and instructions for the five continuous sampling strategies of MIL-STD-1235C are included in the SCMH (see chapter on Statistical Product Acceptance). Those five strategies are labelled CSP-1, CSP-F, CSP-T, CSP-2, and CSP-V. The SCMH gives recommendations on the optimal uses of these strategies. 43BS EN 9138:2019 EN 9138:2019 (E) Jot sampling Skip-lot sampling uses continuous sampling procedures to decide which lots to inspect out of a process and uses lot sampling procedures to decide which units in those lots to inspect. 9.2.1 General conditions for skip-lot sampling Skip-lot sampling shall be permitted provided: a) there is objective evidence that the minimum quality levels defined by this document have been obtained; b) the required objective evidence shall be re-established, when a lot is rejected and there is no documented evidence of a special cause with corrective action; ©) _ any skipped lots shall be retrievable at the point in time when the results of the next inspected lot become known. See Annex H (H.1.1), ANSI/ASQC S1, and ISO 2859-3 for guidance. 9.2.2 Quality protection equivalency Each skip-lot sampling plan shall show its AQQL quality parameter value. If the quality requirement is stated as an AQL, Table 1 provides the equivalent AOQL requirement. The AOQL shall be less than or equal to 1-IRR, if an IRR has been specified, 9.2.3 Considerations for subsequent lots As with other continuous sampling plans, if quality results dictate an increase in inspection, then the units on which that inspection would be performed shall be inspectable. For that reason, if there is a time delay for an inspection process, all subsequent lots presented for inspection (during that time) shall be held where they are available for the increased inspection, if that becomes necessary. NOTE This stipulation can place additional storage and inventory information requirements on the skip-lot inspection process. 9.3. Sampling strategies for special cases ‘The sampling strategies in this clause are provided as valuable alternatives for efficiently meeting technical requirements (see Annex H for guidance). 9.3.1 Special sampling customer approvals The customer can require notification or approval, prior to use of these special sampling techniques (see 5.9). In each case, if the organization chooses to use sampling strategies for special cases, the organization shall demonstrate the statistical validity of these approaches before they are used. 9.3 addresses what is, needed to establish that statistical valBS EN 9138:2019 EN 9138:2019 (E) 9.3.3 Predicting quality from correlated variables The quality of a product feature may be accepted on the basis of other measurements, if these measurements are shown to be predictive. Under these conditions, there are two sets of characteristics, and both shall be measurable or countable with large enough counts to be treated as measurable (eg, particle counts). a) One of the sets of characteristics is directly subject to engineering requirements and shall be accepted as conforming before the product having those characteristics is allowed to be used in aviation, space, and defence deliverables. This is the “dependent” set of characteristics. b) The other set (ie, the “independent” characteristics) might or might not be directly subject to engineering requirements, but is related so closely to the first set that their values may be used to predict the acceptability of the first set of characteristics. The conditional probability of the “dependent” characteristics being acceptable, given the measure or count of the “independent” characteristics, shall satisfy the required quality parameter. ©) Studies showing this conditional probability shall be sufficient to understand the ongoing relationship of such parameters and their resulting influence on the deliverable property or characteristic. An adequate relationship shall show that the probability is high (relative to the quality parameter) the inspected feature will predict when the uninspected feature is non- conforming, Records of these correlation studies or other evidence of validity shall be maintained for as long as this method is used in the acceptance of product. See Annex H (H.2.1) for an example of this concept. 9.3.4 Continuous manufacturing process - First and last unit for a lot produced Statistical evidence for using auto-regressive or auto-correlated statistical analysis, if used, shall compare the prediction interval for the last measurement in a lot to specifications, when the first measurement is given. The time-series model used for analysis shall be documented together with parameter estimates and prediction intervals. See Annex H (H.2.2) for guidance. a) Any “first and last” model application shall require showing the exact sequence of unit quality in lots that are of mixed quality. Attribute data from lots that are 100% conforming or 100% non-conforming does not demonstrate that the first and last units in the lot are more likely to be non-conforming than the other units. b) When doing this on an attribute basis, the number of mixed lots in which the sequence of conforming and non-conforming units is shown depends on both the quality requirement and the degree to which the actual data matches the model. The model shall show evidence that among lots in which the fraction non-conforming equals 1-IRR, the fraction of those lots which would pass this sample (i.e, both the first and last units are conforming) shall be less than or equal to 50 %. 9.3.5 Sequential sampling Sequential sampling is used for destructive or costly tests, with a subgroup size of one thereby making it an item-by-item plan. Item-by-item sequential sampling is based on the concept of the Sequential Probability Ratio Test (SPR); for further definition see "Sequential Analysis” by Abraham Wald. 9.4 Other methods for special cases ‘When a sampling strategy for a special case is developed that is not described in this document, it shall satisfy the requirements of Clause 4 and Clause 5. Its mathematical basis shall be documented and available for review (see Annex H for examples and guidance). New methods may be added periodically to the SCMH when they are justified on the basis of recognized statistical principles, 45BS EN 9138:2019 EN 9138:2019 (E) Annex A (informative) Guidelines for assigning Initial Reliability Requirement quality levels A.1 Introduction ‘The values in this Annex for protection levels common to aviation, space, and defence applications are recommended as defaults to be used when the type design drawings and specifications do not contain sampling instructions or quality parameter values. The authorship of this standard believes that following these recommendations to the extent possible will enhance the success of statistical applications that affect, improve, and maintain product quality. A.2_ Assigning quality levels Some representative examples of how to correctly assign quality levels are recommended to be included in the statistical product acceptance procedure(s). Showing examples in the organization's procedure(s) of how to correctly assign quality levels can be used to remove any ambiguity in the assignment of quality levels. This can take the form of classification of characteristics, quality parameters, lists of inspection characteristics, checklists, etc. Showing how quality parameters are determined is also suggested for inclusion in the procedure(s) or instruction(s). Multiple quality parameters are usually necessary in order to cover the various products, features, and processes at the organization's facility. Characteristics and parts can be grouped in the documentation, if they are inspected to the tightest requirement for the group. A.3 Initial Reliability Requirements Table A.1 contains IRR values often used by the engineering and reliability disciplines for various product and process applications in aviation, space, and defence manufacturing; this Annex recommends some minimum values for selected applications. Engineers, quality departments, or customers can specify values that are different than these values. Furthermore, any values given in type design drawings or specifications take precedence over these default recommendations. Other forms and formats to display this or any other assigned quality parameters in the organization's documented procedure(s) can be used. NOTE See Annex E for an editorial discussion on the relationship between quality parameters and reliability. The examples in Table A.1 do not include all possible applications and cannot authorize bypassing any engineering requirement. 46BS EN 9138:2019 EN 9138:2019 (E) Table A.1 — Minimum probability conformance values Part, Process, Characteristic, or Mechanical Recommended Minimum. Requirements Probability of Conformance Parts: Machined parts (general features, except as noted under 97% mechanical characteristics below) Machined parts; special applications (e.g, key 99% characteristics, location holes) Conductivity 99% Hardness 99% Composite part - structural 98,7 % Composite part - nonstructural 95% Cable form 98% Connector assemblies 99% Printed circuit boards 97% Conformance of printed circuit board configuration 99% Processes: Sheet metal fabrication 95% Receiving inspection of standards 97% Non-Destructive Testing (NDT), where permitted 98,7 % Paint, alodine, anodize, and plating 92% Electrical wire wrap 98% Cleanliness test for printed circuit boards 97% Mechanical Characteristics: Fastener hole diameter tolerances 0,025 mm/0.001 inch or 99% less Fastener hole diameters tolerances 95% 0,025 mm/0.001 inch, location on non-locating holes of > £0,25 mm/0.01 inch linear or equivalent diametral tolerance ‘Thickness with tolerance of +0,127 mm/0.005 inch or less 95% ‘Thickness with tolerance of greater than 92% £0,127 mm/0.005 inch Linear dimension tolerance < +0,25 mm/0.01 inch 95% Linear dimension tolerance greater than 92% £0,25 mm/0.01 inch NOTE The 98,7 % values are derived from specific usage of the 1,22 % AOQL in Table 1, which is used in continuous sampling tables. a7BS EN 9138:2019 EN 9138:2019 (E) Table A.1 — Minimum probability conformance values Part, Process, Characteristic, or Mechanical Recommended Minimum Requirements Probability of Conformance Radi total tolerance range less than 0,51 mm/0.02 inch 97% Radi total tolerance range greater than 0,51 mm/0.02 inch 92% Angle total tolerance range less than or equal to +0,5° 97% Angle total tolerance range greater than +0,5° 92% Concentricity, non-controlled contours, diameter, or fillet 92% radius tolerances greater than +0,25 mm/0.01 inch; flatness, straightness, parallelism, perpendicularity, and profile Controlled contour [e.g, Master Dimension Definition (MDD), 95% Master Dimension Surface (MDS)], diameter or fillet radius tolerances less than or equal to -+0,25 mm/0.01 inch Surface finish roughness (Ra) less than 0,8 micron or 32 yin 99% Surface finish roughness (Ra) greater than or equal to 95% 0,8 micron or 32 nin Flatness (carbon face seals) using optical flats 99% Threaded parts: Internal profile pre-load locking 99% Straight screw threads - total tolerance on pitch diameter 99% Standard electrical connectors [eg, US Military Standards 97% (Ms) part numbers) All other thread applications /characteristics 99% Gears and splines (all characteristics) 99% Dimensions with only maximum or minimum values which 99% directly control material thickness of weldments, castings, or forgings; and at least one surface is machined Absence of magnetism (ferrous parts) 99% Spring rate, when specified by Engineering 99% 48AAQG ABNT AC AFL AIAG ANOVA A0Q AOQL AQL ARL. ARP. ASD-STAN ASQ ATI CEN cL CNC E0Q ERP FAA FMEA Gage R&R cuM BS EN 9138:2019 EN 9138:2019 (E) Annex B (informative) Acronym log ‘Americas Aerospace Quality Group Brazilian Association for Technical Norms Advisory Circular Average Fraction Inspected Automotive Industry Action Group Analysis of Variance Average Outgoing Quality Average Outgoing Quality Limit Acceptable Quality Level Average Run Length Aerospace Recommended Practice Association ~ Standardization Aerospace and Defence Indust American Society for Quality Average Total Inspected European Committee for Standardization Confidence Level Computer Numerically Controlled Expected Outgoing Quality Equal Risk Point Federal Aviation Administration Failure Modes and Effects Analysis Gage Repeatability and Reproducibility Guide to the Expression of Uncertainty in Measurement (reference ISO/IEC Guide 98-3) 49BS EN 9138:2019 EN 9138:2019 (E) 1AQG IRR Iso jis JIT JSA LQ LQL LP LsL. LTpD MDD MDs Ms MSA MTTF NDI NDT oc 0g ogc. PAH PEMEA POD Pog PoQL PPM 50 International Aerospace Quality Group Initial Reliability Requirement International Organization for Standardization Japanese Industrial Standard Just-in-Time Japanese Standards Association Limiting Quality Limiting Quality Level Lot Quality Protection Lower Specification Limit Lot Tolerance Percent Defective Master Dimension Definition Master Dimension Surface Military Standards Measurement System Analysis Mean Time to Failure Non-Destructive Inspection Non-Destructive Testing Operating Characteristic Outgoing Quality Outgoing Quality Confidence Limit Production Approval Holder Process Failure Modes and Effects Analysis Probability of Detection Probability Outgoing Quality Probability Outgoing Quality Limit Parts Per MillionRQL SAC SCMH SJAC SPC SPRT uQL usL BS EN 9138:2019 EN 9138:2019 (E) Rejectable Quality Level Standardization Administration of the People’s Republic of China Supply Chain Management Handbook Society of Japanese Aerospace Companies Statistical Process Control Sequential Probability Ratio Test Unacceptable Quality Level Upper Specificat ion Limit 51BS EN 9138:2019 EN 9138:2019 (E) Annex C (informative) Mathematics for accepting product under lot-by-lot inspection This Annex provides mathematics relative to sampling individual lots. In particular, statistics relating to the probability of conformance for delivered product is emphasized. Mathematics relative to traditional isolated lot sampling is also included; variables referenced in the mathematical formulas are listed in Table C.. Table C.1 — Table of mathematical variables Variable Average Outgoing Quality- The average quality of outgoing product after sampling inspection for a given steady value of incoming product quality AOQL | Average Outgoing Quality Limit - The maximum average percent non-conforming after inspection operations C__ | Maximum number of non-conforming units in a sample that allows acceptance of the lot CL__| Confidence Level Number of non-conforming units detected by assessment of n units D__| Actual number of non-conforming units in a lot of size N | Number of non-conforming units delivered to a customer p | Actual rate or probability of non-conformance in the process producing the lot(s) p _| Estimate of non-conformance rate or probability - pronounced ‘p-hat™ Pa __| Probability of accepting a sample and its lot 0QcL | Outgoing Quality Confidence Limit- The maximum confidence limit on the percent of non-conforming delivered units at the defined CL q__ | Actual yield rate @ _| Estimate of a yield - pronounced ‘q-hat’ in | Number of assessed units N | Number of units in a lot or batch, before inspection N, Number of units in a lot or batch delivered to customer C1 Yield estimation A yield is the probability that a produced unit will be conforming. The yield can be estimated from data and a confidence limit on the estimate can be computed. The yield estimation statistics are discussed in C.1.1 and the confidence limit assessments are discussed in C.1.2. 52BS EN 9138:2019 EN 9138:2019 (E) C11 Yield estimates The yield (q) and the non-conformance rate (p) are the probabilities of an individual part or unit passing or failing, The two probabilities are mutually exclusive and have a sum equal to one. Because of the summation equality, the non-conforming rate can be estimated from a yield or the yield can be estimated from a non-conformance rate. AE (cy p=1-q (c2) q=1-p (c.3) Estimating a yield or a conformance rate uses statistical methods for estimating a probability or proportion. The data for estimating the statistic of yield or conformance rate are accumulated from assessments. The estimate of the yieldq is the ratio of the number of units confirmed as being conforming divided by the number of units actually assessed. confirmed conforming units _n—d units assessed n (c4) detected nonconforming units _d units assessed n 6 (cs) Likewise, the non-conformance rate p is estimated by the ratio of the number of detected non- conforming units divided by the number of units assessed. ‘The assessed units are dispositioned as either conforming or non-conforming (Le, pass or fal). units assessed = confirmed conforming units+detected nonconforming units (c6) 1.2 Confidence limits on yield estimates Estimates of yields and non-conformance rates usually miss the true yield or rate because of sampling error. A lower confidence limit on the yield gives a minimum value that is below the true yield a large percentage of the time. The percentage of the time that this confidence limit is below the true yield is its CL. An upper confidence limit on the non-conformance rate gives an indication of how large the sampling error might be by providing a maximum value that is above the true non-conformance rate a large percentage of the time. The percentage of the time that the upper confidence limit is above the true rate is its CL. C.1.2.1 Mathematics of yield confidence limits The probability of non-conformances given a non-conformance rate is dependent upon the number of units examined (n) and the binomial distribution. The binomial distribution for the probability of a number of non-conforming units (d) for any given yield (q) and sample size n is shown in Formula (C.7). nt Pr(diq.n) = Grace} xqh*x (1-q)4,d = 012.0 (c7) 53BS EN 9138:2019 EN 9138:2019 (E) Given a yield q and the number of units assessed, the observed non-conformances are distributed as the binomial distribution, nt -(—— _) g-tx aq) a (axaca*4 x(1-@)4,d=0,12. (ca) Equivalently, the binomial can use the non-conformance rate p as its basis. .)x pt x (1—p)y"-4,d =01,2...0 (c9) ‘The upper confidence limit is the solution of Formula (C.10). The true probability pg, is such that the sum of the individual probabilities equals 1-CL. Each individual probability in the sum depends on the true probability of non-conformance pc, and the sample size. The quality levels are adjusted until the sum equals 1-Cb, 4 (@exlnd,ct)] 3 )"[(") x pls. x A= pew") = 1- ch (c10) 1.2.2 Yield confidence limit from zero observed non-conformances If the number of observed non-conforming units d is zero, the solution for the upper confidence limit is computed by Formula (C.11). VI-CLita Pew (c.11) How much data is needed for proof of yield can be determined by algebraically manipulating Formula (C.11) to find the sample size n as a function of the rate of non-conforming units pg, and the CL, if no non-conformities are detected. et — CL) ,0 ay (c.17) The final term can be simplified further substituted in for C, because one of the combination factors goes to one. Formula (C.18) summarizes the resultant probability of acceptance. 1 , D=0 (N=D)! x (N=n)! Payox(onnc: WoDoml ent OSDSN-2 (c.18) 0 , D>N=n 2.2 Legacy Pa lot acceptance mathematics Some sampling plans are based upon the probability of acceptance. In those cases, the sample size is selected such that the probability of acceptance for an acceptable number of non-conforming units is less than 1-CL for an intolerable number of non-conforming units. Acceptance of any given lot requires that all inspected units in the sample conform to the specified acceptance requirement. min[n] 3 Pa[Dmax] $ (1 — CL) (c.19) The number of non-conforming units is estimated by the number of units in the lot times the complement of the IRR. Dax = (N)(1 — IRR) (C.20) Non-conforming units can only come in whole numbers, so the maximum number of non-conformances will be a number between whole numbers for most sample sizes. The algorithm below interpolates between the whole numbers when Dinax is not whole. The lower bounding whole number of non- conforming units is Dp.ay Founded down and the upper bounding whole number is Djyax Founded up. D, Wax (c.21) Dy = [Dmaxl (c.22) ‘The probabilities of acceptance associated with the bounding number of non-conforming units are given by Formula (C.23) and Formula (C.24). 1 (N= Dy)! x =n)! (N= D, =n)! x NI 0 1 D>N- 5 D,=0 » ON=n ‘The algorithm is then expressed by Formula (C25). Pa, < (1-CL) vifD; = Dz : “> Dryax ~ Dy min[n; IRRN.CL] >} yy C2) (Paz—Pay) < 1-CL ,ifD, #D, (c25) ‘The CL is replaced by .50 for ERP calculations, .95 for RQL calculations, and .90 for LQ calculations. C.2.3 Yield rates, discrete units, and conformance probabilities In practice, the number D of non-conforming units in a lot is not known, at least not before inspecting the entire lot. Ifthe lot was produced by a process with a constant yield q, the probability of the number of non-conforming units in the lot, from zero to the lot size N, is described by the binomial distribution. N) Pr(Dla.N) = (;,)-a)?a'-?,0 sD @((f)a-are) pealches SY w-py@w-nim Pajor(q.N.n,c=0) = » (qian) a ore | (c.29) & Pawigvncen= 0) >. |(q@ Aa) a -ora-r-| (€30) & 57BS EN 9138:2019 EN 9138:2019 (E) Paior(qninc=0) = 4" (c31) Formula (C.31) is the probability that a process will produce all conforming units for the sample of size n. This is true for any given lot, regardless of whether it is "isolated" or from a series of lots. 2.4 Impact lot sampling methods on delivered yield Delivered yield is the probability of conformance for delivered product. Delivered yields are the result of the process yields and any corrective actions performed on the product before it is delivered to the customer. The emphasis on the conformance for delivered product is to prevent escapements from a supplier to a customer. Escapements are non-conformities that were produced, not detected, not remedied, and subsequently sent to the customer. ‘Two policies of sampling plans that affect the probability of conformance for delivered products are: a) whether lots are "screened" when the lot's sample fails because it has one or more non-conforming units and b) whether detected non-conforming units are rectified or scrapped. Screening is the assessment of 100 % of the lot that is represented by the sample. The screening of lots is contraindicated when inspection tests destroy the unit. Rectifying operations can either replace a non-conforming unit with a known good one or repair it. Escapement rates are called “Outgoing Quality” (0Q). The expected number of escapements divided by the expected number of delivered units defines the A0Q. Ele] 400 = Fi] (32) ‘The number of units delivered (Noy) depends upon the original number of units N, the incoming yield, and whether a sampling plan rectifies or simply removes detected non-conformities. The number of escapements e is influenced by the sampling policy of whether or not to inspect the remainder of lots, after one or more non-conforming units were detected within a sample from a lot. ‘The expected number of escapements and the expected number of delivered units are computed by the aw of total probability, conditioned on the probability (Pr) that an inspected lot of units passes or fails. Ee] = Efe [Pass] x Pr[Pass] + Efe [Fail] x Pr[Fail] (C33) E[Nout] = EfNout |Pass] x Pr[Pass] + E[Noue [Fail] x Pr{Fail] (C34) ‘The probability of passing or failing a lot has been discussed. The number of escapements and the number of delivered units are affected by the probability of acceptance and the pending actions that result after a lot or batch is rejected. Mathematical expressions for the conditional expectations are given in Table C3 for the five viable combinations of the three policy questions listed, 58BS EN 9138:2019 EN 9138:2019 (E) Table C.3 — Expected escapements and delivered units Screen or Scrap Suspect 7 alts te Vanier Screen Screen Scrap Scrap Scrap Rectify or Scrap All Detected Non- Rectify Scrap Rectify Scrap Scrap conforming Units? Deliver or Scrap Conforming Units in Deliver Deliver Deliver Deliver Scrap Sample? Ele|Pass} @MW=) | OW=) | MW=-r) | OW= | &)W=) Ele|Fail] 0 0 0 0 0 E[Noutl Pass] N N N N N E[NoutlPail] N (Gp) n (Gp) 0 ‘The policy of scrapping failed lots precludes delivery and rectification options, and makes screening ineffective (see C.3.6 for similar results involving destructive tests). When lots pass because there were no non-conformities in the sample, the number of delivered units (Nous) i8 equal to the original number of units (N) and whatever non-conformities that are in the portion not sampled (N-n) become escapements (e). When lots fail because one or more non- conformities were in the sample, the number of delivered units is reduced to the number of confirmed conforming units according to the policy selections and the number of escapements is zero. Table C4 presents the A0Q formulas for the five sampling policies that result from substituting into Formula (C.32) the Table C.3 expectations together with the following expressions for the probability of passing or failing a lot. Pr[Pass] = (1 q)",ifC (c35) Pr(Fail] = 1-(1-q)",if =0 (c:36) ‘The A0Q is dependent on the incoming yield and the fraction inspected. For sampling plans where there is a policy to form a sub-lot of only good parts when the original lot fails, (ie,, Method 1 to Method 4), increases in the rate of incoming non-conformities are compensated by lowering the probability of acceptance. That, in turn, reduces the frequency that uninspected and potentially non-conforming units are delivered. Thus the sampling plan has a mechanism for limiting the A0Q. For sampling plans with the policy of scrapping failed lots (i.e., Method 5), the AQ increases linearly with the rate of non-conformities for any combination of lot size and sample size. 59BS EN 9138:2019 EN 9138:2019 (E) Table C.4 — Average Outgoing Quality formulas for lot sampling categories Sereenor | Rectifyor | pativer or Scrap Scrap eo a i Method | Suspect | All Detected | conforming erage Outgoing Units in Non- aa Quality Formula nits in Failed | conforming | Snitsin Lot? Units? “ 1 Screen Rectify Deliver | @)(1-p)" C a *) (c.37) 1=pyt_\ wv 2 Screen Scrap Deliver (oro) (c.38) Cs (@)G =p)" =n) 3 Scrap Rectify Deliver | eV (€.39) 4 Serap Serap Deliver | 5 (€.40) 5 Scrap Scrap sep | (~F *) (c.41) C3. Controlling the probability of conformance for delivered product This clause provides the mathematics and formulas to compute sample sizes for evaluating isolated or individual lots to control the probability of conformance for delivered product. This clause is organized into five sub-clauses of computational methods: a) Methods 1 and 2 that screen failed lots are presented in C.3.2; b) Method 5 that scraps the entirety of failed lots is presented in C.3.3; ©) Methods 3 and 4 that salvage conforming units from the samples of failed lots are presented after Method 5 in C.3.4, because the mathematics uses concepts from Method 5; d)_ the probability of acceptance technique is presented in C.3.5; e) the case when evaluations are destructive is presented in C.3.6. €3.1 Average Outgoing Quality Limits Table C5 shows the placement of the five methods presented in C2 into three groups by the consideration of two limits. The first limit is whether a maximum AQQ or AOQL occurs somewhere between 0 % and 100 % yield for any given lot size N and sample size n. The second limit considered is. whether there is a theoretical limit on the sample size as the lot size gets very large. 60