《抽样检验过程中的相关定义（Relevant definitions in the process of sampling inspection）.doc》由会员分享，可在线阅读，更多相关《抽样检验过程中的相关定义（Relevant definitions in the process of sampling inspection）.doc（22页珍藏版）》请在三一文库上搜索。
1、抽样检验过程中的相关定义（Relevant definitions in the process of sampling inspection）Definitions related to sampling inspection1 batch (under certain conditions produced a certain number of units of products made up of groups), batch of units: volume, box, package, Zhang, Zhang, kg)2 inspection batch (for the im
2、plementation of the inspection and collection of units, easy to sampling, facilitate sampling results more representative)3 batch N4 samples and sample size n5 sample6 for the specification accuracy is not high (such as + 12mm), destructive testing and other projects, the general use of sampling 15
3、special methods, can be corresponding to MIL - STD - 105E sampling of S-I.7, the number of conformity (AC), the number of unqualified (RE)8 sampling inspection: the method of drawing out a certain number of samples in batch size from inspection batch.9 sampling criteria: forms supported by a samplin
4、g plan, such as MIL-STD-105E, GB2828-87.10 sampling methods: methods of sampling samples from inspection batches.11 sampling plan (plan): a combination of sample size or sample size series and decision array (n, /AC, RE).12 determine the array: the number of qualified judgments and the number of unq
5、ualified combination.13 inspection level (IL): check the rank relation between the batch and the sample size.14 qualified quality level (AQL): the highest (large) undesirable rate of customer satisfaction.15, consumer risk rate (beta) (type II error), poor quality, but the quality of the sample qual
6、ified, but sentenced to the whole batch of qualified smaller probability, general beta =10%.16 producer risk rate (alpha) (type I error), good quality, but the quality of substandard samples, but failed to count the whole batch of less probability, general alpha =5%.17 rejection of quality level (LT
7、PD) (RQL): consumers consider the worst quality inspection batches containing the lowest No.Yield.18 serious bad products: contain one or more serious substandard products.A product that contains one or more of the major, substandard, but seriously substandard products.A secondary product that conta
8、ins one or more secondary failures but does not have other nonconforming products.19, the bad rate (P) (hundred pieces of substandard goods)2 thousand number of faults (c)21 transfer principle22. sampling characteristic curve -OC curveA sampling plan for the quality of products to distinguish betwee
9、n the ability to change the sampling characteristics of the program.A sampling plan has corresponds to a OC curve (strictly speaking is a N, N, c) said its quantitative product quality condition (P) and the relationship between the received probability within the said receiving probability L (P) whe
10、n the batch failure rate of P function.1. when the (n, c) is certain, the batch size has little effect on the OC curve.2. when C is certain, the larger the N, the steeper the OC curve.3. when a batch of P fails to be fixed, the larger the N, the more stringent the sampling scheme.Thirteen. Sampling
11、inspection type1, according to quality data categories: metering value sampling, numerical value samplingclassificationProject count sampling inspectionQuantitative sampling inspectionThe quality representation is represented by or expressed by the difference between (good) and (bad) two, and the nu
12、mber of faults indicates the characteristic valueInspection method 1. test without proficiency2., the time required for inspection is short3. the inspection equipment is simple and the inspection cost is low4., the calculation record is simple5. simple calculation, almost no need to calculate 1., us
13、ually in the test should be skilled.2. the time required for the inspection is long3. the inspection equipment is complex and the inspection cost is high4. the inspection record is complex5. computational complexityThe number of sampling plans, the conditions of application, and the product of each
14、breed, a sampling plan should be worked out.Randomization is required for sampling. For each quality characteristic, a sampling plan should be worked out.The characteristic values need to belong to the normal distribution sampling time randomizationWhen the judgment ability and the sample number hav
15、e the same judgment ability, the required sample number is more(1) it is not easy to improve the quality.(2) it is difficult to detect the errors of the inspection instruments. The number of tests at the same time, the ability to judge low. In order to get the same judgment, the required number of s
16、amples is small(1) can lead to improvement in quality(2) the errors of the inspection instruments can be found; the number of checks is same at the same time, and the judgment ability is high.The use of inspection records, low inspection record utilization, high availability of inspection records, d
17、ata feedback, improved process reference. At the same time, the value relation between data points and specifications, to provide reference for the specification of the lenient.The range of applications, sampling of all products, can be applied to destructive tests or samples of valuable productsThe
18、 number of defective products in the sample of the rejected lot exceeds the allowable number of defective products, and the estimated rate of the excess of the batch exceeds the allowable bad rate, and the determination of the possible undesirable products in the rejected batch.(I) sampling by numer
19、ical value;Regular type, single selection type, single adjustment type, single continuous sampling type, successive sampling typeDouble, double, double timesmany times(two) sampling rule of value measurement sampling (single) GB8053-87Adjustable (single) GB6387-86 ISO3951-812, according to the inspe
20、ction method and level points: adjustment type sampling, non adjustment samplingAdjustment type sampling: plus strict sampling, normal sampling, relaxation sampling (ultra wide sampling)Non regulating sampling: standard sampling, selective sampling, continuous sampling, sampling by batchStandard sam
21、pling: the spirit of this sampling plan takes into account both the interests of the seller and the buyer, namely, the risk rate of both parties, the alpha and beta are certain numbers (general alpha =0.05, beta =0.1)Selective sampling: be found to be unqualified, must be all tick, reject the defect
22、ive goods, and make up for good products and accept. This way in a number of long-term continuous sampling sampling, acceptance batch quality will work better than the original average, can ensure that the average quality acceptance batch. For nondestructive testing only.Continuous sampling: suitabl
23、e for continuous production of products, in the process of sampling, sampling a certain number of consecutive samples at the beginning, is a good product to every certain number of pumping a sample inspection, if the bad products, namely when the way to return to the starting test.Batch by batch sam
24、pling: the plan is to take one sample at a time during each inspection batch, check the amount of good and bad products, and then decide whether to accept or reject it in accordance with the formula or look-up table.This plan is the only not to predict the number of sampling, sampling shall be deter
25、mined after qualified or not known until after the number, the number of samples in general, the sampling method of the sample number.3, according to the sampling frequency points: a sampling, two sampling, repeated sampling.4, according to control points: AQL, RQL (LTPD), AOQL typeAOQL: average out
26、going quality limit5 normal inspection, two sampling (32, 32/2, 6)6 plus strict inspection, single samplingUsually in the decision array strict, AQL constant7 single sampling for quantity reduction inspectionIn the reduction, the sample size of AQL unchanged. As a normal test single sample N=5000, I
27、L=G II, n=L=200, AQL=1.0 (5,6) to determine the array, and reduced inspection single sampling n=L=80, determine the array (2,5). The unqualified number is 2 AC; 3 and 4 AC but, under normal forward sampling; 5 RE.Fourteen. Sampling techniques (sampling methods)1. random sampling: completely random,
28、without limitation; usually use random numbers or sampling ball2. system sampling: sampling at a certain time / quantity interval3.: the first layer of stratified sampling after re sampling4. zigzag sampling: the method used to reduce systematic sampling errors due to periodicity. It can be regarded
29、 as random sampling, but it is more complex and regular.5. regional sampling: groups such as a large box of goods, there are ten boxes in the box, each small box containing a number of items. For the convenience of sampling, a number of sample boxes can be randomly extracted from dozens of small box
30、es, and then all samples are checked. This method, such as social investigation, is used for sampling in towns or villages, so it is called regional sampling.The premise is that the variation is large within the region, and the variation between regions is small. In contrast to stratified sampling.6
31、. segment sampling: first, regional sampling is used, and random sampling is taken from the sample units. It can be divided into two sections and many sections.7. repeated sampling: sampling methods to infer population quality in more than one sample within the same inspection lot. Double, multiple,
32、 or successive sampling used in sampling inspectionFifteen. Short term countermeasures for nonconforming products:1., identification, placing should be different (to consider the procurement, inventory, finished products, idle goods, shipped, pending products)2. the defective products found in the f
33、actory are found to be taken out or supplied in good quantities. The inspection found at the guests place should be taken out or made up in good quantities3. audits of nonconforming batches (usually in the form of MRB: material review meetings): rework, rework, selection, selection, acceptance, rele
34、ase, demotion, rejectionSixteen. Sampling standards1. sampling standards approved by China have more than 20 kinds, and only GB2828 and GB2829 are the common standards in industrial enterprises. Of course, beyond these two standards, many international standards are almost universal equivalent.2.GB2
35、828-87: numerical adjustment sampling type, divided into twelve kinds;GB2829-87: count value, successive sampling type (less used)Adjusted numerical sampling type: (1989) MIL-STD-105E, (1963) MIL-STD-105D, (1974) ISO2859(1993) ANSI/ASQCZ1.4, (1963) BS9001Notes on the revision of the name of the year
36、February 1942 JAN-STD-105In September 1950, MIL-STD-105A became the common standard of land, sea and air force, JAN-STD- abolished 105Provisions for small sample inspection with Appendix in April 1955In December 1958, the contents of MIL-STD-105B, 105A and Appendix were arranged as one, and the aver
37、age estimate of the project was precise. In principle, the 5 batch was used to calculate the abolition of 105AIn July 1961, MIL-STD-105C increased the number of individual quality protection provisionsThe presumption of Engineering average is changed into a minor formula, and the number of calculate
38、d objects is restored to the 10 batch, and the 105B is abolishedIn 1963, the 4.29 MIL-STD-105D was changed to a strict adjustment method, which was easier to implementThe efficiency of the double sampling inspection has been improvedAll kinds of tables make 105C abolishedIn 1963, ABC-STD-105 became
39、the first international quality control standard through the military efforts of the United States, Britain and Canada. 105D became the first international quality control standardIn 1974, ISO2859 MIL-STD-105D was recognized as the ISO standardIn 1989, 5.10 MIL-STD-105E MIL-STD-105E promulgated, 105
40、D abolishedIn 1993, the ANSI/ASQCZ1.4 ANSI/ASQCZ1.4 became the national standard for the United States, instead of the MIL-STD-105E 105E, which was abolished in the United States3.C=0 sampling planIn 1965, the C=0 sampling plan was first published1981, C=0 sampling plan, Second EditionIn 1983, the t
41、hird edition, many European and American enterprises are using C=0 sampling planMIL-STD-105E, AQL value, does not mean that the unqualified product unit supplier is authorized to provide any known.When C=0 was qualified; when C 0, sentenced to withhold group (Withhold the lot), project management fo
42、r MRB / decision.4. pipeline sampling plan: GB8052 (MIL-STD-1235B)A. is nominally count continuous sampling, in fact, the piece test, corresponding to the continuous sampling program CSP.Number of defective products during inspectionB. the average non-performing rate = _, defined as the production o
43、f a specified periodTotal product at inspectionThe average rate of adverse effects of the product during initial inspectionThe average process adverse rate in batch sampling in C.GB2828 is defined as a series of initial improvement testsThe rate of nonconformity in the batch can refer to the rate of
44、 defective goods or the number of defects per hundred.D. initial inspection: the first (initial submission) inspection of the product after processing is relative rework, rework, or selectionIn terms of retest.There is a difference between E.GB8052 and AQL in GB2828 (not outlined)F. nonconforming pr
45、oduct rate refers to the ratio of the number of nonconforming products to the number of production in the whole process of production.Average quality of inspection (average rate of rejected products) (AQL): unqualified products in inspection (spot check or full inspection) are replaced by qualified
46、products, average quality of products treated by such treatment, percentage of nonconforming products, P.5. hop batch sampling plan: GB/T3263-91When the test meets certain conditions, after the implementation of skip batch inspection, if the f=1/3 is the 3 batch of 1 batches of seized, that is, ther
47、e are two batches of non inspection, when the inspection has problems, then batch by batch inspection6.CNS2779Z4006 is the same as MIL-STD-105E (CNS is the code of the Republic of China)7.JISZ9002, JISZ9003, JISZ9004, JISZ9009, JISZ9010 Japan standard sampling planJISZ9006 Japan selective sampling planJISZ9008 Japan continuous production sampling planMIL-STD-414ANSIZ1.9Seventeen. Sampling planThe 1. generation of sample size table2.AQL table3. sampling plan (N / AC, RE)4. sample size variability and composite sampling plan4.1 variability in sample size4.2 finished product inspection, w