Sampling procedures and charts for inspection by variables for percent nonconforming(Applied to inspection of successive lots or batches)
Some standard content:
National Standard of the People's Republic of China
Sampling procedures and charts for inspection by variables for percent nonconforming(Applied to inspection of successive lots or batches)UDC 311.213.2
: 658. 562. 012
GB6378-86
This standard specifies the sampling plan and inspection procedure for a single measurement with acceptable quality level as the quality indicator, and is applicable to the batch-by-batch inspection of successive lots. This standard expresses the quality of small batches by nonconforming rate, and is used to control the process nonconforming rate not exceeding a certain specified value. This standard can be used in relevant documents such as product design, process documents, order contracts, product technical standards, inspection rules, certificates of conformity, quality management, etc. When using this standard, the "responsible department" for the implementation of the technical standard should be specified in the above documents. The following cases must be met when using this standard: The batches submitted for inspection are continuous. A series of batches that are continuously changed during the production process can generally be regarded as continuous batches as long as the product design, structure, process, main raw materials and other production conditions and the producer's operating level are basically the same. b. The value of a certain quality characteristic of the inspected product is a single measurement value that can be measured with a continuous scale. C. The inspected quality characteristic follows an approximately normal distribution. d. The specification limits used to determine whether the inspected quality characteristic is qualified or not are specified, including single-sided specification limits and double-sided specification limits. Double-sided specification limits also include discrete double-sided specification limits, comprehensive double-sided specification limits and composite double-sided specification limits. The drag sampling plan must be used together with the transfer rule. This standard refers to the international standard S () 3951-1981 "Measurement sampling inspection procedure and diagram for non-conforming product rate". This standard corresponds to GB2828-81 "Batch-by-batch inspection counting sampling procedure and sampling table (applicable to continuous batch inspection)". 1 Terms and symbols
1.1 Terms
1.1.1 Unit product
The basic unit of product for sampling inspection. 1.1.2 Metrological quality characteristics
Quality characteristics of the inspected product that can be measured by continuous scale 1.1.3 Metrological inspection
The process of comparing the observed value of the unit quality characteristics of the unit product with its technical requirements under specified conditions by measurement, test or other methods, and judging whether the unit product is qualified. 1.1.4 Metrological sampling inspection
The process of randomly selecting some unit products from the batch according to the specified sampling plan for metrological inspection, and judging whether the batch is acceptable. 1.T.5 Batch inspection
The process of inspecting each batch of products and judging whether the batch is acceptable. 1.1.6 Unqualified
The quality characteristics of the single-quality product do not meet the requirements specified in the relevant documents. Issued by the National Bureau of Standards on May 13, 1986
Implemented on May 1, 1987
1.1.7 Nonconforming product
GB6378-86
A unit product that has at least one quality characteristic that is nonconforming. 1.1.8 Specification limit
The limit value used to determine whether a certain metrological quality characteristic of a unit product is qualified. 1.1.9 Upper specification limit
The maximum value of a qualified metrological quality characteristic. 1.1.10 Lower specification limit
The minimum value of a qualified metrological quality characteristic. 1. 1.1.11 One-sided specification limit
The specification limit of the acceptable quality level is specified for only the upper or lower specification limit. 1.1.12 Two-sided specification limit
The specification limit of the acceptable quality level is specified for both the upper or lower specification limit. 1.1.13 Discrete two-sided specification limits
Two-sided specification limits for acceptable quality levels are specified for the upper and lower specification limits. 1.1.14 Composite two-sided specification limits
Two-sided specification limits for an overall acceptable quality level are specified for the upper and lower specification limits. 1. 1.1.15 Composite two-sided specification limits
One overall acceptable quality level is specified for the upper and lower specification limits, and one of the specification limits is specified for an acceptable quality level.
1.1.16 Acceptance or rejection of a batch
The information obtained from the sample is used to determine whether the batch meets the acceptance criteria specified in the sampling plan. Batches that meet the acceptance criteria should be accepted, otherwise they should be rejected.
1. 1. 17 *s” Method
A method to determine whether a batch is accepted or not by using the sample mean and the sample standard deviation. 1.1.18*g\Method
A method to determine whether a batch is accepted or not by using the sample mean and the process standard deviation. 1.1.19 Inspection methods
This standard has the following ten inspection methods:
“s\Methods include:
Upper specification limit
②Lower specification limit
Separate two-sided specification limit
Comprehensive two-sided specification limit
Composite two-sided specification limit
“Methods include:
Upper specification limit
|Lower specification limit
Discrete two-sided specification limit
Combined two-sided specification limit
Composite two-sided specification limit
1.1.20 Upper quality statistic
Function of upper specification limit, sample mean and sample standard deviation (or process standard deviation). 1,1.21 Lower quality statistic
Function of lower specification limit, sample mean and sample standard deviation (or process standard deviation). 1.1.22 Inspection level
The level that determines the relationship between batch size and sample size. 1.1.23 Sample size code (code)
GB 6378—86
The letter used for inspection sample size determined by batch and inspection level. 1.1.24 Batch defective rate
The ratio of the number of defective products in a batch to the number of batches, that is, batch defective rate = the total number of defective products in this batch × 100(%)
1. 1. 25 Process average defective rate
The average defective rate of a series of consecutive batches submitted. This standard uses it to express the average quality of a series of consecutive batches submitted. 1.1.26 Acceptable quality level
For the purpose of sampling inspection, the upper limit of the process average defective rate that is considered acceptable for a series of consecutive batches. 1. 1.27 Acceptance constant
A constant determined by the acceptable quality level and sample size for judging whether a batch is accepted or not. It gives the minimum value of the upper quality statistic and/or the lower quality statistic of an acceptable batch. 1.1.28 Acceptance curve
A curve determined by the acceptable quality level and sample size for judging whether a batch is accepted or not. It represents part of the boundary of the acceptance region. 1.1.29 Sampling plan
A set of rules specified for determining the sample size and judging whether a batch is acceptable or not. It can be composed of sample size and acceptance constant or sample size and acceptance curve (the former is a numerical sampling plan, using a numerical method procedure. The latter is called a graphical sampling plan, using a graphical method procedure). 1.1.0 Strictness of sampling plan
The degree of strictness of the sampling plan adopted for the batch submitted for inspection. The non-standard regulations include three sampling plans: normal inspection, strict inspection and relaxed inspection.
1.1.31 Sampling characteristic curve (0C curve) The functional relationship curve between the expected acceptance rate of the batch and the batch (or process) defective rate. 1.1. Maximum sample standard deviation
The maximum value of the sample standard deviation of the acceptable batch under given conditions. 1. 1. Maximum process standard deviation
The maximum value of the process standard deviation of the acceptable batch under given conditions. 1.1.34 Transfer rules
The rules used when transferring sampling plans (including suspended inspection) with different degrees of width. 1.2 Symbols
: upper specification limit
L: lower specification limit
N: batch size
\: sample size
Note, y and y correspond to the sample size of the upper and lower specification limits, respectively.
Note, y and y correspond to the acceptance constants of the upper and lower specification limits, respectively.
X: sample observation value (measured value).
She: When the measurement sequence needs to be distinguished, Xi represents the first observation value.
x: sample mean
: sample standard value
a: rate of passing the standard
QriE quality statistic
QL: lower quality statistic
AQL: acceptable quality level.
GB 6978-86
QX- or Qt
Note: AQI: and AQLL are the acceptable quality levels of the upper specification limit and the lower specification, respectively. MSSD: Maximum sample standard deviation.
MPSD: Maximum process standard deviation.
fs: Coefficient used in calculating MSSD
MSSD = fs (!-L)
fr: Coefficient used in calculating MPSI).
MPSd = (U- 1)
P: Batch failure rate.
Product: The fertility rate of a batch expected to be judged as accepted. Other terms and symbols refer to GB3358-82 "Statistical Terms and Symbols". Implementation procedures
The procedures for implementing water standards are as follows:
Choose inspection method1
Specify minimum level
Specify acceptable quality level:
Specify the strictness of sampling plan,
Submit product,
Inspect sampling plan:
Draw samples:
Inspect sample +
Determine whether the batch is accepted or not:
Inspect the batch.
8 Preparation for inspection implementation
Choose inspection method
Selection of specification limit
The selection of single-sided or double-sided specification limit depends on the specification of the design quality characteristic specification limit in the technical standards, order contract and other documents.
GB 6378—88
For double-sided specification limits, when it is necessary to control the defective rate exceeding the upper specification limit and below the lower specification limit respectively, use the discrete double-sided specification limits. When it is only necessary to control the total defective rate, use the comprehensive double-sided specification limits. If it is necessary to control not only the total defective rate, but also the defective rate of the ten-sided specification limits, use the composite double-sided specification limits. 3. 1.2 Selection of "S" method and "" method
Generally, you should start with the "S" method. The selection of "" method and "" method shall be in accordance with the provisions in. When applying the "" method, the process standard deviation must be determined in advance. When the OC curves are basically the same, the sample size drawn by using the "" method is smaller than that of the "" method: 3.2 Provisions on inspection levels
This standard specifies three general inspection levels, S-1, S-2, and S-3 and S-4. Inspection level I is usually used; when it is allowed to reduce the discrimination power of the sampling plan for batch quality, inspection level 1 is used; when it is necessary to improve the discrimination power of the sampling plan for batch quality, inspection level III is used.
Special inspection levels S-3 and S-4 are mainly used in situations where the sample size is required to be minimized, but the risk of misjudgment will be increased. 3.8 Provisions on acceptable quality levels
Appropriate acceptable quality levels should be specified for the metrological quality characteristics of various non-conforming types, and clearly specified in documents such as product technical standards or order contracts. This standard has 11 preferred acceptable quality levels from 0.10% to 10.0%. If other values outside this standard are selected, the charts and tables of this standard are not applicable. 3.3. 1 Classification of nonconformities
Before specifying the acceptable quality level for each quality characteristic, first, classify each quality characteristic into different nonconformity classes according to the severity of the nonconformity. Generally, they can be divided into three classes: A, B, and C, but do not include metrological quality characteristics that require an acceptable quality level of 0. 3.3.1.1 Class A
Class A nonconformities have the most serious impact on the quality of the product. In sampling inspection, a very small acceptable quality level should be used.
3.8.1.2 Class B
Class B nonconformities have the most serious impact on the quality of the product. In sampling inspection, an acceptable quality level that is larger than Class A but smaller than Class C can be used. wwW.bzxz.Net
3.3.1.3 Class C
Class C nonconformities have only a slight impact on the quality of the product. In sampling inspection, an acceptable quality level that is larger than both Class A and Class B can be used.
3.3.2 Provision of Acceptable Quality Level
For single-measurement specification limits, only the acceptable quality level AQL value shall be specified for the specification limits that need to be controlled. For discrete double-sided specification limits, the acceptable quality level AQL value shall be specified for the upper and lower specification limits respectively. These two acceptable quality level values may be the same or different. For comprehensive double-sided specification limits, only a total acceptable quality level AQL value shall be specified for the upper and lower specification limits. For composite double-sided specification limits, in addition to specifying a total acceptable quality level AQL value for the upper and lower specification limits, an acceptable quality level AQL value shall also be specified for the specification limits on one side that need to be controlled.
The supplier has no right to provide any known non-conforming products due to the specified acceptable quality level. 3.4 Provisions for product submission inspection
Unit products shall be submitted in batches, and the submission order shall be in the production order. The submitted batch may be the same as or different from the production batch, sales batch, and transportation batch. When necessary, the supplier shall provide sufficient and suitable storage space, equipment and personnel and other necessary conditions to facilitate the management, sampling and testing of the inspection batches. 4 Provisions on the strictness of the sampling plan
4.1 General provisions
If there are no special provisions, the normal inspection sampling plan shall be used from the beginning. In special circumstances, upon the designation of the responsible department, a stricter inspection or a relaxed inspection sampling plan may also be used from the beginning. GB 637B86
Except when the strictness of the sampling plan needs to be changed according to the transfer rules, the next inspection shall continue to use the same sampling plan with the same strictness as the previous inspection.
4.2 Transfer rules
In the sampling inspection, the strictness shall be adjusted according to the following transfer rules according to the quality changes. 4.2.1 Change from normal inspection to strict inspection
When conducting normal inspection, if two batches of no more than five consecutive batches are accepted after initial inspection, strict inspection shall be implemented from the next batch.
4.2.2 From Strict Inspection to Normal Inspection
When performing Strict Inspection, if five consecutive batches are accepted after initial inspection, normal inspection will be performed from the next batch. 4. 2.3 From Normal Inspection to Relaxed Inspection When performing normal inspection, if the following conditions are met, relaxed inspection will be performed from the next batch.
b. After the acceptable quality level is tightened one level, these batches are still accepted!
d. The responsible department intends to transfer to relaxed inspection. 4. 2.4 From Relaxed Inspection to Normal Inspection
When performing relaxed inspection, if any of the following conditions occur, normal inspection will be performed from the next batch.
b. The production is not normal,
d. The responsible department considers it necessary to return to the belt inspection. 4.2.5 From stricter inspection to suspended inspection
After stricter inspection is started, when the number of rejected batches reaches five, the inspection according to this standard shall be temporarily suspended. 4. 2.6 Resuming inspection after suspension of inspection After the inspection is suspended, if the supplier has indeed taken effective improvement measures and improved the quality of the submitted products, the sampling inspection specified in this standard may be resumed. In principle, it should start with stricter inspection, but it is not allowed to start with relaxed inspection. 5 Provisions for conversion between the “\ method and the “α” method 5.1 General provisions
At the beginning of the inspection, the “S” method is generally used. If there are sufficient reasons to show that the process standard deviation is known, with the consent of the responsible department, it can also start with the “” method.
The “S” method and the “α\ method each use independent transfer rules. 5.2 From the "α" method to the "S" method
If the inspection is normal or relaxed when the "S" method is implemented, and the S control chart and quality analysis report show that the sample standard deviation S value is in a statistical control state, the responsible department shall designate and use the "α" method from the appropriate discussion. At this time, the normal inspection of the "α" method shall be started.
When the determined process standard deviation value is greater than the corresponding maximum process standard deviation MPS1) value in Table 10 in 8, it is not allowed to switch from the "3" method to the "R" method. The method for determining the process standard deviation value used in the "G" method is shown in 7.7.2. 5.3 From the "α" method to the "S" method
5.3.1 When the "O" method is implemented, if one of the following situations occurs, the "S" method shall be implemented immediately. a: The 5 control chart is abnormal;
b: There is reason to explain that the process standard deviation has changed, C.Although the s control chart does not show any abnormality, a certain point exceeds the corresponding maximum process standard deviation MPSD. The estimated process standard deviation using the sample standard deviation corresponding to this point and the adjacent nine consecutive points is still greater than the above-mentioned maximum process standard deviation MPSL. CB6878-86
5..zWhen the s control chart shows that the process standard deviation is small, the "" method can continue to be used. When the "a" method before conversion is in normal inspection or relaxed inspection, the converted "s" method starts from normal inspection. When the "a" method before conversion is in tightened inspection: or the estimated new process standard deviation exceeds the maximum process standard deviation of the priority acceptable quality level AQL of the tightened level, the converted "s" method must start from tightened inspection. 6 Retrieval and implementation of sampling plan
6.1 \g\ method
6.1.1 Numerical method
The numerical method procedure of "" method specified in this standard is the same as that of single-sided specification limit and discrete double-sided specification limit. , sampling plan inspection
(1) According to the batch size and inspection level, check the sample size code from Table 1 in Figure 8. (2) According to the sample size code and the acceptable quality level, check the sampling plan (n,) or (n., k) and ny, ku) from Table 3 in Figure 8.
The sampling plans for tightening inspection and relaxing inspection are respectively found in Tables 4 and 5 in Figure 8. If the sample sizes of the two sampling plans for both sides of the discrete bilateral specification limit are different, the larger one is taken as the common sample size, and the acceptance constant remains unchanged. For example, AQLL=0.15%, AQ1=0.25%, the code is E, and the normal inspection sampling plans found in Table 3 in Figure 8 are (10, 2.24) and (7, 2.00). m actually uses the two sampling plans (10.2.24) and (10, 2.00).
b. Judgment of whether to accept or not of a batch
Draw a sample of size n, measure the metrological quality characteristic value of each unit product, and then calculate the sample mean X and the sample standard deviation s.
For the one-sided upper specification limit, calculate the upper quality statistic Qu=UX
If Q is less than 1, then accept the batch; if Q is less than 1, then reject the batch. For the one-sided lower specification limit, calculate the lower quality statistic Qr =
If Q is less than 1, then accept the batch: If Q is less than 1, then reject the batch. For the separate two-sided specification limits, calculate the upper and lower quality statistics Qu at the same time
If Qk and Qu are less than 1, then accept the batch: If Qz is less than 1 or Q is less than 1, then reject the batch. Example 1: The tensile strength of a certain aluminum tube after quenching is required to be no less than 435.V/mm. When conducting this inspection, the inspection batch consists of 100 unit products. After confirmation, the inspection method of the single-side lower specification limit of the “:” method in this standard is used, and the inspection level is taken, and the acceptable quality level is taken as 0.4%. In the process of batch-by-batch inspection, according to the transfer rule, it is assumed that the current batch of the table-batch is inspected, and try to judge whether the batch is accepted based on the sampling results:
Already:
N:. 100,L = 435, AQI. -D.1%. (1) From N=, inspection level, 8Table 1, we get the code F. GB 6378--86
(2) By the "s" method, normal inspection, code F, AQL=0.4%, check 8rtTable 3 to get the sampling plan [n,) is (10,1.98). (3) Ten units of product are randomly selected from the batch to form a sample. The tensile strengths are measured in N/mm2: 475, 470, 465, 466, 475, 465, 465, 48, 455, 435. (4) Calculate the sample value and the standard deviation of the sample. The quality statistic is (5) Because the batch is accepted. (Xx)2 Note: When using the strict inspection or relaxed inspection sampling method, the following table 4 or table 5 in step (2) shows that: In the sampling case (10, 2.11) The relaxed inspection sampling rate (5, 1.65) The other steps remain unchanged.
Calculation value
Example 2: The hardness value of a certain high-speed steel knife after heat treatment is specified as 58～62 HRC. When conducting this inspection, each batch of products is submitted for inspection. After determination, the inspection method of "." method with two-sided specification limits in this standard is used, and the inspection level is taken as II, the acceptable quality level for the lower specification limit is taken as 0.25%, and the acceptable quality level for the upper specification limit is taken as 1.0%. In the process of batch inspection, according to the transfer rule, assuming that a certain batch should be subject to strict inspection, try to judge whether the batch should be changed according to the sampling results. :
N: 8, 1=58, =62, AQLL =0.25%, AQLu = 1.0%. (1) N8, pick up water 』, lose 1 get Ning code. (2) "s" method plus strict inspection, Ning code F, AQL-0.25%, AQL=1.0%, check Table 4 in 8 to get: AQl.= (, 25% sampling plan (n,) is (10, 2.24)) AQL-1.0% sampling plan (n,,) is (7, 1.75). According to 6.1.1a, the sampling plan (ny,) should be changed to (10, 1.75). (3) Ten units of product are randomly selected from the batch and the hardness values HRC are measured as follows: 58.5, 58.0, 60.0, 61.0, 60.5
59.5.60.5，58.0,61.5,60.0
(4) Calculate the sample mean
sample standard case
1 quality statistic
Qu=Ux
Calculate the following quality statistics
(5) Judgment
Because Q1., the batch is rejected.
6.1.2 Graphical method
6.1.2.1 One-sided specification limit
a: Retrieval of sampling plan
GB 6378—86
Use the same method as 6.1.1.@ to retrieve the numerical sampling plan (n,). Then, on the coordinate paper with the horizontal axis representing the sample standard deviation and the vertical axis representing the sample mean x, draw the following acceptance curve X=U - ks
For the upper specification limit
For the lower specification limit
This constitutes a graphical sampling plan and determines the acceptance domain (including the limit) shown in the shaded part of Figure 1 or Figure 2. b. Acceptance domain of specification limit on “,” method
c. Judgment of whether batch is accepted or not
Tu-hs
Connect Rucheng
GB 6378—86
Chapter 2 Acceptance domain of specification limit on “s” method
Draw a sample of size , measure the metrological quality characteristic value X of each unit product, and then calculate the sample mean X and sample standard deviation s. Draw the point (s, X) in Figure 1 or Figure 2. If the point falls within the acceptance domain, the batch is accepted, otherwise the batch is rejected. Example: Using the data of Example 1 in 6.1,1, draw (,) = (12.69, 465.0) in Figure 3. Since the point falls within the acceptance domain, the batch is accepted. 600
Acceptance domain
T-435+1.985
s - 12.69
Figure 3 "s" method single-sided F specification limit, normal inspection, code F, AQL-0.4%6.1.2.2 Discrete double-sided specification limit
GB 8978-86
a, Retrieval of sampling plan
Use the same method as 6.1.1.α to retrieve the numerical sampling plan (n,) and (n,). Then, on the coordinate paper with the horizontal axis representing the sample standard deviation s and the vertical axis representing the sample mean ×, draw the following acceptance curve x=t- a, s
X=L + hr. s
This constitutes the graphical throwing plan and obtains the acceptance domain (including the boundary) shown in the shaded part of Figure 4. Kus
According to the acceptance method of "s" in Figure 4, the acceptance domain b of the bilateral specification limit is divided. The judgment of whether the batch is acceptable or not
Draw a sample of a certain size, measure the metrological quality characteristic value X of each unit product, and then calculate the sample mean X and sample standard deviation s. Put the point (s, X) in Figure 4. If the point falls within the acceptance domain, the batch is accepted, otherwise the batch is rejected. Example: Using the data of Example 2 in 6.1.1, (5, X) = (1.23, 59.75) is drawn in Figure 5. Because the point falls outside the acceptance domain, the batch is rejected. X-62-1.75#
T-59 +2.24
GB 637886
Figure 5 "." method separates the two-sided specification limits, adds strict inspection, code E, AQL1 = U.263, AQl.u = 1.0%
6.1.2. Comprehensive two-sided specification limits
a, the selection of sampling plans
(1) According to the batch size and perfume, the sample size is given in Table 1. (1.2H, 59.76)
(2) Based on the sample code, the sample size of the normal inspection sampling plan is set in Table 2 or Table 3 of 81, and then the acceptance curve marked with the specified acceptable quality level in the figure of the corresponding code in Figure SD~Figure s-.I Figure: EH and Figure s-FG[ in 9.1 is found. This constitutes the normal inspection sampling plan and obtains the acceptance domain (including boundaries) shown in the shaded part of Figure 6. The tightened inspection acceptance curve with the same code and the same AQL value is the normal inspection acceptance curve corresponding to the acceptable quality level one level lower than the AQL in the same figure, that is, the curve close to the normal inspection acceptance curve corresponding to the original AQL value. The relaxed inspection acceptance curve with the same code and the same AQL value can be found by the following method: (1) According to the code and the acceptable quality level used for normal inspection, the sample size and acceptance number of the relaxed inspection sampling plan are obtained from Table 5 in 8.
（2）8Table 3, Sample size code and acceptable quality AQL corresponding to (1 sample size and acceptance constant
（3）The code and indirect acceptance quality AQL obtained from (2) can be used to store the required relaxed inspection acceptance curve from Figure 91: -1)~Figure -P, Figure -)EI and Figure -FGI.
Tip: This standard content only shows part of the intercepted content of the complete standard. If you need the complete standard, please go to the top to download the complete standard document for free.