Single sampling procedures and tables for product quality audit by variables for mean value
other information
Release date:1994-01-22
Review date:2004-10-14
drafter:Yu Zhenfan, Ma Yilin, Chen Zhitian, etc.
Drafting unit:China Institute of Standardization and Information Classification and Coding, Institute of Systems Science, Chinese Academy of Sciences, Computing Center, Chinese Academy of Sciences, Beijing University of Technology
Focal point unit:National Technical Committee for Application of Statistical Methods and Standardization
Proposing unit:State Bureau of Technical Supervision
Publishing department:State Bureau of Technical Supervision
competent authority:National Standardization Administration
Some standard content:
National Standard of the People's Republic of China
Product quality average value measurement...
Single sampling procedures and tables for product quality audit byarlableg for mean value1Subject content and scope of application
GB/T14900--94
This standard specifies the measurement single supervision sampling inspection procedure for the average value of a quality characteristic of the supervision population as the quality indicator. This standard applies\Quality supervision departments shall regularly or irregularly implement quality supervision sampling inspection on the overall product that has passed the acceptance. The quality characteristics of the inspected products shall obey or approximately obey the normal distribution. This standard stipulates that the risk of misjudgment a=0.05.
2 Reference standards
GB/T3358 Statistical terms
GB/T4086 Table of statistical distribution values
GB/T4889 Statistical processing and interpretation of data Estimation and test methods for mean and variance of F-state distribution GB/T 6583 -ISO 8402 Quality - Terminology GB/T10111 Methods of random sampling using random number segments 3 Terminology symbols
3.1 Terminology
3. 1. 1 Surveillance population
The collection of products to be supervised.
3.1.2 Sample
A group of unit products drawn from the surveillance population according to a certain procedure. 3.1.3 Overall mean
The arithmetic mean of the values of a certain quality characteristic of each unit product in the surveillance population. 3.1.4 Population variance
The sum of the squares of the differences between a certain quality characteristic value of each unit product in the supervision population and the population mean divided by the population size minus 1: N
3.1.5 Population standard deviation
The positive square root of the population variance.
3.1.6 Sample mean
Approved by the State Administration of Technical Supervision on 1994-01-22 (X: p)
Implemented on 1994-08-01
GB/T 14900-94
The average value of a certain quality characteristic value of the unit products in the sample. 3.1.7 Sample standard deviation
The sum of the squares of the differences between a certain quality characteristic value of each unit product in the sample and the sample mean divided by the sample size minus 1: (X, - x)
3.1.8 Sample standard deviation
The positive square root of the variance.
3.1.9 Supervision sampling inspection
Sampling inspection conducted independently by the first party to determine whether the supervision population is acceptable. 3.1.10 Acceptable quality
The average value of the basic characteristics of the supervision population that is considered satisfactory with a high acceptance rate in the blue-sampling inspection. 3.1.11 Supervision quality level
The limit value of acceptable quality specified by the specification
3.1.12 The maximum value of acceptable quality specified by the upper specification limit
3.1.13 The minimum value of acceptable quality specified by the lower specification limit
3.1.14 The maximum and minimum values of acceptable quality specified in the two-sided specification limit. 3.1.15 Efficacy of supervision sampling inspection
The probability that the supervision population is judged as unacceptable when the actual quality level of the supervision population does not meet the requirements of the supervision quality level. 3.1.16 Supervision test level
The hierarchical correspondence between sample size and test power in supervision sampling test. 3.1.17 Missed judgment risk
The probability of judging a supervision population that actually passes as unpassed. 3.1.18 Missed judgment risk
The probability of judging a supervision population that actually fails as passed. 3.1.19 Quality statistic
A function composed of specification limit, sample mean and population standard deviation (or sample standard deviation), divided into upper specification limit and lower specification limit.
3.1.20 Failure judgment value
When the supervision population is judged as unpassed, the maximum value allowed by the quality statistic. 3.1.21 One-time supervision sampling test
A sampling test that determines whether the supervision population passes based on the test results of only one sample. 3.1.22 Supervision sampling plan
The combination of sample size and failure judgment value is called supervision sampling plan. 3.1.23 Supervision sampling inspection type
Based on whether the standard deviation of the supervision population is known, it is divided into two types of supervision sampling inspection: known standard deviation and unknown standard deviation. When the standard deviation of the supervision population is known, it is called \.\ method supervision sampling inspection. When the standard deviation of the supervision population is unknown, it is called \S\ method supervision sampling inspection. 3.2 Symbols
: Population value.
: Population standard deviation.
; Estimated value of the population standard plan.
: Lower specification limit.
mu: Upper specification limit.
n: Sample size.
GB/T 14900-94
x represents the quality characteristic value of the th unit product.
: Sample mean.
S: Sample standard deviation.
Q:: Upper specification limit quality statistic.
QL: Lower specification limit quality statistic.
—or Q
: Failure judgment value.
a; Risk of misjudgment.
β: Risk of missed judgment.
[n+]: Sampling plan.
(α), the passing probability of the supervision sampling plan. Procedure for supervision sampling inspection
The inspection procedure specified in this standard is as follows:
Determine the supervision population:
Determine the quality basis characteristics;
Determine the type of supervision sampling inspection;
Determine the supervision sampling inspection method,
Specify the supervision quality level,
Specify the supervision inspection level;
Retrieve the supervision sampling plan;
Extract samples!
Test the sample and make calculation results;
j: Determine whether the monitoring population can pass:
5 Implementation of supervision sampling inspection
5.1 Determine the supervision population
CB/T 14900—94
According to the supervision, the supervision population should be determined. The products in the supervision population can be the products produced by the same manufacturer, the same model, and the same production cycle: when necessary, they can also be products produced by different manufacturers, different models, and different production periods. 5.2 Determine the quality characteristics of the supervision population to be inspected According to the needs of the supervision, determine one or more quality characteristics of the supervision population to be inspected. 5.3 Determine the type of supervision sampling inspection
If the standard deviation of the supervision population can be determined in advance, the "." method can be used; otherwise, the \S\ method should be used. Tables 1 and 2 in this standard are used for the \a\ method, and Tables 3 and 1 are used for the \S" method. 5.4. Specifying the supervision sampling inspection method
This standard includes two different supervision sampling inspection methods: upper specification limit, lower specification limit and bilateral specification limit. They can be selected according to the different requirements of the product standard for the specification limits of product quality characteristics. 5.5 Specifying the supervision quality level
Specify the supervision quality level according to the supervision needs. The supervision quality level should be compatible with the sampling inspection method. 5.6 Specifying the supervision inspection level
Tables 1 to 4 of this standard all give 15 =One supervisory inspection level. Considering the difference between the \a\ method and the s\ method, the sample size series corresponding to the supervisory inspection level in Table 1 and Table 2 are not completely the same as the sample size series corresponding to Table 3 and Table 4. The higher the supervisory inspection level, the larger the sample size and the higher the inspection efficiency. Once the supervisory inspection level is selected, it shall not be changed during implementation.
5.7 Determine the supervisory sampling plan
5.7.1 "\ method one-sided specification limit
In the case of \ method one-sided specification limit, use Table 1. The table lookup method is: directly read the sample size and the failure judgment value from the row where the supervisory inspection level is located.
5.7.2 "\ method two-sided specification limit
In the case of "\ method two-sided specification limit, use Table 2. The table lookup method is: first read the sample size from the row where the supervisory inspection level is located, and then read the failure judgment value from the intersection of the column where the value of \ is located and the row where the sample size is located. a
5.7.3\5 method one-sided specification limit
Use Table 3. The table lookup method is the same as 5.7.1.5.7.4 "5\ method two-sided specification limit
Use Table 4. The table lookup method is: first read the sample size + from the row where the supervision inspection level is located, and then read the failure judgment value from the intersection of the column where the value of —
is located and the row where the sample
/ /n—1. 64
is located. 5.8 Sample selection
The sample should be randomly selected from the supervision population. Simple random sampling can be performed using methods such as GBT10111, or stratified proportional random sampling or multi-level sampling can be used as needed or as appropriate. 5.9 Testing and calculation of samples
The samples shall be tested one by one according to the test, measurement or other methods specified in the product standards or relevant documents. The test results shall be recorded completely and accurately, and the mean and standard deviation of the samples shall be calculated. 5.10 Judgment on whether the supervision population can pass 5.10.1 Judgment rules
8: When the upper specification limit is given
GB/T 14900-94
If Qu is forgotten, the supervision population is judged to be unpassable; if Qu>, the supervision population is judged to be passable.
Where Q--x
h. When the lower specification limit is given
If ≤., the supervision population is judged to be unpassable:
If Q>, the supervision population is judged to be passable.
Where Q--x
h. When the lower specification limit is given
If ≤., the supervision population is judged to be unpassable:
If Q>, the supervision population is judged to be passable.
where;
When the two-sided specification limit is given
if or ≤, then the supervision population is judged to be unpassable+if u> and QL, then the supervision population is judged to be passable. Where u
5.10.2 "S\ method judgment rule
. When the upper specification limit is given
if Q is true, then the supervision population is judged to be unpassable: if ≥, then the supervision population is judged to be passable.
where: Qu=-X
b, when the lower specification limit is given
if Q≤, then the supervision population is judged to be unpassable; if Q.>, then the supervision population is judged to be passable.
where: Q-
When the two-sided specification limit is given
if or ≤, then the supervision population is judged to be unpassable: if Qu> table and Q>, then the supervision population is judged to be passable.uX
where: Qu=some
5.11 Supervision sampling inspection results The statistical explanation of the theory: The quality supervision department confirms that the supervision population that fails the supervision sampling inspection is unqualified, and the supervision population can be reasonably traced. For products that were not present during the supervision sampling, if there is sufficient evidence to prove that they belong to the supervision population (for example, they belong to the same inspection batch), they should also be treated as unqualified.
Because the supervision sampling plan judged as a supervision population that can pass has a greater risk of missed judgment. Therefore, the supervision population that passes the supervision sampling inspection is not confirmed as qualified.
6 Application Examples
6.1 "\ Method
6.1.1 Given Upper Specification Limit
Example: A company produces solid caustic soda and requires that the iron oxide content in it be low. The upper specification limit of the population mean is specified to be 0.0045%, and the supervision inspection level is II. The population standard deviation d-0.0006% is known, and the supervision sampling plan is to be determined. Determination steps:
. Check Table 1, and find out from the line where the supervision inspection level is: GB/T 14900-94
[n,,-—[4,— 0,822]
b, and judge. Randomly select 4 units of products from the population, calculate the sample mean after testing, and the judgment rule is: Qu -4-X 0.004 5 - X
If Qu≤--0.822, then the supervision population cannot pass: If QL>-U.822, then the supervision population can pass. 6.1.2 Given the lower specification limit
Example: The tensile strength of a certain material is better, and the supervision quality level of the population mean is specified.-45X10°Pa, and the supervision inspection level is x. Given the overall standard deviation a = 4 × 10° Pa, try to determine the supervision sampling plan. Determination steps:
Look up Table 1 and find from the row where the supervision inspection level 1 is located, a
[n,] - [14, - 0. 440]
b. Judgment. Randomly select 11 units of products from the population, calculate the sample mean after testing, and the judgment rule is: Q = X 4 -X = 45 × 10
4 × 101
If QL≤-0. 440, then the supervision population is judged to be unpassable; if Q.≥-0. 440, then the supervision population is judged to be passable. 6.1.3 Given bilateral specification limits
Example: The standard size of a certain product is 100 mm, and the upper specification limit uu=100.1 mm and the lower specification limit uu=99. 9 mm of the population mean are specified as the supervision quality level, and the inspection level is the proof. Given that the population standard deviation a=0.3 m, try to determine the sampling plan. Determination steps:
Look up Table 2, and find out from the row where the supervision and inspection level certificate is located: Month 9
hCalculated shoots
100. 1 - 99. 9
u——
Look up Table 2 again, and find out k--0. 549 from the intersection of the column where the calculated value 2. 000 is located and the row where n=9 is located. So the sampling plan is:
[α,k]--[9. 0. 549]
d. Judgment. 9 units of products are randomly selected from the population, and the ladder mean X is calculated after testing. The judgment rule is: Qu = -
100.1— X
X-丝_X- 99.9
If Q:≤ -0.549 or Q1.≤—0. 549, the supervision population is judged to be unpassable; if Qu2—0. 549 and QL>—0.519, the supervision population is judged to be passable.
6.2*s\ method
6.2. 1 Given the upper specification limit
Example: The chemical composition of a raw material is S0, and the requirement is low base content. The supervisory quality level of the overall mean is specified to be Can=1.52%, and the supervisory inspection level is. Since there is no recent quality management or sample inspection data, the overall standard deviation cannot be estimated in advance. The "\ method" with unknown standard deviation is used to determine the sampling plan.
Determination steps:
CB/T14900—94
, check Table 3, and find from the row where the supervisory inspection level V is: [,_ = [8, — 0. 670]
b. Judgment. Randomly select 8 units of product from the population, determine S, and then calculate the sample mean X and sample standard deviation S after the content. The judgment rule is:
Qu ==X_ 1 52-X
If Qt is less than -0.670, the overall supervision is judged to be unacceptable; if Qu>-0.670, the overall supervision is judged to be acceptable. 6.2.2 Given the lower specification limit
Example: A certain steel material is required to have a high Rockwell hardness value. The overall mean supervision quality level is now determined to be 75, and the supervision inspection level is X. Since the overall standard deviation cannot be estimated in advance, the "S\ method" with unknown standard deviation is used to determine the polishing plan. Determination steps:
Check Table 3, and check the blue supervision inspection level X to obtain:
[m+]-_14,—0. 473]
Judge. Randomly select 14 units of products from the population, measure the hardness, calculate the sample mean and sample standard deviation S, and judge b.
The rule is:
XM-X-75
If Q is less than 0.473, then the overall supervision is judged to be unpassable: If QL>-0.473, then the overall supervision is judged to be passed. 6.2.3 Given bilateral specification limits
Example: Assume that the standard deviation of example 6.1.3 is unknown, and it is agreed that 6 is 0.35. Use the *s\ method for unknown standard deviation. Try to determine the sampling plan. Determination steps:
Look up Table 4, and find the following from the line where the inspection level VⅡ is: a
b. Calculate
MLF F41.
/ n - 1. 64
100,1 - 95.9
0. 35/ V9. 36
Then check Table 4. From the intersection of the column where the calculated value 1.748 is located and the row where n=11 is located, we can find that n=-0. 531, which is the sampling plan c.
[n,+] ..L11.—0. 5317
d. Judge. 11 units of products are randomly selected from the population. After testing, the sample mean is calculated and then compared with the sample standard S. The judgment rule is
Q=4=X_100.1-X
If Q:-0.581 or Q1,≤-0.531, then the blue warning population is judged to be unacceptable; if Qu>-0.531 and 2L--0.531, then the supervision population is judged to be acceptable.
Note: The sample cannot be agreed upon.The available sample standard difference 5 in the measurement step h can be replaced. The inspection power of the supervised sampling plan
GB/T 14900—94
After the supervised sampling plan is determined, the passing probability L(μ) can be calculated according to the calculation formula of the passing probability of the sampling plan given in Appendix A. The inspection power is:
1 - L(μ)
8Sampling table
Table 1 gives the sampling table of the single-sided specification limit of the \\ method; Table 2 gives the sampling table of the double-sided specification limit of the "\ method; Table 3 gives the sampling table of the single-sided specification limit of the *S\ method; Table 4 gives the sampling table of the double-sided specification limit of the "S\ method. 1“.\Method one-sided specification limit sampling table
Supervision inspection level
—0. 950
-- 0. 622
—0. 496
—0. 475
9E2'0—
096 0
C9T *E -bZxz.net
IIf o—
96*0—
029:0—
ts6 *0 -
~iss't
6Sf *0—
66t *0 -
929°0-
691'T-
29s*0-
1890—
889*0-
0st*0-
29*0—
299~0-
965*0-
889°0-||tt| |6260—
960—
209\0-
102 *0-
89 *0-
98f *0
619'0-
299*0—
928'0—
IIo'T-
96~0—
997*0-
229 'D -
918,0-
ETG'O-
29*0—||tt| |66# *0—
099*0—
902 *0—
EEG'O--
TTS'0-
929 0 —
-19"
98*0—
989 *0-
t89*0-
698:0-
599*0-
Tst'T-
S8t 'T-
Blue Supervisor Inspection Level
GB/T14900--94
Table 3 "S\ method one-sided specification limit sampling table
GB/T14900—94
769°0-
T98 '0-
1o 'T-
LoS\0-
0950—
869°0-
69 *0-
+1O -T -.
8++*0—
269:0—
692*0—
898-0—
509—
S29*0-
-158*0
96t *0-
2IS'0-||tt| |699*0-
T9*0—
889 *0—
tos 0
F990—
259'0-
6920—
228'0-
666 0-
t61 :1-
619'0-
69 0—
F06\0-
TE9\a-
at8'a -
Itg'n-
g'0—
lo1-t-
~900*0
219 '0 --
059*0 --
696 *0 -
sIt 't-
A1“α\method
GB/T14900—94
Appendix A
Calculation formula for the probability of passing the supervision sampling plan (supplement)
a. Upper specification (low plastic characteristic value)
L(p) = V(_
Lower specification limit (desired characteristic value is high)
L(m) = Vn("2-R))
Bilateral specification limit (desired characteristic value is within a certain range)
(u) = Vn(_ - k) vn(K -
A2"s\ method
For the convenience of application, the approximate calculation formula is listed: a.
Upper specification limit (desired characteristic value is low)
L(μ) = Φ
Lower specification limit (desired characteristic value is high)
L(μ) =
2(n — 1)
Bilateral specification limit (desired characteristic value is within a certain range)1
L(μ) =
2( 1)
2(n 1)
After the sampling plan is determined, the passing probability L() corresponding to the overall mean can be calculated by the above calculation formula.0—
692*0--
898-0—
509—
S29*0-
-158*0||tt| |29 *0-
96t *0-
2IS'0-
699*0-
T9*0—
889 *0—
tos 0
F990—
259'0-
6920—
228'0-
666 0-||tt ||t61 :1-
619'0-
69 0—
F06\0-
TE9\a-
at8'a -
Itg'n-
g'0—||tt ||lo1-t-
~900*0
219 '0 --
059*0 --
696 *0 -
sIt 't-
A1“α\method
GB/T14900—94
Appendix A
Calculation formula for the probability of passing the supervision sampling plan (supplement)||tt ||a, upper specification limit (low plastic property value)
L(p) = V(_
lower specification limit (high plastic property value)
L(m) =Vn("2-R))
bilateral specification limit (it is hoped that the characteristic value is within a certain range) class 2 drop
(u) =Vn(_ - k) vn(K -
A2“s\ method
For ease of application, the approximate calculation formula is listed: a.
Upper specification limit (desired low characteristic value)
L(μ ) = Φ
Lower specification limit (desired characteristic value is high)
L(μ) =
2(n — 1)
Both-sided specification limits (desired characteristic value is high) Value is within a certain range) 1
L(μ) =
2(=1)
2(n 1)
After the monitoring sample plan is determined , the passing probability L() corresponding to the population mean can be calculated by the above calculation formula.0—
692*0--
898-0—
509—
S29*0-
-158*0||tt| |29 *0-
96t *0-
2IS'0-
699*0-
T9*0—
889 *0—
tos 0
F990—
259'0-
6920—
228'0-
666 0-||tt ||t61 :1-
619'0-
69 0—
F06\0-
TE9\a-
at8'a -
Itg'n-
g'0—||tt ||lo1-t-
~900*0
219 '0 --
059*0 --
696 *0 -
sIt 't-
A1“α\method
GB/T14900—94
Appendix A
Calculation formula for the probability of passing the supervision sampling plan (supplement)||tt ||a, upper specification limit (low plastic property value)
L(p) = V(_
lower specification limit (high plastic property value)
L(m) =Vn("2-R))
bilateral specification limit (it is hoped that the characteristic value is within a certain range) class 2 drop
(u) =Vn(_ - k) vn(K -
A2“s\ method
For ease of application, the approximate calculation formula is listed: a.
Upper specification limit (desired low characteristic value)
L(μ ) = Φ
Lower specification limit (desired characteristic value is high)
L(μ) =
2(n — 1)
Both-sided specification limits (desired characteristic value is high) Value is within a certain range) 1
L(μ) =
2(=1)
2(n 1)
After the monitoring sample plan is determined , the passing probability L() corresponding to the population mean can be calculated by the above calculation formula.
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.