Method and system for identifying key quality control points in multi-station assembly process based on measurement data of different point locations

Abstract

The invention relates to the field of product quality control, in particular to a multi-station assembly process key quality control point identification method and system based on different point location measurement data. The method is a method for identifying key control points based on a generalized belief inference method and a process quality control principle. According to the method, a data generation process which is independent and identically distributed with observation data is constructed, and the data generation process comprises known distribution, a data generation structure function and a process capability index Cpk. In order to deduce the parameter, the solution of the parameter can be deduced based on the observation value, and the solution of the parameter can be understood as the generalized belief distribution of the parameter.

Description

Technical Field

[0001] The present application relates to the technical field of product quality control, in particular to a method and system for identifying key quality control points in a multi-station assembly process based on measurement data at different points. Background Art

[0002] With the development of industrial production automation and precision, higher requirements are put forward for the monitoring and control of product quality. Traditional quality control methods often rely on fixed control charts and basic process quality characteristics, and these quality characteristics may not have good frequency properties when making statistical inferences about the parameter distribution models of quality data with small and medium sample sizes. The process capability index can measure the stability and consistency of the product production process and can reflect the ability of production to meet the specified requirements. By calculating the process capability index, the performance of the product production process is evaluated whether it meets the requirements of design or customers, and it is judged whether the production process is stable and consistent.

[0003] Therefore, a method for statistical inference of process quality characteristics that can adapt to quality data with small and medium samples, accurately identify key control points, and effectively monitor product quality is needed. Summary of the Invention

[0004] The present invention provides a method and system for identifying key quality control points in a multi-station assembly process based on measurement data at different points, adopting a process capability index estimation technology based on generalized fiducial inference and combining with a statistical control chart to improve the accuracy and real-time performance of quality control. The specific technical solutions are as follows:

[0005] In a first aspect, the present invention provides a method for identifying key quality control points in a multi-station assembly process based on measurement data at different points. The method is applied to the statistical inference of process quality characteristics of small-sample quality data, and the method includes:

[0006] Generating a random variable with the same distribution based on a predefined quality data distribution, where the quality data distribution is the distribution state of product quality characteristics in known data, and the random variable with the same distribution is a random variable related to product quality characteristics;

[0007] Determining the parameter value under generalized fiducial inference based on the random variable with the same distribution;

[0008] Determining the process capability index based on the parameter value, and identifying the key control points based on the numerical range of the process capability index.

[0009] Further, before generating the random variable with the same distribution based on the predefined quality data distribution, it further includes:

[0010] Obtaining the measurement data of different points of the product to be evaluated during the assembly process, and preprocessing the measurement data.

[0011] Further, the preprocessing of the measurement data includes:

[0012] Determine the number of samples of the product to be evaluated and the control points of the samples;

[0013] Determine the mean value and variance of the quality data of each control point in the samples of the determined number of samples.

[0014] Further, the generation of the identically distributed random variables based on the predefined quality data distribution includes:

[0015] Define the quality data distribution based on the normal distribution, and the product quality characteristics follow the normal distribution;

[0016] Under the normal distribution condition, determine the identically distributed random variables based on the quality data distribution.

[0017] Further, the determination of the parameter values under the generalized fiducial inference based on the identically distributed random variables includes:

[0018] Determine the fiducial distributions of the parameters μ and σ based on the identically distributed random variables.

[0019] In a second aspect, the present invention also provides a multi-station assembly process key quality control point identification system based on measurement data at different points. The system is applied to the statistical inference of process quality characteristics of small-sample quality data, and the system includes: a definition and inference unit, a parameter determination unit, and a result analysis unit;

[0020] The definition and inference unit is used to generate identically distributed random variables based on a predefined quality data distribution. The quality data distribution is the distribution state of product quality characteristics in known data, and the identically distributed random variables are random variables related to product quality characteristics;

[0021] The parameter determination unit is used to determine the parameter values under the generalized fiducial inference based on the identically distributed random variables;

[0022] The result analysis unit is used to determine the process capability index based on the parameter values and identify the key control points based on the numerical range of the process capability index.

[0023] In another embodiment of the present invention, the system further includes: a preprocessing unit;

[0024] The preprocessing unit is used to obtain the measurement data at different points during the assembly process of the product to be evaluated and preprocess the measurement data.

[0025] In another embodiment of the present invention, the definition and inference unit is specifically used for:

[0026] Determine the sample quantity of the product to be evaluated and the control points of the samples;

[0027] Determine the mean value of the quality data and the variance of the quality data for each control point in the samples of the determined sample quantity.

[0028] In another embodiment of the present invention, the parameter determination unit is specifically configured to:

[0029] Define the quality data distribution based on the normal distribution, and the product quality characteristics follow the normal distribution;

[0030] Under the normal distribution condition, determine the identically distributed random variables based on the quality data distribution.

[0031] In another embodiment of the present invention, the parameter determination unit is further configured to:

[0032] Determine the belief distributions of parameter μ and parameter σ based on the identically distributed random variables.

[0033] The beneficial effects of the present invention are as follows:

[0034] The present invention provides a method and a system for identifying key quality control points in a multi-station assembly process based on measurement data at different positions. By adopting the process capability index estimation technology based on generalized belief inference and combining with statistical control charts, the accuracy and real-time performance of quality control are improved. It has the following advantages:

[0035] ① Based on the method of generalized belief inference of the present invention, point estimation is performed on the process capability index, thereby identifying key quality control points. The calculation is simple, and the evaluation accuracy is guaranteed for quality data with a small to medium sample size, reflecting its superiority in reliability evaluation of small samples.

[0036] ② The key quality control point identification method proposed by the present invention can be applicable to different types of quality data distributions by selecting appropriate pivot quantities, and the method has strong scalability. Description of the Drawings

[0037] Figure 1 is a flowchart of a method for identifying key quality control points in a multi-station assembly process based on measurement data at different positions;

[0038] Figure 2 is a diagram of the kernel density estimation of the measurement data at the first position. Detailed Embodiments

[0039] To make the objectives, technical solutions, and advantages of this specification clearer, the technical solutions of this application will be clearly and completely described below in conjunction with the specific embodiments of this application and their corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, rather than all of them. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of this document.

[0040] The following is a detailed description in conjunction with Figure 1-2 , of the technical solutions provided in each embodiment of this specification. Specific Embodiment 1:

[0041] To solve the problem of deviation in the estimation of process capability indices under small-sample measurement data at different points, the present invention proposes a method for identifying key quality control points in a multi-station assembly process based on small-sample measurement data at different points. This method uses a process capability index estimation technique based on generalized fiducial inference and combines it with a process quality analysis method to improve the accuracy and consistency of quality control.

[0042] To further illustrate the concept of the present invention, the following definitions are given:

[0043] 1. Data collection: Collect measurement data from multiple key points in the production process, including but not limited to key quality parameters such as temperature, pressure, and dimensions.

[0044] 2. The present invention is a method for identifying key quality control points in a multi-station assembly process based on measurement data at different points, and the following definitions need to be given:

[0045] Definition 1: When the output center μ does not coincide with the specification center (USL + LSL) / 2, the process capability index C pk , and is defined as

[0046]

[0047] where (LSL, USL) represents the upper and lower specification limits. The process capability index C pk is applicable to the state where the output center is offset from the specification center.

[0048] Definition 2: Assume that the random variable E exists in the space ε, and the distribution of E is known. There exists a function h(η, e) from the space Ω×ε to χ, where Ω and χ are the spaces where the random variables η and X are located respectively. For any η ∈ Ω,

[0049] X = h(η, E) (2)

[0050] Furthermore, for any observed values x ∈ χ and e ∈ ε, the equation x = h(η, e) has a unique solution ηx (e). Then η x (e) and θ(η x (e))'s probability distributions are respectively called the distributions of parameters η and θ.

[0051] 3. Based on the above definitions, the present invention is implemented through the following steps:

[0052] Step 1: Collect measurement data at different points

[0053] Collect measurement data from multiple key points during the multi-station assembly process, including but not limited to key quality parameters such as temperature, pressure, and dimensions. These data reflect the actual output of the process.

[0054] Step 2: Judge the distribution of multi-station quality data

[0055] Before calculating the quality model parameters, it is necessary to judge the distribution of multi-station quality data to construct a suitable belief distribution function model. For the medium and small sample size quality data measured at different stations, use the K-S test method to judge whether it follows a normal distribution. Establish a hypothesis test:

[0056]

[0057] Among them, H 0 represents the null hypothesis, and H 1 represents the alternative hypothesis.

[0058] Assume that the quality characteristic X follows a normal distribution N(μ,σ 2 ), where μ and σ respectively represent the mean and standard deviation of the normal distribution. Then the density function of the normal distribution is expressed as

[0059]

[0060] First, calculate the mean μ and standard deviation σ of the sample quality data. Calculate the empirical distribution function ECDF as

[0061]

[0062] where I(·) represents the indicator function. Calculate the cumulative distribution function CDF under the normal distribution as

[0063]

[0064] Further calculate the K-S statistic as

[0065] D n = sup x |F 0 (x) - CDF(x)|(7)

[0066] Finally, the critical value β of the one-sample K-S test statistic is referred to determine whether the sample quality data at different points follows a normal distribution: If D n <β, it means that the normal distribution is satisfied; otherwise, it is not satisfied.

[0067] Step 3: Construct the belief distribution of the parameters based on the distribution type of the quality data

[0068] Case 1: The multi-station quality data follows a normal distribution. Then the mean and standard deviation of the sample quality data {x 1 , x 2 ,..., x n} are respectively

[0069]

[0070] and

[0071]

[0072] Obviously, the mean and the variance are independent.

[0073] Under the normal distribution condition, generate identically distributed random variables. The random variable is defined as

[0074]

[0075] and where χ 2 (v) is the chi-square distribution with degree of freedom v.

[0076] Next, estimate the parameter values based on the generated variables. In order to obtain the belief distribution of the parameters μ and σ, transform formula (10) to get

[0077]

[0078] Case 2: The multi-station quality data follows an unknown distribution.

[0079] First, use kernel density estimation (KDE) to estimate the distribution of the data. Kernel density estimation is a non-parametric method that can estimate the probability density function without assuming the form of the data distribution. The formula is as follows:

[0080]

[0081] where, is the estimated density function, n is the sample size, h is the bandwidth parameter, and K(·) is the kernel function (usually the Gaussian kernel). The bandwidth h controls the width of the kernel function and affects the smoothness of the estimation. Here, the Gaussian kernel function is selected, and the formula is as follows:

[0082]

[0083] To represent the uncertainty in the data, a belief function is constructed based on the kernel density estimation results. Specifically, the kernel density estimation results are normalized to obtain the belief degree of each data point. The formula for the belief function is as follows:

[0084]

[0085] where m(B) is the basic probability assignment value. In this technical solution, it is simplified to directly use the kernel density estimation results as the belief function, that is:

[0086]

[0087] Thus, the relative belief degree of each data point is obtained for subsequent belief inference.

[0088] The belief function (16) and the generalized belief inference method are used to estimate the mean and standard deviation of the process. First, the mean and standard deviation of the quality data are calculated through the belief function:

[0089]

[0090] where x i is the sample quality data, and Bel(x i ) is the belief degree. Thus, through the processing method of kernel density estimation, the parameter estimation values are obtained under the condition that the multi-station quality data distribution is unknown.

[0091] Step Four: Calculate the inferred process capability index C pk

[0092] Based on the estimated mean and standard deviation, calculate the process capability index. When the multi-station quality data follows a normal distribution, substituting the result of equation (11) into equation (1) gives,

[0093]

[0094] When the multi-station quality data follows an unknown distribution, substituting the result of equation (17) into equation (1) gives,

[0095]

[0096] From equations (18) and (19), the process capability index obtained based on belief inference can be obtained, and further quality control procedures can be carried out.

[0097] Step Five: Process capability analysis and identification of key control points

[0098] A higher process capability index value indicates better process performance, being able to meet the specified requirements, and having smaller deviations and variances. On the contrary, a lower process capability index means poorer process performance, likely having larger deviations and variances, and being unable to stably meet the design or customer requirements. Calculate the process capability index C for the samples at each control point pk , and conduct process capability analysis:

[0099] ① When the value of C pk is greater than or equal to 1.33, it indicates that the process capability of this control point is sufficient, and the quality control is qualified, and it should not be listed as a key control point;

[0100] ② When the value of C pk is 1.00 - 1.33, it indicates that the process capability is poor and the quality control is not strict enough, and it should be listed as a secondary control point;

[0101] ③ When the value of C pk is less than 1.00, it indicates insufficient process capability, and strict quality control measures should be taken to ensure the consistency of assembly quality, so it is listed as a key control point.

[0102] Step Six: Method Verification

[0103] To demonstrate the superiority of the process capability index estimation technology based on fiducial inference in estimating quality data with a small sample size, this patent defines an average relative deviation to describe it, and the formula is as follows:

[0104]

[0105] where and respectively represent the process capability indices corresponding to the jth control point obtained from quality data with normal and small sample sizes, and m is the number of control points. The smaller the average relative deviation ε, the closer the estimated process capability index is to the actual value, and the better the estimation effect.

[0106] (3) Advantages and Innovations

[0107] The specific advantages of the present invention are as follows:

[0108] ① Based on the fiducial inference method of the present invention, point estimation is performed on the process capability index, thereby identifying key quality control points. The calculation is simple, and the evaluation accuracy is ensured for quality data with medium and small sample sizes, reflecting its superiority in small sample reliability evaluation.

[0109] ② The key quality control point identification method proposed by the present invention constructs a fiducial distribution of parameters based on the distribution types of quality data at different points, and can be applicable to different quality data distribution types. The method has strong expandability and estimation accuracy. Specific Example 2:

[0111] The present invention will be further described in detail below with reference to examples.

[0112] Step 1: Collect measurement data at different points

[0113] For a certain D product, measurement data at different points are measured during the assembly process. A total of 150 samples are taken. Taking the first 8 control points as an example, the data of the first 8 control points of one sample are shown in Table 1.

[0114] Table 1 Measurement data of a certain D product sample

[0115]

[0116]

[0117] Step 2: Judge the distribution of multi-station quality data

[0118] Conduct the following hypothesis test to judge the data distribution:

[0119]

[0120] Calculate the statistic D of the K-S test for the quality data distribution n , as shown in Table 2.

[0121] Table 2 Statistic of the K-S test

[0122] 1) 2) 3) 4) 5) 6) 7) 8) <![CDATA[D n > 0.117 0.139 0.124 0.059 0.089 0.083 0.075 0.067

[0123] Table 3 Critical value reference table of the single-sample K-S test statistic

[0124]

[0125] Look up the critical value reference table 3 of the single-sample K-S test statistic, and determine the critical value as According to the K-S test criterion, if D n < β, it means that it satisfies the normal distribution; otherwise, it does not. Therefore, it can be judged that the quality data measured from the 1st, 2nd, and 3rd points do not follow the normal distribution, and the remaining 5 points all satisfy.

[0126] Step 3: Construct the belief distribution of parameters according to the quality data distribution type

[0127] In the first case, assume that all multi-station quality data follow the normal distribution and conduct belief inference of parameters. Calculate the and variance of the quality data means of 150 samples for each control point according to Equations (8) and (9), as shown in Table 4.

[0128] Table 4 Quality data statistics

[0129]

[0130] Next, generate random variables with the same normal distribution according to Equation (10). The results are shown in Table 5.

[0131] Table 5 Random variables with the same distribution

[0132]

[0133] Next, based on the generated variables and the results calculated previously, estimate the mean μ and variance σ under fiducial inference using Equations (11) and (12). 2 , and the results are shown in Table 6.

[0134] Table 6 Estimated parameter values

[0135]

[0136] In the second case, assume that the multi-station quality data does not follow a normal distribution. In this case, further fiducial inference of the parameters is carried out. First, estimate the distribution of the data using kernel density estimation (KDE) with reference to Equations (13) and (14). The kernel density estimation of the measurement data at the first point is as shown Figure 2 by Figure 2 it can be concluded that it is unreasonable to estimate the parameters by defaulting to a normal distribution for the sample data at this point. Then, construct a fiducial function based on the kernel density estimation results with reference to Equations (15) and (16). Finally, estimate the mean and standard deviation of the process data according to Equation (17). The estimation results are shown in Table 7.

[0137] Table 7 Parameter values of kernel density estimation

[0138]

[0139] Step 4: Calculate the inferred process capability index C pk

[0140] Calculate the process capability index based on the mean and standard deviation estimated previously. The process capability index C is calculated respectively under the assumptions of normal distribution and non-normal distribution of the data by Equations (18) and (19). pk , and the results are summarized in Table 8.

[0141] Table 8 Process capability index C pk Summary

[0142]

[0143] Step 5: Process Capability Analysis and Identification of Critical Control Points

[0144] The process capability index C inferred based on belief pk is used for process capability analysis. Among them, for the 1st and 3rd control points, the C pk value is greater than 1.33, indicating that the process capability of this control point is sufficient and the quality control is qualified, so it should not be listed as a critical control point; for the 2nd, 6th, 7th, and 8th control points, the C pk value is between [1, 1.33], indicating that the process capability is poor and the quality control is not strict enough, so it should be listed as a secondary control point; for the 4th and 5th control points, the C pk value is less than 1.00, the process capability is seriously insufficient, and immediate improvement is required, so it is identified as a critical control point.

[0145] Observing Table 8, it can be found that if the distribution of quality data is not judged and it is defaulted to be a normal distribution for belief inference, the results of key point identification will be biased. Specifically, the 2nd control point will be listed as a critical control point. Therefore, it is necessary to construct the belief distribution of parameters based on the type of quality data distribution.

[0146] Step 6: Method Verification

[0147] To highlight the superiority of the process capability index estimation technology based on belief inference for quality data with small sample sizes at different positions, the quality data with small sample sizes of n = 50 and n = 25 are intercepted for further analysis. To reduce the randomness of single calculation, 5 calculation experiments are carried out respectively under different sample sizes, and the process capability index estimated after averaging 5 times and the average relative error are given. The comparison is shown in Table 9.

[0148] Table 9 Comparison Results of C Estimated by Belief Inference and Direct Calculation pk Comparison Results

[0149]

[0150] It can be clearly seen from Table 9 that when dealing with quality data with small sample sizes, the process capability index estimated by belief inference is closer to the actual value than direct calculation, and the average relative deviation ε rises more slowly as the sample size becomes smaller, and the estimation effect is better. Thus, it reflects the superiority of the process capability index estimation technology based on belief inference for quality data with small sample sizes at different positions. Therefore, this patent proves the effectiveness of the proposed method for identifying critical control points based on measurement data with small sample sizes at different positions, and has significant guiding significance for the quality control of multi-station assembly processes.

[0151] In summary, the present invention provides a method for identifying key quality control points in a multi-station assembly process based on measurement data at different points, which is a product quality control method based on the belief inference theory and the process quality control principle. The specific steps of the method are as follows: First, collect measurement data at different points; Second, judge the distribution of multi-station quality data; Third, construct the belief distribution of parameters according to the type of quality data distribution; Fourth, calculate the inferred process capability index C pk ; Fifth, process capability analysis and identification of key control points; Sixth, method verification. The present invention is applicable to the field of quality control based on measurement data of small and medium sample sizes at different points, and has strong practicability and operability. Specific Embodiment 3:

[0153] The present invention provides a method and a system for identifying key quality control points in a multi-station assembly process based on measurement data at different points, which adopt a process capability index estimation technology based on generalized belief inference and combine statistical control charts to improve the accuracy and real-time performance of quality control. The specific technical solutions are as follows:

[0154] In the first aspect, the present invention provides a method for identifying key quality control points in a multi-station assembly process based on measurement data at different points. The method is applied to the statistical inference of process quality characteristics of small sample quality data, and the method includes:

[0155] Generate a random variable with the same distribution based on a predefined quality data distribution, where the quality data distribution is the distribution state of product quality characteristics in known data, and the random variable with the same distribution is a random variable related to product quality characteristics;

[0156] Determine the parameter value under generalized belief inference based on the random variable with the same distribution;

[0157] Determine the process capability index based on the parameter value, and identify the key control points based on the numerical range of the process capability index.

[0158] Further, before generating the random variable with the same distribution based on the predefined quality data distribution, it further includes:

[0159] Obtain the measurement data of different points of the product to be evaluated during the assembly process, and preprocess the measurement data.

[0160] Further, preprocessing the measurement data includes:

[0161] Determine the sample quantity of the product to be evaluated and the control points of the sample;

[0162] Determine the mean value and variance of the quality data of each control point in the samples with the determined sample quantity.

[0163] Further, the generation of the identically - distributed random variables based on the predefined quality data distribution includes:

[0164] Define the quality data distribution based on the normal distribution, where the product quality characteristics follow the normal distribution;

[0165] Under the condition of the normal distribution, determine the identically - distributed random variables based on the quality data distribution.

[0166] Further, the determination of the parameter values under the generalized fiducial inference based on the identically - distributed random variables includes:

[0167] Determine the fiducial distributions of parameter μ and parameter σ based on the identically - distributed random variables.

[0168] In a second aspect, the present invention also provides a key quality control point identification system for the multi - station assembly process based on measurement data at different points. The system is applied to the statistical inference of process quality characteristics of small - sample quality data, and the system includes: a definition and inference unit, a parameter determination unit, and a result analysis unit;

[0169] The definition and inference unit is used to generate identically - distributed random variables based on the predefined quality data distribution. The quality data distribution is the distribution state of product quality characteristics in known data, and the identically - distributed random variables are random variables related to product quality characteristics;

[0170] The parameter determination unit is used to determine the parameter values under the generalized fiducial inference based on the identically - distributed random variables;

[0171] The result analysis unit is used to determine the process capability index based on the parameter values and identify the key control points based on the numerical range of the process capability index.

[0172] In another embodiment of the present invention, the system further includes: a pre - processing unit;

[0173] The pre - processing unit is used to obtain the measurement data of different points of the product to be evaluated during the assembly process and pre - process the measurement data.

[0174] In another embodiment of the present invention, the definition and inference unit is specifically used for:

[0175] Determine the sample quantity of the product to be evaluated and the control points of the samples;

[0176] Determine the mean value and variance of the quality data of each control point in the samples of the sample quantity.

[0177] In another embodiment of the present invention, the parameter determination unit is specifically used for:

[0178] Define the quality data distribution based on the normal distribution, and the product quality characteristics follow the normal distribution;

[0179] Under the normal distribution condition, determine the i.i.d. random variable based on the quality data distribution.

[0180] In another embodiment of the present invention, the parameter determination unit is further configured to:

[0181] Determine the belief distributions of the parameter μ and the parameter σ based on the i.i.d. random variable.

[0182] Example embodiments have been disclosed herein, and although specific terms are employed, they are used for and are to be interpreted only for general illustrative purposes and not for the purpose of limitation. In some embodiments, it will be apparent to those skilled in the art that, unless otherwise expressly specified, the features, features, and / or elements described in connection with a particular embodiment may be used alone or in combination with the features, features, and / or elements described in connection with other embodiments. Accordingly, those skilled in the art will understand that various forms and details may be changed without departing from the scope of the invention as set forth in the appended claims.

Claims

1. A method for identifying critical quality control points in a multi-station assembly process based on measurement data at different points, which is applied to statistical inference of process quality characteristics of small sample quality data, and is characterized in that: The method comprises: Generate identically distributed random variables based on a predefined quality data distribution, wherein the quality data distribution is a distribution state of product quality characteristics in known data, and the identically distributed random variables are random variables related to the product quality characteristics; Determining parameter values ​​under generalized belief inference based on the identically distributed random variables; A process capability index is determined based on the parameter value, and a critical control point is identified based on a range of values ​​of the process capability index.

2. The method for identifying critical quality control points in a multi-station assembly process based on different point measurement data according to claim 1, characterized in that: Before generating the same distribution random variables based on the predefined quality data distribution, the method further includes: Acquire measurement data of different points of the product to be evaluated during the assembly process, and preprocess the measurement data.

3. The method for identifying critical quality control points in a multi-station assembly process based on different point measurement data as claimed in claim 2, characterized in that: Preprocessing the measurement data includes: Determine the number of samples of the product to be evaluated and the control points of the samples; The quality data mean and quality data variance of each control point in the sample number are determined.

4. The method for identifying critical quality control points in a multi-station assembly process based on different point measurement data according to claim 1, characterized in that: The generating of identically distributed random variables based on a predefined quality data distribution comprises: Defining quality data distribution based on normal distribution, wherein the product quality characteristics obey normal distribution; In a normal distribution condition, the identically distributed random variables are determined based on the mass data distribution.

5. The method for identifying critical quality control points in a multi-station assembly process based on different point measurement data according to claim 1, characterized in that: Determining the parameter value under generalized belief inference based on the identically distributed random variables includes: The belief distribution of the parameters μ and σ is determined based on the identically distributed random variables.

6. A critical quality control point identification system for a multi-station assembly process based on measurement data at different points. The system is applied to statistical inference of process quality characteristics of small sample quality data, and is characterized by: The system comprises: a definition and estimation unit, a parameter determination unit, and a result analysis unit; The definition inference unit is used to generate identically distributed random variables based on a predefined quality data distribution, wherein the quality data distribution is a distribution state of product quality characteristics in known data, and the identically distributed random variables are random variables related to product quality characteristics; The parameter determination unit is used to determine the parameter value under generalized belief inference based on the identically distributed random variables; The result analysis unit is used to determine a process capability index based on the parameter value, and to identify a critical control point based on a numerical range of the process capability index.

7. The system for identifying critical quality control points in a multi-station assembly process based on different point measurement data as claimed in claim 6, characterized in that: The system further comprises: a pre-processing unit; The preprocessing unit is used to obtain measurement data of different points of the product to be evaluated during the assembly process, and preprocess the measurement data.

8. The system for identifying critical quality control points in a multi-station assembly process based on different point measurement data as claimed in claim 7, characterized in that: The definition inference unit is specifically used for: Determine the number of samples of the product to be evaluated and the control points of the samples; The quality data mean and quality data variance of each control point in the sample number are determined.

9. The system for identifying critical quality control points in a multi-station assembly process based on different point measurement data as claimed in claim 6, characterized in that: The parameter determination unit is specifically used for: Defining quality data distribution based on normal distribution, wherein the product quality characteristics obey normal distribution; In a normal distribution condition, the identically distributed random variables are determined based on the mass data distribution.

10. The system for identifying critical quality control points in a multi-station assembly process based on different point measurement data according to claim 6, characterized in that: The parameter determination unit is also used for: The belief distribution of the parameters μ and σ is determined based on the identically distributed random variables.

Images