A power system material sampling inspection management method and device

Abstract

The application discloses a kind of power system material sampling management method and device, method includes: obtaining the list of the power system material to be sampled, the historical quality data of the material to be sampled is retrieved from database;According to first preset parameter and second preset parameter, the failure function of the material to be sampled is established;The distribution of failure function is predicted by Bayesian network, and the historical quality data of the material to be sampled is imported, the posterior distribution of first preset parameter and second preset parameter is determined, and the average value of first preset parameter and second preset parameter is generated as quality result;According to the numerical information of first preset parameter and second preset parameter, the material to be sampled of different manufacturers is counted and sorted, and the sorting result is generated;According to sorting result and preset sampling proportion coefficient, the material to be sampled of different manufacturers is sampled, and the sampling result is generated, to realize the differentiating sampling of the material to be sampled, improve the precision of material sampling.

Description

Technical Field

[0001] This invention relates to the field of material sampling and management technology, and in particular to a method, device and storage medium for material sampling and management in power systems. Background Technology

[0002] In recent years, with the rapid development of China's economy, electricity demand has increased dramatically, leading to a growing need for power grid construction. The safe, stable, and efficient operation of the power grid depends on the reliability of power supply materials. Therefore, power supply material management must accurately identify quality defects, prevent quality accidents, and improve equipment reliability. Sampling inspection is a crucial tool for product quality control; effective sampling inspection strategies can reduce the cost and time of quality inspection and provide statistical control over the quality of shipped products. However, currently, most power grid material sampling inspection strategies are determined based on experience, lacking clear theoretical support. Material sampling inspection is a vital link in the supply chain, and reliable power grid materials are the foundation for the safe and stable operation of the power grid.

[0003] Currently, there are many problems with the technology for quality inspection and quality control management of power grid materials:

[0004] (1) The sampling inspection work is not standardized and lacks relevant theoretical and standard support. The traditional national standard GB / T13264-2008 sampling theory is not applicable to the high cost and destructive nature of power grid material testing.

[0005] (2) Lack of analysis and mining of historical data on sampled materials. There is a lack of relevant standards for rigorous sampling of categories and suppliers with prominent quality problems.

[0006] (3) The differentiated management strategy for different sampled materials is not reflected. The sampling inspection upon arrival usually emphasizes the comprehensiveness of the sampled materials, and the distribution of the number of materials tested across all categories is unreasonable.

[0007] (4) The failure to take into account the impact of multiple factors such as sampling costs and losses caused by non-conforming products makes it impossible to maximize the return on investment in quality control resources. Summary of the Invention

[0008] This invention provides a method and apparatus for managing the random inspection of materials in a power system. By analyzing the historical quality data of the materials to be inspected, and performing differentiated random inspections based on the analysis results, the accuracy of material random inspection is improved.

[0009] To improve the accuracy of material sampling inspection, this invention provides a method for managing material sampling inspection in a power system, comprising: obtaining a list of materials to be sampled in the power system, and retrieving historical quality data of the materials to be sampled from a database according to the category of the materials to be sampled; the historical quality data includes historical fault data and historical truncated data of commissioning data; the historical truncated data includes the operating time of the materials to be sampled that have not yet experienced a fault.

[0010] Based on the first preset parameter and the second preset parameter, a fault function for the material to be inspected is established; wherein the distribution of the first preset parameter conforms to the Gamma distribution and the distribution of the second preset parameter conforms to the normal distribution; the distribution of the fault function is predicted by a Bayesian network, and historical quality data of the material to be inspected is imported to determine the posterior distribution of the first preset parameter and the second preset parameter; based on the posterior distribution, the average value of the first preset parameter and the second preset parameter is generated as the quality result.

[0011] Based on the quality results, numerical information of the first preset parameter and the second preset parameter is generated; based on the numerical information, the materials to be inspected from different manufacturers are statistically analyzed and sorted to generate a sorting result; based on the sorting result and a preset sampling ratio coefficient, the materials to be inspected from different manufacturers are sampled and inspected to generate an inspection result.

[0012] As a preferred embodiment, this invention uses historical quality data of the materials to be inspected to predict the fault distribution of the materials through a Bayesian network, generating quality results. Based on these quality results, materials from different manufacturers are statistically analyzed and ranked, and then sampled using a preset sampling ratio coefficient. By using different sampling ratios to sample materials of different qualities, differentiated sampling of the materials is achieved, improving the accuracy of material sampling.

[0013] As a preferred embodiment, before obtaining the list of materials to be inspected from the power system and retrieving historical quality data of the materials to be inspected from the database according to their categories, the process further includes:

[0014] Obtain material information for all primary materials, classify the material information, generate corresponding category tags, and store them in the database; the material information includes: material list, historical quality data, annual sampling inspection results data, and warehousing information list;

[0015] Based on the material information, calculate the first failure rate, the first defect rate, and the first inspection pass rate of the materials to be inspected, and generate the first correlation data between the first failure rate, the first defect rate, the first inspection pass rate, and the quantity of materials entering the warehouse.

[0016] As a preferred embodiment, this invention collects data reflecting the quality of power grid materials, establishes a full life-cycle quality database, classifies and processes material information, and assigns specific label information to all categories of materials and suppliers. This enables the rapid retrieval of quality inspection data and operational quality data of materials to be inspected from the database for automatic statistical and correlation analysis, thereby improving the efficiency of material sampling inspection.

[0017] As a preferred embodiment, a fault function for the materials to be inspected is established based on the first preset parameter and the second preset parameter, specifically as follows:

[0018] Let the failure occurrence time of the materials to be inspected follow a two-parameter Weibull distribution, and calculate the probability density function and failure function of the materials to be inspected:

[0019]

[0020]

[0021] Where f(t) is the fault probability density function, F(t) is the fault function, t is the fault occurrence time, the first preset parameter is β, which is a shape parameter used to represent the dispersion of the fault distribution, and the second preset parameter is η, which is a scale parameter used to represent the average fault life of the material.

[0022] As a preferred embodiment, this invention utilizes the fact that the failure occurrence time of the materials to be inspected follows a two-parameter Weibull distribution. By calculating the probability density function and failure function of the materials, the operational quality of the materials is quantified based on the failure distribution. The failure distribution of the materials is determined according to the shape and scale parameters in the formula, thereby assessing the operational quality risk. This invention predicts and analyzes the operational quality risk of the materials to be inspected, thus enabling differentiated sampling and improving the accuracy of material sampling.

[0023] As a preferred embodiment, the distribution of the first preset parameter conforms to a Gamma distribution and the distribution of the second preset parameter conforms to a normal distribution, specifically:

[0024]

[0025] Where a and b are constants reflecting the β distribution, and x is a variable;

[0026]

[0027] Where μ and σ 2 These are the mean and variance of η, respectively.

[0028] As a preferred embodiment, this invention, based on historical statistical experience, determines that the distribution of the scale parameter β conforms to a Gamma distribution, and its value represents the dispersion of the fault distribution; the distribution of the shape parameter η conforms to a normal distribution, and its value represents the average fault life of the materials. By judging the fault distribution of the materials to be inspected based on the shape and scale parameters, the operational quality risk of the materials can be assessed, thereby enabling differentiated sampling inspection of the materials and improving the accuracy of material sampling inspection.

[0029] As a preferred embodiment, before predicting the distribution of the fault function using a Bayesian network and importing historical quality data of the materials to be inspected to determine the distribution probabilities of the first and second preset parameters, the following steps are also included:

[0030] Based on the probability density function and failure function of the materials to be inspected, a likelihood function for historical quality data is constructed. n sample data points are extracted from the operational data of the materials to be inspected. These sample data points include r historical failure data points and (nr) historical truncated data points. The operational time of the historical failure data points or historical truncated data points of the materials to be inspected is t. i The likelihood function of the material is obtained based on the parameters η and β to be determined as follows:

[0031]

[0032] In the formula, σ i The truncation indicator variable takes the value 0 when the i-th sample data is historical truncation data, and takes the value 1 when the i-th sample data is historical fault data; L(D|β,η) is the likelihood function, where D represents the sample status of the historical quality data of the materials to be inspected.

[0033] As a preferred embodiment, the present invention constructs a likelihood function based on the probability density function and fault function of the materials to be inspected, and then imports the historical quality data of the materials to be inspected into the likelihood function, so as to calculate the posterior distribution of shape parameters and scale parameters in the form of the likelihood function, and judge the operational quality risk of the materials.

[0034] As a preferred approach, a Bayesian network is used to predict the distribution of the fault function, and historical quality data of the materials to be inspected are imported to determine the posterior distributions of the first and second preset parameters, specifically:

[0035] Based on the likelihood function, the distribution of the fault function is predicted using a Bayesian network, and the posterior distributions of η and β are calculated:

[0036]

[0037] Where π(β,η) is the joint distribution of η and β, and f(β,η|D) is the posterior distribution of η and β.

[0038] As a preferred embodiment, the present invention calculates the posterior distribution of η and β through a Bayesian network based on the constructed likelihood function, determines the distribution of η and β, obtains the joint distribution of η and β, and then judges the fault status of the materials to be inspected.

[0039] As a preferred embodiment, the average value of the first preset parameter and the second preset parameter is generated as the quality result based on the posterior distribution, specifically as follows:

[0040] Based on the joint distribution of η and β in the posterior distribution, calculate the average values ​​of η and β respectively:

[0041]

[0042]

[0043] Where E(η) is the average value of η, and E(β) is the average value of β.

[0044] As a preferred embodiment, the present invention calculates the average values ​​of η and β based on the posterior distribution, where η represents the characteristic value of the material's service life. The larger the average value of η, the lower the operational risk and the higher the quality of the material. β represents the dispersion of the fault distribution. The smaller the value, the more dispersed the defect distribution. The larger the value of β, the more concentrated the fault distribution. Based on η and β, the fault distribution of the material to be inspected is determined, and the operational quality risk of the material is judged. This enables differentiated sampling inspection of the material to be inspected, thereby improving the accuracy of material sampling inspection.

[0045] As a preferred option, random inspections are conducted on materials from different manufacturers to be inspected, and inspection results are generated, specifically as follows:

[0046] If the number of non-compliant materials to be inspected is not less than the first threshold, then a non-compliant inspection result will be generated.

[0047] If the number of non-compliant items in the sampling inspection is 0, then a sampling inspection result of passing inspection will be generated.

[0048] If the number of non-compliant materials to be inspected is lower than the first threshold and not equal to 0, then a preset number of materials to be inspected will be added.

[0049] If the number of non-conforming items in the pre-set sample of materials to be inspected is 0, then a sample inspection result of passing inspection will be generated.

[0050] If the number of non-compliant items in the pre-set sampled materials is not less than the second threshold, then a sampling result indicating that the sampling failed will be generated.

[0051] As a preferred embodiment, this invention prioritizes the materials to be inspected based on their operational quality risks and uses different sampling coefficients to inspect materials of varying quality, thus achieving differentiated sampling. Furthermore, based on this differentiated sampling, a preset pass / fail standard is used to adjust the threshold for the number of non-conforming materials according to their quality status, enabling preliminary sampling. The preliminary sampling results are then analyzed to determine whether additional sampling of some non-conforming materials is necessary, thereby improving the accuracy of the material sampling.

[0052] As a preferred option, a sampling pass rate is generated based on the sampling results; the numerical information of the second preset parameter is used as the material risk prediction value, and the second correlation data between the material risk prediction value and the sampling pass rate is calculated.

[0053] As a preferred approach, this invention evaluates the accuracy of quality risk prediction and the effect of differentiated sampling of materials by using correlation data between the material quality risk prediction results and the pass rate of random inspections. The closer the correlation data is to 1, the more accurate the risk prediction and the more significant the effect of differentiated sampling.

[0054] Accordingly, the present invention also provides a power system material sampling inspection management device, comprising: a data acquisition module, a quality prediction module, and a sampling inspection module;

[0055] The data acquisition module is used to acquire a list of materials to be inspected in the power system, and retrieve historical quality data of the materials to be inspected from the database according to the category of the materials to be inspected; the historical quality data includes historical fault data and historical truncated data of the commissioning data; the historical truncated data includes the operating time of the materials to be inspected that have not yet experienced a fault.

[0056] The quality prediction module is used to establish a fault function of the material to be inspected based on a first preset parameter and a second preset parameter; wherein the distribution of the first preset parameter conforms to a Gamma distribution and the distribution of the second preset parameter conforms to a normal distribution; the distribution of the fault function is predicted by a Bayesian network, and historical quality data of the material to be inspected is imported to determine the posterior distribution of the first preset parameter and the second preset parameter, and the average value of the first preset parameter and the second preset parameter is generated as the quality result based on the posterior distribution;

[0057] The sampling inspection module is used to generate numerical information of the first preset parameter and the second preset parameter based on the quality results; to statistically analyze and sort the materials to be inspected from different manufacturers based on the numerical information, and generate a sorting result; and to conduct sampling inspections on the materials to be inspected from different manufacturers based on the sorting result and a preset sampling ratio coefficient, and generate sampling inspection results.

[0058] As a preferred embodiment, the quality prediction module of this invention predicts the fault distribution of the materials to be inspected using a Bayesian network based on historical quality data, generating quality results. The sampling module then uses these quality results to statistically analyze and rank the materials from different manufacturers, and performs sampling inspections based on preset sampling ratios. By using different sampling ratios to inspect materials of varying quality, differentiated sampling is achieved, improving the accuracy of material sampling.

[0059] Accordingly, the present invention also provides a computer-readable storage medium, the computer-readable storage medium including a stored computer program; wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform a power system material sampling management method as described in the present invention. Attached Figure Description

[0060] Figure 1 This is a flowchart illustrating an embodiment of a method for random inspection and management of materials in a power system provided by the present invention.

[0061] Figure 2 This is a schematic diagram showing the distribution of the first preset parameter β in an embodiment of a power system material sampling management method provided by the present invention;

[0062] Figure 3 This is a schematic diagram showing the distribution of the second preset parameter η in one embodiment of a power system material sampling management method provided by the present invention;

[0063] Figure 4 This is a schematic diagram of the joint distribution of η and β in an embodiment of a power system material sampling management method provided by the present invention;

[0064] Figure 5 This is a schematic diagram of an embodiment of a power system material sampling and management device provided by the present invention. Detailed Implementation

[0065] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] Example 1

[0067] Please refer to Figure 1 The present invention provides a method for random inspection and management of power system materials, comprising steps S101-S103:

[0068] Step S101: Obtain the list of materials to be inspected in the power system, and retrieve the historical quality data of the materials to be inspected from the database according to the category of the materials to be inspected; the historical quality data includes historical fault data and historical truncated data of the commissioning data; the historical truncated data includes the running time of the materials to be inspected that have not yet experienced a fault.

[0069] In this embodiment, before obtaining the list of materials to be inspected from the power system and retrieving historical quality data of the materials to be inspected from the database according to their categories, the method further includes:

[0070] Obtain material information for all primary materials, classify the material information, generate corresponding category tags, and store them in the database; the material information includes: material list, historical quality data, annual sampling inspection results data, and warehousing information list;

[0071] Based on the material information, calculate the first failure rate, the first defect rate, and the first sampling pass rate of the first material, and generate the first correlation data between the first failure rate, the first defect rate, the first sampling pass rate, and the quantity of goods entering the warehouse.

[0072] Furthermore, in this embodiment, the material information of all first materials is obtained, and the material information is classified and processed to generate corresponding category tags and store them in the database, specifically as follows:

[0073] Based on the list of primary materials in the power system, establish a full-category list of materials and assign corresponding information to the material names and supplier names in the list; obtain relevant fault information and defect lists of primary materials from the power grid fault and defect statistics list; obtain the power grid material sampling inspection list of primary materials over the years from the power grid sampling inspection units; obtain the power grid material warehousing information list of primary materials over the years.

[0074] Extract quality information such as the type of failure, production year, cause of failure, time of failure, and manufacturer of the primary materials from the list of material failures and defects over the years; extract sampling information such as the type of materials sampled, supplier, quantity sampled, results sampled, and sampling ratio of the primary materials from the material sampling inspection list; extract material purchase information of the primary materials from the power grid material warehousing list, and determine warehousing information such as material type, supplier information, warehousing time, quantity sampled, and amount sampled.

[0075] The quality information of primary materials is categorized and processed to establish a complete list of primary materials. Based on the importance of the materials, the complete list is divided into "critical control materials" and "routine control materials." The materials are further categorized into three levels of directories, from broad to narrow. The first level includes primary equipment, intelligent equipment, and automated equipment; the second level includes equipment categories such as transformers, cables, and surge arresters; and the third level includes equipment specifications such as 10kV pole-mounted load switches and 10kV pole-mounted circuit breakers. All primary materials' sampling inspection information, operational quality information, production information, and warehousing information are compiled and categorized, then organized into their respective information tags and stored in the database.

[0076] In this embodiment, based on the contents of the materials database, information such as the failure rate, defect rate, and sampling pass rate of the first material are statistically analyzed, and the calculation formula is shown below:

[0077]

[0078] In the formula: κ is the material failure rate, defect rate, or sampling pass rate, n is the corresponding number of failures, defects, or sampling pass numbers, and N is the total amount of materials.

[0079] Establish a correlation analysis of material failure, defect, sampling inspection, and warehousing data to obtain the correlation information between material failure rate, defect rate, material sampling pass rate, and warehousing quantity. The specific expressions are as follows:

[0080]

[0081] In the formula, m is the amount of data entered into the database, and x i y is the correlation value of any two of the four data points: material failure rate, defect rate, material sampling pass rate, and quantity received. i Let y be the correlation value of any two of the four data points: material failure rate, defect rate, material sampling pass rate, and quantity received. i The two selected data points and x i The two selected data points are not exactly the same. In this case, the Pearson correlation coefficient P(x,y) ranges from -1 to 1. The closer the correlation coefficient is to 1, the higher the correlation, indicating a better correlation between the two data points in the correlation analysis of material failure rate, defect rate and material sampling pass rate, and quantity entering the warehouse.

[0082] In this embodiment, the first material includes all materials to be inspected. After obtaining the list of materials to be inspected in the power system, based on the category of the materials to be inspected, historical quality data of the materials to be inspected is retrieved from the data of different labels of all the first materials recorded in the database. This includes the number of materials entering the warehouse, the year of entry, the number of failures, the time of failure, the number of inspections, and the inspection ratio, etc., to generate historical failure data and historical truncated data. The historical truncated data includes the operating time of the materials to be inspected that have not yet experienced failures.

[0083] Step S102: Based on the first preset parameter and the second preset parameter, establish the fault function of the material to be inspected; wherein the distribution of the first preset parameter conforms to the Gamma distribution and the distribution of the second preset parameter conforms to the normal distribution; predict the distribution of the fault function through a Bayesian network, import the historical quality data of the material to be inspected, determine the posterior distribution of the first preset parameter and the second preset parameter, and generate the average value of the first preset parameter and the second preset parameter as the quality result based on the posterior distribution;

[0084] In this embodiment, a fault function for the material to be inspected is established based on a first preset parameter and a second preset parameter, specifically as follows:

[0085] Let the failure occurrence time of the materials to be inspected follow a two-parameter Weibull distribution, and calculate the probability density function and failure function of the materials to be inspected:

[0086]

[0087]

[0088] Where f(t) is the failure probability density function, F(t) is the failure function, t is the failure occurrence time, the first preset parameter is β, which is a shape parameter used to represent the dispersion of the failure distribution, and the second preset parameter is η, which is a scale parameter used to represent the average failure life of the materials. The distribution of material failures is affected by the parameters η and β, thus determining the operational quality risk of the materials.

[0089] In this embodiment, the distribution of the first preset parameter conforms to a Gamma distribution and the distribution of the second preset parameter conforms to a normal distribution, specifically:

[0090]

[0091] Where a and b are constants reflecting the β distribution, and x is a variable;

[0092]

[0093] Where μ and σ 2These are the mean and variance of η, respectively.

[0094] In this embodiment, based on the statistical results of historical operational quality data, with a set to 1 and b set to 2 respectively, the β distribution is obtained as: β~Gamma(x,1,2)=xe x Its distribution results are as follows Figure 2 As shown.

[0095] Similarly, based on historical statistical information, let μ and σ 2 When the values ​​are 40 and 0.1 respectively, the distribution of η is as follows: Its distribution results are as follows Figure 3 As shown.

[0096] In this embodiment, before predicting the distribution of the fault function using a Bayesian network and importing historical quality data of the materials to be inspected to determine the distribution probabilities of the first preset parameter and the second preset parameter, the method further includes:

[0097] Based on the probability density function and failure function of the materials to be inspected, a likelihood function for historical quality data is constructed. n sample data points are extracted from the operational data of the materials to be inspected. These sample data points include r historical failure data points and (nr) historical truncated data points. The operational time of the historical failure data points or historical truncated data points of the materials to be inspected is t. i The likelihood function of the material is obtained based on the parameters η and β to be determined as follows:

[0098]

[0099] In the formula, σ i The truncation indicator variable takes the value 0 when the i-th sample data is historical truncation data, and takes the value 1 when the i-th sample data is historical fault data; L(D|β,η) is the likelihood function, where D represents the sample status of the historical quality data of the materials to be inspected.

[0100] Since η and β are linked to operational quality data, their actual values ​​cannot be directly calculated. Therefore, the distribution probability of η and β can be analyzed by predicting the distribution of the fault function using a Bayesian network. Based on the prior distributions of η and β, the actual distribution probability of both is determined using a Bayesian network, where the Bayesian network is:

[0101]

[0102] Among them, H j It describes the distribution of the prior probability, H. i To test the observation time, E is the test result, P(X) is the probability of X, and X is H. i or H jP(A|B) is the probability that A occurs given that B has occurred; A is either E or H. j B is E, H i or H j ;

[0103] In this embodiment, the distribution of the fault function is predicted using a Bayesian network, and historical quality data of the materials to be inspected is imported to determine the posterior distribution of the first preset parameter and the second preset parameter, specifically as follows:

[0104] Based on the likelihood function, the distribution of the fault function is predicted using a Bayesian network, and the posterior distributions of η and β are calculated:

[0105]

[0106] Where π(β,η) is the joint distribution of η and β, and its distribution result is as follows: Figure 4 As shown; f(β,η|D) is the posterior distribution of η and β.

[0107] In this embodiment, based on the posterior distribution, the average value of the first preset parameter and the second preset parameter is generated as the quality result, specifically as follows:

[0108] Based on the joint distribution of η and β in the posterior distribution, calculate the average values ​​of η and β respectively:

[0109]

[0110]

[0111] Where E(η) is the average value of η, and E(β) is the average value of β.

[0112] In this embodiment, the average values ​​of η and β are calculated to express the quality status of the materials. η represents the characteristic value of the material's service life. The larger the average value of η, the lower the operational risk and the higher the quality of the materials. β represents the dispersion of the fault distribution. The smaller the value, the more dispersed the defect distribution. Since faults may exist throughout the entire service life, the quality status of the materials is determined first based on the value of η, making it easier to take efficient countermeasures against fault risks. Therefore, when η is consistent, the larger β is, the lower the risk of material loss and the higher the quality status.

[0113] Step S103: Based on the quality results, generate numerical information of the first preset parameter and the second preset parameter; statistically analyze and sort the materials to be inspected from different manufacturers based on the numerical information, and generate sorting results; based on the sorting results and the preset sampling ratio coefficient, conduct sampling inspections on the materials to be inspected from different manufacturers, and generate sampling inspection results.

[0114] In this embodiment, based on the quality results, a dataset reflecting the quality of each category of materials in the sampled materials is established. The numerical information of η and β is statistically analyzed, where η is a characteristic value reflecting the years of material operation quality information. The η values ​​of different manufacturers for each category of materials are statistically analyzed and sorted from high to low material quality, i.e., the η values ​​are sorted from high to low. For example, if there are n1 production suppliers for the first category of materials, there are n2 production methods for the first category of materials, resulting in n2 sorted η values. β is a dispersion index, an auxiliary parameter obtained synchronously. η and β need to be calculated synchronously, and the information of β is stored in the database as auxiliary information.

[0115] The distribution of sampling inspection ratios for materials is as follows:

[0116] Where P is the average sampling rate, m is the number of historical samplings, and n and N are the number of samplings and the total number of materials, respectively. The average sampling rate of this type of material within the numerical range is obtained through statistical analysis.

[0117] Each sorted η value is assigned a preset sampling ratio coefficient, and its initial sampling ratio is: P1=k×P; where: k is the preset sampling ratio coefficient, and P1 is the initial sampling ratio.

[0118] For materials ranked in the top 10% by risk quality (i.e., materials with n2 η values ​​ranking in the top 10%), k = 0.8 is set; for materials ranked in the 10%-80% risk quality range, k = 1 is set; and for materials ranked in the bottom 20% risk quality range, k = 1.2 is set to achieve differentiated sampling inspection of material quality.

[0119] In this embodiment, a sampling ratio coefficient of k>1 is assigned to data with poor quality (ranked in the bottom 20% of risk quality) to ensure the quality level of materials and reduce the risk of use; for data with good historical quality analysis results (ranked in the top 10% of risk quality), a ratio coefficient of k<1 is assigned to reduce unnecessary work redundancy and improve work efficiency.

[0120] In this embodiment, random inspections are conducted on materials from different manufacturers to generate inspection results, specifically as follows:

[0121] If the number of non-compliant materials to be inspected is not less than the first threshold, then a non-compliant inspection result will be generated.

[0122] If the number of non-compliant items in the sampling inspection is 0, then a sampling inspection result of passing inspection will be generated.

[0123] If the number of non-compliant materials to be inspected is lower than the first threshold and not equal to 0, then a preset number of materials to be inspected will be added.

[0124] If the number of non-conforming items in the pre-set sample of materials to be inspected is 0, then a sample inspection result of passing inspection will be generated.

[0125] If the number of non-compliant items in the pre-set sampled materials is not less than the second threshold, then a sampling result indicating that the sampling failed will be generated.

[0126] In this embodiment, materials that pass all random inspections are deemed qualified and directly approved.

[0127] When the number of non-conforming samples is 1, a suitable additional sampling plan is formulated; if the number of non-conforming samples after the additional sampling is 0, the sampling is deemed to have passed; when the number of non-conforming samples after the additional sampling is greater than or equal to 1, the sampling is deemed to have failed.

[0128] If, during the initial sampling inspection, the number of non-compliant items is greater than or equal to 2, the materials are directly deemed to be of substandard quality, no additional sampling strategy will be formulated, and the batch of materials will be directly rejected.

[0129] In this embodiment, a sampling pass rate is generated based on the sampling results; the second correlation data between the material risk prediction value and the sampling pass rate is calculated based on the numerical information of the second preset parameter as the material risk prediction value.

[0130] Furthermore, after the sampling inspection, data such as the category of materials to be inspected, supplier information, production date of materials, sampling date, sampling time, sampling cost, total quantity of materials, and additional sampling plan information are collected and imported into the database.

[0131] A correlation analysis was performed on the statistical results to analyze the correlation between the predicted material quality risks and the sampling pass rate. The statistical scheme is shown in the formula:

[0132]

[0133] In the formula, m represents the amount of data on the incoming materials to be inspected, and x represents the amount of data on the incoming materials. i Let η be the predicted value of material risk, and y be the predicted value of material risk. i P(x,y) represents the pass rate of the materials to be inspected. The larger the value of P(x,y), the higher the correlation coefficient P(x,y) is. The Pearson correlation coefficient P(x,y) ranges from -1 to 1. The closer the correlation coefficient is to 1, the higher the correlation, indicating that the correlation between the predicted risk value of the materials and the pass rate of the inspection is better. The closer the correlation between the predicted risk value of the materials and the pass rate of the inspection is to 1, the more accurate the risk prediction and the more significant the effect of differentiated inspection.

[0134] Based on the results of the random inspection, a conclusion is drawn on the quality of the materials, and the information is sent to the corresponding power grid unit and suppliers. For materials that fail the initial random inspection, improvement requirements are put forward to the suppliers to ensure the quality of the next random inspection.

[0135] Implementing the embodiments of the present invention has the following effects:

[0136] This invention uses historical quality data of materials to be inspected and a Bayesian network to predict the fault distribution of these materials, generating quality results. Based on these results, materials from different manufacturers are statistically analyzed and ranked, and then sampled using a preset sampling ratio. By applying different sampling ratios to materials of varying quality, differentiated sampling is achieved, improving the accuracy of material inspection.

[0137] Example 2

[0138] Please refer to Figure 5 The present invention provides a power system material sampling inspection management device, comprising: a data acquisition module 201, a quality prediction module 202, and a sampling inspection module 203;

[0139] The data acquisition module 201 is used to acquire a list of materials to be inspected in the power system, and retrieve historical quality data of the materials to be inspected from the database according to the category of the materials to be inspected; the historical quality data includes historical fault data and historical truncated data of the commissioning data; the historical truncated data includes the running time of the materials to be inspected that have not yet experienced a fault.

[0140] The quality prediction module 202 is used to establish a fault function of the material to be inspected based on a first preset parameter and a second preset parameter; wherein the distribution of the first preset parameter conforms to a Gamma distribution and the distribution of the second preset parameter conforms to a normal distribution; the distribution of the fault function is predicted by a Bayesian network, and historical quality data of the material to be inspected is imported to determine the posterior distribution of the first preset parameter and the second preset parameter, and the average value of the first preset parameter and the second preset parameter is generated as the quality result based on the posterior distribution;

[0141] The sampling inspection module 203 is used to generate numerical information of the first preset parameter and the second preset parameter based on the quality results; to statistically analyze and sort the materials to be inspected from different manufacturers based on the numerical information, and generate a sorting result; and to conduct sampling inspections on the materials to be inspected from different manufacturers based on the sorting result and a preset sampling ratio coefficient, and generate sampling inspection results.

[0142] The quality prediction module 202 includes: a fault function construction unit, a likelihood function construction unit, a posterior distribution calculation unit, and a quality result calculation unit;

[0143] The fault function construction unit is used to make the fault occurrence time of the material to be inspected follow a two-parameter Weibull distribution, and to calculate the probability density function and fault function of the material to be inspected.

[0144]

[0145]

[0146] Where f(t) is the fault probability density function, F(t) is the fault function, t is the fault occurrence time, the first preset parameter is β, which is a shape parameter used to represent the dispersion of the fault distribution, and the second preset parameter is η, which is a scale parameter used to represent the average fault life of the material.

[0147] The distribution of the first preset parameter conforms to a Gamma distribution, and the distribution of the second preset parameter conforms to a normal distribution, specifically as follows:

[0148]

[0149] Where a and b are constants reflecting the β distribution, and x is a variable;

[0150]

[0151] Where μ and σ 2 These are the mean and variance of η, respectively.

[0152] The likelihood function construction unit is used to construct a likelihood function for historical quality data based on the probability density function and fault function of the materials to be inspected, and to extract n sample data from the operation data of the materials to be inspected. The sample data includes r historical fault data and (nr) historical truncated data, and the operation time of the historical fault data or historical truncated data of the materials to be inspected is t. i The likelihood function of the material is obtained based on the parameters η and β to be determined as follows:

[0153]

[0154] In the formula, σ i The truncation indicator variable takes the value 0 when the i-th sample data is historical truncation data, and takes the value 1 when the i-th sample data is historical fault data; L(D|β,η) is the likelihood function, where D represents the sample status of the historical quality data of the materials to be inspected.

[0155] The posterior distribution calculation unit is used to predict the distribution of the fault function through a Bayesian network, and imports historical quality data of the materials to be inspected to determine the posterior distribution of the first preset parameter and the second preset parameter, specifically:

[0156] Based on the likelihood function, the distribution of the fault function is predicted using a Bayesian network, and the posterior distributions of η and β are calculated:

[0157]

[0158] Where π(β,η) is the joint distribution of η and β, and f(β,η|D) is the posterior distribution of η and β.

[0159] The quality result calculation unit is used to calculate the average values ​​of η and β respectively based on the joint distribution of η and β in the posterior distribution:

[0160]

[0161]

[0162] Where E(η) is the average value of η, and E(β) is the average value of β.

[0163] The sampling inspection module 203 includes: a sampling inspection processing unit;

[0164] The sampling inspection processing unit is used to generate a sampling inspection result indicating that the sampling inspection has failed if the number of non-conforming samples of the materials to be inspected is not less than a first threshold.

[0165] If the number of non-compliant items in the sampling inspection is 0, then a sampling inspection result of passing inspection will be generated.

[0166] If the number of non-compliant materials to be inspected is lower than the first threshold and not equal to 0, then a preset number of materials to be inspected will be added.

[0167] If the number of non-conforming items in the pre-set sample of materials to be inspected is 0, then a sample inspection result of passing inspection will be generated.

[0168] If the number of non-compliant items in the pre-set sampled materials is not less than the second threshold, then a sampling result indicating that the sampling failed will be generated.

[0169] The power system material sampling and management device also includes: a database construction module and an evaluation module;

[0170] The database construction module is used to obtain material information of all first materials, classify the material information, generate corresponding category tags and store them in the database; the material information includes: material list, historical quality data, annual sampling inspection results data and warehousing information list;

[0171] Based on the material information, calculate the first failure rate, the first defect rate, and the first inspection pass rate of the materials to be inspected, and generate the first correlation data between the first failure rate, the first defect rate, the first inspection pass rate, and the quantity of materials entering the warehouse.

[0172] The evaluation module is used to generate a sampling pass rate based on the sampling results; and to calculate the second correlation data between the material risk prediction value and the sampling pass rate based on the numerical information of the second preset parameter as the material risk prediction value.

[0173] The aforementioned power system material sampling inspection management device can implement the power system material sampling inspection management method described in the above method embodiments. The options in the above method embodiments are also applicable to this embodiment, and will not be detailed here. The remaining content of this application embodiment can be referred to the content of the above method embodiments, and will not be repeated in this embodiment.

[0174] Implementing the embodiments of the present invention has the following effects:

[0175] The quality prediction module of this invention predicts the fault distribution of the materials to be inspected based on historical quality data using a Bayesian network, generating quality results. The sampling module then uses these quality results to statistically analyze and rank the materials from different manufacturers, and performs sampling inspections based on preset sampling ratios. By using different sampling ratios to inspect materials of varying quality, differentiated sampling is achieved, improving the accuracy of material sampling.

[0176] Example 3

[0177] Accordingly, the present invention also provides a computer-readable storage medium, the computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the power system material sampling management method as described in any of the above embodiments.

[0178] For example, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0179] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0180] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0181] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function, etc.; the data storage area may store data created based on the use of the mobile terminal, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, RAM, plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0182] Wherein, if the modules / units integrated in the terminal device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, it can implement the steps of the various method embodiments described above. Wherein, the computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0183] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A method for managing random inspections of materials in a power system, characterized in that, include: Obtain a list of materials to be inspected in the power system, and retrieve historical quality data of the materials to be inspected from the database according to the category of the materials to be inspected; the historical quality data includes historical fault data and historical truncated data of the commissioning data; the historical truncated data includes the operating time of the materials to be inspected that have not yet experienced a fault. Based on the first preset parameter and the second preset parameter, a fault function for the material to be inspected is established; wherein the distribution of the first preset parameter conforms to the Gamma distribution and the distribution of the second preset parameter conforms to the normal distribution; the distribution of the fault function is predicted by a Bayesian network, and historical quality data of the material to be inspected is imported to determine the posterior distribution of the first preset parameter and the second preset parameter; based on the posterior distribution, the average value of the first preset parameter and the second preset parameter is generated as the quality result. Based on the quality results, generate numerical information of the first preset parameter and the second preset parameter; based on the numerical information, statistically analyze and sort the materials to be inspected from different manufacturers, and generate sorting results; based on the sorting results and the preset sampling ratio coefficient, conduct sampling inspections on the materials to be inspected from different manufacturers, and generate sampling inspection results. The step of establishing a fault function for the material to be inspected based on the first preset parameter and the second preset parameter is as follows: Let the failure occurrence time of the materials to be inspected follow a two-parameter Weibull distribution, and calculate the probability density function and failure function of the materials to be inspected: ； ； Where f(t) is the fault probability density function, F(t) is the fault function, t is the fault occurrence time, the first preset parameter is β, which is a shape parameter used to represent the dispersion of the fault distribution, and the second preset parameter is η, which is a scale parameter used to represent the average fault life of the material. The distribution of the first preset parameter conforms to a Gamma distribution, and the distribution of the second preset parameter conforms to a normal distribution, specifically: ； Where a and b are constants reflecting the β distribution, and x is a variable; ； where μ and σ 2 are the mean and variance size of η, respectively; Before predicting the distribution of the fault function using a Bayesian network and importing historical quality data of the materials to be inspected to determine the distribution probabilities of the first and second preset parameters, the method further includes: Based on the probability density function and failure function of the materials to be inspected, a likelihood function for historical quality data is constructed. n sample data points are extracted from the operational data of the materials to be inspected. These sample data points include r historical failure data points and (nr) historical truncated data points. The operational time of the historical failure data or historical truncated data points of the materials to be inspected is t. i The likelihood function of the material is obtained based on the parameters η and β to be determined as follows: ； In the formula, σ i The truncation indicator variable takes the value 0 when the i-th sample data is historical truncation data, and takes the value 1 when the i-th sample data is historical fault data; L(D|β,η) is the likelihood function, where D represents the sample status of the historical quality data of the materials to be inspected; The process involves predicting the distribution of the fault function using a Bayesian network and importing historical quality data of the materials to be inspected to determine the posterior distributions of the first and second preset parameters. Specifically: Based on the likelihood function, the distribution of the fault function is predicted using a Bayesian network, and the posterior distributions of η and β are calculated: ； Where π(β, η) is the joint distribution of η and β, and f(β, η|D) is the posterior distribution of η and β; The step of generating the average value of the first preset parameter and the second preset parameter as the quality result based on the posterior distribution specifically involves: Based on the joint distribution of η and β in the posterior distribution, calculate the average values ​​of η and β respectively: ； ； Where E(η) is the average value of η, and E(β) is the average value of β.

2. The method for random inspection management of power system materials as described in claim 1, characterized in that, Before obtaining the list of materials to be inspected from the power system and retrieving historical quality data of the materials to be inspected from the database according to their categories, the process also includes: Obtain material information for all primary materials, classify the material information, generate corresponding category tags, and store them in the database; the material information includes: material list, historical quality data, annual sampling inspection results data, and warehousing information list; Based on the material information, calculate the first failure rate, the first defect rate, and the first inspection pass rate of the materials to be inspected, and generate the first correlation data between the first failure rate, the first defect rate, the first inspection pass rate, and the quantity of materials entering the warehouse.

3. The method for random inspection management of power system materials as described in claim 1, characterized in that, The process of randomly sampling materials from different manufacturers and generating sampling results is as follows: If the number of non-compliant materials to be inspected is not less than the first threshold, then a non-compliant inspection result will be generated. If the number of non-compliant materials to be inspected is 0, then a passing inspection result will be generated. If the number of non-compliant materials to be inspected is lower than the first threshold and not equal to 0, then a preset number of materials to be inspected will be added. If the number of non-conforming items in the pre-set sample of materials to be inspected is 0, then a sample inspection result of passing inspection will be generated. If the number of non-compliant items in the pre-set sampled materials is not less than the second threshold, then a sampling result indicating that the sampling failed will be generated.

4. The method for random inspection and management of power system materials as described in claim 3, characterized in that, After sampling and inspecting materials from different manufacturers and generating inspection results, the process also includes: The sampling pass rate is generated based on the sampling results; the second correlation data between the material risk prediction value and the sampling pass rate is calculated based on the numerical information of the second preset parameter as the material risk prediction value.

5. A power system material sampling inspection management device, characterized in that, The method for random inspection and management of power system materials as described in any one of claims 1 to 4 is adopted. The power system material sampling management device includes: a data acquisition module, a quality prediction module, and a sampling module; The data acquisition module is used to acquire a list of materials to be inspected in the power system, and retrieve historical quality data of the materials to be inspected from the database according to the category of the materials to be inspected; the historical quality data includes historical fault data and historical truncated data of the commissioning data; the historical truncated data includes the operating time of the materials to be inspected that have not yet experienced a fault. The quality prediction module is used to establish a fault function of the material to be inspected based on a first preset parameter and a second preset parameter; wherein the distribution of the first preset parameter conforms to a Gamma distribution and the distribution of the second preset parameter conforms to a normal distribution; the distribution of the fault function is predicted by a Bayesian network, and historical quality data of the material to be inspected is imported to determine the posterior distribution of the first preset parameter and the second preset parameter, and the average value of the first preset parameter and the second preset parameter is generated as the quality result based on the posterior distribution; The sampling inspection module is used to generate numerical information of the first preset parameter and the second preset parameter based on the quality results; to statistically analyze and sort the materials to be inspected from different manufacturers based on the numerical information, and generate a sorting result; and to conduct sampling inspections on the materials to be inspected from different manufacturers based on the sorting result and a preset sampling ratio coefficient, and generate sampling inspection results.

Images