Bad station detection method, device, equipment and storage medium

Abstract

The application discloses a kind of bad work station detection method, device, equipment and storage medium.The present application is classified according to the product of each over-station product in the period to be measured according to the bad information, and the bad sample and the qualified sample in each over-station product are determined according to the product classification result, each bad sample and each qualified sample are matched respectively, and sample pair is obtained, the sample pair is input into probability analysis model to carry out probability analysis, and the bad probability information of each work station is obtained, each work station is screened according to the bad probability information, and the bad work station in each work station is determined according to the screening result;Since the present application inputs the sample pair between the bad sample and the qualified sample in each over-station product into probability analysis model to carry out probability analysis, obtains the bad probability information of each work station, and then screens the bad work station from each work station according to the bad probability information, so as to accurately detect the work station with bad problem on process line, and realize the rapid positioning of product bad reason.

Description

Technical Field

[0001] This invention relates to the field of automation technology, and in particular to a method, apparatus, equipment and storage medium for detecting defective workstations. Background Technology

[0002] In recent years, with the gradual development of my country's manufacturing sector, smart factories and intelligent manufacturing have become new directions in the industry. In the production and assembly of electronic and semiconductor products, the increasing precision and complexity of products makes it increasingly difficult to manually track defective products. Currently, the method for analyzing defective products and finding the root cause involves disassembling the device to inspect and test the performance of each module, and then using the test results to locate the workstation with the defect. However, disassembly testing may introduce new problems during the process, making it currently impossible to accurately detect the workstation with the defect, thus hindering the accurate identification of the cause of the product defect.

[0003] The above content is only used to help understand the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention

[0004] The main objective of this invention is to provide a method, apparatus, equipment, and storage medium for detecting defective workstations, aiming to solve the technical problem that existing technologies cannot accurately detect workstations with defects, thus making it impossible to accurately locate the cause of product defects.

[0005] To achieve the above objectives, the present invention provides a method for detecting defective workstations, the method comprising the following steps:

[0006] Based on the adverse information within the testing period, each product passing through the station is classified, and the defective and qualified samples in each product passing through the station are determined based on the product classification results.

[0007] Each defective sample is matched with each qualified sample to obtain sample pairs;

[0008] The sample pairs are input into a probability analysis model for probability analysis to obtain the defect probability information of each workstation;

[0009] Each workstation is screened based on the defect probability information, and the defective workstations in each workstation are determined based on the screening results.

[0010] Optionally, the step of matching each defective sample with each qualified sample to obtain sample pairs includes:

[0011] Identify the adverse factors corresponding to each adverse sample based on the adverse information;

[0012] The adverse samples are classified according to the aforementioned adverse factors;

[0013] Multiple sets of defective samples are constructed based on the classification results, wherein each set of defective samples includes at least one defective sample.

[0014] Each set of defective samples is matched with each set of qualified samples to obtain sample pairs.

[0015] Optionally, the step of matching each set of defective samples with each set of qualified samples to obtain sample pairs includes:

[0016] Obtain product information for each product passing through a station during the time period to be tested;

[0017] Based on the product information, determine the production batch number and material information of each product passing through the station;

[0018] Based on the production batch number and the material information, each defective sample in each defective sample set is matched with each qualified sample to obtain a sample pair, wherein the sample pair contains the defective sample and the qualified sample corresponding to the defective sample.

[0019] Optionally, before inputting the sample pairs into the probability analysis model for probability analysis to obtain the defect probability information of each workstation, the method further includes:

[0020] Obtain process information for each workstation;

[0021] Determine the node location information corresponding to each workstation based on the process information;

[0022] The process structure of each workstation is determined based on the node location information;

[0023] A probabilistic analysis model is constructed based on the process structure.

[0024] Optionally, before classifying the products passing through each station based on the defect information within the testing time period, and determining the defective and qualified samples among the products passing through each station based on the product classification results, the method further includes:

[0025] The time period to be tested is determined according to the preset detection strategy;

[0026] Obtain the record information of each workstation within the time period to be tested;

[0027] The recorded information is then cleaned.

[0028] Extract the station transit information of each workstation within the time period to be tested from the data cleaning and recorded information;

[0029] Based on the transit information, determine the transit products within the test time period.

[0030] Optionally, the step of inputting the sample pairs into a probability analysis model for probability analysis to obtain the defect probability information for each workstation includes:

[0031] Each sample in the sample pair is labeled according to the product classification results;

[0032] Obtain the workstation name corresponding to each sample;

[0033] Each sample in the sample pair is encoded based on the labeling results and the workstation name;

[0034] The encoded sample pairs are input into a probability analysis model for probability analysis to obtain the defect probability information of each workstation.

[0035] Optionally, the step of screening each workstation based on the defect probability information and determining the defective workstations in each workstation based on the screening results includes:

[0036] The probability information of failure is input into a pre-built confidence analysis model for confidence analysis to obtain the confidence analysis results;

[0037] Determine whether the confidence analysis results meet the preset conditions;

[0038] If so, obtain the pass rate information for each workstation;

[0039] Based on the pass rate information and the confidence analysis results, each workstation is screened, and the defective workstations in each workstation are identified based on the screening results.

[0040] Optionally, before inputting the adverse probability information into a pre-built confidence analysis model for confidence analysis and obtaining the confidence analysis results, the method further includes:

[0041] Obtain information on adverse events within a historical time period;

[0042] Based on the adverse event information, identify the historical adverse events that occurred at each workstation within the historical time period;

[0043] Determine the historical failure probability information for each workstation based on each historical adverse event;

[0044] The confidence parameters for each workstation are determined based on the historical failure probability information.

[0045] A confidence analysis model is constructed based on each workstation and its corresponding confidence parameters.

[0046] Furthermore, to achieve the above objectives, the present invention also proposes a defective workstation detection device, the defective workstation detection device comprising:

[0047] The sample selection module is used to classify products passing through each station based on the defect information within the test time period, and to determine the defective and qualified samples in each product passing through the station based on the product classification results.

[0048] The sample matching module is used to match each defective sample with each qualified sample to obtain sample pairs.

[0049] The probability analysis module is used to input the sample pairs into the probability analysis model to perform probability analysis and obtain the defect probability information of each workstation.

[0050] The workstation screening module is used to screen each workstation according to the defect probability information, and to determine the defective workstations in each workstation based on the screening results.

[0051] Furthermore, to achieve the above objectives, the present invention also proposes a defective workstation detection device, which includes: a memory, a processor, and a defective workstation detection program stored in the memory and executable on the processor, wherein the defective workstation detection program is configured to implement the steps of the defective workstation detection method described above.

[0052] In addition, to achieve the above objectives, the present invention also proposes a storage medium storing a defective workstation detection program, which, when executed by a processor, implements the steps of the defective workstation detection method described above.

[0053] This invention categorizes products at each station based on defect information within the testing period. Based on the product classification results, it identifies defective and qualified samples at each station. Each defective sample is then matched with each qualified sample to obtain sample pairs. These sample pairs are input into a probability analysis model for probability analysis to obtain the defect probability information for each station. Based on this defect probability information, each station is screened, and the defective stations within each station are identified based on the screening results. Because this invention categorizes products at each station based on defect information within the testing period, and determines defective and qualified samples based on the product classification results, it achieves product classification and sample selection within the testing period. By matching each defective sample with each qualified sample to obtain sample pairs, and inputting these sample pairs into a probability analysis model for probability analysis to obtain the defect probability information for each station, it accurately obtains the defect probability distribution of each station. Based on this defect probability information, each station is screened, and the defective stations within each station are identified based on the screening results. This allows for the accurate detection of stations with defects on the production line, enabling rapid location of product defects. Attached Figure Description

[0054] Figure 1This is a schematic diagram of the structure of a faulty workstation detection device for the hardware operating environment involved in the embodiments of the present invention;

[0055] Figure 2 This is a flowchart illustrating the first embodiment of the defective workstation detection method of the present invention;

[0056] Figure 3 This is a process structure diagram of the first embodiment of the defective workstation detection method of the present invention;

[0057] Figure 4 This is a flowchart illustrating the second embodiment of the defective workstation detection method of the present invention;

[0058] Figure 5 This is a flowchart illustrating the third embodiment of the defective workstation detection method of the present invention;

[0059] Figure 6 This is a structural block diagram of the first embodiment of the defective workstation detection device of the present invention.

[0060] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0061] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0062] Reference Figure 1 , Figure 1 This is a schematic diagram of the structure of a faulty workstation detection device for the hardware operating environment involved in the embodiments of the present invention.

[0063] like Figure 1 As shown, the defect detection station equipment may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen and an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wireless-Fidelity (Wi-Fi) interface). The memory 1005 may be high-speed random access memory (RAM) or stable non-volatile memory (NVM), such as a disk storage device. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0064] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on defective workstation inspection equipment and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0065] like Figure 1 As shown, the memory 1005, which serves as a storage medium, may include an operating system, a network communication module, a user interface module, and a faulty workstation detection program.

[0066] exist Figure 1 In the defective workstation detection device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the defective workstation detection device of the present invention can be set in the defective workstation detection device, and the defective workstation detection device calls the defective workstation detection program stored in the memory 1005 through the processor 1001 and executes the defective workstation detection method provided in the embodiment of the present invention.

[0067] This invention provides a method for detecting defective workstations, referring to... Figure 2 , Figure 2 This is a flowchart illustrating the first embodiment of a defective workstation detection method according to the present invention.

[0068] In this embodiment, the defective workstation detection method includes the following steps:

[0069] Step S10: Classify the products passing through each station according to the defect information within the test time period, and determine the defective and qualified samples in each product passing through the station based on the product classification results.

[0070] It should be understood that the execution subject of the method in this embodiment can be a defective workstation detection device with data processing, network communication and program running functions, such as a computer, or other devices or equipment that can achieve the same or similar functions. Here, the defective workstation detection device (hereinafter referred to as the workstation detection device) is used as an example for explanation.

[0071] It should be noted that the testing time period can be a time period set based on the current inspection time. For example, the testing time period can be 1 hour, 2 hours, or 3 hours. The workstation inspection equipment can set the testing time period according to the timestamp of the defective product, thereby effectively avoiding the problem of missed inspections. The aforementioned products that have passed through the workstation on the process line or product line during the testing time period can be semi-finished products or finished products. The aforementioned defective samples can be products with defects produced during the testing time period, and the aforementioned qualified samples can be products without defects produced during the testing time period. The workstation inspection equipment distinguishes between positive and negative samples based on the defect information. Defective samples are products containing defect information, while qualified samples are products that do not contain defect information and were produced within the same time period as the defective products.

[0072] It should be understood that when the workstation testing equipment in this embodiment detects a defective product, it obtains the timestamp of the defective product based on the product information of the defective product, determines the test time period based on the timestamp and the current time, obtains the station information within the test time period, determines the station products that have passed through the station within the test time period based on the station information, classifies each station product based on the defect information within the test time period, determines the defective and qualified products among each station product based on the product classification results, obtains the product code of each defective product, obtains the product code of each qualified product, re-numbers the defective products based on the product codes of each defective product, re-numbers the qualified products based on the product codes of each qualified product, and uses the defective products as defective samples based on the defective product numbering results, and uses the qualified products as qualified samples based on the qualified product numbering results, thereby effectively selecting samples from each station product.

[0073] In practice, the selection of qualified samples is determined based on the transit timestamp of defective samples. The minimum and maximum timestamps of defective samples at each station are statistically analyzed. The timestamp of the station corresponding to the qualified sample should meet the selection range of defective samples. The station testing equipment determines the defective products among the transit products, obtains the timestamp of the defective products, and statistically analyzes the minimum and maximum timestamps of defective samples at each station based on the timestamp of the defective products. The timestamp of the station corresponding to the qualified sample should meet the selection range of defective samples.

[0074] Furthermore, in order to accurately select the transit products that need to be tested, the following may be included before step S10:

[0075] The time period to be tested is determined according to the preset detection strategy;

[0076] Obtain the record information of each workstation within the time period to be tested;

[0077] The recorded information is then cleaned.

[0078] Extract the station transit information of each workstation within the time period to be tested from the recorded information after data cleaning;

[0079] The transit products within the time period to be tested are determined based on the transit information.

[0080] It should be noted that the preset detection strategy can be a pre-set strategy for dividing the detection time period. For example, the preset detection strategy could be that when the workstation detection equipment detects a defective product, it calculates the minimum and maximum timestamps of each workstation the defective product has passed through based on the product's timestamp, and then determines the time period to be tested based on the minimum and maximum timestamps. The aforementioned station information can be relevant information about the products processed by each workstation or relevant information about the materials processed.

[0081] It should be understood that the workstation inspection equipment extracts all transit records of the product within the current time period, all defective product information within the current time period, and sub-material information of the product within the current time period from the database of the Manufacturing Execution System (MES). It then removes duplicate data and data containing null values ​​through data cleaning, and finally links all the information together through product code information.

[0082] Step S20: Match each defective sample with each qualified sample to obtain sample pairs.

[0083] It should be noted that a sample pair can be a set of defective samples and qualified samples. The set of samples may contain defective samples and at least one qualified sample corresponding to the defective sample.

[0084] It should be understood that in this embodiment, the workstation testing equipment acquires product information of each defective sample, determines the historical workstation information of each defective sample and the sub-material information used in the workstations through which each defective sample passed based on the product information, and performs sample matching between each defective sample and each qualified sample based on the historical workstation information and sub-material information, so as to match the defective sample with the qualified sample that is consistent with its production conditions, thereby obtaining a sample pair.

[0085] For example, each station has defective sample A and defective sample B, as well as qualified sample X, qualified sample Y, and qualified sample Z. The product information for each station is obtained. Based on this information, it is determined that defective sample A belongs to the same process as qualified samples Y and Z, and defective sample B belongs to the same process as qualified sample X. Therefore, based on the product information, each defective sample is matched with each qualified sample to obtain two sample pairs: one consisting of defective sample A and qualified samples Y and Z, and the other consisting of defective sample B and qualified sample X.

[0086] Step S30: Input the sample pairs into the probability analysis model to perform probability analysis and obtain the defect probability information of each workstation.

[0087] It should be noted that the probability analysis model can be a network model pre-built based on the process structure on the production line. For example, the probability analysis model can be a Bayesian network model, and the above-mentioned defect probability information can be the probability of defective products produced at each workstation.

[0088] In practical implementation, the workstation inspection equipment can select a Bayesian-Markov chain based on the production process routing. Markov chains have characteristics such as single-chain propagation and only being related to the previous node, which meets practical needs. Since finished products are often not assembled from a single semi-finished product, additional sub-chains are added according to actual requirements. Finally, positive and defective samples are used as input to the Bayesian network, and the probability distribution of each workstation is output.

[0089] Furthermore, in order to perform probability analysis accurately, the following may be included before step S30:

[0090] Obtain process information for each workstation;

[0091] Determine the node location information corresponding to each workstation based on the process information;

[0092] The process structure of each workstation is determined based on the node location information;

[0093] A probabilistic analysis model is constructed based on the process structure.

[0094] It should be noted that process information can be relevant information about the production process in which each workstation is located. For example, process information can include the location information of each workstation within the production process. The aforementioned node location information can also refer to treating each workstation as a node in the process, and the node location information can be the location of each node within the process.

[0095] Furthermore, in order to improve the analysis efficiency of each workstation, step S30 above may include:

[0096] Each sample in the sample pair is labeled according to the product classification results;

[0097] Obtain the workstation name corresponding to each sample;

[0098] Each sample in the sample pair is encoded based on the labeling results and the workstation name;

[0099] The encoded sample pairs are input into a probability analysis model for probability analysis to obtain the defect probability information of each workstation.

[0100] It should be noted that the workstation testing equipment determines the defective workstations containing defective products and the qualified workstations without defective products in the sample pair based on the product classification results. Each workstation is coded according to the marking results, with qualified workstations coded as 0 and unqualified workstations coded in ascending order. For example, qualified workstation A, defective workstation B, defective workstation C, and defective workstation D are coded as 0, 1, 2, and 3, respectively.

[0101] It should be understood that the workstation inspection equipment extracts the record information of each workstation within the test period from the database of the manufacturing execution system. In order to remove invalid information, the record information is cleaned to remove duplicate data and data containing null values. The data string after integrating the cleaned record information has each line containing all the information of a product. Each line includes the product's SN, all the workstations it passed through, the timestamp of the workstations it passed through, the sub-material information used by the workstations it passed through, and if the product is defective, it includes error information and the workstation that reported the error.

[0102] In the specific implementation, refer to Figure 3 , Figure 3 The diagram shows the process structure. Workstations 1 through n are the main processes, each workstation being a node. Due to the single-chain propagation characteristic of product processing, each process can use a Markov chain as the network model. Processes 1 and 2, as semi-finished products, have their own independent production processes and are inserted into the main route for assembly after their independent production processes. After all processes in the main route, the output reaches the result node. If the product is good, the result is marked as 0; if the product is defective and an error occurs at some intermediate workstation, the result node is marked as x, where x is determined by the defect number of the defective sample. When a user selects a defective item as input, the algorithm automatically selects the subset containing that defective item and the corresponding set of qualified samples from the defective set. Simultaneously, all workstations in that subset are automatically converted to new codes. This embodiment uses Bayesian networks and Markov chains to effectively solve the problem of linear sequential production line operations. It clearly simulates the actual production line operation and allows for model changes based on production line adjustments.

[0103] Step S40: Filter each workstation according to the defect probability information, and determine the defective workstations in each workstation according to the filtering results.

[0104] It should be noted that a defective workstation can be a workstation with a high probability of producing defective products.

[0105] It should be understood that the workstation testing equipment determines the defect probability information of each workstation based on the output of the probability analysis model, removes workstations with low defect probability based on the defect probability information, and records workstations with high defect probability, and treats the workstations with high defect probability as defective workstations.

[0106] This embodiment classifies products at each station based on defect information within the testing period. Based on the classification results, it identifies defective and qualified samples at each station. Each defective sample is matched with each qualified sample to obtain sample pairs. These sample pairs are then input into a probability analysis model for probability analysis to obtain the defect probability information for each station. Based on this defect probability information, each station is screened, and the defective stations within each station are identified based on the screening results. Because this embodiment classifies products at each station based on defect information within the testing period and identifies defective and qualified samples based on the classification results, it achieves product classification and sample selection within the testing period. By matching each defective sample with each qualified sample to obtain sample pairs and inputting these pairs into a probability analysis model for probability analysis to obtain the defect probability information for each station, it accurately obtains the defect probability distribution of each station. Based on this defect probability information, each station is screened, and the defective stations within each station are identified based on the screening results. This accurately detects stations with defects on the production line and enables rapid location of product defects.

[0107] refer to Figure 4 , Figure 4 This is a flowchart illustrating a second embodiment of a defective workstation detection method according to the present invention.

[0108] Based on the first embodiment described above, in this embodiment, step S20 includes:

[0109] Step S201: Determine the adverse factors corresponding to each adverse sample based on the adverse information.

[0110] It should be noted that a defective factor can be any problem present in each defective sample, such as a battery problem, a circuit problem, or other types of problems.

[0111] It should be understood that the workstation testing equipment determines the defect information of each product passing through the station based on the error information of each workstation, and determines the corresponding defect factors for each defective sample based on the defect information.

[0112] Step S202: Classify each defective sample according to the aforementioned adverse factors.

[0113] In the specific implementation, the workstation inspection equipment treats all defective samples as a set, which is then further divided into multiple subsets based on the defect information and the workstation that reported the error. The same error information and the same workstation are considered as belonging to the same set. The defective sample subsets are sorted according to the number of elements within each subset. Different defect information is recorded and assigned a new number, and the defect information is then replaced with the number. The numbering starts from 0 and increments automatically. The selection of qualified samples is determined by the timestamp of the defective sample passing through the workstation. The minimum and maximum timestamps of each workstation for defective samples are calculated. The timestamps of the workstations corresponding to qualified samples should meet the selection range for defective samples. Based on the organized sample set, the names of each workstation are counted, recorded, and reassigned with a new number, starting from 0 and incrementing automatically. After recording, the workstation name is replaced with the new number.

[0114] Step S203: Construct multiple sets of bad samples based on the classification results, wherein the sets of bad samples include at least one bad sample.

[0115] It should be noted that the set of defective samples can be a set containing at least one defective sample, and the defective factors of each defective sample in the set are the same.

[0116] It should be understood that the workstation testing equipment classifies each defective sample according to the cause of the defect, and constructs a defective sample set based on defective samples with the same defective factors.

[0117] In the specific implementation, there are defective samples A, B, C and D in the product category. The defective factors of defective samples A and D are battery defects, while the defective factors of defective samples B and C are line defects. Therefore, a battery defective sample set is constructed based on defective samples A and D, and a line defective sample set is constructed based on defective samples B and C.

[0118] Step S204: Match each set of defective samples with each set of qualified samples to obtain sample pairs.

[0119] It should be noted that a sample pair can be a set of defective samples and qualified samples. The set of samples may contain defective samples and at least one qualified sample corresponding to the defective sample.

[0120] It should be understood that in this embodiment, the workstation testing equipment acquires product information of each defective sample, determines the historical workstation information of each defective sample and the sub-material information used in the workstations through which each defective sample passed based on the product information, determines the production conditions of each defective sample based on the historical workstation information and sub-material information, and performs sample matching between each defective sample in the defective sample set and each qualified sample based on the production conditions, so as to match the defective sample with the qualified sample whose production conditions are consistent with it, thereby obtaining a sample pair.

[0121] Furthermore, to improve the accuracy of sample matching, step S204 above may include:

[0122] Step S2041: Obtain product information for each product passing through a station during the time period to be tested;

[0123] Step S2042: Determine the production batch number and material information of each product passing through the station based on the product information;

[0124] Step S2043: Match each defective sample in each defective sample set with each qualified sample according to the production batch number and the material information to obtain sample pairs, wherein each sample pair contains a defective sample and a qualified sample corresponding to the defective sample.

[0125] It should be noted that the production batch number can be the production code and product batch number of each product passing through the station. The above material information can be the sub-material information that makes up the product passing through the station, and the product passing through the station is composed of multiple sub-materials.

[0126] It should be understood that the workstation testing equipment obtains the product information of each product passing through the station within the time period to be tested based on the station's transit records, determines the production batch number of each product passing through the station based on the product information, as well as the sub-material information that makes up the product passing through the station, determines the production conditions of each product passing through the station based on the production batch number and the sub-material information, and performs sample matching between each defective sample in each defective sample set and each qualified sample based on the production conditions of each product passing through the station to obtain sample pairs.

[0127] This embodiment determines the adverse factors corresponding to each adverse sample based on the adverse information, classifies each adverse sample according to the adverse factors, and constructs multiple adverse sample sets based on the classification results. Each adverse sample set includes at least one of the adverse samples. Each adverse sample set is then matched with each qualified sample to obtain sample pairs. Because this embodiment classifies each adverse sample according to the adverse factors and constructs multiple adverse sample sets based on the classification results, the efficiency of sample matching is improved. Matching each adverse sample set with each qualified sample to obtain sample pairs provides accurate input samples for probability analysis and ensures the accuracy of the output results of the probability analysis model.

[0128] refer to Figure 5 , Figure 5 This is a flowchart illustrating a third embodiment of a defective workstation detection method according to the present invention.

[0129] Based on the first embodiment described above, in this embodiment, after step S40, the method further includes:

[0130] Step S50: Input the probability of failure information into the pre-built confidence analysis model for confidence analysis to obtain the confidence analysis results;

[0131] Step S60: Determine whether the confidence analysis result meets the preset conditions;

[0132] Step S70: If yes, then obtain the pass rate information for each workstation;

[0133] Step S80: Screen each workstation according to the pass rate information and the confidence analysis results, and determine the defective workstations in each workstation according to the screening results.

[0134] It should be noted that the confidence analysis model can be used to store historical confidence judgment experience. The workstation testing equipment can input the failure probability information into the confidence analysis model so that the confidence analysis model can analyze the failure probability information based on historical confidence judgment experience and judge whether the failure probability information is accurate based on the analysis results.

[0135] Furthermore, to improve the accuracy of the confidence analysis model, the following may be included before step S50:

[0136] Obtain information on adverse events within a historical time period;

[0137] Based on the adverse event information, identify the historical adverse events that occurred at each workstation within the historical time period;

[0138] Determine the historical failure probability information for each workstation based on each historical adverse event;

[0139] The confidence parameters for each workstation are determined based on the historical failure probability information.

[0140] A confidence analysis model is constructed based on each workstation and its corresponding confidence parameters.

[0141] It should be noted that the historical time period can be a sampling period preset by the workstation's testing equipment. By sampling the product information produced by the workstation during the historical time period, defective event information in the process can be determined, thereby identifying the historical defective events generated by each workstation within the historical time period. The aforementioned historical defective events can be events in which defective products were produced by each workstation within the historical time period. Refer to Formula 1 below, which is the formula for calculating the confidence level parameter, where δ is the confidence level parameter, and p(x) = 1 / 2. i ) represents the historical probability of failure.

[0142]

[0143] Referring to Formula 2 below, which is the formula for the confidence analysis model, the graph network consists of G nodes, and N represents the names of each workstation node. Workstation names should conform to the principles of universality, uniqueness, and identifiability. Each network node should contain numerous features, which are used to describe the node's universality and uniqueness. In the expert system, feature values ​​include, but are not limited to, workstation functions, the correlation and non-correlation between the workstation and other workstations, workstation yield, and workstation parameters. The expert system is an undirected graph, with bidirectional edges E connecting each node. Here, G is the graph network, N is the nodes, E is the edges, name is the workstation name, ok_rate is the workstation's product qualification rate, feature is the workstation's production feature, parameter is the parameter, α indicates whether it is a relevant workstation, β indicates whether it is a semi-finished product access workstation, γ represents the workstation's yield, δ is the confidence parameter based on probability improvement, and N... A and N B These represent the preceding and following workstations, respectively.

[0144]

[0145] in,

[0146] if, Then α = 1, otherwise α = 0.

[0147] When N B To intervene in the semi-finished product workstation, β = 1; otherwise, β = 0.

[0148] in,

[0149] p(x i () represents the probability xi of a random event X.

[0150] because

[0151] so,

[0152] according to

[0153] Then, p(x) any )*δ=1,any∈I.

[0154] Based on Formula 1 above, the Bayesian-Markov chain can be combined with the confidence analysis model to obtain Formula 3 below. Formula 3 is the combined formula, where P(A|B) represents the likelihood probability, indicating the probability that the current station will produce a defect given a previous station. Since P(B|A) indicates the connection between the preceding and following stations, we can introduce the expert system node parameter δ, which represents the correlation between the previous station B and the current station A.

[0155]

[0156] It should be understood that the workstation inspection equipment filters out workstations with low failure probabilities based on the failure probability information and a preset threshold, records the names and labels of workstations with failure probabilities exceeding the preset threshold, and identifies these workstations as defective workstations. The names of defective workstations are then input into the confidence analysis model. Based on the records in the confidence analysis model, the workstation with the highest correlation P(A|B) of the error reporting workstation is selected and cross-compared with the workstation in the MES system.

[0157] This embodiment inputs the defect probability information into a pre-built confidence analysis model for confidence analysis to obtain the confidence analysis results. It then determines whether the confidence analysis results meet preset conditions. If so, it obtains the pass rate information for each workstation, filters each workstation based on the pass rate information and the confidence analysis results, and identifies defective workstations based on the filtering results. Because this embodiment performs confidence analysis on the defect probability information output by the probability analysis model, it ensures that the output result of the probability analysis model is a high-confidence result. The high-confidence result is then combined with the pass rate information of each workstation for comprehensive judgment, filtering each workstation, and identifying defective workstations based on the filtering results, thereby accurately detecting defective workstations with problems.

[0158] Furthermore, this embodiment of the invention also proposes a storage medium storing a defective workstation detection program, which, when executed by a processor, implements the steps of the defective workstation detection method described above.

[0159] Since this storage medium adopts all the technical solutions of all the above embodiments, it has at least all the beneficial effects brought about by the technical solutions of the above embodiments, which will not be repeated here.

[0160] Reference Figure 6 , Figure 6 This is a structural block diagram of the first embodiment of the defective workstation detection device of the present invention.

[0161] like Figure 6 As shown, the defective workstation detection device proposed in this embodiment of the invention includes:

[0162] The sample selection module 10 is used to classify the products passing through each station according to the defect information within the test time period, and to determine the defective and qualified samples in each product passing through the station based on the product classification results.

[0163] The sample matching module 20 is used to match each defective sample with each qualified sample to obtain sample pairs.

[0164] The probability analysis module 30 is used to input the sample pairs into the probability analysis model for probability analysis to obtain the failure probability information of each workstation;

[0165] The workstation screening module 40 is used to screen each workstation according to the defect probability information, and to determine the defective workstations in each workstation according to the screening results.

[0166] Furthermore, the sample matching module 20 is also used to determine the adverse factors corresponding to each adverse sample based on the adverse information; classify each adverse sample according to the adverse factors; construct multiple adverse sample sets based on the classification results, wherein each adverse sample set includes at least one of the adverse samples; and perform sample matching between each adverse sample set and each qualified sample to obtain sample pairs.

[0167] Furthermore, the sample matching module 20 is also used to obtain product information of each product passing through the station during the test time period; determine the production batch number and material information of each product passing through the station based on the product information; and perform sample matching between each defective sample in each defective sample set and each qualified sample based on the production batch number and the material information to obtain sample pairs, wherein the sample pairs contain defective samples and qualified samples corresponding to the defective samples.

[0168] Furthermore, the probability analysis module 30 is also used to acquire process information of each workstation; determine the node position information corresponding to each workstation based on the process information; determine the process structure of the process in which each workstation is located based on the node position information; and construct a probability analysis model based on the process structure.

[0169] Furthermore, the sample selection module 10 is also used to determine the test time period according to a preset detection strategy; acquire the record information of each workstation within the test time period; perform data cleaning on the record information; extract the station passing information of each workstation within the test time period from the data-cleaned record information; and determine the station passing products within the test time period based on the station passing information.

[0170] Furthermore, the sample selection module 10 is also used to filter out the processes to be numbered in each process according to the station information; obtain the product output information of the processes to be numbered, and the station name of each station in the processes to be numbered; and number each station according to the preset numbering order, the product output information and the station name.

[0171] Furthermore, the workstation screening module 40 is also used to input the defect probability information into a pre-constructed confidence analysis model for confidence analysis to obtain confidence analysis results; determine whether the confidence analysis results meet preset conditions; if so, obtain the pass rate information of each workstation; screen each workstation according to the pass rate information and the confidence analysis results, and determine the defective workstations in each workstation according to the screening results.

[0172] Furthermore, the workstation screening module 40 acquires adverse event information within a historical time period; determines the historical adverse events generated by each workstation within the historical time period based on the adverse event information; determines the historical adverse probability information of each workstation based on each historical adverse event; determines the confidence level parameter corresponding to each workstation based on the historical adverse probability information; and constructs a confidence analysis model based on each workstation and its corresponding confidence level parameter.

[0173] This embodiment classifies products at each station based on defect information within the testing period. Based on the classification results, it identifies defective and qualified samples at each station. Each defective sample is matched with each qualified sample to obtain sample pairs. These sample pairs are then input into a probability analysis model for probability analysis to obtain the defect probability information for each station. Based on this defect probability information, each station is screened, and the defective stations within each station are identified based on the screening results. Because this embodiment classifies products at each station based on defect information within the testing period and identifies defective and qualified samples based on the classification results, it achieves product classification and sample selection within the testing period. By matching each defective sample with each qualified sample to obtain sample pairs and inputting these pairs into a probability analysis model for probability analysis to obtain the defect probability information for each station, it accurately obtains the defect probability distribution of each station. Based on this defect probability information, each station is screened, and the defective stations within each station are identified based on the screening results. This accurately detects stations with defects on the production line and enables rapid location of product defects.

[0174] It should be understood that the above are merely illustrative examples and do not constitute any limitation on the technical solutions of the present invention. In specific applications, those skilled in the art can make settings as needed, and the present invention does not impose any restrictions on this.

[0175] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of this invention. In practical applications, those skilled in the art can select some or all of the workflow to achieve the purpose of this embodiment according to actual needs, and no restrictions are imposed here.

[0176] In addition, for technical details not described in detail in this embodiment, please refer to the defective workstation detection method provided in any embodiment of the present invention, which will not be repeated here.

[0177] Furthermore, it should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0178] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0179] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory (ROM) / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0180] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for detecting defective workstations, characterized in that, The defective workstation detection method includes: Based on the adverse information within the testing period, each product passing through the station is classified, and the defective and qualified samples in each product passing through the station are determined based on the product classification results. Each defective sample is matched with each qualified sample to obtain sample pairs; The sample pairs are input into a probability analysis model based on a Bayesian-Markov chain to perform probability analysis and obtain the defect probability information of each workstation. Based on the defect probability information, each workstation is screened, and the defective workstations in each workstation are determined based on the screening results. The step of screening each workstation based on the defect probability information and determining the defective workstations in each workstation based on the screening results includes: The probability of failure is input into a pre-built confidence analysis model to perform confidence analysis and obtain the confidence analysis results. Determine whether the confidence analysis results meet the preset conditions; If so, obtain the pass rate information for each workstation; Based on the pass rate information and the confidence analysis results, each workstation is screened, and the defective workstations in each workstation are determined based on the screening results. The step of matching each defective sample with each qualified sample to obtain sample pairs includes: Obtain product information for each product passing through a station during the time period to be tested; Based on the product information, determine the production batch number and material information of each product passing through the station; Based on the production batch number and the material information, each defective sample is matched with each qualified sample to obtain a sample pair, wherein the sample pair contains the defective sample and the qualified sample corresponding to the defective sample. Before inputting the sample pairs into the probability analysis model for probability analysis to obtain the defect probability information of each workstation, the method further includes: Obtain process information for each workstation; Determine the node location information corresponding to each workstation based on the process information; The process structure of each workstation is determined based on the node location information; A probabilistic analysis model is constructed based on the process structure.

2. The defective workstation detection method as described in claim 1, characterized in that, The step of matching each defective sample with each qualified sample to obtain sample pairs includes: Identify the adverse factors corresponding to each adverse sample based on the adverse information; The adverse samples are classified according to the aforementioned adverse factors; Multiple sets of defective samples are constructed based on the classification results, wherein each set of defective samples includes at least one defective sample. Each set of defective samples is matched with each set of qualified samples to obtain sample pairs.

3. The defective workstation detection method as described in claim 1, characterized in that, Before classifying the products passing through each station based on the defect information within the testing time period, and determining the defective and qualified samples in each product passing through station based on the product classification results, the process also includes: The time period to be tested is determined according to the preset detection strategy; Obtain the record information of each workstation within the time period to be tested; The recorded information is then cleaned. Extract the station transit information of each workstation within the time period to be tested from the recorded information after data cleaning; The transit products within the time period to be tested are determined based on the transit information.

4. The defective workstation detection method according to any one of claims 1 to 3, characterized in that, The step of inputting the sample pairs into a probability analysis model for probability analysis to obtain the defect probability information for each workstation includes: Each sample in the sample pair is labeled according to the product classification results; Obtain the workstation name corresponding to each sample; Each sample in the sample pair is encoded based on the labeling results and the workstation name; The encoded sample pairs are input into a probability analysis model for probability analysis to obtain the defect probability information of each workstation.

5. The defective workstation detection method as described in claim 1, characterized in that, Before inputting the adverse probability information into a pre-built confidence analysis model for confidence analysis and obtaining the confidence analysis results, the method further includes: Obtain information on adverse events within a historical time period; Based on the adverse event information, identify the historical adverse events that occurred at each workstation within the historical time period; Determine the historical failure probability information for each workstation based on each historical adverse event; The confidence parameters for each workstation are determined based on the historical failure probability information. A confidence analysis model is constructed based on each workstation and its corresponding confidence parameters.

6. A defective workstation detection device, characterized in that, The defective workstation detection device includes: The sample selection module is used to classify products passing through each station based on the defect information within the test time period, and to determine the defective and qualified samples in each product passing through the station based on the product classification results. The sample matching module is used to match each defective sample with each qualified sample to obtain sample pairs. The sample matching module is further configured to perform sample matching between each defective sample and each qualified sample to obtain sample pairs, including: obtaining product information of each product passing through the station during the test time period; determining the production batch number and material information of each product passing through the station based on the product information; and performing sample matching between each defective sample and each qualified sample based on the production batch number and the material information to obtain sample pairs, wherein each sample pair contains a defective sample and a qualified sample corresponding to the defective sample. The probability analysis module is used to input the sample pairs into a probability analysis model based on a Bayesian-Markov chain to perform probability analysis and obtain the defect probability information of each workstation. The probability analysis module, before inputting the sample pairs into the probability analysis model to perform probability analysis and obtain the defect probability information of each workstation, further includes: acquiring the process information of each workstation; determining the node position information corresponding to each workstation based on the process information; determining the process structure of the process in which each workstation is located based on the node position information; and constructing a probability analysis model based on the process structure. The workstation screening module is used to screen each workstation according to the defect probability information, and to determine the defective workstations in each workstation according to the screening results. Specifically, the workstation screening module is used to input the defect probability information into a pre-built confidence analysis model for confidence analysis to obtain confidence analysis results; determine whether the confidence analysis results meet preset conditions; if so, obtain the pass rate information of each workstation; screen each workstation according to the pass rate information and the confidence analysis results, and determine the defective workstations in each workstation according to the screening results.

7. A defective workstation detection device, characterized in that, The defective workstation detection device includes: a memory, a processor, and a defective workstation detection program stored in the memory and executable on the processor, wherein the defective workstation detection program is configured to implement the defective workstation detection method as described in any one of claims 1 to 5.

8. A storage medium, characterized in that, The storage medium stores a defective workstation detection program, which, when executed by a processor, implements the defective workstation detection method as described in any one of claims 1 to 5.

Images