Process map-based industrial production quality root cause positioning method and system

Abstract

This invention provides a method and system for root cause localization of industrial production quality based on process graphs, relating to the field of industrial production technology. The method includes: acquiring raw process data from a manufacturing execution system; parsing the raw process data to obtain standard process data; establishing a mapping relationship from product type to process chain based on the standard process data; constructing a process graph based on the mapping relationship; establishing a parameter normal range benchmark table; acquiring the batch number of the finished product to be traced; performing reverse tracing of the batch number of the finished product to be traced based on the process graph to obtain the defect batch tracing result; calculating the anomaly score of each defect batch parameter using an anomaly detection algorithm based on the defect batch tracing result and the parameter normal range benchmark table; calculating the comprehensive suspicion score of each defect batch parameter based on the anomaly score to obtain a root cause candidate list; and filtering the root cause candidate list to obtain the industrial production quality root cause localization result.

Description

Technical Field

[0001] This invention relates to the field of industrial production technology, and in particular to a method and system for locating the root causes of industrial production quality issues based on process maps. Background Technology

[0002] In continuous and batch production processes in industry, manufacturing execution systems generate massive amounts of data on processes, equipment, materials, parameters, and quality inspection. The production process is characterized by long links, numerous process nodes, and complex parameter coupling. Once a quality defect occurs in the finished product, it is necessary to quickly complete batch traceability, anomaly location, and root cause analysis to support process optimization, defect containment, and closed-loop quality control. Therefore, full-link quality traceability and accurate root cause location in the production process have become core requirements in intelligent manufacturing and lean production.

[0003] Existing technologies typically employ traditional ledger retrieval or single-dimensional data query methods for quality traceability. These methods can only provide simple information on the source of raw materials or the destination of finished products based on batch numbers. While some solutions introduce simple data statistical comparisons, their data organization is fragmented and lacks strong correlation. They fail to form a structured process correlation model and rely solely on manual experience to screen key parameters and determine the degree of anomalies. Consequently, the traceability process is fragmented and the analysis methods are crude.

[0004] However, existing technologies suffer from poor correlation of multi-source data, insufficient traceability integrity and automation, strong subjectivity in anomaly identification and single evaluation dimensions, resulting in low root cause localization accuracy and difficulty in implementing and reusing analysis results. Summary of the Invention

[0005] To address the technical problems of existing technologies, such as poor correlation of multi-source production data, insufficient traceability integrity and automation, strong subjectivity in anomaly identification and single evaluation dimensions, resulting in low root cause localization accuracy and difficulty in implementing and reusing analysis results, this invention provides an industrial production quality root cause localization method and system based on process maps.

[0006] The technical solutions provided by the embodiments of the present invention are as follows: The first aspect of this invention provides a method for identifying the root causes of industrial production quality issues based on process maps, comprising: S1: Obtain raw process data from the manufacturing execution system; S2: Analyze the original process data to obtain standard process data; S3: Based on standard process data, establish a mapping relationship between product type and process chain; S4: Construct a process map based on the mapping relationship; S5: Establish a reference table for normal parameter ranges; S6: Obtain the batch number of the finished product to be traced; S7: Based on the process diagram, reverse trace the batch number of the finished product to be traced to obtain the traceability results of the defective batch; S8: Based on the defect batch traceability results and the parameter normal range benchmark table, calculate the abnormality score of each defect batch parameter through the anomaly detection algorithm; S9: Based on the anomaly score, calculate the comprehensive suspicion score of each defect batch parameter to obtain the root cause candidate list; S10: Screen the root cause candidate list to obtain the root cause localization results for industrial production quality.

[0007] A second aspect of this invention provides an industrial production quality root cause localization system based on process maps, comprising: processor; The memory stores computer-readable instructions, which, when executed by a processor, implement the industrial production quality root cause localization method based on process maps as described in the first aspect.

[0008] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the industrial production quality root cause localization method based on process map as described in the first aspect.

[0009] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: In this embodiment of the invention, a process map is constructed based on the mapping relationship between product type and process chain, and reverse tracing is performed on the batch number of the finished product to be traced based on the process map. This can improve the correlation of multi-source data, traceability completeness and automation. Combining the traceability results of defective batches and the normal range benchmark table of parameters, the abnormality score of each defective batch parameter is calculated by the anomaly detection algorithm. This can reduce the subjectivity of anomaly judgment. Then, the comprehensive suspicion score of each defective batch parameter is calculated based on the anomaly score to obtain the root cause candidate list. This can enrich the evaluation dimensions, improve the positioning accuracy and help improve the current situation where the analysis results are not easy to implement and reuse. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a schematic flowchart of a method for locating the root causes of industrial production quality based on process diagrams, provided in an embodiment of the present invention.

[0012] Figure 2This is a schematic diagram of the structure of an industrial production quality root cause localization system based on process diagrams, provided in an embodiment of the present invention. Detailed Implementation

[0013] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0014] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0015] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0016] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0017] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0018] Reference manual attached Figure 1 The diagram shows a flowchart of an industrial production quality root cause localization method based on process diagrams provided by an embodiment of the present invention.

[0019] This invention provides a method for identifying the root causes of quality issues in industrial production based on process maps. This method can be implemented using a process map-based industrial production quality root cause identification device, which can be a terminal or a server. The processing flow of the process map-based industrial production quality root cause identification method may include the following steps: S1: Obtain raw process data from the manufacturing execution system.

[0020] S2: Analyze the original process data to obtain standard process data.

[0021] It should be noted that the data access module connects to the Manufacturing Execution System (MES) database, parses the MES system data tables, and extracts the process definition table, process route table, daily production reports for each process, equipment status table for each process, inventory flow table, material reports for each process, and quality inspection record table. A data dictionary is established, and field naming and data types are standardized.

[0022] In this embodiment of the invention, a dedicated data access module is developed to connect to the Manufacturing Execution System (MES) database. This module is used to analyze and extract full-dimensional production data tables, including process definitions, process routes, daily production reports, equipment status, inventory flow, material reports, and quality inspection records. At the same time, a unified data dictionary is established and field naming and data types are standardized. This not only achieves efficient integration and accurate extraction of scattered process data in the MES system, ensuring the comprehensiveness and integrity of the original process data, but also eliminates the field differences and data format chaos between various data tables through standardization processing, forming standardized and unified process data.

[0023] S3: Based on standard process data, establish a mapping relationship between product type and process chain.

[0024] It should be noted that the process route definition table is read and the standard process chain for each product type is parsed. The process chain defines the sequence of processes that a batch should go through, such as batching, mixing, settling, pressing, drying, firing, and sorting. A mapping relationship between product type and process chain is established to support differentiated process routes for multiple product types.

[0025] In one possible implementation, S3 specifically includes sub-steps S301 to S304: S301: Determine the standard process route for each product based on standard process data.

[0026] Specifically, using the product type code as the primary key, the product master data table is linked to the process route master table to determine the currently effective standard process route for each product. If multiple route versions exist for the same product, a unique and valid route definition is selected by combining the activation status, version number, and effective time.

[0027] S302: Determine the standard process chain based on the standard process route.

[0028] Specifically, based on the established process route, all process records under that route are extracted from the process route details table and sorted in ascending order of process sequence number to reconstruct a complete standard process chain, such as "ingredient preparation, mixing and grinding, settling, pressing, drying, firing, and sorting".

[0029] S303: Associate the process codes and process definition tables in the standard process chain to obtain a structured process chain.

[0030] Specifically, the codes of each process in the process chain are further linked with the process definition table, and basic attributes such as process name, process category, work section, corresponding equipment type and key control parameters are completed, so that the process chain is transformed from a simple sequential list into a structured process chain that can be traced and analyzed.

[0031] S304: Construct a mapping relationship based on the structured process chain.

[0032] It should be noted that the compiled results are formed into a mapping table or mapping dictionary of "product type corresponding to process chain", that is, each product type corresponds to a standard process chain arranged in sequence. In this way, different product types can be mapped to different process routes to meet the requirements of differentiated production.

[0033] Furthermore, if there are experimental processes, customer-customized processes, or route variations, the standard process chain is replaced, inserted, or trimmed based on the standard mapping and conditions such as order attributes, formula type, production line, and time version to generate the actual process sequence to be executed for a specific batch. After the mapping relationship is established, the system also needs to check whether each product can be mapped to a unique and valid process route, and check whether the process sequence is complete and whether there are any broken chains, duplicates, or missing key processes. After verification, the mapping relationship is used as the basis for subsequent map construction and batch traceability.

[0034] In this embodiment of the invention, by analyzing standard process chains for each product type based on standard process data and establishing a mapping relationship between product type and process chain, and by associating product master data with process route master table to filter unique valid routes, extracting process records to restore standard process chains, and associating process definition table to complete attributes to form structured process chains, a series of operations to finally construct a standardized mapping table / dictionary can not only match unique, standardized process chains with complete process attributes for each product type, achieving accurate adaptation of differentiated process routes for multiple product types, but also flexibly adjust standard process chains according to actual production conditions such as orders and formulas to generate actual execution process sequences for specific batches. At the same time, by verifying the uniqueness of the mapping relationship and the completeness of the process sequence, problems such as broken chains, duplication, and missing key processes are effectively avoided. This transforms the process route from a simple list of process sequences into traceable and analyzable structured data, providing a unified and standardized process framework for the accurate construction of subsequent production data maps, and laying a clear process logic foundation for process traversal and material chain tracking during reverse tracing of defective batches, ensuring the accuracy, completeness, and flexibility of process link tracing in the subsequent quality root cause localization process.

[0035] S4: Construct a process map based on the mapping relationship.

[0036] Specifically, the daily production data for each process is traversed, and data is aggregated by batch number. For each batch, its processing records for each process are extracted, including processing time, equipment used, process parameters, and input materials. Equipment operation data in the equipment status table is linked to the batch node through timestamp matching. Operator information is linked through shift numbers. Flow edges between batch nodes are constructed to form a complete process graph.

[0037] In this embodiment of the invention, a process graph is constructed based on the mapping relationship from product type to process chain. By aggregating daily production data of each process by batch number, core processing records such as processing time, equipment used, process parameters, and input materials are extracted. Equipment operation data and operator information are associated with timestamps and shift numbers respectively. At the same time, flow edges between batch nodes are constructed to form a complete graph. This not only realizes the deep association and integration of production, equipment, personnel, and material data scattered in various data tables with batch as the core, breaking down data silos, but also transforms the abstract process route into a concrete and structured graph model. This allows the flow trajectory of the batch in the entire process and the status information of each link to form a traceable and traversable complete link. This provides a clear node-based data carrier for subsequent reverse tracing of defective batches, enabling rapid and accurate extraction of batch-wide process association information. It also provides complete and unified end-to-end data support for anomaly calculation and root cause localization, ensuring the efficiency and relevance of data retrieval in subsequent quality analysis work, and significantly improving the accuracy and efficiency of quality root cause localization.

[0038] S5: Establish a baseline table for normal parameter ranges.

[0039] Specifically, recent qualified batches (e.g., qualified batches from the last 6 months) are selected from quality inspection records as the normal batch group. The parameter distribution characteristics of the normal batch group are statistically analyzed by process and parameter dimensions, including statistics such as mean, standard deviation, and quantiles. A parameter normal range benchmark table is established as a reference benchmark for subsequent anomaly detection.

[0040] In one possible implementation, S5 specifically includes sub-steps S501 to S503: S501: Select normal batches from quality inspection records.

[0041] S502: Statistical distribution characteristics of normal batch groups.

[0042] Optionally, the parameter distribution characteristics specifically include: mean, standard deviation, and quantiles.

[0043] S503: Establish a reference table for the normal range of parameters based on the parameter distribution characteristics.

[0044] In this embodiment of the invention, a normal sample group is formed by selecting qualified batches within a preset period from quality inspection records, and statistically analyzing the distribution characteristics such as mean, standard deviation, and quantiles according to process and parameter dimensions. This establishes a parameter normal range benchmark table, which can form a quantitative judgment standard that fits the process conditions based on compliant data from actual production. This avoids the subjectivity and bias caused by manual experience thresholds, achieving standardization and objectivity of the process parameter judgment basis. Simultaneously, this benchmark table accurately corresponds to each process and parameter, providing a unified, reliable, and adaptable reference basis for subsequent anomaly detection. It effectively improves the accuracy and stability of parameter deviation identification, shields against interference from operating condition fluctuations, and makes anomaly calculation and parameter root cause judgment more scientific and reliable. Furthermore, it provides clear and standardized data support for subsequent tracing and comparative analysis, further optimizing the accuracy and reliability of overall quality root cause localization.

[0045] S6: Obtain the batch number of the finished product to be traced.

[0046] S7: Based on the process diagram, reverse trace the batch number of the finished product to be traced to obtain the traceability results of the defective batch.

[0047] It should be noted that when a defective batch number is entered, the batch tracking module is invoked for reverse tracing. The data map is traversed backwards along the process route to extract the complete parameter vectors for that batch at each process node. A full-chain traceability result for the defective batch is constructed, including equipment information, process parameters, material information, and personnel information for each process.

[0048] In one possible implementation, S7 specifically includes sub-steps S701 to S706: S701: Initialize the traceability chain and the current traceability batch.

[0049] It should be noted that the process involves reading the batch number of the finished product to be traced, initializing the traceability chain and the current traceability batch, and verifying whether the process diagram and process route are available.

[0050] S702: Construct a backtracking process sequence based on the product type corresponding to the batch number of the finished product to be traced.

[0051] Specifically, the standard process route is read according to the product type, the process list is obtained and reversed to form a backtracking process sequence from the finished product end to the raw material end.

[0052] S703: Based on the backtracking process sequence, find the node corresponding to the batch number of the finished product to be traced in the process diagram.

[0053] S704: After organizing the nodes corresponding to the batch numbers of the finished products to be traced into a unified structure, add them to the traceability chain to obtain the process nodes.

[0054] It should be noted that, in reverse order, the process diagram is searched for the node corresponding to "current batch number + current process number". If found, the batch number, process code, process name, start and end time, equipment number, equipment status, process parameters, input materials and operators are extracted, organized into a unified structure and added to the traceability chain; if not found, the missing process is recorded and the backtracking continues.

[0055] S705: Check if there are material splits, batch mappings, or intermediate batch conversions at the process node. If so, update the current traceable batch to the source material batch. Otherwise, continue tracing back along the current batch number.

[0056] Specifically, check if there is material splitting, batch mapping, or intermediate batch conversion at the current process node. If so, update the current traceable batch to the source material batch to ensure that subsequent backtracking continues along the real material chain; if not, continue backtracking along the current batch number.

[0057] S706: Adjust and statistically analyze the traceability chain to obtain the traceability results of defective batches.

[0058] It should be noted that after all processes are completed, the traceability chain is adjusted to a forward display order from the raw material end to the finished product end, the actual number of processes hit is counted, and the target batch number, traceability chain, total number of processes and traceability time are encapsulated to generate the final traceability result output.

[0059] In this embodiment of the invention, reverse tracing is performed on the batch number of the finished product to be traced based on the constructed process map, and defect batch tracing results are generated. By initializing tracing parameters, verifying the validity of the map, generating a reverse backtracking process sequence according to product type, matching map nodes step by step and extracting full-dimensional information on equipment, process parameters, materials and personnel, identifying material splitting and batch conversion relationships to follow the real material chain backtracking, and finally organizing the tracing link sequence and encapsulating and outputting complete tracing results, it can achieve accurate tracing and full coverage of the entire process, process and elements of defect batches, effectively avoiding problems such as broken tracing chains, missing information and misaligned paths, and completely restoring the historical working conditions and production conditions in the defect generation process. It can not only provide a clear, unified and complete original data foundation for subsequent anomaly detection and parameter comparison, but also automatically distinguish missing process nodes and batch flow relationships, greatly improving tracing efficiency and accuracy, reducing the difficulty of manual investigation, and enabling subsequent root cause analysis to quickly locate the source of anomalies and lock key production links, thereby enhancing the practicality and reliability of the overall quality root cause localization method.

[0060] S8: Based on the defect batch traceability results and the parameter normal range benchmark table, calculate the abnormality score of each defect batch parameter using an anomaly detection algorithm.

[0061] It should be noted that the anomaly propagation analysis module is invoked to compare each parameter of the defective batch with the baseline of the normal batch group. The anomaly score for each parameter is calculated using Z-Score or quantile methods. Parameters with anomalies exceeding the threshold are identified and marked as potential anomalies. The propagation relationship of anomalous parameters between processes is analyzed to identify the source process of the anomaly.

[0062] Specifically, the process begins by reading the defect batch traceability results, normal batch samples, and anomaly detection methods. An anomaly analysis result container is initialized, and the integrity of the traceability chain, the number of normal samples, and the legality of the anomaly detection method parameters are verified. Each step in the defect batch traceability chain is traversed, the current step code is extracted, and the corresponding parameter sample set for the same step is extracted from the normal batch, creating an anomaly analysis result container for the current step. The process continues to traverse each parameter under the current step, reading the parameter name and value, extracting the corresponding normal sample sequence, and performing null value removal, illegal value cleaning, and numerical type unification operations on the samples. If the processed sample is empty, the parameter is skipped. The parameter anomaly degree is calculated based on the selected anomaly detection method. If the Z-Score algorithm is used, the standardized deviation of the parameter value is calculated based on the mean and standard deviation of the normal samples. If the quantile algorithm is used, the quantile position of the parameter value in the normal samples is calculated and converted into the deviation relative to the median. If the Mahalanobis distance algorithm is used, a parameter vector is constructed at the step level, and the Mahalanobis distance is calculated by combining the mean vector and the covariance matrix. The parameters are judged to be abnormal according to the threshold rules corresponding to the current anomaly detection method. The current value of the parameter, the normal mean, the normal standard deviation, the anomaly score and the anomaly mark are written into the anomaly analysis results. Finally, the anomaly analysis results at the process plus parameter level are generated and output.

[0063] Optionally, the anomaly detection algorithms include: Z-Score algorithm, quantile algorithm, and Mahalanobis distance algorithm.

[0064] The Z-Score algorithm calculates the standard deviation distance between the parameter values ​​of the defective batch and the mean of the normal group. Parameters whose absolute Z-Score values ​​exceed the threshold (usually 2 or 3) are marked as abnormal.

[0065] The quantile algorithm calculates the quantile position of the defective batch parameter values ​​in the normal population distribution. Parameters located in extreme quantile intervals (such as below 5% or above 95%) are marked as abnormal.

[0066] Among them, the Mahalanobis distance algorithm considers the correlation between parameters, calculates the Mahalanobis distance between the parameter vector of the defective batch and the center of the normal group, and the batches whose distance exceeds the threshold are judged as abnormal.

[0067] It should be noted that if the Z-Score method is used to calculate the anomaly score, firstly, based on the normal sample data in the parameter normal range benchmark table, the normal sample mean and normal sample standard deviation are calculated. The normal sample mean is the ratio of the sum of all normal sample values ​​to the number of normal samples. The normal sample standard deviation is the square root of the ratio of the sum of the squares of the differences between each normal sample value and the normal sample mean to the number of normal samples. Then, using the normal sample mean and normal sample standard deviation calculated above, the Z-Score anomaly score of the current defect batch parameter is calculated. The Z-Score anomaly is the ratio of the absolute value of the difference between the current parameter value and the mean of the normal samples to the standard deviation of the normal samples. The current parameter value is the actual value of the parameter corresponding to the defective batch. The mean of the normal samples represents the center position of the normal samples, and the standard deviation of the normal samples represents the fluctuation range of the normal samples. The larger the Z-Score anomaly value, the more obvious the deviation of the current parameter value from the normal samples. When the Z-Score anomaly value is greater than the preset threshold, it is judged as a significant anomaly. If the calculated result of the standard deviation of the normal samples is 0, it is processed according to the preset fallback rule.

[0068] It should be noted that if the quantile method is used to calculate the anomaly score, firstly, the normal sample data in the parameter normal range benchmark table are sorted in ascending order of value. Then, the quantile position of the parameter value of the current defect batch in the sorted normal samples is calculated. This quantile position, i.e., the quantile, is the ratio of the number of normal samples less than or equal to the current parameter value to the total number of normal samples. Then, the calculated quantile is converted into quantile anomaly, which is the result of multiplying the absolute value of the difference between the quantile and 0.5 by 2. Here, 0.5 represents the median position of the normal sample data. The closer the quantile value is to 0 or 1, the higher the quantile anomaly; the closer the quantile value is to 0.5, the lower the quantile anomaly.

[0069] Furthermore, after the anomaly calculation is completed, the current parameter value, normal mean, normal standard deviation, quantile position, and anomaly score are written into the anomaly analysis results as input for subsequent suspicion calculation.

[0070] In this embodiment of the invention, by combining the complete traceability results of defective batches with a pre-established benchmark table of normal parameter ranges, the abnormality scores of each process parameter are automatically quantified and calculated using a variety of switchable anomaly detection algorithms. First, the legality of the traceability link, sample quantity, and algorithm parameters is verified and the data is cleaned and organized. Then, the normal sample distribution is matched process by process and parameter by parameter. Anomalies are accurately identified from the dimensions of single parameter deviation, distribution position, and multi-parameter correlation based on Z-Score, quantiles, or Mahalanobis distance. This avoids the subjectivity, errors, and omissions caused by manual judgment, while also taking into account independent parameter deviations and coupling relationships between parameters, forming structured anomaly location results at the process and parameter levels. At the same time, anomaly thresholds are identified, quantitative indicators are recorded, and the transmission relationship of anomalies between processes is sorted out, quickly locking the initial anomaly source process. This provides objective, quantitative, and standardized input basis for subsequent normalization processing, comprehensive suspicion calculation, and root cause ranking, significantly improving the comprehensiveness of anomaly identification, the rigor of the calculation process, and the accuracy and automation of quality deviation tracing. It effectively shortens the investigation cycle, reduces on-site dependence, and enhances the stability and versatility of the entire root cause location method.

[0071] S9: Based on the anomaly score, calculate the comprehensive suspicion score of each defect batch parameter to obtain the root cause candidate list.

[0072] Specifically, load the parameter importance weight configuration (which can be obtained through expert experience or historical data analysis). Query historical defect data and calculate the historical correlation strength between each parameter and the current defect type. Calculate the overall suspicion score for each parameter by combining three factors: anomaly, importance, and historical correlation. Sort the parameters in descending order of suspicion score to generate a root cause candidate list.

[0073] In one possible implementation, S9 specifically includes sub-steps S901 to S903: S901: Calculate the historical correlation strength between the parameters of each defect batch and the current defect type.

[0074] The historical correlation strength can be calculated using the basic form, which is the ratio of the number of batches with abnormal parameters belonging to the current defect type to the total number of batches of the current defect type in history. A higher historical correlation strength indicates a higher frequency of the parameter occurring simultaneously with the current type of quality defect in historical production processes. To enhance the ability of historical correlation strength to distinguish the relationship between parameters and defect types, an enhanced form can be used. The enhanced form is the ratio of the number of batches with abnormal parameters belonging to the current defect type to the total number of batches of the current defect type in history, divided by the ratio of the total number of batches with abnormal parameters to the total number of batches in history. This calculation method directly reflects the concentration of abnormal parameter occurrences within the current defect type.

[0075] Furthermore, after obtaining the anomaly degree, parameter importance, and historical correlation strength of the parameters, these three factors are first normalized to ensure they are on the same scale. The calculated anomaly degree is denoted as the anomaly degree factor, the parameter importance as the importance factor, and the calculated historical correlation strength as the historical correlation factor. The normalized anomaly degree factor is calculated using the formula min(Z-Score anomaly degree / maximum reference anomaly value, 1). If the quantile method is used to calculate the anomaly degree, no additional normalization is required, and the quantile anomaly degree itself can be directly used as the anomaly degree factor.

[0076] S902: Calculate the comprehensive suspicion score for each defect batch parameter based on historical correlation strength and anomaly score.

[0077] Specifically, the overall suspicion level is calculated using a weighted summation method. The formula for overall suspicion level is: Overall Suspicion Level = Anomaly Weight × Anomaly Factor + Importance Weight × Importance Factor + Historical Correlation Weight × Historical Correlation Factor, where the sum of the anomaly weight, importance weight, and historical correlation weight is 1. A higher overall suspicion level indicates a more significant deviation of the corresponding parameter in the current production process, higher importance at the process level, and a higher frequency of association between this parameter and current quality defects in historical production. Such parameters should be prioritized for root cause analysis.

[0078] S903: Sort the comprehensive suspicion scores in descending order to obtain the root cause candidate list.

[0079] In this embodiment of the invention, based on the calculated parameter anomaly score, combined with configurable process parameter importance weights and the historical correlation strength between parameters and defect types obtained from historical defect data, the correlation degree is quantified through basic or enhanced forms. Then, the anomaly, importance, and historical correlation strength are normalized to eliminate dimensional differences. A weighted summation method is then used to fuse the three factors to calculate the comprehensive suspicion score for each defect batch parameter. Finally, a structured root cause candidate list is generated by arranging the scores in descending order. This method takes into account the actual deviation, the inherent importance of the process, and the historical fault correlation patterns, overcoming the one-sidedness of single-indicator judgment and the subjectivity of human experience. This makes the scoring logic more comprehensive, objective, and aligned with actual production, accurately prioritizing high-risk, highly correlated, and high-weight parameters. It automatically forms a clear and focused candidate root cause list, providing quantitative basis and orderly reference for subsequent screening and identification of the final root cause. This significantly reduces the difficulty of manual investigation and significantly improves the intelligence, accuracy, and efficiency of quality root cause localization.

[0080] S10: Screen the root cause candidate list to obtain the root cause localization results for industrial production quality.

[0081] In one possible implementation, S10 specifically involves: screening the root cause candidate list, extracting root cause candidates with a comprehensive suspicion level below a preset value, and obtaining the root cause localization result for industrial production quality.

[0082] Those skilled in the art can set the size of the preset value according to actual needs, and the present invention does not limit this.

[0083] It should be noted that the top N (e.g., Top-5) root cause candidates with the highest suspicion level are extracted. A structured traceability report is generated, including basic information on the defective batch, a traceability diagram, anomaly parameter analysis, a root cause ranking list, and improvement suggestions. Visualized reports in PDF or HTML format are supported for export.

[0084] In this embodiment of the invention, by extracting key candidates with high comprehensive suspicion from the root cause candidate list based on preset screening conditions, the core causes are quickly identified and the final industrial production quality root cause localization results are output. At the same time, a structured traceability report is automatically generated, which includes basic information of defective batches, a complete traceability chain diagram, abnormal parameter analysis, a root cause ranking list, and optimization and improvement suggestions. It also supports visual export in PDF, HTML and other formats. This not only allows for flexible configuration of screening thresholds according to on-site process requirements and adaptation to different production scenarios, but also simplifies the scope of investigation, focuses on high-risk core factors, avoids redundant analysis, and significantly reduces manual screening costs and judgment errors. Furthermore, it transforms complex traceability and calculation results into intuitive, easy-to-understand, archiveable, traceable, and reusable standardized visual documents, facilitating process rectification, quality review, on-site implementation, and management traceability. This achieves a closed loop from data calculation and intelligent ranking to result output and business application, further improving the overall practicality, ease of use, and scalability of the method.

[0085] Reference manual attached Figure 2 The diagram shows a structural schematic of an industrial production quality root cause localization system based on process maps provided by the present invention.

[0086] The present invention also provides an industrial production quality root cause localization system 20 based on process maps, applied to the above-mentioned industrial production quality root cause localization method based on process maps, comprising: Processor 201.

[0087] The memory 202 stores computer-readable instructions. When the computer-readable instructions are executed by the processor 201, they implement the industrial production quality root cause localization method based on process map as described in the method embodiment.

[0088] The industrial production quality root cause localization system 20 based on process map provided by the present invention can execute the above-mentioned industrial production quality root cause localization method based on process map and achieve the same or similar technical effects. To avoid repetition, the present invention will not elaborate further.

[0089] It should be understood that the processor in the embodiments of the present invention can be a central processing unit (CPU), or it can be 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. The general-purpose processor can be a microprocessor or any conventional processor.

[0090] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0091] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0092] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0093] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0094] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0095] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0096] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0097] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0098] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0099] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0100] If the aforementioned functions 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, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0101] This invention provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the industrial production quality root cause localization method based on process map as described in the method embodiment.

[0102] The present invention provides a computer-readable storage medium that can implement the steps and effects of the above-described method embodiment of the industrial production quality root cause localization method based on process map. To avoid repetition, the present invention will not repeat them.

[0103] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0104] The following points need to be explained: (1) The accompanying drawings of the embodiments of the present invention only involve the structures involved in the embodiments of the present invention. Other structures can refer to the general design.

[0105] (2) For clarity, the thickness of layers or regions is enlarged or reduced in the drawings used to describe embodiments of the invention, i.e., these drawings are not drawn to scale. It is understood that when an element such as a layer, film, region or substrate is referred to as being “above” or “below” another element, the element may be “directly” located “above” or “below” the other element or there may be intermediate elements.

[0106] (3) Where there is no conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other to obtain new embodiments.

[0107] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. The scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for locating the root causes of quality issues in industrial production based on process maps, characterized in that, include: S1: Obtain raw process data from the manufacturing execution system; S2: Parse the original process data to obtain standard process data; S3: Based on the standard process data, establish a mapping relationship between product type and process chain; S4: Construct a process map based on the mapping relationship; S5: Establish a reference table for normal parameter ranges; S6: Obtain the batch number of the finished product to be traced; S7: Based on the process diagram, the batch number of the finished product to be traced is traced backward to obtain the defect batch traceability result; S8: Based on the defect batch traceability results and the parameter normal range benchmark table, calculate the abnormality score of each defect batch parameter using an anomaly detection algorithm; S9: Based on the anomaly score, calculate the comprehensive suspicion score of each defect batch parameter to obtain the root cause candidate list; S10: Filter the root cause candidate list to obtain the root cause localization results for industrial production quality.

2. The method for identifying the root causes of industrial production quality issues based on process maps according to claim 1, characterized in that, S3 specifically includes: S301: Determine the standard process route for each product based on the aforementioned standard process data; S302: Determine the standard process chain according to the aforementioned standard process route; S303: Associate the process codes and process definition tables in the standard process chain to obtain a structured process chain; S304: Construct the mapping relationship based on the structured process chain.

3. The method for identifying the root causes of industrial production quality issues based on process maps according to claim 1, characterized in that, S5 specifically includes: S501: Select normal batches from quality inspection records; S502: Statistically analyze the parameter distribution characteristics of the normal batch group; S503: Based on the parameter distribution characteristics, establish a reference table for the normal range of the parameters.

4. The method for identifying the root causes of industrial production quality issues based on process maps according to claim 3, characterized in that, The specific distribution characteristics of the parameters include: mean, standard deviation, and quantiles.

5. The method for identifying the root causes of industrial production quality issues based on process maps according to claim 1, characterized in that, Specifically, S7 includes: S701: Initialize the traceability chain and the current traceability batch; S702: Construct a backtracking process sequence based on the product type corresponding to the batch number of the finished product to be traced; S703: Based on the backtracking process sequence, find the node corresponding to the batch number of the finished product to be traced in the process diagram; S704: After organizing the nodes corresponding to the batch number of the finished product to be traced into a unified structure, add them to the traceability chain to obtain the process node; S705: Check if there is material splitting, batch mapping, or intermediate batch conversion at the process node; if so, update the current traceable batch to the source material batch; otherwise, continue tracing back along the current batch number. S706: Adjust and statistically analyze the traceability chain to obtain the traceability results of the defective batch.

6. The method for identifying the root causes of industrial production quality issues based on process maps according to claim 1, characterized in that, The anomaly detection algorithms specifically include: Z-Score algorithm, quantile algorithm, and Mahalanobis distance algorithm.

7. The method for identifying the root causes of industrial production quality issues based on process maps according to claim 1, characterized in that, S9 specifically includes: S901: Calculate the historical correlation strength between the parameters of each defect batch and the current defect type; S902: Based on the historical correlation strength and the anomaly score, calculate the comprehensive suspicion score for each defect batch parameter; S903: Sort the comprehensive suspicion scores in descending order to obtain the root cause candidate list.

8. The method for identifying the root causes of industrial production quality issues based on process maps according to claim 1, characterized in that, Specifically, S10 is: The root cause candidate list is screened, and root cause candidates with a comprehensive suspicion level before a preset value are extracted to obtain the root cause localization results of industrial production quality.

9. An industrial production quality root cause localization system based on process maps, characterized in that, include: processor; A memory storing computer-readable instructions, which, when executed by the processor, implement the industrial production quality root cause localization method based on process maps as described in any one of claims 1 to 8.

10. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the industrial production quality root cause localization method based on process maps as described in any one of claims 1 to 8.

Images