A supply chain data traceability monitoring method and system and a storage medium

Abstract

The application provides a supply chain data traceability monitoring method and system and a storage medium, relates to the technical field of supply chain quality traceability, and comprises the following steps: collecting production data of different material batches in each link of a supply chain, including raw material performance, processing process parameters and quality inspection results; calculating process capability indexes of each node through an industrial internet gateway, synchronously uploading the production data to the cloud, converting the production data into a vector format, and generating a cross-organization data sequence with material batches as the main line; analyzing link correlation based on the sequence, and constructing a quality influence topology graph dynamically reflecting the influence weight of process parameters on the final quality; when a quality inspection finds a batch quality defect, combining the topology graph and the parameter deviation of the defective batch, calculating the prediction probability of the upstream link as the source of the defect, and generating a visual report showing the most possible source link and the conduction path of the defect, so that accurate traceability of the supply chain quality defect and visual presentation of the influence path can be realized.

Description

Technical Field

[0001] This application relates to the field of supply chain quality traceability technology, and in particular to a supply chain data traceability and monitoring method, system and storage medium. Background Technology

[0002] Supply chain information traceability and monitoring is a key means to ensure product quality and improve supply chain management efficiency. It is widely used in manufacturing, food processing, and other fields, and has promising application prospects and practical value. Efficient traceability and monitoring methods can help companies quickly locate quality problems, reduce quality risks, and enhance the overall competitiveness of the supply chain.

[0003] Currently, a common method for supply chain data traceability and monitoring in existing technologies is to collect basic production data from each link of the supply chain through manual recording or simple data collection terminals, and to classify and store the data according to material batches. When quality problems occur, the data from each link is checked manually to investigate the cause.

[0004] However, this existing technology can only achieve simple storage and retrieval of production data, and cannot analyze the correlation and impact between parameters in each stage, nor can it quickly locate the root cause and transmission path of quality defects. This results in low efficiency and insufficient accuracy in quality problem investigation, making it difficult to meet the needs of refined supply chain management. Therefore, existing technologies suffer from low efficiency and insufficient accuracy in tracing the source of supply chain quality defects. Summary of the Invention

[0005] The purpose of this application is to provide a supply chain data traceability and monitoring method, system and storage medium to solve the problems of low efficiency and insufficient accuracy in tracing supply chain quality defects in the prior art.

[0006] To address the aforementioned technical problems, in a first aspect, this application provides a supply chain data traceability and monitoring method, comprising:

[0007] Collect production data corresponding to different material batches at each stage of the supply chain. The production data includes raw material performance data from suppliers, real-time process parameters of each processing station in the plant, and measurement results from the quality inspection station.

[0008] Through the industrial Internet of Things (IIoT) gateway of the supply chain, based on the real-time process parameters and preset specification limits, the process capability indicators of each node in the production process are calculated, and the process capability indicators and the production data are synchronously uploaded to the cloud.

[0009] In the cloud, the process capability indicators and the production data are converted into vector format to generate cross-organizational data sequences with material batches as the main line.

[0010] Based on the cross-organizational data sequence, the correlation between various links in the supply chain is analyzed. Taking each link in the supply chain as a node and the flow relationship between materials and parameters as an edge, a quality impact topology diagram of the supply chain is constructed. The quality impact topology diagram dynamically describes the impact weight of process parameters in different links on the final quality.

[0011] When a batch quality defect is identified during the quality inspection process, based on the quality impact topology map and combined with the specific parameter deviations of the defective batch in the cross-organizational data sequence, the predicted probability of different upstream links as the root cause of the defect is calculated, and a visualization report is generated based on the predicted probability. The visualization report shows the most likely root cause of the quality defect and its transmission path in the supply chain.

[0012] Optionally, based on the cross-organizational data sequence, the relationships between various links in the supply chain are analyzed. Using each link in the supply chain as a node and the flow relationships of materials and parameters as edges, a quality impact topology diagram of the supply chain is constructed. This quality impact topology diagram dynamically describes the influence weights of process parameters at different links on the final quality, including:

[0013] From the cross-organizational data sequence, the processing node parameter set and quality inspection node feature set of adjacent links are extracted. When it is detected that the direction of change of the value of a specific process parameter in the previous link is consistent with the direction of change of the value of the quality feature in the subsequent link, it is determined that there is a correlation.

[0014] A one-way influence edge is established for adjacent links with the aforementioned relationship, and the frequency of occurrence of the relationship in all production batches is statistically analyzed as the dynamic weight value of the influence edge.

[0015] Using supply chain links as nodes, the unidirectional influence edges as connections, and the dynamic weight values ​​as edge weights, an initial quality influence topology graph is constructed.

[0016] Whenever a new batch of cross-organizational data sequences is added, the frequency of occurrence of synchronization changes is recalculated to update the dynamic weight values ​​of the initial quality impact topology, resulting in a dynamically updated quality impact topology.

[0017] Optionally, when it is detected that the direction of change of a specific process parameter value in the preceding stage is consistently consistent with the direction of change of the quality characteristic value in the following stage, a correlation is determined to exist, including:

[0018] From the cross-organizational data sequence, a first sequence of specific process parameter values ​​from the previous stage is extracted and arranged in chronological order, and a second sequence of quality characteristic values ​​from the next stage is extracted and arranged in chronological order. The first and second sequences cover the same multiple consecutive production batches.

[0019] For each production batch in the first sequence and the second sequence, except for the earliest batch, the values ​​of specific process parameters of the current batch are compared with those of the adjacent previous batch to generate a parameter change direction indicator, and the values ​​of quality characteristics of the current batch are compared with those of the adjacent previous batch to generate a characteristic change direction indicator.

[0020] For each production batch, compare the parameter change direction flag with the feature change direction flag. When both flags indicate an increase in value or a decrease in value, record a direction consistency event.

[0021] Within a moving window containing a predetermined number of production batches, the number of occurrences of the direction alignment event is counted, and the ratio of the number of occurrences to the total number of batches within the moving window is calculated.

[0022] When the ratio exceeds a preset consistency threshold, it is determined that there is a correlation between the previous step and the next step.

[0023] Optionally, when a batch quality defect is identified during the quality inspection process, based on the quality impact topology map and combined with the specific parameter deviations of the defective batch in the cross-organizational data sequence, the predicted probability of different upstream links as the root cause of the defect is calculated, and a visualization report is generated based on the predicted probability. The visualization report shows the most likely root cause of the quality defect and its transmission path in the supply chain, including:

[0024] When a batch quality defect is identified during the quality inspection process, the out-of-limit quality characteristics are extracted from the cross-organizational data sequence of the corresponding batch.

[0025] Taking the quality inspection node as the endpoint, in the quality influence topology graph, all upstream paths pointing to the quality inspection node are traversed in reverse to extract the actual values ​​of process parameters of the processing nodes in the defective batch in each upstream path.

[0026] Calculate the deviation between the actual value of the process parameter and the historical normal batch average value of the processing node. Based on the deviation and the dynamic weight value from the node to the downstream node, calculate the path contribution of the processing node.

[0027] The path contribution of the same processing node in all paths is accumulated, and the predicted probability of the corresponding processing stage as the source of quality defects is calculated according to the proportion of the total contribution.

[0028] Based on the processing node with the highest predicted probability and its transmission path in the supply chain, a tree-structured visualization report is generated.

[0029] Optionally, the deviation between the actual value of the process parameter and the historical normal batch average of the processing node is calculated. Based on the deviation and in conjunction with the dynamic weight value of the edge from the node to the downstream node, the path contribution of the processing node is calculated, including:

[0030] From the cross-organizational data sequence, the process parameter values ​​of the processing node in a preset number of historical normal batches are selected, and the average value of the process parameter values ​​in the historical normal batches is calculated as the historical normal batch average.

[0031] The relative offset is calculated based on the actual values ​​of the process parameters of the processing nodes in the defective batch and the average values ​​of the historical normal batches.

[0032] In the quality impact topology graph, the dynamic weight value of the directed edge from the current processing node to the direct downstream node is read, and the relative offset magnitude is multiplied by the dynamic weight value to obtain the single-step contribution from the current processing node to its direct downstream node.

[0033] In the upstream path from the current processing node to the quality inspection node, the single-step contribution of all adjacent nodes on the path is multiplied together to obtain the path contribution of the processing node on the current path.

[0034] Optionally, through the industrial IoT gateway of the supply chain, based on the real-time process parameters and preset specification limits, the process capability indicators of each node in the production process are calculated, and the process capability indicators and the production data are synchronously uploaded to the cloud, including:

[0035] Deploy industrial IoT gateways at key nodes in the supply chain, with each gateway having a built-in edge computing unit;

[0036] The edge computing unit calculates the percentage of values ​​in the real-time process parameters that fall within the range of the upper and lower thresholds according to the upper and lower thresholds in the preset specification limits, and uses this percentage as the first indicator.

[0037] The absolute offset between the average value of the real-time process parameters and the center value of the preset specification limit is calculated synchronously and used as a second indicator;

[0038] The first indicator and the second indicator are combined according to preset weights to generate process capability indicators for each processing node.

[0039] The process capability indicators, real-time process parameters, and corresponding material batch identifiers are bound into data units and uploaded to the cloud via an industrial IoT gateway.

[0040] Optionally, the process capability indicators and the production data are converted into vector format in the cloud to generate a cross-organizational data sequence based on the material batches, including:

[0041] The production data and process capability indicators for each batch are extracted from the cloud according to the material batch identifier.

[0042] The production data is reorganized into three basic data units: raw material performance dataset, process parameter dataset, and measurement result feature set; and the process capability index is converted into independent numerical units.

[0043] According to the time sequence of material batches in the supply chain, the three types of basic data units and the independent numerical units of each link are spliced ​​together to form a link data block; all the link data blocks are connected in the link sequence to form a cross-organization data sequence.

[0044] Secondly, this application provides a supply chain information traceability and monitoring system, including:

[0045] The data acquisition module is used to collect production data corresponding to different batches of materials at each stage of the supply chain. The production data includes raw material performance data of suppliers, real-time process parameters of each processing station in the plant, and measurement results of the quality inspection station.

[0046] The calculation module is used to calculate the process capability indicators of each node in the production process based on the real-time process parameters and preset specification limits through the industrial Internet of Things gateway of the supply chain, and to synchronously upload the process capability indicators and the production data to the cloud.

[0047] The generation module is used to convert the process capability indicators and the production data into vector format in the cloud, and generate cross-organizational data sequences based on material batches.

[0048] The construction module is used to analyze the correlation between various links in the supply chain based on the cross-organizational data sequence. Taking each link in the supply chain as a node and the flow relationship between materials and parameters as edges, a quality impact topology diagram of the supply chain is constructed. The quality impact topology diagram dynamically describes the impact weight of process parameters in different links on the final quality.

[0049] The prediction module is used to calculate the predicted probability of different upstream links as the root cause of the defect when the quality inspection process identifies a batch quality defect. Based on the quality impact topology map and combined with the specific parameter deviation of the defective batch in the cross-organization data sequence, the module generates a visualization report based on the predicted probability. The visualization report shows the most likely root cause of the quality defect and its transmission path in the supply chain.

[0050] Thirdly, this application provides an electronic device, comprising:

[0051] Memory, used to store computer programs;

[0052] A processor is used to implement the steps of the supply chain data traceability and monitoring method as described in the first aspect above when executing the computer program.

[0053] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps of the supply chain data traceability and monitoring method described in the first aspect above.

[0054] The supply chain data traceability and monitoring method provided in this application can comprehensively acquire quality-related data throughout the entire supply chain by collecting production data corresponding to different material batches at each stage of the supply chain, providing a complete data foundation for subsequent traceability and monitoring. By calculating process capability indicators at each node through the industrial IoT gateway of the supply chain and simultaneously uploading them to the cloud, it can achieve real-time processing and centralized storage of production process parameters, facilitating subsequent unified analysis. By converting relevant data into vector format in the cloud to generate cross-organizational data sequences based on material batches, it can standardize data formats and achieve orderly connection of material batch data throughout the entire process. By constructing a quality impact topology map based on the cross-organizational data sequence, it can clearly present the degree of influence of process parameters at each stage on the final quality and clarify the correlation between each stage. When a quality defect is detected, by combining the topology map and parameter deviation to calculate the predicted probability of the defect source and generating a visual report, it can quickly locate the root cause of the defect and its transmission path, improving the efficiency and accuracy of quality problem investigation.

[0055] Furthermore, processing parameter sets and quality inspection feature sets of adjacent links are extracted from cross-organizational data sequences. The correlation between links is determined by the consistency of parameter change directions. One-way influence edges are established for related links, and the frequency of correlation occurrence is used as a dynamic weight to construct an initial quality influence topology map. When new batches of data are added, the frequency is recalculated to update the topology map weights, achieving dynamic updates. This system can accurately identify the correlation between various links in the supply chain, dynamically adjust the influence weights of process parameters at each link on the final quality, ensuring the quality influence topology map always aligns with actual production conditions. This improves the accuracy and practicality of the topology map, providing a more reliable basis for subsequent quality defect tracing, further enhancing the accuracy of defect source location, and helping enterprises more efficiently identify quality problems and optimize production processes. Attached Figure Description

[0056] To more clearly illustrate the technical solutions of the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0057] Figure 1 A flowchart illustrating a supply chain data traceability and monitoring method provided in this application embodiment;

[0058] Figure 2 A flowchart illustrating the specific implementation of a supply chain data traceability and monitoring method provided in this application embodiment;

[0059] Figure 3 This is a schematic diagram of the structure of a supply chain information traceability and monitoring system provided in an embodiment of this application. Detailed Implementation

[0060] In existing supply chain information traceability and monitoring, the common method is to manually record or simply collect and store production data at each stage. This only enables basic data storage and retrieval, and cannot analyze the correlation and impact between parameters at each stage. When product quality defects occur, the only option is to manually check the data one by one to find the root cause. This is not only inefficient, but also makes it difficult to accurately find the source of the defect and its transmission path. It cannot meet the needs of enterprises for refined supply chain management, nor can it quickly resolve quality problems and reduce quality risks.

[0061] To address the aforementioned issues, this application proposes a supply chain data traceability and monitoring method. The core of this method is to collect and analyze data from each stage of the supply chain to achieve precise traceability and path visualization of quality defects. Specifically, the method first collects production data corresponding to each batch of materials at each stage, processes it through an industrial IoT gateway, and uploads it to the cloud. This data is then transformed into a standardized data sequence and a quality impact topology map is constructed, clearly showing the influence of parameters at each stage on the final quality. When a quality defect is detected, the probability of the defect source is quickly calculated by combining the topology map and parameter deviations, and a visual report is generated, clearly identifying the root cause and transmission path of the defect. This method overcomes the limitations of manual inspection, quickly and accurately locating the source of defects through data-driven and visual methods. It solves the problems of low traceability efficiency and insufficient accuracy in existing technologies, effectively improving the efficiency of supply chain quality control and reducing quality risks.

[0062] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0063] Example 1

[0064] The core of this invention is to provide a supply chain data traceability and monitoring method, the flowchart of one specific implementation of which is shown below. Figure 1 As shown, the method includes:

[0065] S101. Collect production data corresponding to different material batches at each stage of the supply chain.

[0066] Among them, production data is the quality-related data corresponding to material batches in each link of the supply chain, including three parts: raw material performance data provided by suppliers, real-time production process parameters of each processing station within the enterprise, and quality measurement results of materials or products by quality inspection stations.

[0067] In one specific implementation, various production data are acquired separately through dedicated data acquisition equipment at corresponding stages, simultaneously collecting relevant data from different material batches at each stage to ensure a one-to-one correspondence between data and material batches. Raw material performance data is provided by suppliers after testing using professional testing equipment; for example, specific data such as the tensile strength of steel and the melting point of plastics can be collected. Real-time process parameters at each processing station in the factory are collected in real time through parameter recorders built into the processing equipment; for example, specific values ​​such as mold temperature during injection molding and pressure during stamping can be collected. Measurement results from the quality inspection station are collected after testing with quality inspection instruments; for example, specific data such as dimensional deviations of parts and the pass rate of finished products can be collected.

[0068] S102. Through the industrial IoT gateway of the supply chain, based on real-time process parameters and preset specification limits, calculate the process capability indicators of each node in the production process, and upload the process capability indicators and production data to the cloud synchronously.

[0069] The Industrial Internet of Things (IIoT) gateway is a core device connecting data acquisition equipment at each node of the supply chain to the cloud. It primarily receives real-time data from each node, processes the data locally, and uploads the processed data to the cloud, enabling data transfer and preliminary analysis. Process capability indicators are core parameters for measuring the stability of processes at each production node and their ability to meet quality requirements. They reflect whether real-time process parameters meet preset standards and the degree of compliance. Preset specification limits are reasonable ranges of process parameters set in advance based on product quality requirements, used to determine whether real-time process parameters are qualified.

[0070] Optionally, step S102 may specifically include the following steps:

[0071] S1021. Deploy industrial IoT gateways at key nodes in the supply chain. The industrial IoT gateways have built-in edge computing units.

[0072] Critical nodes refer to the links in the supply chain that have a significant impact on product quality, including supplier raw material supply nodes, in-house processing stations, and quality inspection nodes. Edge computing units are small data processing modules integrated within the industrial IoT gateway. They are used to quickly process collected real-time data locally, eliminating the need to directly upload all raw data to the cloud, thus reducing data transmission pressure and improving data processing efficiency.

[0073] S1022. Using the edge computing unit, calculate the percentage of values ​​in the real-time process parameters that fall within the range of the upper and lower thresholds according to the upper and lower thresholds in the preset specification limits, and use this as the first indicator.

[0074] The upper limit threshold is the maximum allowable value among the preset specification limits, and the lower limit threshold is the minimum allowable value among the preset specification limits. Together, they constitute the acceptable range of the process parameters. The first indicator is used to reflect the acceptance rate of real-time process parameters. The higher the acceptance rate, the better the stability and the higher the acceptance level of the process parameters at that node.

[0075] S1023. Simultaneously calculate the absolute offset between the average value of the real-time process parameters and the center value of the preset specification limit, and use it as the second indicator.

[0076] Among them, the center value is the middle value of the preset specification limits, that is, the average value of the upper and lower thresholds, representing the ideal standard value of the process parameters. The absolute offset is the absolute value of the difference between the average value of the real-time process parameters and the center value. The second indicator is used to reflect the degree to which the real-time process parameters deviate from the ideal standard. The smaller the offset, the closer the process parameters are to the ideal state.

[0077] S1024. Combine the first indicator and the second indicator according to the preset weights to generate process capability indicators for each processing node.

[0078] The preset weights are set in advance based on the supply chain quality control requirements. They are used to reflect the importance of the first and second indicators to the process capability assessment, and the sum of their weights is 1.

[0079] S1025. Bind process capability indicators, real-time process parameters and corresponding material batch identifiers into data units and upload them to the cloud through industrial IoT gateways.

[0080] The material batch identifier is a unique code used to distinguish different material batches, and can be composed of letters and numbers. A data unit is a unified data set formed by integrating related data, facilitating data classification, storage, and management in the cloud.

[0081] In this embodiment, industrial IoT gateways with built-in edge computing units are first deployed at key nodes in the supply chain. Then, the first indicator and the second indicator are calculated sequentially through the edge computing units. Combined with preset weights, process capability indicators are generated. Finally, the relevant data are bound to the material batch identifier and uploaded to the cloud, forming a complete data processing and uploading process.

[0082] First, through step S1021, industrial IoT gateways are deployed at key nodes in the supply chain. These gateways have built-in edge computing units. After deployment, the gateways can receive various types of data transmitted by the corresponding node acquisition devices in real time, and the edge computing units can perform local fast processing on the received data, laying the foundation for subsequent indicator calculations.

[0083] For example, industrial IoT gateways with built-in edge computing units can be deployed at key nodes such as supplier raw material inspection stations, in-house injection molding stations, and quality inspection stations. Taking the injection molding station as an example, the deployed industrial IoT gateway can receive process parameters such as mold temperature and pressure transmitted by the injection molding equipment in real time, and quickly receive and preliminarily process this real-time data through the built-in edge computing unit, without having to directly upload the raw data to the cloud.

[0084] Secondly, through step S1022, the edge computing unit deployed in S1021 is used to call the upper and lower threshold values ​​in the preset specification limits to statistically analyze the collected real-time process parameters, calculate the percentage of parameter values ​​within the range of the upper and lower threshold values, and use this percentage as the first indicator to reflect the qualified ratio of process parameters.

[0085] For example, taking the mold temperature parameters of an injection molding station as an example, the edge computing unit of the industrial IoT gateway at this node calls the upper limit threshold of 180℃ and the lower limit threshold of 160℃ of the preset specification limits for the injection molding process, and statistically analyzes the real-time mold temperature data collected over a period of time during the injection molding process of a certain batch of materials. The percentage of values ​​within the range of 160℃-180℃ is calculated as the first indicator. A total of 100 sets of real-time mold temperature data were collected for this batch, of which 88 sets of data were within the acceptable range, therefore the first indicator is 0.88.

[0086] Next, in step S1023, while calculating the first index, the average value of the collected real-time process parameters is calculated simultaneously, and then the absolute offset of the average value from the center value of the preset specification limit is calculated. This absolute offset is used as the second index to reflect the degree to which the overall process parameters deviate from the ideal standard.

[0087] In practical applications, the above 100 sets of real-time mold temperature data are calculated simultaneously. First, the average value of these 100 sets of data is 172℃. Then, based on the upper limit threshold of 180℃ and the lower limit threshold of 160℃ of the preset specification limit, the center value is calculated as (160+180) / 2=170℃. Finally, the absolute deviation between the average value and the center value is calculated, that is, 172℃-170℃=2℃. This 2℃ is the second indicator.

[0088] Next, in step S1024, the preset weights are called, and the first index calculated in S1022 and the second index calculated in S1023 are combined and calculated according to the weights. Through a reasonable calculation method, the process capability index of the processing node is finally generated, so as to realize the quantitative evaluation of the node's process level.

[0089] For example, if the weight of the first indicator is preset to 0.6 and the weight of the second indicator is 0.4, and the calculation method of "process capability index = first indicator × 0.6 + (1 - second indicator / (upper limit threshold - center value)) × 0.4" is adopted, the first indicator 0.88 and the second indicator 2℃ are substituted into the calculation, that is, 0.88 × 0.6 + (1 - 2 / 20) × 0.4 = 0.528 + 0.36 = 0.888. This 0.888 is the process capability index of the batch of materials in the injection molding station.

[0090] Finally, through step S1025, the process capability indicators generated in S1024 and the real-time process parameters collected in each step are bound with the unique identifier corresponding to the batch of materials, and integrated into a unified data unit. Then, through the industrial IoT gateway deployed in S1021, the data unit is uploaded to the cloud to complete the centralized storage of data.

[0091] For example, the calculated process capability index of 0.888, the aforementioned 100 sets of real-time mold temperature data, and the unique identifier A12345 of the batch of materials are bound together to form a complete data unit. This data unit is then uploaded to the cloud through an industrial IoT gateway deployed at the injection molding station, ensuring that the data accurately corresponds to the material batch. The above example is merely one example of this application. In practical applications, the preset weights, calculation methods, and data collection quantities can be adjusted according to requirements, and this application does not limit these aspects.

[0092] This application achieves quantitative assessment of the process capabilities of each production node through efficient processing and analysis of supply chain production data, completes centralized management and precise traceability of production-related data, and effectively reduces data transmission pressure and improves data processing efficiency.

[0093] S103. Convert process capability indicators and production data into vector format in the cloud to generate cross-organizational data sequences based on material batches.

[0094] Vector format transforms various unstructured or fragmented numerical data into a unified, ordered combination of values, facilitating data computation, analysis, and correlation in the cloud. Cross-organizational data sequences are ordered data sets that integrate relevant data from different organizations and stages of the supply chain, with material batches as the core thread. They can fully present the data trajectory of a material batch throughout the entire supply chain process.

[0095] Optionally, step S103 may specifically include the following steps:

[0096] S1031. Extract production data and process capability indicators for all stages of each batch from the cloud according to the material batch identifier.

[0097] S1032. Reorganize the production data into three basic data units: raw material performance dataset, process parameter dataset, and measurement result feature set, and transform the process capability index into independent numerical units.

[0098] The raw material performance dataset is a collection that integrates all raw material performance-related data from the same batch of materials, used to centrally present raw material quality-related information. The process parameter dataset is a collection that integrates real-time process parameters from each processing stage of the same batch of materials, used to comprehensively record parameter changes during processing. The measurement result feature set is a collection that integrates measurement results from each quality inspection stage of the same batch of materials, used to centrally reflect the quality inspection status at each stage. Basic data units are the smallest data sets formed by integrating similar data, facilitating data classification, processing, and assembly. Independent numerical units treat a single quantitative parameter, such as a process capability index, as a separate data unit, facilitating combination with various basic data units.

[0099] S1033. According to the time sequence of material batches in the supply chain, the three types of basic data units and independent numerical units of each link are spliced ​​together to form a link data block; all link data blocks are connected in the link sequence to form a cross-organization data sequence.

[0100] Among them, the process data block is a data block formed by splicing and integrating the three types of basic data units of a material batch in a single link of the supply chain with the independent numerical units of the process capability indicators of that link, which can completely present all relevant data of that link. The time sequence refers to the sequential flow of the material batch in the supply chain from raw material supply, processing and production to quality inspection completion.

[0101] In one specific implementation, corresponding data is extracted from the cloud according to the material batch identifier, the data is classified and reorganized into basic data units and independent numerical units, and then spliced ​​and integrated according to time and process sequence to finally form a cross-organizational data sequence.

[0102] First, through step S1031, production data and process capability indicators of all stages of each batch are extracted from the cloud according to the material batch identifier. Relying on the uniqueness of the material batch identifier, all production data and corresponding process capability indicators of a certain material batch uploaded to the cloud at each stage of the supply chain are accurately matched and extracted, ensuring that the extracted data completely corresponds to the material batch without omission or confusion.

[0103] For example, for the injection molding material batch with the identifier A12345, all production data and process capability indicators corresponding to the batch are extracted from the cloud, including raw material performance data provided by the supplier, process parameter data of the injection molding station, measurement results of the quality inspection station, and process capability indicators corresponding to each link, ensuring that the extracted data fully covers the entire process of the batch from raw materials to quality inspection.

[0104] Secondly, in step S1032, the extracted production data of this batch is classified and reorganized. All data related to raw material performance are integrated into a raw material performance dataset, and all process parameter data of the injection molding process are integrated into a process parameter dataset, including mold temperature, pressure, speed, etc. All measurement results of the quality inspection process are integrated into a measurement result feature set, including finished product size deviation, pass rate, etc. At the same time, the process capability index calculated by each process is converted into an independent numerical unit, such as 0.888 in the injection molding process.

[0105] In practical applications, for batch A12345, data such as the melting point of 150℃ and purity of 99.5% of the plastic raw materials in this batch were integrated into a raw material performance dataset; 100 sets of mold temperature data and pressure data collected during injection molding were integrated into a process parameter dataset; data such as the finished product size deviation of ±0.1mm and the pass rate of 98% obtained in the quality inspection process were integrated into a measurement result feature set; and the process capability index of 0.888 in the injection molding process and the process capability index of 0.92 in the raw material inspection process were treated as independent numerical units.

[0106] Finally, in step S1033, according to the time sequence of the material batch in the supply chain, the raw material performance dataset, process parameter dataset, and measurement result feature set of the raw material supply link are first concatenated with the independent numerical unit 0.92 of that link to form a raw material supply link data block. The raw material supply link has no processing link, and the process parameter dataset is an empty set. Then, the three types of basic data units of the injection molding process are concatenated with the independent numerical unit 0.888 to form an injection molding process data block. Finally, the three types of basic data units of the quality inspection link are concatenated with the independent numerical unit 0.95 of that link to form a quality inspection link data block. The three link data blocks are connected sequentially according to the link sequence of raw material supply, injection molding, and quality inspection to form the cross-organizational data sequence of this batch.

[0107] For example, following the flow time sequence of batch A12345, the three types of basic data units and independent numerical unit 0.92 from the raw material supply stage are first spliced ​​together to form the raw material stage data block; then, the three types of basic data units and independent numerical unit 0.888 from the injection molding stage are spliced ​​together to form the processing stage data block; finally, the three types of basic data units and independent numerical unit 0.95 from the quality inspection stage are spliced ​​together to form the quality inspection stage data block. Connecting the three stage data blocks sequentially forms a complete cross-organizational data sequence for batch A12345, fully presenting the entire process data trajectory of this batch from raw materials to finished products. The above example is merely one example of this application. In practical applications, the data reorganization and splicing methods can be adjusted according to the number of supply chain stages and data types; this application does not limit this.

[0108] This application achieves standardized integration and orderly connection of data throughout the entire supply chain, clearly presenting the data trajectory of material batches at each stage of the supply chain. It provides a standardized, complete, and traceable data foundation for analyzing the relationships between each stage and constructing a quality impact topology diagram, thereby improving data availability and analysis efficiency.

[0109] S104. Based on cross-organizational data sequences, analyze the relationships between various links in the supply chain. Using each link in the supply chain as a node and the flow relationship between materials and parameters as edges, construct a quality impact topology diagram of the supply chain.

[0110] The Quality Impact Topology (QITOP) is a graphical model that uses nodes and edges as its core, dynamically reflecting the degree of influence of process parameters at each stage on the final product quality through edge weights. Relationships refer to the corresponding connections between adjacent or related stages in the supply chain, where changes in process parameters in one stage correspond to changes in quality characteristics in the next. Nodes are the basic units representing a single stage in the supply chain within the QITOP. Edges are lines connecting two related nodes, representing the flow of materials and parameters. Influence weights are numerical values ​​that quantify the degree of impact of process parameters at each stage on the final quality.

[0111] Optionally, such as Figure 2 As shown, step S104 may specifically include the following steps:

[0112] S1041. Extract the processing node parameter set and quality inspection node feature set of adjacent links from the cross-organization data sequence. When it is detected that the direction of change of the specific process parameter value of the previous link is consistent with the direction of change of the quality feature value of the subsequent link, it is determined that there is a correlation.

[0113] Specifically, step S1041 may include the following process: Extracting a first sequence of specific process parameter values ​​from the previous stage, arranged chronologically, from the cross-organizational data sequence; extracting a second sequence of quality characteristic values ​​from the next stage, arranged chronologically, where the first and second sequences cover the same multiple consecutive production batches; for each production batch in the first and second sequences, except for the earliest batch, comparing the specific process parameter values ​​of the current batch with those of the adjacent previous batch to generate a parameter change direction indicator; comparing the quality characteristic values ​​of the current batch with those of the adjacent previous batch to generate a characteristic change direction indicator; for each production batch, comparing the parameter change direction indicator with the characteristic change direction indicator; recording a direction consistency event when both indicators indicate an increase or decrease; counting the number of direction consistency events within a moving window containing a predetermined number of production batches, and calculating the ratio of the number of events to the total number of batches within the moving window; determining a correlation between the previous and next stages when the ratio exceeds a preset consistency threshold.

[0114] In the above steps, adjacent links refer to two links closely connected in the material flow sequence of the supply chain. The processing node parameter set is a set of specific process parameters related to processing and production at a certain link, extracted from cross-organizational data sequences. The quality inspection node feature set is a set of quality characteristic values ​​related to quality inspection at a certain link, extracted from cross-organizational data sequences. The parameter change direction indicator is an identifier indicating the direction of change by comparing the values ​​of the same process parameter in two adjacent production batches. The feature change direction indicator is an identifier indicating the direction of change by comparing the values ​​of the same quality characteristic in two adjacent production batches. A direction consistency event refers to a situation where the change direction of the process parameter in the preceding link is the same as that of the quality characteristic in the following link. The moving window is a preset statistical range containing a fixed number of consecutive production batches. The preset consistency threshold is a critical ratio value used to determine whether there is a correlation between links.

[0115] S1042. Establish unidirectional influence edges for adjacent links that have a relationship, and count the frequency of the relationship in all production batches as the dynamic weight value of the influence edge.

[0116] In this context, a unidirectional influence edge is a line connecting adjacent nodes with a relationship, pointing from the preceding influencing node to the following affected node. Occurrence frequency is the ratio of the number of times, across all production batches, that the process parameters of the preceding stage and the quality characteristics of the following stage change in the same direction. The dynamic weight value is a numerical value calculated based on the occurrence frequency, quantifying the degree of influence of the unidirectional influence edge, and is updated with each production batch.

[0117] S1043. Using supply chain links as nodes, unidirectional influence edges as connections, and dynamic weight values ​​as edge weights, construct an initial quality influence topology graph.

[0118] The initial quality impact topology graph is a newly constructed graph based on existing production batch data, containing nodes, unidirectional impact edges, and dynamic weight values. The edge weights are the dynamic weight values ​​labeled on the unidirectional impact edges.

[0119] S1044. Whenever a new batch of cross-organizational data sequences is added, the frequency of occurrence of synchronous changes is recalculated to update the dynamic weight values ​​of the initial quality impact topology map, thus obtaining a dynamically updated quality impact topology map.

[0120] The frequency of synchronous changes is the ratio of the number of times the related links in all batches show the same direction of change after a new production batch is added, to the total number of batches. The dynamically updated quality impact topology graph is a topology graph that is updated with the latest production data after adjusting the edge weights.

[0121] In one specific implementation, relevant data of adjacent links are extracted from cross-organizational data sequences, the correlation is determined through multi-batch comparative analysis, one-way influence edges are established for the related links and dynamic weight values ​​are calculated to construct an initial topology graph, and the topology graph is continuously updated by combining new batch data to complete the construction and dynamic optimization of the quality influence topology graph.

[0122] First, through step S1041, the processing node parameter set and quality inspection node feature set of adjacent links are extracted from the cross-organization data sequence. According to the subdivided process, the first sequence and the second sequence are extracted. The change direction of parameters and features of each batch is compared and analyzed. The ratio of consistent events in the direction is statistically analyzed by moving the window. When the ratio exceeds the preset consistency threshold, it is determined that there is a correlation between the two.

[0123] For example, taking injection molding as the preceding stage and quality inspection as the following stage, mold temperature parameters from the injection molding stage are extracted as specific process parameters from cross-organizational data sequences of each batch, forming a processing node parameter set. Finished product dimensional deviations from the quality inspection stage are extracted as quality features, forming a quality inspection node feature set. Mold temperature values ​​from 10 consecutive production batches are extracted and arranged chronologically to form the first sequence. Simultaneously, the corresponding finished product dimensional deviation values ​​from these 10 batches are extracted and arranged chronologically to form the second sequence. The preset moving window contains 5 consecutive batches, and the preset consistency threshold is 0.7.

[0124] For each batch, except the first batch, the mold temperature values ​​are compared with the previous batch. An increase in mold temperature generates an "increase" parameter change direction indicator, while a decrease generates a "decrease" indicator. Simultaneously, the finished product dimensional deviation values ​​are compared with the previous batch. An increase in deviation generates an "increase" feature change direction indicator, while a decrease generates a "decrease" indicator. In each batch, if both indicators are either increasing or both are decreasing, a directional consistency event is recorded. In the first moving window (batches 1-5), a total of 4 directional consistency events were generated, with a ratio of 4 / 5 = 0.8, exceeding the preset consistency threshold of 0.7, indicating a correlation between the injection molding process and the quality inspection process.

[0125] Secondly, through step S1042, a unidirectional influence edge is established from the injection molding process to the quality inspection process for those processes that are determined to be related. Among all 10 production batches, the number of batches where the mold temperature change direction is consistent with the finished product dimensional deviation change direction is 8 batches. The frequency of occurrence is calculated as 8 / 10 = 0.8, and this 0.8 is used as the dynamic weight value of this unidirectional influence edge.

[0126] In practical applications, in addition to the injection molding and quality inspection processes, the relationship between the raw material supply and injection molding processes is analyzed simultaneously. After determining that there is a relationship between the two, a one-way influence edge is established from the raw material supply to the injection molding process. The frequency of the relationship in all 10 batches is 0.75, and 0.75 is used as the dynamic weight value of this one-way influence edge.

[0127] Next, through step S1043, the raw material supply link, injection molding processing link and quality inspection link in the supply chain are taken as three independent nodes, and the two established unidirectional influence edges are used as the connection relationship. The edge weights of the corresponding edges are 0.75 and 0.8 respectively, to construct an initial quality influence topology graph, which intuitively presents the quality influence intensity between the three links.

[0128] For example, in the initial quality influence topology diagram, the raw material supply node is connected to the injection molding node through a one-way influence edge with an edge weight of 0.75, and the injection molding node is connected to the quality inspection node through a one-way influence edge with an edge weight of 0.8. It can be clearly seen that the quality influence of the injection molding stage on the quality inspection stage is slightly greater than the influence of the raw material supply stage on the injection molding stage.

[0129] Finally, in step S1044, after adding a new cross-organizational data sequence for a production batch, the frequency of synchronous changes in the corresponding relationships between the two unidirectional influence edges in all 11 production batches is recalculated. Specifically, the frequency of occurrence between the raw material supply stage and the injection molding stage is updated to 0.76, and the frequency of occurrence between the injection molding stage and the quality inspection stage is updated to 0.82. The updated frequencies are used to replace the weight values ​​of the corresponding edges in the initial topology graph, resulting in a dynamically updated quality influence topology graph.

[0130] For example, after adding one production batch, the number of batches with consistent relationships between the raw material supply and injection molding processes becomes 9, with an occurrence frequency of 9 / 11≈0.82. Similarly, the number of batches with consistent relationships between the injection molding and quality inspection processes also becomes 9, with an occurrence frequency of 9 / 11≈0.82. After updating the edge weights, the topology graph can accurately reflect the quality impact relationships of each process under the latest production data. The above example is merely one illustration of this application. In practical applications, parameters such as the moving window size and preset consistency thresholds can be adjusted according to requirements, and this application does not impose any limitations on this.

[0131] This application enables the accurate identification and quantification of the relationships between various links in the supply chain, constructs a dynamically updatable quality impact topology map, and intuitively presents the impact weight of process parameters at each link on the final quality, ensuring that the quality impact analysis is aligned with the latest production data and improving the accuracy and timeliness of supply chain quality control.

[0132] S105. When a batch quality defect is identified in the quality inspection process, based on the quality impact topology diagram and combined with the specific parameter deviations of the defective batch in the cross-organizational data sequence, the predicted probability of different upstream links as the root cause of the defect is calculated, and a visualization report is generated based on the predicted probability.

[0133] The visualization report showcases the most likely root causes of quality defects and their transmission paths within the supply chain. A batch quality defect refers to a batch of materials failing to meet preset quality standards after quality inspection. Parameter deviation refers to the difference between the process parameters or quality characteristics of a defective batch and the corresponding values ​​of a normal batch. Predicted probability quantifies the likelihood that different upstream links are the root cause of defects. The visualization report presents information such as the predicted probability of defect sources and transmission paths in an intuitive graphical format. The transmission path refers to the flow of quality defects from the source link, through various intermediate links, to the quality inspection link.

[0134] Optionally, step S105 may specifically include the following steps:

[0135] S1051. When a batch quality defect is identified during the quality inspection process, the out-of-limit quality characteristics are extracted from the cross-organizational data sequence of the corresponding batch.

[0136] Among them, the quality characteristics exceeding the limit refer to the quality inspection-related characteristics in the defective batch where the value exceeds the preset quality standard limit.

[0137] S1052. Taking the quality inspection node as the endpoint, in the quality influence topology graph, traverse all upstream paths pointing to the quality inspection node in reverse and extract the actual values ​​of process parameters of the processing nodes in the defective batch in each upstream path.

[0138] Reverse traversal refers to starting from the quality inspection node and searching backwards for all upstream links and paths that point to that node. An upstream path refers to the connection path between various links in the supply chain that precede the quality inspection stage and can influence the quality inspection results. Actual process parameter values ​​refer to the actual process parameter values ​​collected at each processing node for defective batches.

[0139] S1053. Calculate the deviation between the actual value of the process parameter and the historical normal batch average value of the processing node. Based on the deviation and the dynamic weight value from the node to the downstream node, calculate the path contribution of the processing node.

[0140] Specifically, step S1053 may include the following processes: From the cross-organizational data sequence, filter out the process parameter values ​​of the processing nodes in a preset number of historical normal batches, calculate the average value of the process parameter values ​​in the historical normal batches, and use this as the historical normal batch average; calculate the relative offset amplitude based on the actual values ​​of the process parameters of the processing nodes in the defective batches and the historical normal batch average; in the quality influence topology graph, read the dynamic weight value of the directed edge from the current processing node to the direct downstream node, multiply the relative offset amplitude by the dynamic weight value to obtain the single-step contribution of the current processing node to its direct downstream node; in the upstream path from the current processing node to the quality inspection node, multiply the single-step contributions between all adjacent nodes on the path to obtain the path contribution of the processing node on the current path.

[0141] In the above steps, the historical average of normal batches refers to the average value of process parameters among a predetermined number of batches without quality defects at a certain processing node. The deviation magnitude refers to the degree of difference between the actual value of the process parameters in the defective batch and the historical average of normal batches, expressed here as a relative deviation magnitude. The relative deviation magnitude is a numerical value that quantifies the degree to which the actual value of the process parameters deviates from the historical average of normal batches. The path contribution is a numerical value that quantifies the degree of influence a certain processing node has on quality defects along a certain upstream path. The direct downstream node refers to the next node in the supply chain path that is closely connected to and directly affected by the current processing node. The single-step contribution is a numerical value that quantifies the degree of influence the current processing node has on its direct downstream nodes.

[0142] S1054. Accumulate the path contribution of the same processing node in all paths, and calculate the predicted probability of the corresponding processing stage as the source of quality defects based on the proportion of the total contribution.

[0143] The total contribution refers to the sum of the path contributions of the same processing node across all upstream paths. The total contribution ratio is the ratio of the total contribution of a particular processing node to the sum of the total contributions of all upstream processing nodes.

[0144] S1055. Generate a tree-like visualization report based on the processing node with the highest prediction probability and its transmission path in the supply chain.

[0145] Among them, the tree-structured visualization report refers to a report that clearly presents the defect source node, each intermediate transmission link and quality inspection node in the form of a tree diagram, which can intuitively show the defect transmission logic.

[0146] In one specific implementation, after the quality inspection process identifies defects, it extracts the out-of-limit quality features, reversely traverses the topology map to obtain the actual values ​​of the upstream node process parameters, calculates the offset magnitude and path contribution, accumulates the total contribution and calculates the prediction probability, and finally generates a tree-shaped visualization report to complete the defect source location and path presentation.

[0147] First, through step S1051, when the quality inspection process identifies a batch quality defect, all out-of-limit quality characteristics are accurately extracted from the cross-organizational data sequence of the corresponding batch, and the quality inspection indicators related to the defect are identified.

[0148] For example, during the quality inspection process, a batch of injection molding materials with the identifier A12345 was found to have a quality defect. The dimensional deviation of the finished product in this batch exceeded the preset limit. From the cross-organizational data sequence of this batch, the quality feature that exceeded the limit was extracted as a dimensional deviation of 0.3 mm, while the preset specification limit was ±0.1 mm.

[0149] Secondly, through step S1052, taking the quality inspection node as the endpoint, in the constructed quality influence topology graph, all upstream paths pointing to the quality inspection node are traversed in reverse to find all upstream processing nodes that may affect the quality inspection results, and the actual values ​​of process parameters of these processing nodes in the defective batch are extracted.

[0150] For example, taking the quality inspection node as the endpoint, the quality influence topology is traversed in reverse to obtain the unique upstream path of injection molding process → quality inspection process. The actual values ​​of the process parameters of the defective batch in the injection molding process are extracted, namely the actual value of mold temperature 185℃ and the actual value of pressure 12MPa.

[0151] Next, in step S1053, the historical normal batch process parameters of the processing node are first screened, the historical normal batch average is calculated, and then the relative deviation of the actual value of the defective batch process parameters from the average is calculated. Combined with the dynamic weight value in the topology graph, the single-step contribution and path contribution are calculated in sequence.

[0152] In practical applications, taking the injection molding process as an example, the mold temperature data of 10 historical normal batches for this node are selected from the cross-organization data sequence. The average temperature of the historical normal batches is calculated to be 170℃; the actual mold temperature of the defective batches is 185℃, and the relative offset is calculated to be (185-170) / 170≈0.088. From the quality influence topology diagram, the dynamic weight value of the unidirectional influence edge from the injection molding process to the quality inspection process is read as 0.8. Multiplying the relative offset by the dynamic weight value, the single-step contribution is 0.088×0.8≈0.0704. This upstream path only contains one single step from the injection molding process to the quality inspection process; therefore, the path contribution of the injection molding process on this path is 0.0704.

[0153] Next, by accumulating the path contribution of the same processing node in all upstream paths through step S1054, the proportion of the total contribution of the processing node to the sum of the total contributions of all upstream nodes is calculated, thereby obtaining the predicted probability of the processing node as a source of defects.

[0154] For example, if there is only one upstream path for this defective batch, the total contribution of the injection molding process is 0.0704, and the sum of the total contributions of all upstream nodes is also 0.0704, then the total contribution percentage is calculated as 0.0704 / 0.0704=1, meaning that the predicted probability of the injection molding process being the source of the quality defect in this batch is 100%.

[0155] Finally, in step S1055, the processing node with the highest predicted probability is selected as the root cause of the defect. The supply chain transmission path from this node to the quality inspection stage is sorted out, and a visual report is generated in a tree structure to clearly present the relevant information.

[0156] For example, the processing node with the highest predicted probability is the injection molding process. The transmission path from injection molding to quality inspection is injection molding → quality inspection. Based on this node and transmission path, a tree-structured visual report is generated. The top of the report shows the defect source, the injection molding process; the middle shows the transmission path; and the bottom shows the quality inspection process and out-of-limit quality characteristics. The predicted probability of the injection molding process is also marked as 100%. The above example is merely one illustration of this application. In practical applications, parameters such as the number of historical normal batches and the calculation method can be adjusted according to requirements. This application does not limit these adjustments.

[0157] This application enables precise location and probability quantification of the root cause of quality defects, and can intuitively present the transmission path of defects in the supply chain, improve the efficiency and accuracy of quality defect investigation, help to quickly identify problem links, take targeted corrective measures, and reduce the losses caused by quality defects.

[0158] Example 2

[0159] Figure 3 This is a schematic diagram of a specific implementation of a supply chain information traceability and monitoring system provided in Embodiment 2 of this application, with reference to... Figure 3 The system may include:

[0160] The data acquisition module 31 is used to collect production data corresponding to different material batches in each link of the supply chain. The production data includes raw material performance data of suppliers, real-time process parameters of each processing station in the plant, and measurement results of the quality inspection station.

[0161] The calculation module 32 is used to calculate the process capability indicators of each node in the production process based on the real-time process parameters and preset specification limits through the industrial Internet of Things gateway of the supply chain, and to upload the process capability indicators and the production data to the cloud synchronously.

[0162] The generation module 33 is used to convert the process capability indicators and the production data into vector format in the cloud and generate a cross-organization data sequence with material batches as the main line.

[0163] Module 34 is used to analyze the correlation between various links in the supply chain based on the cross-organizational data sequence. It constructs a quality impact topology diagram of the supply chain with each link in the supply chain as a node and the flow relationship between materials and parameters as an edge. The quality impact topology diagram dynamically describes the impact weight of process parameters in different links on the final quality.

[0164] The prediction module 35 is used to calculate the predicted probability of different upstream links as the root cause of the defect when the quality inspection process identifies a batch quality defect. Based on the quality impact topology map and combined with the specific parameter deviation of the defective batch in the cross-organization data sequence, the module generates a visualization report based on the predicted probability. The visualization report shows the most likely root cause of the quality defect and its transmission path in the supply chain.

[0165] The supply chain information traceability and monitoring system of this application embodiment is used to implement the aforementioned supply chain data traceability and monitoring method. Therefore, the specific implementation of the supply chain information traceability and monitoring system can be found in the embodiment section of the supply chain data traceability and monitoring method above. The specific implementation can be referred to the description of the corresponding embodiments, which will not be repeated here.

[0166] This application also provides an electronic device, including: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of any of the above-described supply chain data traceability and monitoring methods.

[0167] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of any of the above-described supply chain data traceability and monitoring methods.

[0168] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, random access memory, portable hard drives, magnetic disks, or optical disks.

[0169] Embodiments of the present invention also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above embodiments of the supply chain data traceability and monitoring method.

[0170] Those skilled in the art will further 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, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. 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.

[0171] The above provides a detailed description of a supply chain data traceability and monitoring method, system, and storage medium provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A supply chain data traceability and monitoring method, characterized in that, include: Collect production data corresponding to different material batches at each stage of the supply chain. The production data includes raw material performance data from suppliers, real-time process parameters of each processing station in the plant, and measurement results from the quality inspection station. Through the industrial Internet of Things (IIoT) gateway of the supply chain, based on the real-time process parameters and preset specification limits, the process capability indicators of each node in the production process are calculated, and the process capability indicators and the production data are synchronously uploaded to the cloud. In the cloud, the process capability indicators and the production data are converted into vector format to generate cross-organizational data sequences with material batches as the main line. Based on the cross-organizational data sequence, the correlation between various links in the supply chain is analyzed. Taking each link in the supply chain as a node and the flow relationship between materials and parameters as an edge, a quality impact topology diagram of the supply chain is constructed. The quality impact topology diagram dynamically describes the impact weight of process parameters in different links on the final quality. When a batch quality defect is identified during the quality inspection process, based on the quality impact topology map and combined with the specific parameter deviations of the defective batch in the cross-organizational data sequence, the predicted probability of different upstream links as the root cause of the defect is calculated, and a visualization report is generated based on the predicted probability. The visualization report shows the most likely root cause of the quality defect and its transmission path in the supply chain.

2. The method according to claim 1, characterized in that, Based on the cross-organizational data sequence, the relationships between various links in the supply chain are analyzed. Using each link in the supply chain as a node and the flow relationships of materials and parameters as edges, a quality impact topology diagram of the supply chain is constructed. This quality impact topology diagram dynamically describes the influence weights of process parameters at different links on the final quality, including: From the cross-organizational data sequence, the processing node parameter set and quality inspection node feature set of adjacent links are extracted. When it is detected that the direction of change of the value of a specific process parameter in the previous link is consistent with the direction of change of the value of the quality feature in the subsequent link, it is determined that there is a correlation. A one-way influence edge is established for adjacent links with the aforementioned relationship, and the frequency of occurrence of the relationship in all production batches is statistically analyzed as the dynamic weight value of the influence edge. Using supply chain links as nodes, the unidirectional influence edges as connections, and the dynamic weight values ​​as edge weights, an initial quality influence topology graph is constructed. Whenever a new batch of cross-organizational data sequences is added, the frequency of occurrence of synchronization changes is recalculated to update the dynamic weight values ​​of the initial quality impact topology, resulting in a dynamically updated quality impact topology.

3. The method according to claim 2, characterized in that, When the direction of change of a specific process parameter in the preceding stage is consistently consistent with the direction of change of a quality characteristic value in the following stage, a correlation is determined to exist, including: From the cross-organizational data sequence, a first sequence of specific process parameter values ​​from the previous stage is extracted and arranged in chronological order, and a second sequence of quality characteristic values ​​from the next stage is extracted and arranged in chronological order. The first and second sequences cover the same multiple consecutive production batches. For each production batch in the first sequence and the second sequence, except for the earliest batch, the values ​​of specific process parameters of the current batch are compared with those of the adjacent previous batch to generate a parameter change direction indicator, and the values ​​of quality characteristics of the current batch are compared with those of the adjacent previous batch to generate a characteristic change direction indicator. For each production batch, compare the parameter change direction flag with the feature change direction flag. When both flags indicate an increase in value or a decrease in value, record a direction consistency event. Within a moving window containing a predetermined number of production batches, the number of occurrences of the direction alignment event is counted, and the ratio of the number of occurrences to the total number of batches within the moving window is calculated. When the ratio exceeds a preset consistency threshold, it is determined that there is a correlation between the previous step and the next step.

4. The method according to claim 1, characterized in that, When a batch quality defect is identified during quality inspection, based on the aforementioned quality impact topology map and combined with the specific parameter deviations of the defective batch in the cross-organizational data sequence, the predicted probability of different upstream links as the root cause of the defect is calculated. A visualization report is then generated based on this predicted probability. This visualization report displays the most likely root cause of the quality defect and its transmission path within the supply chain, including: When a batch quality defect is identified during the quality inspection process, the out-of-limit quality characteristics are extracted from the cross-organizational data sequence of the corresponding batch. Taking the quality inspection node as the endpoint, in the quality influence topology graph, all upstream paths pointing to the quality inspection node are traversed in reverse to extract the actual values ​​of process parameters of the processing nodes in the defective batch in each upstream path. Calculate the deviation between the actual value of the process parameter and the historical normal batch average value of the processing node. Based on the deviation and the dynamic weight value from the node to the downstream node, calculate the path contribution of the processing node. The path contribution of the same processing node in all paths is accumulated, and the predicted probability of the corresponding processing stage as the source of quality defects is calculated according to the proportion of the total contribution. Based on the processing node with the highest predicted probability and its transmission path in the supply chain, a tree-structured visualization report is generated.

5. The method according to claim 2, characterized in that, Calculate the deviation between the actual value of the process parameter and the historical normal batch average of the processing node. Based on the deviation and the dynamic weight value of the edge from the node to the downstream node, calculate the path contribution of the processing node, including: From the cross-organizational data sequence, the process parameter values ​​of the processing node in a preset number of historical normal batches are selected, and the average value of the process parameter values ​​in the historical normal batches is calculated as the historical normal batch average. The relative offset is calculated based on the actual values ​​of the process parameters of the processing nodes in the defective batch and the average values ​​of the historical normal batches. In the quality impact topology graph, the dynamic weight value of the directed edge from the current processing node to the direct downstream node is read, and the relative offset magnitude is multiplied by the dynamic weight value to obtain the single-step contribution from the current processing node to its direct downstream node. In the upstream path from the current processing node to the quality inspection node, the single-step contribution of all adjacent nodes on the path is multiplied together to obtain the path contribution of the processing node on the current path.

6. The method according to claim 1, characterized in that, Through the industrial IoT gateway of the supply chain, based on the real-time process parameters and preset specification limits, the process capability indicators of each node in the production process are calculated, and the process capability indicators and the production data are synchronously uploaded to the cloud, including: Deploy industrial IoT gateways at key nodes in the supply chain, with each gateway having a built-in edge computing unit; The edge computing unit calculates the percentage of values ​​in the real-time process parameters that fall within the range of the upper and lower thresholds according to the upper and lower thresholds in the preset specification limits, and uses this percentage as the first indicator. The absolute offset between the average value of the real-time process parameters and the center value of the preset specification limit is calculated synchronously and used as a second indicator. The first indicator and the second indicator are combined according to preset weights to generate process capability indicators for each processing node. The process capability indicators, real-time process parameters, and corresponding material batch identifiers are bound into data units and uploaded to the cloud via an industrial IoT gateway.

7. The method according to claim 1, characterized in that, The process capability indicators and production data are converted into vector format in the cloud to generate a cross-organizational data sequence based on the material batches, including: The production data and process capability indicators for each batch are extracted from the cloud according to the material batch identifier. The production data is reorganized into three basic data units: raw material performance dataset, process parameter dataset, and measurement result feature set; and the process capability index is converted into independent numerical units. According to the time sequence of material batches in the supply chain, the three types of basic data units and the independent numerical units of each link are spliced ​​together to form a link data block; all the link data blocks are connected in the link sequence to form a cross-organization data sequence.

8. A supply chain information traceability and monitoring system, characterized in that, include: The data acquisition module is used to collect production data corresponding to different batches of materials at each stage of the supply chain. The production data includes raw material performance data of suppliers, real-time process parameters of each processing station in the plant, and measurement results of the quality inspection station. The calculation module is used to calculate the process capability indicators of each node in the production process based on the real-time process parameters and preset specification limits through the industrial Internet of Things gateway of the supply chain, and to synchronously upload the process capability indicators and the production data to the cloud. The generation module is used to convert the process capability indicators and the production data into vector format in the cloud, and generate cross-organizational data sequences based on material batches. The construction module is used to analyze the correlation between various links in the supply chain based on the cross-organizational data sequence. Taking each link in the supply chain as a node and the flow relationship between materials and parameters as edges, a quality impact topology diagram of the supply chain is constructed. The quality impact topology diagram dynamically describes the impact weight of process parameters in different links on the final quality. The prediction module is used to calculate the predicted probability of different upstream links as the root cause of the defect when the quality inspection process identifies a batch quality defect. Based on the quality impact topology map and combined with the specific parameter deviation of the defective batch in the cross-organization data sequence, the module generates a visualization report based on the predicted probability. The visualization report shows the most likely root cause of the quality defect and its transmission path in the supply chain.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the supply chain data traceability and monitoring method as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, enables the implementation of the supply chain data traceability and monitoring method as described in any one of claims 1 to 7.

Images