An engine part production tracing method and system based on bar code data

Abstract

The application relates to the technical field of data processing, in particular to an engine part production tracing method and system based on bar code data, which comprises the following steps: acquiring a bar code scanning event sequence of an engine part; reconstructing a link of the bar code scanning event sequence by using an improved time sequence diagram attention network model and outputting a tracing result of the engine part; wherein the improved time sequence diagram attention network model comprises a frequency parameter. The application solves the problems of low accuracy and continuity of the tracing result.

Description

Technical Field

[0001] This invention relates to the field of data processing technology. More specifically, this invention relates to a method and system for tracing the production of engine parts based on barcode data. Background Technology

[0002] In modern engine manufacturing, there are numerous types of components and complex production processes involving precision machining, assembly, and inspection. Each component undergoes multiple operations throughout its production cycle. Traditional production traceability methods rely primarily on manual recording or static barcode scanning, lacking comprehensive and real-time monitoring capabilities for the dynamic status of components during the production process. This approach not only suffers from data update delays and incomplete records but also struggles to accurately reflect the dynamic changes of components at different stages of the process. This can easily lead to omissions in component production information, incomplete traceability links, or delayed information updates, thereby affecting quality management, production anomaly analysis, and the reliability of subsequent maintenance services. Furthermore, with the development of intelligent manufacturing and the increasing automation of production lines, higher demands are placed on real-time traceability and refined management of component production processes. This makes traditional traceability methods insufficient to meet the production monitoring needs in the context of industrial big data.

[0003] To improve the accuracy and intelligence of component production traceability, a traceability method based on temporal graph attention network has been proposed in recent years. By constructing a temporal graph of component barcode scanning events, the time information and process category information of components in each process link are mapped into the graph structure. Attention mechanism is used to capture the temporal dependencies between components and process nodes, thereby realizing the reconstruction and traceability of the component production chain.

[0004] However, in existing methods, time-series graph attention networks typically use fixed frequency parameters to weight features and allocate attention to each process node. These fixed parameters cannot adapt to the differences in execution time, operation rhythm, and fluctuation range of different process categories. This can easily lead to insufficient attention to abnormal processes or over-response to stable processes, making it difficult to effectively identify potential production anomalies or process fluctuations. Consequently, the accuracy and continuity of traceability results are not high. Summary of the Invention

[0005] To address the data processing problems mentioned in the background section, the present invention provides solutions in several aspects.

[0006] In a first aspect, the present invention provides a method for tracing the production of engine parts based on barcode data, comprising: acquiring a barcode scanning event sequence of engine parts; reconstructing the links of the barcode scanning event sequence using an improved temporal graph attention network model and outputting the traceability result of the engine parts; wherein, the improved temporal graph attention network model includes a frequency parameter, the frequency parameter being positively correlated with an adaptive bandwidth factor; the adaptive bandwidth factor specifically comprises: acquiring the time interval between any two adjacent barcode scanning events in the barcode scanning event sequence to obtain a time interval sequence, and grouping the time interval sequence according to process category to obtain a grouped time interval sequence corresponding to each process category; acquiring the separation degree of the target process category, the separation degree being positively correlated with the standard deviation of the grouped time interval sequence corresponding to the target process category and negatively correlated with the mean of the grouped time interval sequence corresponding to the target process category; acquiring the anomaly degree of the target process category, the anomaly degree being positively correlated with the separation degree and the standard deviation of the grouped time interval sequence corresponding to the target process category, and negatively correlated with the median of the grouped time interval sequence corresponding to the target process category; the target process category being any process category among all process categories, and the adaptive bandwidth factor being positively correlated with the anomaly degree.

[0007] The above technical solution enables dynamic traceability of engine parts throughout the production and assembly process. By adaptively adjusting the network's focus weight on each process, it achieves higher accuracy in identifying and reconstructing key processes and links with abnormal fluctuations, while reducing redundant focus on stable processes. This significantly improves the completeness, continuity, and reliability of the parts traceability results, enhances sensitivity to production anomalies, and enables visualized monitoring, intelligent scheduling, and quality management of the production process. It provides highly reliable data support for production optimization, anomaly warning, and subsequent maintenance.

[0008] Furthermore, the first Frequency parameters for each process category for: , For the first Adaptive bandwidth factor for each process category This represents the maximum value of the frequency parameter. This represents the total number of process categories.

[0009] The above technical solution dynamically adjusts the frequency parameters of each process in the network, so that different processes receive attention weights that match their importance and degree of abnormality during link reconstruction. This improves the identification accuracy of key processes and the detection capability of abnormal processes, while ensuring the continuity and integrity of the overall traceability process. It significantly improves the reliability and accuracy of the traceability results of parts production, and provides highly reliable data support for production process monitoring, quality management and intelligent scheduling.

[0010] Furthermore, the first Adaptive bandwidth factor for each process category for: , For the first The degree of abnormality of each process category For the first The dominant frequency of the grouping time interval sequence corresponding to each process category This represents the mean of the dominant frequencies of the grouped time interval sequences corresponding to all process categories. , The first The standard deviation and median of the time interval sequence corresponding to each process category.

[0011] The above technical solution adaptively adjusts the network's attention weight to each process based on the fluctuation characteristics and abnormality of each process over time intervals. This allows processes with significant deviations or abnormal behaviors to receive higher identification priority during the traceability process, while suppressing excessive attention to stable processes. As a result, it significantly improves the accuracy, continuity, and reliability of the component production chain reconstruction, effectively enhances the detection capability of abnormal processes, and provides highly reliable data support for production process monitoring, quality management, and intelligent scheduling.

[0012] Furthermore, the dominant frequency specifically refers to: the frequency of the first... Each process category corresponds to a grouped time interval sequence which is then subjected to a Fast Fourier Transform to obtain the frequency value of each frequency point. The frequency corresponding to the frequency value is used as the first frequency. The dominant frequency of the grouping time interval sequence corresponding to each process category.

[0013] The above technical solution extracts the dominant frequency of the time interval sequence of each process, enabling the network to accurately capture the periodic characteristics and regular fluctuations of the process in the time dimension. This improves the ability to identify key and abnormal processes, enhances the accuracy and continuity of the component production chain reconstruction, and achieves the reliability and completeness of traceability results. At the same time, it provides stable and highly reliable data support for production process monitoring, quality control and intelligent scheduling.

[0014] Furthermore, the first The degree of abnormality of each process category for: , , The first The, the The mean of the time interval sequence corresponding to each process category. , The first The, the The standard deviation of the time interval sequence corresponding to each process category The mean of the separation degree for all process categories. In order to be with the first The number of adjacent process categories for each process category.

[0015] The above technical solution effectively identifies processes that deviate significantly from the norm in terms of production rhythm or execution behavior by quantifying the time interval differences and fluctuation characteristics of each process relative to adjacent processes. This enhances the network's sensitivity to abnormal processes, improves the accuracy and continuity of component production link reconstruction, and ensures the integrity and reliability of traceability results, providing highly reliable data support for production process monitoring, quality management, and intelligent scheduling.

[0016] Furthermore, the first Separation of process categories for: , , The first The standard deviation and mean of the time interval sequence corresponding to each process category. For the first The kurtosis value of the time interval sequence corresponding to each process category is the kurtosis value of the time interval sequence.

[0017] The above technical solution, by comprehensively considering the fluctuation range, average level and distribution sharpness of the time interval sequence of each process, can accurately reflect the significance and difference of the process in the overall production rhythm, thereby enhancing the network's ability to identify abnormal or critical processes, improving the accuracy and continuity of the component production link reconstruction, ensuring the integrity and reliability of traceability results, and providing highly reliable data support for production process monitoring, quality management and intelligent scheduling.

[0018] Furthermore, the kurtosis value of the time interval sequence is obtained by calculating the normal distribution of the time interval sequence.

[0019] Furthermore, it also includes performing missing completion processing on the barcode scanning event sequence.

[0020] Furthermore, the barcode scanning event sequence includes: engine component ID, timestamp, and process category.

[0021] In a second aspect, the present invention provides an engine component production traceability system based on barcode data, including a memory and a processor. The memory stores computer program instructions, which, when executed by the processor, implement any of the above-described engine component production traceability methods based on barcode data.

[0022] The beneficial effects of this invention are as follows: This invention achieves precise reconstruction and full traceability of the parts production chain by comprehensively acquiring and dynamically analyzing the barcode scanning event sequence of engine parts, combined with adaptively adjusted network parameters and multi-dimensional process characteristics. It can effectively identify key processes and abnormal fluctuation links, improve the accuracy, continuity and reliability of traceability results, and ensure the integrity and availability of production process information. Through the completion processing of missing data and in-depth mining of time series characteristics, it achieves efficient monitoring of parts processing, assembly and inspection links, providing highly reliable data support and decision-making basis for production process optimization, quality management, anomaly early warning and intelligent scheduling. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating an engine component production traceability method based on barcode data according to an embodiment of the present invention; Figure 2 This is a schematic illustration of a heatmap showing the reconstruction accuracy of different process links in an engine component production traceability method based on barcode data according to an embodiment of the present invention. Figure 3 This is a schematic diagram illustrating the structural block of an engine component production traceability system based on barcode data according to an embodiment of the present invention. Detailed Implementation

[0024] An embodiment of a method for tracing the production of engine parts based on barcode data.

[0025] like Figure 1 The flowchart shown is a production traceability method for engine parts based on barcode data according to an embodiment of the present invention, which includes the following steps: S1: Obtain the barcode scanning event sequence of engine parts.

[0026] In a preferred embodiment, the barcode scanning event sequence includes key data items such as a unique identifier (ID) for each engine component, corresponding timestamp information, and process category. The engine component ID uniquely identifies each component, ensuring traceability during production and assembly. The timestamp information records the specific time of occurrence of each component in each process stage, enabling precise monitoring of production progress and assembly flow. The process category information distinguishes the specific process stage of the component, facilitating subsequent data analysis, anomaly detection, and production process optimization. To ensure the integrity and reliability of the barcode scanning event sequence data, missing data completion processing is performed. This includes automatically inferring and filling in missing barcode scanning records due to scanning failures, equipment malfunctions, or data transmission delays, ensuring continuous and uninterrupted sequence data. This missing data completion processing not only enhances data integrity and consistency but also improves the accuracy of subsequent analysis algorithms, making component status monitoring, production anomaly diagnosis, and quality traceability more reliable, thereby achieving refined management and intelligent control of the production process. Implementing this step can effectively reduce analytical biases caused by missing data, while improving the visualization management level of the entire engine production line, providing a solid data foundation for production optimization, quality improvement, and intelligent decision-making.

[0027] S2: The improved temporal graph attention network model is used to reconstruct the links of the barcode scanning event sequence and output the traceability results of engine parts.

[0028] like Figure 2 As shown, this is a heatmap illustrating the reconstruction accuracy of different process links in an engine component production traceability method based on barcode data, according to an embodiment of the present invention.

[0029] In a preferred embodiment, the improved temporal graph attention network model includes a frequency parameter, the first... Frequency parameters for each process category for: , For the first Adaptive bandwidth factor for each process category This represents the maximum value of the frequency parameter. This represents the total number of process categories.

[0030] By introducing an improved temporal graph attention network model to reconstruct the chain of barcode scanning event sequences for engine components, precise tracking of components during production and assembly was achieved. When processing events of different process categories, the temporal graph attention network model dynamically adjusts the attention weight of each process in the network based on its characteristics, thereby more accurately capturing the correlation and time dependency characteristics of components between different processes. Through this adaptive weight adjustment mechanism, the model can strengthen its ability to identify key process nodes and important events, while reducing interference from noise or minor events, resulting in more complete and reliable traceability results. This improves the accuracy and continuity of component traceability, enables transparent management of the production process, and provides highly reliable data support for subsequent production anomaly analysis, quality control, and intelligent scheduling, thus significantly enhancing the intelligence and controllability of the production line.

[0031] The adaptive bandwidth factor specifically involves: obtaining the time interval between any two adjacent barcode scanning events in the barcode scanning event sequence to obtain a time interval sequence, and grouping the time interval sequence according to the process category to obtain a grouped time interval sequence corresponding to each process category; No. Adaptive bandwidth factor for each process category for: , For the first The degree of abnormality of each process category For the first The dominant frequency of the grouping time interval sequence corresponding to each process category This represents the mean of the dominant frequencies of the grouped time interval sequences corresponding to all process categories. , The first The standard deviation and median of the time interval sequence corresponding to each process category. The adaptive bandwidth factor not only considers the abnormal situation of each process, but also integrates the dominant frequency deviation and fluctuation amplitude of the process time interval. This allows the model to give higher sensitivity to processes with abnormalities or significant changes, while reducing the weight of normal and stable processes, thereby achieving accurate identification and reconstruction of the component process chain.

[0032] The dominant frequency is specifically: for the first Each process category corresponds to a grouped time interval sequence which is then subjected to a Fast Fourier Transform to obtain the frequency value of each frequency point. The frequency corresponding to the frequency value is used as the first frequency. The dominant frequency of the grouping time interval sequence corresponding to each process category.

[0033] By collecting and analyzing the time intervals of adjacent events in the component barcode scanning event sequence, and grouping the time intervals by process category, a grouped time interval sequence for each process is obtained. Based on this, frequency domain analysis is performed on the time interval sequence of each process to extract its dominant frequency. Then, an adaptive bandwidth factor is calculated by combining the anomaly degree of each process and the fluctuation characteristics of the time intervals, which is used to adjust the model's focus weight in link reconstruction. This allows the model to pay more attention to abnormal fluctuations and critical processes when reconstructing the component process link, while reducing interference with stable or routine processes. This improves the accuracy, continuity, and reliability of traceability results, enabling refined monitoring and anomaly identification of the production process, and providing highly reliable data support for production optimization, quality control, and intelligent scheduling.

[0034] No. The degree of abnormality of each process category for: , , The first The, the The mean of the time interval sequence corresponding to each process category. , The first The, the The standard deviation of the time interval sequence corresponding to each process category The mean of the separation degree for all process categories. In order to be with the first The number of adjacent process categories for each process category.

[0035] By calculating the ratio of the difference in the mean time interval between each process category and its adjacent process categories to the fluctuation characteristics, and combining this with the mean of the overall separation degree between process categories, the degree of anomaly in each process category is quantified, thereby achieving an accurate assessment of the abnormal characteristics of processes in the time series. Its construction logic lies in highlighting those processes that significantly deviate from the norm in execution rhythm or processing behavior by comparing the statistical characteristics of the time intervals of adjacent processes. This allows the model to adaptively identify key or abnormal processes during link reconstruction or traceability. This enhances the detection capability of abnormal processes, improves the accuracy and continuity of component process link reconstruction, and provides reliable data support for anomaly early warning, quality control, and intelligent management of the production process, ensuring the integrity and credibility of traceability results.

[0036] No. Separation of process categories for: , , The first The standard deviation and mean of the time interval sequence corresponding to each process category. For the first The kurtosis value of the time interval sequence corresponding to each process category The kurtosis value is the kurtosis value of the time interval sequence, which is obtained by calculating the normal distribution of the time interval sequence.

[0037] By combining the fluctuation amplitude and average level of the time interval sequence for each process category, as well as the kurtosis characteristics of the sequence and its difference from the kurtosis of the overall time interval sequence, a separation index for process categories is constructed to quantify the significance and deviation of each process in the time distribution. By simultaneously considering the instability and sharpness of process execution, processes that are significantly different from the overall rhythm or exhibit abnormal fluctuations can be identified, allowing the model to assign higher weights to key processes during link reconstruction and tracing. This improves the ability to distinguish process features and the accuracy of anomaly identification, enhances the accuracy and continuity of component process link reconstruction, and provides a reliable quantitative basis for anomaly detection, quality monitoring, and intelligent management in the production process.

[0038] The solution of this invention enables full dynamic traceability of engine parts during the production process by intelligently analyzing and reconstructing the barcode scanning event sequence. This accurately identifies key and abnormal processes, improves the accuracy, continuity, and reliability of traceability results, enhances the ability to detect production anomalies, and ensures the integrity of information by filling in missing data. It also enables visualized monitoring, quality management, and intelligent scheduling of the production process, providing highly reliable data support for production optimization, anomaly warning, and subsequent maintenance, and significantly improving the intelligence level and decision-making efficiency of engine part production management.

[0039] An example of an engine component production traceability system based on barcode data: like Figure 3 As shown in the diagram, a structural block diagram of an engine component production traceability system based on barcode data according to an embodiment of the present invention includes a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement the engine component production traceability method based on barcode data according to the present invention.

[0040] The aforementioned engine component production traceability system based on barcode data also includes other components well known to those skilled in the art, such as communication interfaces. Their settings and functions are known in the art and will not be described in detail here.

[0041] In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used or combined with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store desired information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device. Any application or module described in this invention can be implemented using computer-readable / executable instructions stored or otherwise maintained by such a computer-readable medium.

[0042] In the description of this specification, "multiple" or "several" means at least two, such as two, three or more, unless otherwise explicitly specified.

[0043] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.

Claims

1. A method for tracing the production of engine parts based on barcode data, characterized in that, include: Obtain the barcode scanning event sequence of engine parts; An improved temporal graph attention network model is used to reconstruct the links of the barcode scanning event sequence and output the traceability results of engine parts; The improved temporal graph attention network model includes a frequency parameter, which is positively correlated with the adaptive bandwidth factor. The adaptive bandwidth factor specifically involves: obtaining the time interval between any two adjacent barcode scanning events in the barcode scanning event sequence to obtain a time interval sequence, and grouping the time interval sequence according to the process category to obtain a grouped time interval sequence corresponding to each process category; The separation degree of the target process category is obtained, which is positively correlated with the standard deviation of the grouping time interval sequence corresponding to the target process category and negatively correlated with the mean of the grouping time interval sequence corresponding to the target process category; the anomaly degree of the target process category is obtained, which is positively correlated with the separation degree and the standard deviation of the grouping time interval sequence corresponding to the target process category and negatively correlated with the median of the grouping time interval sequence corresponding to the target process category; the target process category is any process category among all process categories, and the adaptive bandwidth factor is positively correlated with the anomaly degree.

2. The method for tracing the production of engine parts based on barcode data according to claim 1, characterized in that, No. Frequency parameters for each process category for: , For the first Adaptive bandwidth factor for each process category This represents the maximum value of the frequency parameter. This represents the total number of process categories.

3. The method for traceability of engine component production based on barcode data according to claim 1, characterized in that, No. Adaptive bandwidth factor for each process category for: , For the first The degree of abnormality of each process category For the first The dominant frequency of the grouping time interval sequence corresponding to each process category This represents the mean of the dominant frequencies of the grouped time interval sequences corresponding to all process categories. , The first The standard deviation and median of the time interval sequence corresponding to each process category.

4. The method for traceability of engine component production based on barcode data according to claim 3, characterized in that, The dominant frequency is specifically: for the first Each process category corresponds to a grouped time interval sequence which is then subjected to a Fast Fourier Transform to obtain the frequency value of each frequency point. The frequency corresponding to the frequency value is used as the first frequency. The dominant frequency of the grouping time interval sequence corresponding to each process category.

5. The method for tracing the production of engine parts based on barcode data according to claim 1, characterized in that, No. The degree of abnormality of each process category for: , , The first The, the The mean of the time interval sequence corresponding to each process category. , The first The, the The standard deviation of the time interval sequence corresponding to each process category The mean of the separation degree for all process categories. In order to be with the first The number of adjacent process categories for each process category.

6. The method for traceability of engine component production based on barcode data according to claim 1, characterized in that, No. Separation of process categories for: , , The first The standard deviation and mean of the time interval sequence corresponding to each process category. For the first The kurtosis value of the time interval sequence corresponding to each process category is the kurtosis value of the time interval sequence.

7. The method for traceability of engine component production based on barcode data according to claim 6, characterized in that, The kurtosis value of the time interval sequence is obtained by calculating the normal distribution of the time interval sequence.

8. The method for traceability of engine component production based on barcode data according to claim 1, characterized in that, It also includes missing completion processing for the barcode scanning event sequence.

9. The method for traceability of engine component production based on barcode data according to claim 1, characterized in that, The barcode scanning event sequence includes: engine component ID, timestamp, and process category.

10. A production traceability system for engine parts based on barcode data, characterized in that, It includes a memory and a processor, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, it implements the engine component production traceability method based on barcode data as described in any one of claims 1 to 9.

Images