Product quality monitoring method and system based on multi-source data analysis

Abstract

The invention discloses a product quality monitoring method and system based on multi-source data analysis, and belongs to the technical field of product quality monitoring. According to the method, a multi-source data acquisition device is configured for a production link, a quality abnormal state is extracted, a feature tag is generated, and meanwhile time synchronization processing is carried out to ensure data consistency; after statistics of the data source and the abnormal state completion label, constructing an abnormal local data matrix, and calculating abnormal difference of time and production link dimensions to generate a quality feature matrix; and calculating a quality portrait score based on the matrix, and prompting to adjust production parameters through threshold judgment. The system comprises a data acquisition and synchronization module, a data statistics and label module, a matrix construction and feature module, and a quality evaluation and feedback module. According to the method, the problems of data asynchronization and one-sided evaluation in traditional monitoring are solved, the accuracy and the real-time performance of quality anomaly detection are improved, and the method is suitable for whole-process quality control of a complex production scene.

Description

Technical Field

[0001] This invention relates to the field of product quality monitoring technology, specifically to a product quality monitoring method and system based on multi-source data analysis. Background Technology

[0002] In modern manufacturing, product quality monitoring is a core element in ensuring production stability and product reliability. Traditional quality monitoring methods often rely on single equipment or manual sampling, which has the following limitations: First, the data source is singular, making it difficult to cover all quality-influencing factors in the entire production process (such as raw materials, equipment operation, environmental parameters, etc.); second, the lack of temporal synchronization in multi-equipment monitoring data makes it difficult to analyze the spatiotemporal correlation of abnormal states; third, quality assessment relies heavily on empirical thresholds, making it difficult to quantify the transmission relationship of anomalies in different production stages, which easily leads to missed detections or misjudgments.

[0003] With the development of intelligent manufacturing, multi-source data fusion has become a trend in quality monitoring. However, existing technologies have not yet formed a standardized multi-source data processing process: on the one hand, the time synchronization accuracy of multi-source equipment is insufficient, resulting in distortion of time series analysis of abnormal states; on the other hand, there is a lack of quantitative modeling of abnormal characteristics in multiple links, making it difficult to achieve correlation assessment from local anomalies to overall quality. Summary of the Invention

[0004] The purpose of this invention is to provide a product quality monitoring method and system based on multi-source data analysis to solve the problems mentioned in the background art. Specifically, the core principle of this invention is to achieve full-process quality monitoring through a logical chain of "multi-source data synchronization - anomaly feature labeling - feature matrix quantification - dynamic quality score evaluation," including: Multi-source data synchronization: Multi-source acquisition devices are configured for each stage of production, and time synchronization commands are used to ensure that the monitoring time base of all devices is consistent, laying the foundation for subsequent anomaly correlation analysis; Abnormal feature labeling: Abnormal states in each stage (such as equipment parameters exceeding the standard, process deviations, etc.) are converted into structured labels, which include key information such as the duration of the abnormality and the number of equipment involved, so as to achieve standardized recording of abnormal states; Feature matrix quantification: Construct an abnormal local data matrix, and by calculating the abnormal difference in the time dimension (abnormal differences at different trigger times) and the production process dimension (abnormal transmission between adjacent processes), transform unstructured abnormal features into a quantifiable quality feature matrix, which intuitively reflects the spatiotemporal distribution pattern of abnormalities. Dynamic quality assessment: Based on the quality feature matrix, the overall dispersion (reflecting quality stability) and modulus deviation (reflecting deviation from the pass standard) are calculated to generate a quality profile score, and dynamic early warning of quality risks is achieved through preset thresholds.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: A product quality monitoring system based on multi-source data analysis, the system includes: a data acquisition and synchronization module, a data statistics and tagging module, a matrix construction and feature module, and a quality assessment and feedback module; The data acquisition and synchronization module is used to configure the acquisition device and synchronize the time, and extract abnormal quality states to generate feature labels. The data statistics and labeling module is used to collect data on data sources and abnormal states, and to fill in feature labels; The matrix construction and feature module is used to construct an abnormal local data matrix and quantify abnormal differences to generate a quality feature matrix. The quality assessment and feedback module quantifies the quality profile score based on the quality feature matrix and prompts the user to operate the production parameters according to the quality profile score.

[0006] Furthermore, the data acquisition and synchronization module includes a device configuration unit and a time synchronization unit; The equipment configuration unit is used to configure multi-source data acquisition equipment for the production process, extract quality anomalies and generate feature labels. The time synchronization unit is used to perform time synchronization processing on the monitoring feedback behavior of multi-source data acquisition devices, triggering synchronization instructions at the beginning of the production process to update the device time synchronization status to ensure time synchronization in abnormal states.

[0007] Furthermore, the data statistics and tagging module includes a data source statistics unit and an abnormal status recording unit; The data source statistics unit is used to count the multi-source data acquisition devices configured in each production stage, forming a data source set; The abnormal state recording unit, based on the time synchronization sequence and trigger tags, counts the quality abnormal states in the statistical data source set, records them into feature tags, and improves the tag information.

[0008] Furthermore, the matrix construction and feature module includes a local matrix generation unit and a quality feature generation unit; The local matrix generation unit is used to capture the quality anomaly status of different product batches through the production device port and construct an anomaly local data matrix. The quality feature generation unit is used to calculate the abnormal differences in the time dimension and the production process dimension, centered on the elements of the abnormal local data matrix, to form a local vector and calculate its modulus, and then map the modulus to the corresponding position in the matrix to generate a quality feature matrix.

[0009] Furthermore, the quality assessment and feedback module includes a scoring calculation unit and a threshold judgment unit; The scoring calculation unit is used to calculate the overall dispersion and modulus deviation of quality features based on the quality feature matrix, and to obtain a quality profile score. The threshold judgment unit is used to preset the quality profile scoring threshold. If the score reaches or exceeds the threshold, the staff is prompted to adjust the operating parameters of the production unit; if the score does not reach the threshold, the staff is prompted to maintain the existing parameters.

[0010] A product quality monitoring method based on multi-source data analysis, comprising the following steps: Step S1: Configure multi-source data acquisition equipment for the production process, extract the quality anomaly status monitored by the multi-source data acquisition equipment and generate product quality anomaly feedback feature tags, and at the same time perform time synchronization processing on the monitoring feedback behavior of the multi-source data acquisition equipment to ensure that the quality anomaly status has time synchronization. Step S2: Analyze the multi-source data acquisition devices configured in each production stage to form a data source set. Based on the time synchronization sequence and trigger tags, analyze the quality anomaly status in the data source set and record it in the product quality anomaly feedback feature tag. Step S3: Capture the quality anomaly status of different product batches through the production device port, construct an anomaly local data matrix, calculate the anomaly difference in the time dimension and the production process dimension, and generate a quality feature matrix; Step S4: Based on the quality feature matrix, quantify the overall dispersion and modulus deviation of the quality features to obtain a quality profile score. Based on the quality profile score threshold, prompt the staff to adjust the operating parameters of the production unit or maintain the existing parameters.

[0011] Furthermore, the specific implementation process of step S1 includes: Based on the production process, multi-source data acquisition devices are configured. Several types of multi-source data acquisition devices are configured for each production process. The quality anomaly statuses monitored and fed back by the multi-source data acquisition devices are extracted to generate product quality anomaly feedback feature tags. The multi-source data acquisition devices are used to monitor the operating behavior of the production equipment in the production process. The quality anomaly status refers to the abnormal status fed back by the multi-source data acquisition devices when they detect that the operating behavior leads to quality risks. Each quality anomaly status is configured with an anomaly status code. Based on the execution logic of the production process, the monitoring and feedback behavior of multi-source data acquisition devices is synchronized in time. At the beginning stage of entering the production process, the time synchronization command is executed and the time synchronization status of the multi-source data acquisition devices is updated synchronously, so that the quality anomaly status monitored and fed back by the multi-source data acquisition devices has time synchronization.

[0012] Furthermore, the specific implementation process of step S2 includes: Let any k-th multi-source data acquisition device be denoted as . Let any l-th production stage be denoted as Then, statistical production process The built-in multi-source data acquisition devices constitute a data source set. p represents the total number of multi-source data acquisition devices; After the time synchronization instruction is triggered, a time synchronization sequence is formed, denoted as... , This represents the nth synchronization time node, where N represents the total number of synchronization time nodes. A trigger tag is appended to the time synchronization sequence based on the number of times the time synchronization command is triggered, denoted as... ,and , g represents the trigger sequence number; In trigger tag The following statistical data source set The feedback characteristics of each multi-source data acquisition device that generates quality anomaly feedback are recorded in the product quality anomaly feedback characteristic label. In the middle, and ,in, Indicates that the tag is triggered Downstream production stage The duration, and , and These respectively represent the triggering tags Downstream production The start and end synchronization times. Indicates that the tag is triggered The following statistical data source set The total number of multi-source data acquisition devices that generate quality anomaly status feedback.

[0013] Furthermore, the specific implementation process of step S3 includes: Configure production unit ports to connect to multi-source data acquisition devices and capture quality anomalies in different product batches during production. The m-th production unit port is denoted as... Let the a-th product batch be denoted as , to the production device port Captured product batches Product quality anomaly feedback feature labels mapped to anomaly local data matrix In the matrix, the element corresponding to the g-th row and l-th column is the product quality anomaly feedback feature label. ; Based on the abnormal local data matrix Select product quality anomaly feedback feature tags Using local centers, calculate product quality anomaly feedback feature labels. Time-dimension anomaly discrepancies and production-process-dimension anomaly discrepancies: ; In the formula, Product quality anomaly feedback feature tags Anomalous difference in the time dimension Product quality anomaly feedback feature tags The production process dimension shows abnormal differences; Based on the time-dimension anomaly differences and the production-stage anomaly differences, a local vector is constructed. Obtain local vectors model and the model Mapping to the abnormal local data matrix The matrix position corresponding to the g-th row and l-th column will then be the abnormal local data matrix. Transformed into product batches during the production process The quality characteristic matrix, denoted as .

[0014] Furthermore, the specific implementation process of step S4 includes: Based on the quality feature matrix, product batches during the production process Conduct quality profile scoring: Based on quality feature matrix Calculate the standard deviation of all moduli to reflect the overall dispersion of the quality characteristics. In the formula, Represents the quality characteristic matrix The mean of all modulo 1, and G represents the total number of trigger tags, and L represents the total number of production stages; Based on quality feature matrix Calculate the deviation of each modulus value from the standard modulus value, where the standard modulus value is the mean of the modulus values ​​of the quality characteristic matrix of historical qualified batches. The deviation of the modulus value in the g-th row and l-th column is calculated. In the formula, The standard modulus value; The quality profile score is obtained based on the overall dispersion and the modulus deviation. In the formula, The importance weights of the pre-set production units, The maximum allowable dispersion is preset. Preset quality profile scoring threshold If the quality profile score If the quality profile score is low, the staff will be prompted to adjust the operating parameters of the production unit. If no quality risk is found, the staff will be advised to maintain the operating parameters of the production unit.

[0015] Compared with the prior art, the beneficial effects achieved by the present invention are: Improved data synchronization: By using time synchronization commands, the time error of multi-source devices is controlled within ±0.1s, which solves the problem of "time misalignment" of data from multiple devices in the past and makes spatiotemporal correlation analysis of abnormal states possible; Enhanced comprehensiveness of assessment: Through multi-dimensional anomaly differential quantification, it can not only identify local anomalies in a single link, but also capture the transmission pattern of anomalies between production links (such as the impact of equipment anomalies in the preceding link on the quality of subsequent assembly). Improved decision-making accuracy: Dynamic threshold judgment is achieved based on quantitative quality profile scoring, avoiding the subjectivity of traditional experience thresholds and making quality adjustment suggestions more data-supported (e.g., when the score is below the threshold, abnormal links can be accurately located and the direction of parameter adjustment can be indicated). Attached Figure Description

[0016] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0017] Figure 1 This is a schematic diagram illustrating the steps of a product quality monitoring method based on multi-source data analysis according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] In this first embodiment: a product quality monitoring system based on multi-source data analysis is provided. The system includes: a data acquisition and synchronization module, a data statistics and tagging module, a matrix construction and feature module, and a quality assessment and feedback module. The data acquisition and synchronization module is used to configure the acquisition device and synchronize the time, and extract abnormal quality states to generate feature labels. The data acquisition and synchronization module includes a device configuration unit and a time synchronization unit. The equipment configuration unit is used to configure multi-source data acquisition equipment for the production process, extract quality anomalies and generate feature labels. The time synchronization unit is used to perform time synchronization processing on the monitoring feedback behavior of multi-source data acquisition devices, trigger synchronization instructions at the beginning of the production process, and update the device time synchronization status to ensure the time synchronization of abnormal states. The data statistics and labeling module is used to collect data on data sources and abnormal states, and to fill in feature labels; The data statistics and tagging module includes a data source statistics unit and an abnormal status recording unit. The data source statistics unit is used to count the multi-source data acquisition devices configured in each production stage, forming a data source set; The abnormal state recording unit, based on the time synchronization sequence and trigger tags, counts the quality abnormal states in the statistical data source set, records them into feature tags, and improves the tag information. The matrix construction and feature module is used to construct an abnormal local data matrix and quantify abnormal differences to generate a quality feature matrix. The matrix construction and feature module includes a local matrix generation unit and a quality feature generation unit. The local matrix generation unit is used to capture the quality anomaly status of different product batches through the production device port and construct an anomaly local data matrix. The quality feature generation unit is used to calculate the abnormal differences in the time dimension and the production process dimension with the elements of the abnormal local data matrix as the center, form a local vector and calculate its modulus, and map the modulus to the corresponding position in the matrix to generate a quality feature matrix. The quality assessment and feedback module quantifies the quality profile score based on the quality feature matrix and prompts the operator to operate the production parameters according to the quality profile score. The quality assessment and feedback module includes a scoring calculation unit and a threshold judgment unit. The scoring calculation unit is used to calculate the overall dispersion and modulus deviation of quality features based on the quality feature matrix, and to obtain a quality profile score. The threshold judgment unit is used to preset the quality profile scoring threshold. If the score reaches or exceeds the threshold, the staff is prompted to adjust the operating parameters of the production unit; if the score does not reach the threshold, the staff is prompted to maintain the existing parameters.

[0020] Please see Figure 1 In this second embodiment, a product quality monitoring method based on multi-source data analysis is provided, applicable to the first embodiment described above. Taking the production scenario of automobile engine cylinder blocks as an example, the specific implementation process of the present invention is explained as follows: The production process includes four stages: casting, machining, cleaning, and assembly (L1-L4). The multi-source data acquisition devices configured for each stage are as follows: L1 (casting) is equipped with temperature and pressure sensors; L2 (machining) is equipped with vibration sensors and dimensional measuring instruments; L3 (cleaning) is equipped with pH and flow sensors; and L4 (assembly) is equipped with torque sensors and visual inspection instruments. The preset quality profile scoring threshold is 0.7 (≥0.7 prompts adjustment, <0.7 maintains the current level), and the standard modulus value μ0 for qualified batches is 5.2. The specific implementation process includes the following: Step S1: Configure multi-source data acquisition equipment for the production process, extract the quality anomaly status monitored by the multi-source data acquisition equipment and generate product quality anomaly feedback feature tags, and at the same time perform time synchronization processing on the monitoring feedback behavior of the multi-source data acquisition equipment to ensure that the quality anomaly status has time synchronization. For example, based on the production process, multi-source data acquisition devices are configured, wherein several types of multi-source data acquisition devices are configured for each production process, and the quality anomaly statuses monitored and fed back by the multi-source data acquisition devices are extracted to generate product quality anomaly feedback feature tags. The multi-source data acquisition devices are used to monitor the operating behavior of the production equipment in the production process. The quality anomaly status refers to the anomaly status fed back by the multi-source data acquisition devices when they detect that the operating behavior has caused quality risks. Each quality anomaly status is configured with an anomaly status code. Based on the execution logic of the production process, the monitoring and feedback behavior of multi-source data acquisition devices is synchronized in time. At the beginning stage of entering the production process, the time synchronization instruction is executed and the time synchronization status of the multi-source data acquisition devices is updated synchronously, so that the quality abnormality status monitored and fed back by the multi-source data acquisition devices has time synchronization. For example, when production starts, a time synchronization command is triggered, and the time of each device is synchronized to the same reference (error ≤ 0.1s); the temperature sensor in the casting process (L1) detects that the molten metal temperature exceeds the upper limit (abnormal state), and generates a tag (including the abnormal duration of 12s and 1 device involved); the vibration sensor in the machining process (L2) detects that the tool vibration exceeds the standard, and generates a tag (duration of 8s and 1 device involved).

[0021] Step S2: Analyze the multi-source data acquisition devices configured in each production stage to form a data source set. Based on the time synchronization sequence and trigger tags, analyze the quality anomaly status in the data source set and record it in the product quality anomaly feedback feature tag. For example, let any k-th multi-source data acquisition device be denoted as Let any l-th production stage be denoted as Then, statistical production process The built-in multi-source data acquisition devices constitute a data source set. p represents the total number of multi-source data acquisition devices; After the time synchronization instruction is triggered, a time synchronization sequence is formed, denoted as... , This represents the nth synchronization time node, where N represents the total number of synchronization time nodes. A trigger tag is appended to the time synchronization sequence based on the number of times the time synchronization command is triggered, denoted as... ,and , g represents the trigger sequence number; In trigger tag The following statistical data source set The feedback characteristics of each multi-source data acquisition device that generates quality anomaly feedback are recorded in the product quality anomaly feedback characteristic label. In the middle, and ,in, Indicates that the tag is triggered Downstream production stage The duration, and , and These respectively represent the triggering tags Downstream production stage The start and end synchronization times. Indicates that the tag is triggered The following statistical data source set The total number of multi-source data acquisition devices that generate quality anomaly status feedback; For example, the data source set for L1 is {temperature sensor, pressure sensor}, and for L2 it is {vibration sensor, size measuring instrument}; under the first trigger tag (g=1), the abnormal tag record for L1 is [duration 12s, involving 1 device], and for L2 it is [duration 8s, involving 1 device].

[0022] Step S3: Capture the quality anomaly status of different product batches through the production device port, construct an anomaly local data matrix, calculate the anomaly difference in the time dimension and the production process dimension, and generate a quality feature matrix; For example, a production device port is configured to connect to a multi-source data acquisition device and capture the quality anomalies of different product batches during the production process. The m-th production device port is denoted as... Let the a-th product batch be denoted as , to the production device port Captured product batches Product quality anomaly feedback feature labels mapped to anomaly local data matrix In the matrix, the element corresponding to the g-th row and l-th column is the product quality anomaly feedback feature label. ; Based on the abnormal local data matrix Select product quality anomaly feedback feature tags Using local centers, calculate product quality anomaly feedback feature labels. Time-dimension anomaly discrepancies and production-process-dimension anomaly discrepancies: ; In the formula, Product quality anomaly feedback feature tags Anomalous difference in the time dimension Product quality anomaly feedback feature tags The production process dimension shows abnormal differences;

[0023] Based on the time-dimension anomaly differences and the production-stage anomaly differences, a local vector is constructed. Obtain local vectors model and the model Mapping to the abnormal local data matrix The matrix position corresponding to the g-th row and l-th column will then be the abnormal local data matrix. Transformed into product batches during the production process The quality characteristic matrix, denoted as ; For example, the production unit port captures the abnormal state of the first batch (S1) and constructs an abnormal local data matrix Q(S1|V1) (V1 is the cylinder block production main line port); calculate the time dimension difference: the difference in the duration of L1 abnormality between the first trigger and the second trigger (g=2) is 5s, the difference in the number of equipment involved is 0, and the difference modulus is √(5²+0²)=5; the production link dimension difference: the difference in the duration of L1 abnormality between L2 is 4s, the difference in the number of equipment involved is 0, and the difference modulus is √(4²+0²)=4; map the modulus to the matrix to generate the quality feature matrix q(S1|V1).

[0024] Step S4: Based on the quality feature matrix, quantify the overall dispersion and modulus deviation of the quality features to obtain a quality profile score. Based on the quality profile score threshold, prompt the staff to adjust the operating parameters of the production unit or maintain the existing parameters. For example, based on the quality feature matrix, product batches during the production process... Conduct quality profile scoring: Based on quality feature matrix Calculate the standard deviation of all moduli to reflect the overall dispersion of the quality characteristics. In the formula, Represents the quality characteristic matrix The mean of all modulo 1, and G represents the total number of trigger tags, and L represents the total number of production stages; Based on quality feature matrix Calculate the deviation of each modulus value from the standard modulus value, where the standard modulus value is the mean modulus value of the quality characteristic matrix of historical qualified batches, and the modulus deviation in the g-th row and l-th column. In the formula, The standard modulus value; The quality profile score is obtained based on the overall dispersion and the modulus deviation. In the formula, The importance weights of the pre-set production units, The maximum allowable dispersion is preset. Preset quality profile scoring threshold If the quality profile score If the quality profile score is low, the staff will be prompted to adjust the operating parameters of the production unit. If the result is negative, it is determined that there is no quality risk, and staff are instructed to maintain the operating parameters of the production unit. For example, the mean value of the quality feature matrix is ​​calculated to be μ=4.8, the overall dispersion is δ=1.2 (maximum allowable dispersion δ(max)=3.0), and the deviation of the modulus is γ=|(4.8-5.2) / 5.2|=0.077; the quality profile score is H=1-[0.6×(1.2 / 3.0)+(1-0.6)×0.077]=1-[0.24+0.031]=0.729≥0.7, prompting the staff to adjust the temperature control parameters of L1.

[0025] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0026] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A product quality monitoring method based on multi-source data analysis, characterized in that, The method includes the following steps: Step S1: Configure multi-source data acquisition equipment for the production process, extract the quality anomaly status monitored by the multi-source data acquisition equipment and generate product quality anomaly feedback feature tags, and at the same time perform time synchronization processing on the monitoring feedback behavior of the multi-source data acquisition equipment to ensure that the quality anomaly status has time synchronization. Step S2: Analyze the multi-source data acquisition devices configured in each production stage to form a data source set. Based on the time synchronization sequence and trigger tags, analyze the quality anomaly status in the data source set and record it in the product quality anomaly feedback feature tag. Step S3: Capture the quality anomaly status of different product batches through the production device port, construct an anomaly local data matrix, calculate the anomaly difference in the time dimension and the production process dimension, and generate a quality feature matrix; Step S4: Based on the quality feature matrix, quantify the overall dispersion and modulus deviation of the quality features to obtain a quality profile score. Based on the quality profile score threshold, prompt the staff to adjust the operating parameters of the production unit or maintain the existing parameters.

2. The product quality monitoring method based on multi-source data analysis according to claim 1, characterized in that, The specific implementation process of step S1 includes: Based on the production process, multi-source data acquisition devices are configured. Several types of multi-source data acquisition devices are configured for each production process. The quality anomaly statuses monitored and fed back by the multi-source data acquisition devices are extracted to generate product quality anomaly feedback feature tags. The multi-source data acquisition devices are used to monitor the operating behavior of the production equipment in the production process. The quality anomaly status refers to the abnormal status fed back by the multi-source data acquisition devices when they detect that the operating behavior leads to quality risks. Each quality anomaly status is configured with an anomaly status code. Based on the execution logic of the production process, the monitoring and feedback behavior of multi-source data acquisition devices is synchronized in time. At the beginning stage of entering the production process, the time synchronization command is executed and the time synchronization status of the multi-source data acquisition devices is updated synchronously, so that the quality anomaly status monitored and fed back by the multi-source data acquisition devices has time synchronization.

3. The product quality monitoring method based on multi-source data analysis according to claim 2, characterized in that, The specific implementation process of step S2 includes: Let any k-th multi-source data acquisition device be denoted as . Let any l-th production stage be denoted as Then, statistical production process The built-in multi-source data acquisition devices constitute a data source set. p represents the total number of multi-source data acquisition devices; After the time synchronization instruction is triggered, a time synchronization sequence is formed, denoted as... , This represents the nth synchronization time node, where N represents the total number of synchronization time nodes. A trigger tag is appended to the time synchronization sequence based on the number of times the time synchronization command is triggered, denoted as... ,and , g represents the trigger sequence number; In trigger tag The following statistical data source set The feedback characteristics of each multi-source data acquisition device that generates quality anomaly feedback are recorded in the product quality anomaly feedback characteristic label. In the middle, and ,in, Indicates that the tag is triggered Downstream production The duration, and , and These respectively represent the triggering tags Downstream production The start and end synchronization times. Indicates that the tag is triggered The following statistical data source set The total number of multi-source data acquisition devices that generate quality anomaly status feedback.

4. The product quality monitoring method based on multi-source data analysis according to claim 3, characterized in that, The specific implementation process of step S3 includes: Configure production unit ports to connect to multi-source data acquisition devices and capture quality anomalies in different product batches during production. The m-th production unit port is denoted as... Let the a-th product batch be denoted as , to the production device port Captured product batches Product quality anomaly feedback feature labels mapped to anomaly local data matrix In the matrix, the element corresponding to the g-th row and l-th column is the product quality anomaly feedback feature label. ; Based on the abnormal local data matrix Select product quality anomaly feedback feature tags Using local centers, calculate product quality anomaly feedback feature labels. Time-dimension anomaly discrepancies and production-process-dimension anomaly discrepancies: ； In the formula, Product quality anomaly feedback feature tags Anomalous difference in the time dimension Product quality anomaly feedback feature tags The production process dimension shows abnormal differences; Based on the time-dimension anomaly differences and the production-stage anomaly differences, a local vector is constructed. Obtain local vectors model and the model Mapping to the abnormal local data matrix The matrix position corresponding to the g-th row and l-th column will then be the abnormal local data matrix. Transformed into product batches during the production process The quality characteristic matrix, denoted as .

5. A product quality monitoring method based on multi-source data analysis according to claim 4, characterized in that, The specific implementation process of step S4 includes: Based on the quality feature matrix, product batches during the production process Conduct quality profile scoring: Based on quality feature matrix Calculate the standard deviation of all moduli to reflect the overall dispersion of the quality characteristics. In the formula, Represents the quality characteristic matrix The mean of all modulo 1, and G represents the total number of trigger tags, and L represents the total number of production stages; Based on quality feature matrix Calculate the deviation of each modulus value from the standard modulus value, where the standard modulus value is the mean of the modulus values ​​of the quality characteristic matrix of historical qualified batches. The deviation of the modulus value in the g-th row and l-th column is calculated. In the formula, The standard modulus value; The quality profile score is obtained based on the overall dispersion and the modulus deviation. In the formula, The importance weights of the pre-set production units, The maximum allowable dispersion is preset. Preset quality profile scoring threshold If the quality profile score If the quality profile score is low, the staff will be prompted to adjust the operating parameters of the production unit. If no quality risk is found, the staff will be advised to maintain the operating parameters of the production unit.

6. A product quality monitoring system based on multi-source data analysis, executing the product quality monitoring method as described in any one of claims 1-5, characterized in that, The system includes: a data acquisition and synchronization module, a data statistics and labeling module, a matrix construction and feature module, and a quality assessment and feedback module; The data acquisition and synchronization module is used to configure the acquisition device and synchronize the time, and extract abnormal quality states to generate feature labels. The data statistics and labeling module is used to collect data on data sources and abnormal states, and to fill in feature labels; The matrix construction and feature module is used to construct an abnormal local data matrix and quantify abnormal differences to generate a quality feature matrix. The quality assessment and feedback module quantifies the quality profile score based on the quality feature matrix and prompts the user to operate the production parameters according to the quality profile score.

7. A product quality monitoring system based on multi-source data analysis according to claim 6, characterized in that: The data acquisition and synchronization module includes a device configuration unit and a time synchronization unit; The equipment configuration unit is used to configure multi-source data acquisition equipment for the production process, extract quality anomalies and generate feature labels. The time synchronization unit is used to perform time synchronization processing on the monitoring feedback behavior of multi-source data acquisition devices, triggering synchronization instructions at the beginning of the production process to update the device time synchronization status to ensure time synchronization in abnormal states.

8. A product quality monitoring system based on multi-source data analysis according to claim 6, characterized in that: The data statistics and tagging module includes a data source statistics unit and an abnormal status recording unit; The data source statistics unit is used to count the multi-source data acquisition devices configured in each production stage, forming a data source set; The abnormal state recording unit, based on the time synchronization sequence and trigger tags, counts the quality abnormal states in the statistical data source set, records them into feature tags, and improves the tag information.

9. A product quality monitoring system based on multi-source data analysis according to claim 6, characterized in that: The matrix construction and feature module includes a local matrix generation unit and a quality feature generation unit; The local matrix generation unit is used to capture the quality anomaly status of different product batches through the production device port and construct an anomaly local data matrix. The quality feature generation unit is used to calculate the abnormal differences in the time dimension and the production process dimension, centered on the elements of the abnormal local data matrix, to form a local vector and calculate its modulus, and then map the modulus to the corresponding position in the matrix to generate a quality feature matrix.

10. A product quality monitoring system based on multi-source data analysis according to claim 6, characterized in that: The quality assessment and feedback module includes a scoring calculation unit and a threshold judgment unit; The scoring calculation unit is used to calculate the overall dispersion and modulus deviation of quality features based on the quality feature matrix, and to obtain a quality profile score. The threshold judgment unit is used to preset the quality profile scoring threshold. If the score reaches or exceeds the threshold, the staff is prompted to adjust the operating parameters of the production unit; if the score does not reach the threshold, the staff is prompted to maintain the existing parameters.

Images