A method for monitoring the operating status of an automated production line of a laser gas sensor

By performing production node splitting and monitoring sensor layout analysis on the laser gas sensor automation production line, combined with correlation and weighted fusion analysis, the problem of insufficient early warning accuracy in the existing technology is solved, and more accurate abnormal state recognition and early warning optimization are achieved. .

CN118706792BActive Publication Date: 2024-11-05HUAXIA TIANXIN SENSOR TECH (DALIAN) CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411186205.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-28
Publication Date
2024-11-05
Estimated Expiration
2044-08-28

AI Technical Summary

Technical Problem

The existing laser gas sensor automated production line operating status monitoring methods lack correlation analysis, resulting in insufficient early warning accuracy and it is difficult to identify abnormal states caused by the combined action of multiple factors.

Method used

By splitting the production nodes of the laser gas sensor automation production line, conducting monitoring sensor layout analysis, conducting status monitoring and abnormal detection analysis, conducting correlation analysis and weighted fusion analysis, identifying the correlation between different monitoring points, and optimizing the early warning mechanism.

Benefits of technology

It improves the accuracy of identifying abnormal states, enhances the accuracy of early warning, and ensures the stable operation of the production line.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118706792B_ABST
    Figure CN118706792B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for monitoring the operation status of an automated production line of a laser gas sensor, and relates to the field related to intelligent manufacturing. The method comprises: splitting the production nodes, obtaining a set of sensor production nodes, marking the application attributes, and obtaining a set of application attribute information; performing a monitoring sensor layout analysis, determining a set of sensor layout parameters, monitoring the sensor layout, and obtaining a sensor monitoring module; performing state monitoring, obtaining a set of operation status monitoring data streams, performing anomaly detection analysis, and obtaining an abnormal operation feature set; performing a correlation analysis, obtaining a set of cascade abnormal features, performing a weighted fusion analysis, obtaining an abnormal operation status analysis result, and performing early warning optimization. The method solves the technical problems of the lack of correlation analysis and insufficient early warning accuracy in the existing monitoring of the operation status of the automated production line of laser gas sensors, and achieves the technical effect of more accurately identifying abnormal states and improving the accuracy of early warnings.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to fields related to intelligent manufacturing, and in particular to a method for monitoring the operating status of an automated production line of a laser gas sensor. Background Art

[0002] With the advancement of Industry 4.0 and intelligent manufacturing, automated production lines play an increasingly important role in the manufacturing industry. As a high-precision, high-sensitivity detection tool, laser gas sensors have been widely used in environmental monitoring, industrial process control and other fields. In the existing laser gas sensor automated production line operation status monitoring methods, they often only focus on the data of a single monitoring point and ignore the correlation between different monitoring points. This makes it difficult for the system to identify abnormal conditions caused by the combined effect of multiple factors, thus affecting the accuracy of the early warning.

[0003] Among the current related technologies, there are technical problems in the operation status monitoring of the automated production line of laser gas sensors, such as lack of correlation analysis and insufficient early warning accuracy. Summary of the invention

[0004] The present application provides a method for monitoring the operating status of an automated production line of laser gas sensors. It uses technical means such as splitting production nodes, analyzing the layout of monitoring sensors, performing status monitoring, performing anomaly detection analysis, performing correlation analysis, and performing weighted fusion analysis to comprehensively analyze the data from different monitoring points on the production line, identify the correlation between different factors, and achieve the technical effect of more accurately identifying abnormal conditions and improving the accuracy of early warnings.

[0005] The present application provides a method for monitoring the operating status of an automated production line of a laser gas sensor, comprising:

[0006] S1: Split the production nodes of the laser gas sensor automated production line to obtain a set of sensor production nodes, and apply attribute marking to the sensor production node set to obtain a set of production node application attribute information; S2: Perform monitoring sensor deployment analysis based on the production node application attribute information set to determine a set of production node sensor deployment parameters, and monitor sensor deployment on the sensor production node set through the production node sensor deployment parameter set to obtain a production node sensor monitoring module; S3: Perform status monitoring on the laser gas sensor automated production line through the production node sensor monitoring module to obtain a set of production node operation status monitoring data streams, and perform anomaly detection analysis on the production node operation status monitoring data streams to obtain a set of production node abnormal operation features; S4: Perform correlation analysis based on the production node abnormal operation feature set to obtain a set of production node cascade abnormal features, and perform weighted fusion analysis on the production node cascade abnormal feature set to obtain an abnormal operation status analysis result of the production line, and perform early warning optimization on the laser gas sensor automated production line based on the abnormal operation status analysis result of the production line.

[0007] In a possible implementation, the production node sensor deployment parameter set is determined in S2, and the following processing is performed:

[0008] S21: Perform monitoring function requirement analysis on the production node application attribute information set to obtain a production node monitoring function requirement parameter set, wherein the production node monitoring function requirement parameter set includes the monitoring function parameter type, monitoring function parameter accuracy and monitoring function parameter threshold of each production node; S22: Perform sensor type configuration on the sensor production node set based on the monitoring function parameter type to obtain a production node sensor type parameter set; S23: Perform sensor specification analysis on the production node sensor type parameter set according to the monitoring function parameter accuracy and the monitoring function parameter threshold to obtain a production node sensor specification parameter set; S24: Perform monitoring coverage analysis on the sensor production node set based on the production node application attribute information set to obtain a production node sensor installation location parameter set; S25: Determine the production node sensor deployment parameter set based on the production node sensor type parameter set, the production node sensor specification parameter set and the production node sensor installation location parameter set.

[0009] In a possible implementation, the abnormal operation feature set of the production node is obtained in S3, and the following processing is performed:

[0010] S31: Standardize the production node operation status monitoring data flow set to obtain a production node standard operation status monitoring data flow set; S32: Perform production standard quantification on the sensor production node set in turn according to the sensor production standard to obtain a production node data quantification standard set; S33: Perform anomaly identification on the production node standard operation status monitoring data flow set based on the production node data quantification standard set to obtain a production node abnormal operation status data flow set; S34: Perform abnormal feature analysis on the production node abnormal operation status data flow set to obtain the production node abnormal operation feature set.

[0011] In a possible implementation, the production node data quantization standard set is obtained in S32, and the following processing is performed:

[0012] The associated quality indicators of the sensor production node set are extracted respectively to obtain multiple production node associated quality indicator sets; dimensionality reduction screening is performed based on the multiple production node associated quality indicator sets to obtain multiple production node quality evaluation indicator sets; each quality evaluation indicator in the multiple production node quality evaluation indicator sets is quantified in turn according to the sensor production standards to obtain multiple production node indicator quantization standard sets; based on the multiple production node indicator quantization standard sets, the sensor monitoring data of the sensor production node set is analyzed for association to obtain the production node data quantization standard set.

[0013] In a possible implementation, the abnormal operation feature set of the production node is obtained in S34, and the following processing is performed:

[0014] Acquire a production line abnormal operation status database, the production line abnormal operation status database includes historical production node abnormal operation status data streams and corresponding abnormal detection feature data; parse the abnormal detection feature data to obtain abnormal detection type data features and abnormal detection degree data features; perform clustering and labeling training on the historical production node abnormal operation status data streams based on the abnormal detection type data features and the abnormal detection degree data features to generate a sensor production line abnormal feature recognition network; perform abnormal feature analysis on the production node abnormal operation status data stream set based on the sensor production line abnormal feature recognition network, and output the production node abnormal operation feature set.

[0015] In a possible implementation, the production node cascade abnormal feature set is obtained in S4, and the following processing is performed:

[0016] S41: Perform time series feature extraction on the abnormal operation feature set of the production node to obtain the abnormal trend feature set of the production node; S42: Perform correlation calculation based on the abnormal trend feature set of the production node to obtain the correlation coefficient information of the production node; S43: Perform hysteresis analysis on the abnormal operation feature set of the production node based on the correlation coefficient information of the production node to determine the hysteresis influence coefficient set of the production node; S44: Perform cascade analysis on the abnormal operation feature set of the production node through the hysteresis influence coefficient set of the production node to obtain the cascade abnormal feature set of the production node.

[0017] In a possible implementation, the production node hysteresis influence coefficient set is determined in S43, and the following processing is performed:

[0018] According to the production node correlation coefficient information, the sensor production node set is relevance screened to obtain the abnormal feature associated production node set; based on the production node abnormal operation feature set, the abnormal feature associated production node set is marked upstream and downstream to obtain the upstream and downstream information of the associated production nodes; according to the upstream and downstream information of the associated production nodes, the production node abnormal operation feature set is sequentially analyzed for hysteresis to obtain the node abnormal feature hysteresis influence factor set; the node abnormal feature hysteresis influence factor set is node-fused to determine the production node hysteresis influence coefficient set.

[0019] In a possible implementation, the abnormal operation status analysis result of the production line is obtained in S4, and the following processing is performed:

[0020] S45: Assign weights to the sensor production node set to obtain a production node weight factor set; S46: Based on the production node weight factor set, perform weighted fusion on the production node cascade abnormal feature set to obtain fused feature information of the abnormal operation status of the production line; S47: Analyze the generation cause of the fused feature information of the abnormal operation status of the production line to determine the generation cause of the abnormal operation status; S48: Based on the fused feature information of the abnormal operation status of the production line and the generation cause of the abnormal operation status, obtain the analysis result of the abnormal operation status of the production line.

[0021] A method for monitoring the operation status of an automated production line of a laser gas sensor proposed in the present application first splits the production nodes of the automated production line of the laser gas sensor to obtain a set of sensor production nodes, marks the sensor production node set with application attributes to obtain a set of application attribute information of the production nodes, then performs a monitoring sensor layout analysis based on the production node application attribute information set to determine a set of sensor layout parameters for the production nodes, monitors the sensor layout of the sensor production node set through the set of sensor layout parameters for the production nodes to obtain a production node sensor monitoring module, and then monitors the state of the automated production line of the laser gas sensor through the production node sensor monitoring module to obtain a set of production node operation status monitoring data streams, performs anomaly detection analysis on the production node operation status monitoring data streams respectively to obtain a set of abnormal operation features of the production nodes, and finally performs a correlation analysis based on the set of abnormal operation features of the production nodes to obtain a set of cascade abnormal features of the production nodes, performs a weighted fusion analysis on the set of cascade abnormal features of the production nodes to obtain an analysis result of the abnormal operation status of the production line, and performs early warning optimization on the automated production line of the laser gas sensor based on the analysis result of the abnormal operation status of the production line, thereby achieving a technical effect of more accurately identifying abnormal states and improving the accuracy of early warnings. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the accompanying drawings of the embodiment of the present invention will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the method according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0023] Figure 1 A schematic flow chart of a method for monitoring the operating status of an automated production line of laser gas sensors provided in an embodiment of the present application.

[0024] Figure 2 A schematic diagram of a process for determining a set of production node sensor layout parameters in a method for monitoring the operating status of a laser gas sensor automated production line provided in an embodiment of the present application. DETAILED DESCRIPTION

[0025] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0026] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0027] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict. The terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0028] The present application embodiment provides a method for monitoring the operating status of an automated production line of a laser gas sensor. Figure 1 As shown, the method includes:

[0029] S1: Split the production nodes of the laser gas sensor automated production line to obtain a set of sensor production nodes, and mark the sensor production node set with application attributes to obtain a set of production node application attribute information. Specifically, analyze the overall process flow of the laser gas sensor automated production line to identify key production links or steps, where key production links or steps are parts that directly affect product quality, production efficiency or cost, such as raw material processing, processing and manufacturing, assembly integration, etc. Define the identified key production links or steps as production nodes (equipment or equipment combination that completes specific process steps or tasks), and each production node has clear input, output and specific processing or handling functions. Sort out all the identified production nodes to form a complete set of sensor production nodes. According to the functions and characteristics of the production nodes, the design documents of the production line, the equipment manual, the historical operation data, etc., determine the application attribute categories that need to be marked. The application attributes are parameters or data that describe the characteristics or functions of the production nodes, which are used to identify and understand the operation status and requirements of the production nodes, including but not limited to: the functional description of the production nodes, the equipment type, the number of equipment, the equipment layout, the production capacity, the key performance indicators, etc. For each production node, the corresponding application attribute information is collected and organized, and the application attribute information of all production nodes is organized into a set, namely, the production node application attribute information set.

[0030] S2: Based on the production node application attribute information set, perform monitoring sensor deployment analysis, determine the production node sensor deployment parameter set, and deploy the monitoring sensor on the sensor production node set through the production node sensor deployment parameter set to obtain the production node sensor monitoring module. Specifically, based on the production node application attribute information set, evaluate the monitoring requirements of each production node, including determining the physical quantities to be monitored (such as temperature, humidity, pressure, vibration, etc.) and the accuracy requirements of monitoring. According to the monitoring requirements, determine the type of monitoring sensor. The monitoring sensor is a device used to monitor the operating status of the production node in real time and collect relevant data. For example, if the temperature needs to be monitored, select a temperature sensor; if the vibration needs to be monitored, select a vibration sensor, etc. According to the equipment layout, process flow, and monitoring requirements of the production node, determine the specific deployment location of the monitoring sensor on the production node. Determine a series of parameters required for the deployment of the monitoring sensor, such as sensor model, quantity, connection method, data transmission method, sampling frequency, etc. According to the sensor type, deployment location, and monitoring requirements, set the specific parameter value of each monitoring sensor, and organize all the deployment parameters of the monitoring sensors into a set, namely, the production node sensor deployment parameter set. According to the production node sensor deployment parameter set, purchase the required monitoring sensors, install and connect them according to the set parameters, debug and test the installed monitoring sensors to enable them to work normally and collect data accurately, and integrate all deployed monitoring sensors and related equipment (such as data acquisition cards, data transmission equipment, etc.) into a monitoring module to form a complete production node sensor monitoring module.

[0031] like Figure 2As shown, in a possible implementation, the determination of the production node sensor layout parameter set in S2 further includes S21: performing monitoring function requirement analysis on the production node application attribute information set to obtain the production node monitoring function requirement parameter set, the production node monitoring function requirement parameter set includes the monitoring function parameter type, monitoring function parameter accuracy and monitoring function parameter threshold of each production node. Specifically, the production node application attribute information set is analyzed to identify the main functions and key performance indicators of the production node, and the specific monitoring requirements of each production node are determined according to the functions and performance indicators, as well as reference industry standards or historical data, including the monitoring function parameter type (the type of physical quantity or performance indicator to be monitored, such as temperature, pressure, vibration, etc.), the monitoring function parameter accuracy (the accuracy of the monitoring result, expressed in error range or relative error, such as ±0.1℃, ±1%, etc.), and the monitoring function parameter threshold (the standard value or range for judging whether the monitoring result is normal, such as temperature exceeding 60℃ is abnormal). S22: Based on the monitoring function parameter type, the sensor production node set is configured with the sensor type to obtain the production node sensor type parameter set. Specifically, list all available sensor types and their characteristics, and select suitable sensor types for production nodes based on the monitoring function parameter type as well as factors such as cost and durability. For example, for temperature monitoring, select a temperature sensor; for vibration monitoring, select a vibration sensor.

[0032] S23: According to the monitoring function parameter accuracy and the monitoring function parameter threshold, the sensor specification analysis is performed on the production node sensor type parameter set to obtain the production node sensor specification parameter set. Specifically, according to the monitoring function parameter accuracy and the monitoring function parameter threshold, the specification analysis is performed on the selected sensor type, that is, by consulting the specification data sheet of the sensor type, the technical parameters and performance indicators of the monitoring sensor, such as measurement range, accuracy, response time, etc., are determined, and the accuracy and threshold parameters in the specification data sheet are compared with the requirements of the monitoring function parameter accuracy and threshold, so that the selected monitoring sensor can meet the accuracy and threshold requirements. S24: Based on the production node application attribute information set, the monitoring coverage analysis is performed on the sensor production node set to obtain the production node sensor installation position parameter set. Specifically, a layout diagram of the production node is drawn, and based on the application attribute information of the production node, the coverage of the monitoring sensor during installation is analyzed, that is, the monitoring area and range that the monitoring sensor can cover after installation, and the optimal installation position of the monitoring sensor is determined according to the analysis results to ensure the comprehensiveness and accuracy of the monitoring and avoid blind spots. S25: Determine the production node sensor deployment parameter set according to the production node sensor type parameter set, the production node sensor specification parameter set and the production node sensor installation location parameter set. Specifically, based on the results of the above steps, sort out the parameters such as sensor type, specification and installation location, check whether the parameters are complete and accurate, and meet the monitoring requirements, and form the final production node sensor deployment parameter set. This implementation method parses the application attribute information of the production node and performs monitoring coverage analysis to ensure that the monitoring requirements can be fully and accurately met when the monitoring sensors are deployed on the production line, thereby achieving the technical effect of improving the accuracy and reliability of the monitoring system.

[0033] S3: The laser gas sensor automated production line is monitored by the production node sensor monitoring module, and a set of production node operation status monitoring data streams is obtained. The production node operation status monitoring data streams are respectively subjected to anomaly detection and analysis to obtain a set of abnormal operation characteristics of the production node. Specifically, the laser gas sensor automated production line is continuously monitored by the production node sensor monitoring module. The monitoring sensors in the production node sensor monitoring module collect various operation data of the production node in real time, such as temperature, humidity, pressure, vibration, etc. The monitoring data of each production node are integrated into a data stream set to form a production node operation status monitoring data stream set, wherein a data stream refers to a data sequence generated continuously over a period of time, which is used to describe the dynamic changes of a process. The collected production node operation status monitoring data stream set is preprocessed, such as data cleaning, denoising, and standardization, to improve the accuracy of anomaly detection. According to the normal operation range and historical data of the production node, the thresholds of each monitoring parameter are set. The anomaly detection algorithm (such as a statistical method, a machine learning method, etc.) is used to perform anomaly detection on the preprocessed data stream. When the value of a parameter exceeds the set threshold, it is considered that the production node has an abnormality. For the detected anomalies, their key features are extracted, such as the time when the anomaly occurred, the value of the abnormal parameter, the duration of the anomaly, etc., to form a set of abnormal operation features of the production node.

[0034] In a possible implementation, the abnormal operation feature set of the production node is obtained in S3, further including S31: standardizing the production node operation status monitoring data stream set to obtain a standard operation status monitoring data stream set of the production node. Specifically, identify different monitoring parameter types (such as temperature, pressure, vibration, etc.) in the production node operation status monitoring data stream set, and standardize each monitoring parameter type separately, such as using the Z-score standardization method (i.e., subtracting the mean and dividing by the standard deviation). The dimensional differences between different monitoring sensors or different monitoring parameters are eliminated through standardization, so that all data can be compared and analyzed at the same scale, and the standardized data are integrated into a standard operation status monitoring data stream set of the production node. S32: Perform production standard quantization on the sensor production node set in turn according to the sensor production standard to obtain a standard set of production node data quantization. Specifically, the production standard document of the laser gas sensor (or target product) is consulted to determine the specific requirements of the product (laser gas sensor) for various performance indicators and quality parameters. Then, each production node in the sensor production node set is analyzed to clarify their respective roles and functions in the production process, as well as their relationship with the final performance and quality of the product (laser gas sensor). Based on the analysis of production standards and production nodes, quantitative indicators directly related to the performance and quality of the product (laser gas sensor) are determined, such as certain key production parameters, monitoring data or performance indicators. According to the production standards and quantitative indicators, corresponding data quantification standards are formulated for each production node. These data quantification standards can be specific numerical ranges, thresholds, or judgment criteria based on statistics. The data quantification standards of each production node are integrated together to form a complete set of production node data quantification standards. S33: Based on the production node data quantification standard set, the production node standard operation status monitoring data stream set is abnormally identified to obtain the production node abnormal operation status data stream set. Specifically, traverse each data stream in the production node standard operation status monitoring data stream set, compare the value of each data stream with the corresponding data quantization standard, if the value of the data stream exceeds the data quantization standard, mark the data stream as abnormal, and add it to the production node abnormal operation status data stream set. S34: Perform abnormal feature analysis on the production node abnormal operation status data stream set to obtain the production node abnormal operation feature set. Specifically, analyze each abnormal data stream in the production node abnormal operation status data stream set, extract the abnormal feature information, such as the specific time when the abnormality occurs, the corresponding production node location, the duration of the abnormality, etc., and integrate all abnormal features into the production node abnormal operation feature set. This implementation method ensures that the data of each monitoring sensor has a clear reference standard for abnormal identification by quantifying the standard based on the production standard, thereby achieving the technical effect of improving the accuracy of abnormal identification.

[0035] In a possible implementation, the production node data quantification standard set obtained in S32 further includes: extracting the associated quality indicators of the sensor production node set respectively to obtain multiple production node associated quality indicator sets. Specifically, the production process and flow of the production node corresponding to each monitoring sensor are analyzed, all potential quality indicators related to the production quality of each production node are identified and listed, each potential quality indicator is evaluated, its relevance and importance to the production quality are determined, and the quality indicators that are strongly associated with the production quality and important are screened out to form a set of associated quality indicators for each production node. Dimensionality reduction screening is performed based on the multiple production node associated quality indicator sets to obtain multiple production node quality evaluation indicator sets. Specifically, the associated quality indicator set of each production node is analyzed to find out the redundancy and correlation between them, and the associated quality indicators are screened and simplified by applying dimensionality reduction techniques (such as principal component analysis, factor analysis, etc.), and the core quality evaluation indicators of each production node are determined to form a set of quality evaluation indicators. According to the sensor production standard, each quality evaluation indicator in the multiple production node quality evaluation indicator sets is quantified in turn to obtain multiple production node indicator quantification standard sets. Specifically, the production standards of laser gas sensors are clarified, including specific requirements for product performance, quality, reliability, etc., and multiple production nodes in the production process of laser gas sensors are analyzed to determine the criticality and influence of each production node in the production process. For each production node, the quality evaluation index related to it is identified, and the quality evaluation index of each production node is standardized and quantified according to the production standards of the product (laser gas sensor), that is, specific quantitative standards are set for each quality evaluation index, such as numerical range, threshold, etc. The quantitative standards of the quality evaluation index of each production node are integrated together to form multiple production node index quantitative standard sets. Based on the multiple production node index quantitative standard sets, the sensor monitoring data of the sensor production node set is associated and analyzed to obtain the production node data quantitative standard set. Specifically, the monitoring data of each production node in the sensor production node set is collected, including real-time measurement values, status information, etc., and the collected monitoring data is cleaned to remove noise, outliers or duplicate data. The preprocessed monitoring data is associated with the corresponding production node index quantitative standard set. For each production node, its related quality evaluation index is matched with the corresponding monitoring data, and the monitoring data is analyzed to see whether it meets the corresponding index quantitative standard. Based on the results of association analysis, a data quantification standard set is constructed for each production node, including the normal operating range of the monitoring data, abnormal thresholds and other information. This implementation method establishes a clear data quantification standard set for the monitoring data of the production node through the steps of association quality indicator extraction, dimensionality reduction screening, standard quantification and association analysis, ensuring that the monitoring data is directly related to the actual production quality, and achieving the technical effect of improving the accuracy of abnormal identification.

[0036] In a possible implementation, the abnormal operation feature set of the production node is obtained in S34, further comprising: collecting and acquiring a production line abnormal operation status database, the production line abnormal operation status database including a historical production node abnormal operation status data stream and corresponding abnormal detection feature data. Specifically, historical data on the abnormal operation status of the production line is collected from the internal database, production management system or other related systems of the enterprise, including the abnormal operation status data stream of each production node, and these historical data are sorted to form a database (production line abnormal operation status database) containing historical production node abnormal operation status data streams and corresponding abnormal detection feature data. The abnormal detection feature data is parsed to obtain abnormal detection type data features and abnormal detection degree data features. Specifically, the abnormal detection feature data in the production line abnormal operation status database is read, and the abnormal detection type data features and abnormal detection degree data features are extracted and parsed, wherein the abnormal detection type data features are data features that describe the cause or category of the abnormality, such as mechanical failure, electrical failure, operation error, etc.; the abnormal detection degree data features are data features that describe the severity of the abnormality, such as slight, medium, severe, etc. Based on the abnormal detection type data features and the abnormal detection degree data features, the historical production node abnormal operation status data stream is clustered and labeled for training, and a sensor production line abnormal feature recognition network is generated. Specifically, a clustering algorithm, such as K-means, hierarchical clustering, etc., is selected or designed, and the abnormal detection type data features and the abnormal detection degree data features are used as feature vectors to cluster the historical production node abnormal operation status data stream, and a label is generated for each cluster, and these labels represent different types of abnormal features. Based on a machine learning or deep learning model, a sensor production line abnormal feature recognition network is trained using the clustering results. The sensor production line abnormal feature recognition network is a model or network that can recognize abnormal features of the sensor production line. Based on the sensor production line abnormal feature recognition network, the abnormal feature analysis of the production node abnormal operation status data stream set is performed, and the production node abnormal operation feature set is output. Specifically, the production node abnormal operation status data stream set is input into the sensor production line abnormal feature recognition network, and the sensor production line abnormal feature recognition network analyzes these data, identifies the abnormal features therein, and outputs the production node abnormal operation feature set, which describes the abnormal type and degree of the production node. This implementation method achieves the technical effect of improving the efficiency of anomaly detection and the accuracy of anomaly identification by training a sensor production line abnormal feature recognition network.

[0037] S4: Based on the abnormal operation feature set of the production node, a correlation analysis is performed to obtain a cascade abnormal feature set of the production node, and a weighted fusion analysis is performed on the cascade abnormal feature set of the production node to obtain the abnormal operation status analysis result of the production line, and the laser gas sensor automatic production line is early-warning optimized according to the abnormal operation status analysis result of the production line. Specifically, each feature in the abnormal operation feature set of the production node is identified and understood, and the type of abnormality and the scope of influence it represents are determined. By checking the time sequence of the occurrence of abnormal features, the correlation of abnormal parameters, the physical connection relationship between devices, etc., it is analyzed whether there is a correlation or causal relationship between the abnormal operation features of different production nodes. If it is found that the abnormality of a production node is directly or indirectly caused by the abnormality of other production nodes, or there is a mutual influence between the abnormalities of multiple production nodes, these abnormalities are regarded as cascade abnormalities, and the cascade abnormal feature set of production nodes is extracted. According to the importance of each production node in the production line, the severity and impact range of the abnormal feature, a weight value is set for each abnormal feature, and the abnormal features with similarity or correlation are fused to form a comprehensive abnormal feature. In the fusion process, the weight value of each feature is analyzed and weighted fusion is performed. Through weighted fusion analysis, a comprehensive analysis result reflecting the overall abnormal operation state of the production line is obtained, that is, the analysis result of the abnormal operation state of the production line, including the type, severity, impact range and possible causes of the abnormality. According to the analysis results of the abnormal operation state of the production line, a corresponding early warning mechanism is established. When the production line is monitored to be abnormal, an early warning signal is issued in time so that the operator or automatic control system can take measures quickly. According to the analysis results of the abnormal operation state of the production line, targeted optimization suggestions are generated, including adjusting production parameters, replacing faulty equipment, optimizing process flow, etc., to reduce the abnormal occurrence rate of the production line or improve the efficiency of abnormal handling. The optimization suggestions are converted into specific operation plans for implementation. During the implementation process, the operation state of the production line is continuously monitored, and adjustments and improvements are made according to actual conditions. The embodiment of the present application adopts technical means such as splitting production nodes, performing monitoring sensor layout analysis, performing status monitoring, performing anomaly detection analysis, performing correlation analysis, and performing weighted fusion analysis to comprehensively analyze the data from different monitoring points on the production line, identify the correlation between different factors, and achieve the technical effect of more accurately identifying abnormal conditions and improving the accuracy of early warnings.

[0038] In a possible implementation, the production node cascade abnormal feature set obtained in S4 further includes S41: performing time series feature extraction on the production node abnormal operation feature set to obtain the production node abnormal trend feature set. Specifically, the production node abnormal operation feature set is sorted so that each abnormal feature has a corresponding timestamp, and the time series features (features that describe the characteristics of data that change over time) such as the frequency, duration, and periodicity of the occurrence of abnormalities are determined according to business needs and data characteristics. A time series analysis is performed on the abnormal operation features of each production node to extract abnormal trend features (features that change over time extracted from abnormal operation data), such as abnormal growth trends, periodic fluctuations, etc., and the extracted abnormal trend features are sorted into a production node abnormal trend feature set. S42: Correlation calculation is performed based on the production node abnormal trend feature set to obtain production node correlation coefficient information. Specifically, the abnormal trend feature set of the production node is used as input data, and a suitable correlation measurement method is selected according to the data characteristics, such as the Pearson correlation coefficient, the Spearman rank correlation coefficient, etc., and the abnormal trend features of any two production nodes are correlated to obtain a correlation coefficient matrix, and the correlation coefficient matrix is ​​analyzed to identify the production node pairs with potential correlation. S43: Based on the production node correlation coefficient information, a hysteresis analysis is performed on the abnormal operation feature set of the production node to determine the hysteresis influence coefficient set of the production node. Specifically, a suitable hysteresis analysis method is selected according to the data characteristics and business requirements, such as the Granger causality test, the vector autoregression model, etc., and the hysteresis order to be investigated is determined according to the requirements or experience of the analysis method, and a hysteresis analysis is performed on the production node pairs with potential correlation to determine whether there is a hysteresis in the impact of the abnormality of one production node on the abnormality of another production node. If there is a hysteresis effect, the hysteresis influence coefficient (a coefficient that quantifies the degree of the time lag effect of the abnormality of one production node on the abnormality of another production node) is calculated to form a set of hysteresis influence coefficients of the production node. S44: Perform cascade analysis on the abnormal operation feature set of the production nodes through the hysteresis influence coefficient set of the production nodes to obtain the cascade abnormal feature set of the production nodes. Specifically, based on the hysteresis influence coefficient set of the production nodes, determine the abnormal influence relationship and degree between the production nodes, and construct a cascade analysis model between the production nodes. The cascade analysis model describes how the abnormality of a production node is transmitted to other production nodes through the hysteresis influence and causes the occurrence of cascade abnormalities. In the cascade analysis model, according to the hysteresis influence coefficient of the production node, an additive analysis is performed on the production node that causes the abnormal impact. By weighting the hysteresis influence coefficient or directly applying the hysteresis influence coefficient to the calculation of the abnormal feature, the influence of the production nodes that have an important impact on the abnormality in the cascade abnormal feature set is quantified and improved.On the basis of additive analysis, according to the cascade analysis model, the cascade abnormality characteristics of each production node are calculated. These cascade abnormality characteristics not only take into account the abnormal characteristics of a single production node, but also take into account the abnormal characteristics of other related production nodes transmitted through hysteresis effects. The calculated cascade abnormality characteristics of each production node are integrated into a set of production node cascade abnormality characteristics. The set of production node cascade abnormality characteristics reflects the abnormal correlation and cascade influence degree between the production nodes on the entire production line. This implementation method accurately describes the abnormal correlation and cascade influence between the production nodes on the production line through hysteresis analysis and cascade analysis, achieving the technical effect of providing more accurate and comprehensive support for abnormal warning and fault prediction.

[0039] In a possible implementation, the step of determining the set of hysteresis influence coefficients of the production nodes in S43 further includes: performing correlation screening on the set of sensor production nodes according to the correlation coefficient information of the production nodes, and obtaining a set of production nodes associated with abnormal features. Specifically, according to business requirements and data characteristics, a correlation threshold (a threshold for determining whether there is a significant correlation between two production nodes) is set, and the calculated correlation coefficient is compared with the preset correlation threshold, and the production nodes whose correlation coefficients exceed the correlation threshold are screened out to form a set of production nodes associated with abnormal features. Based on the abnormal operation feature set of the production nodes, the set of production nodes associated with abnormal features is marked with upstream and downstream information to obtain upstream and downstream information of the associated production nodes. Specifically, according to the actual process of the production line, the upstream and downstream relationships between the production nodes are determined, wherein the upstream node is a node that provides raw materials or intermediate products to the downstream node, and the downstream node is a node that relies on the output of the upstream node for production. In the set of production nodes associated with abnormal features, the upstream and downstream nodes of each production node are marked according to the production process by adding additional fields to the data structure. According to the upstream and downstream information of the associated production nodes, the abnormal operation feature set of the production nodes is subjected to hysteresis analysis in turn to obtain a set of hysteresis impact factors of the node abnormal features. Specifically, according to the characteristics of the data and business needs, a suitable hysteresis analysis method is selected, such as Granger causality test, vector autoregression model, etc., and the hysteresis analysis method is used to analyze the impact of the abnormal operation characteristics of the upstream nodes on the abnormalities of the downstream nodes in turn according to the upstream and downstream information of the associated production nodes. For each abnormal feature of the upstream node, the degree of its impact on the abnormality of the downstream node is calculated, that is, the hysteresis impact factor. All the calculated hysteresis impact factors are sorted into a set, that is, the set of hysteresis impact factors of node abnormal features. The set of hysteresis impact factors of node abnormal features contains information on the degree to which each production node is affected by the abnormalities of other production nodes. The set of hysteresis impact factors of node abnormal features is subjected to node fusion to determine the set of hysteresis impact coefficients of the production nodes. Specifically, according to business needs and data characteristics, select appropriate fusion methods, such as weighted summation, weighted average, etc., fuse the hysteresis impact factors of all abnormal features of each production node, obtain the hysteresis impact coefficient of the production node, and organize the hysteresis impact coefficients of all production nodes into a set of production node hysteresis impact coefficients. This implementation method focuses on those production nodes that are truly correlated through correlation screening, avoids the analysis of irrelevant production nodes, and obtains a comprehensive hysteresis impact coefficient through node fusion, achieving the technical effect of improving the efficiency and accuracy of hysteresis analysis.

[0040] In a possible implementation, the step of obtaining the analysis result of the abnormal operation state of the production line in S4 further includes S45: weighting the sensor production node set to obtain a set of production node weight factors. Specifically, according to the actual situation of the production line, the importance of each production node in the production process is analyzed through factors such as the production volume, failure frequency, and the degree of influence on downstream nodes of the node. Based on the analysis results, a weight allocation standard is set, for example, a higher weight is allocated to production nodes with large production volume, high failure frequency, and large downstream influence. According to the set weight allocation standard, a weight factor is allocated to each production node, and this weight factor represents the importance of the production node in the production line. S46: Based on the set of production node weight factors, the set of cascade abnormal features of the production nodes is weighted and fused to obtain the fused feature information of the abnormal operation state of the production line. Specifically, the cascade abnormal feature information of each production node is multiplied by the corresponding weight factor to implement weighted processing, and all weighted abnormal feature information is fused (summed, weighted averaged or other fusion algorithms, etc.) to obtain the fused feature information of the abnormal operation state of the production line. S47: Analyze the generation reasons of the fused characteristic information of the abnormal operation state of the production line to determine the generation reasons of the abnormal operation state. Specifically, analyze the obtained fused characteristic information of the abnormal operation state of the production line, identify the abnormal mode reflected in the fused characteristic information of the abnormal operation state of the production line according to historical data and expert knowledge, and determine the generation reasons of the abnormal operation state of the production line based on the analysis of the abnormal mode, including equipment failure, raw material problems, operation errors and other aspects. S48: Obtain the analysis results of the abnormal operation state of the production line according to the fused characteristic information of the abnormal operation state of the production line and the generation reasons of the abnormal operation state. Specifically, integrate the fused characteristic information of the abnormal operation state of the production line and the generation reasons of the abnormal operation state to obtain the analysis results of the abnormal operation state of the production line. The analysis results of the abnormal operation state of the production line can be presented in the form of a report, including a description of the abnormal operation state, fused characteristic information, generation reasons and possible solutions or suggestions. This implementation method determines the root cause of the abnormal operation of the production line by analyzing the generation reasons of the fused characteristic information, provides strong support for subsequent troubleshooting, optimization and preventive measures, and achieves the technical effect of enhancing the depth and breadth of abnormal analysis.

[0041] The above specific implementation manner does not constitute a limitation to the protection scope of the present application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application. In some cases, the actions or steps recorded in the present application can be performed in an order different from that in the embodiment and can still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A method for monitoring the operating status of an automated production line of a laser gas sensor, characterized in that: The method comprises: S1: splitting the production nodes of the laser gas sensor automated production line to obtain a sensor production node set, marking the sensor production node set with application attributes to obtain a production node application attribute information set; S2: Performing monitoring sensor deployment analysis based on the production node application attribute information set, determining a production node sensor deployment parameter set, and deploying monitoring sensors on the sensor production node set through the production node sensor deployment parameter set to obtain a production node sensor monitoring module; S3: monitoring the status of the laser gas sensor automated production line through the production node sensor monitoring module, obtaining a set of production node operation status monitoring data streams, performing anomaly detection and analysis on the production node operation status monitoring data streams respectively, and obtaining a set of production node abnormal operation characteristics; S4: performing correlation analysis based on the abnormal operation feature set of the production node to obtain a cascade abnormal feature set of the production node, performing weighted fusion analysis on the cascade abnormal feature set of the production node to obtain an abnormal operation status analysis result of the production line, and performing early warning optimization on the laser gas sensor automation production line according to the abnormal operation status analysis result of the production line; The production node cascade abnormal feature set obtained in S4 includes: S41: extracting time series features from the abnormal operation feature set of the production node to obtain an abnormal trend feature set of the production node; S42: performing correlation calculation based on the abnormal trend feature set of the production nodes to obtain production node correlation coefficient information; S43: performing hysteresis analysis on the abnormal operation feature set of the production nodes based on the production node correlation coefficient information to determine a hysteresis influence coefficient set of the production nodes; S44: performing cascade analysis on the abnormal operation feature set of the production nodes through the hysteresis influence coefficient set of the production nodes to obtain the cascade abnormal feature set of the production nodes; Determining the set of production node hysteresis influence coefficients in S43 includes: Perform correlation screening on the sensor production node set according to the production node correlation coefficient information to obtain an abnormal feature associated production node set; Based on the abnormal operation feature set of the production node, the set of production nodes associated with the abnormal feature is marked as upstream and downstream, so as to obtain upstream and downstream information of the associated production nodes; According to the upstream and downstream information of the associated production nodes, a hysteresis analysis is performed on the abnormal operation feature set of the production nodes in turn to obtain a set of hysteresis influencing factors of the node abnormality feature; The node abnormal characteristic hysteresis impact factor set is node-fused to determine the production node hysteresis impact coefficient set.

2. A method for monitoring the operating status of an automated production line of a laser gas sensor as claimed in claim 1, characterized in that: The step S2 determines a set of production node sensor deployment parameters, including: S21: performing monitoring function requirement analysis on the production node application attribute information set to obtain a production node monitoring function requirement parameter set, wherein the production node monitoring function requirement parameter set includes a monitoring function parameter type, a monitoring function parameter accuracy, and a monitoring function parameter threshold of each production node; S22: performing sensor type configuration on the sensor production node set based on the monitoring function parameter type to obtain a production node sensor type parameter set; S23: performing sensor specification analysis on the production node sensor type parameter set according to the monitoring function parameter accuracy and the monitoring function parameter threshold to obtain a production node sensor specification parameter set; S24: performing monitoring coverage analysis on the sensor production node set based on the production node application attribute information set to obtain a production node sensor installation location parameter set; S25: Determine the production node sensor layout parameter set according to the production node sensor type parameter set, the production node sensor specification parameter set and the production node sensor installation location parameter set.

3. A method for monitoring the operating status of an automated production line of a laser gas sensor as claimed in claim 1, characterized in that: The abnormal operation feature set of the production node obtained in S3 includes: S31: performing standardization processing on the production node operation status monitoring data stream set to obtain a production node standard operation status monitoring data stream set; S32: performing production standard quantization on the sensor production node set in sequence according to the sensor production standard to obtain a production node data quantization standard set; S33: performing abnormal identification on the production node standard operation status monitoring data flow set based on the production node data quantification standard set to obtain a production node abnormal operation status data flow set; S34: Perform abnormal feature analysis on the abnormal operation status data stream set of the production node to obtain the abnormal operation feature set of the production node.

4. A method for monitoring the operating status of an automated production line of a laser gas sensor as claimed in claim 3, characterized in that: The production node data quantification standard set obtained in S32 includes: Extracting associated quality indicators from the sensor production node sets respectively to obtain multiple production node associated quality indicator sets; Perform dimensionality reduction screening based on the multiple production node associated quality indicator sets to obtain multiple production node quality evaluation indicator sets; According to the sensor production standard, each quality evaluation indicator in the plurality of production node quality evaluation indicator sets is quantified in turn to obtain a plurality of production node indicator quantification standard sets; Based on the multiple production node indicator quantification standard sets, correlation analysis is performed on the sensor monitoring data of the sensor production node set to obtain the production node data quantification standard set.

5. A method for monitoring the operating status of an automated production line of a laser gas sensor as claimed in claim 3, characterized in that: The abnormal operation feature set of the production node is obtained in S34, including: Acquire a production line abnormal operation status database, wherein the production line abnormal operation status database includes a historical production node abnormal operation status data stream and corresponding abnormal detection feature data; Parsing the anomaly detection feature data to obtain anomaly detection type data features and anomaly detection degree data features; Based on the anomaly detection type data features and the anomaly detection degree data features, clustering and labeling training is performed on the abnormal operation status data stream of the historical production node to generate an abnormal feature recognition network for the sensor production line; Based on the sensor production line abnormal feature recognition network, an abnormal feature analysis is performed on the abnormal operation status data stream set of the production node, and the abnormal operation feature set of the production node is output.

6. A method for monitoring the operating status of an automated production line of a laser gas sensor as claimed in claim 1, characterized in that: The abnormal operation status analysis result of the production line obtained in S4 includes: S45: performing weight allocation on the sensor production node set to obtain a production node weight factor set; S46: performing weighted fusion on the production node cascade abnormal feature set based on the production node weight factor set to obtain fused feature information of abnormal operation state of the production line; S47: Analyze the generation cause of the abnormal operation state fusion feature information of the production line to determine the generation cause of the abnormal operation state; S48: Obtaining an analysis result of the abnormal operation state of the production line according to the fusion feature information of the abnormal operation state of the production line and the cause of the abnormal operation state.