Abnormal state monitoring method and system of chip mounter driving system

Abstract

The invention relates to the technical field of chip mounters, in particular to an abnormal state monitoring method and system for a chip mounter driving system, and the method comprises the following steps: carrying out the real-time state monitoring of the chip mounter driving system, so as to obtain the state monitoring data of the driving system; performing time sequence analysis on the state monitoring data of the driving system to generate a time sequence state monitoring curve of the driving system; performing dynamic behavior analysis on the time sequence state monitoring curve of the driving system, and constructing a dynamic behavior graph of the driving system; performing abnormal state detection on the time sequence state monitoring data of the driving system based on the dynamic behavior graph of the driving system to generate abnormal state marking data of the driving system; performing clustering analysis on the driving system abnormal state mark data to generate driving system abnormal state clustering analysis data; and performing abnormal state path tracing on the driving system abnormal state clustering analysis data to generate a driving system abnormal state area. According to the invention, an efficient and accurate abnormal state monitoring method is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of chip placement machines, and in particular to a method and system for monitoring the abnormal state of a chip placement machine drive system. Background Art

[0002] In SMT drive systems, monitoring abnormal conditions is crucial to ensuring stable equipment operation and product quality. Due to long-term operation, environmental changes, load fluctuations, or component aging, SMT drive systems may experience abnormal conditions, leading to performance degradation, unstable assembly accuracy, and even equipment failure. Traditional abnormal condition monitoring methods typically rely on manual inspections and testing, which are inefficient and inaccurate. Therefore, an intelligent, automated abnormal condition monitoring method and system are needed. Summary of the Invention

[0003] In order to solve the above technical problems, the present invention proposes a system and a method to solve at least one of the above technical problems.

[0004] To achieve the above object, the present invention provides a method comprising the following steps:

[0005] Step S1: performing real-time status monitoring on the placement machine drive system to obtain drive system status monitoring data; performing time series analysis on the drive system status monitoring data to generate a drive system time series status monitoring curve;

[0006] Step S2: Performing dynamic behavior analysis on the drive system timing state monitoring curve to construct a drive system dynamic behavior diagram; performing abnormal state detection on the drive system timing state monitoring data based on the drive system dynamic behavior diagram to generate drive system abnormal state marking data;

[0007] Step S3: performing cluster analysis on the drive system abnormal state marking data to generate drive system abnormal state cluster analysis data; performing abnormal state path tracing on the drive system abnormal state cluster analysis data to generate a drive system abnormal state area;

[0008] Step S4: performing abnormal state trend analysis on the abnormal state area of the drive system to generate abnormal state trend prediction data of the drive system; performing data visualization on the abnormal state trend prediction data of the drive system to generate a visual view of the abnormal state trend of the drive system;

[0009] Step S5: Optimizing monitoring decisions for the drive system abnormal state marking data based on the abnormal state trend visualization view of the drive system, and generating an abnormal state monitoring optimization decision;

[0010] Step S6: Use the recurrent convolutional network to perform dilated convolution on the abnormal state monitoring optimization decision to build a drive system monitoring decision model to perform abnormal state monitoring operations.

[0011] By real-time monitoring of the drive system status, the present invention can obtain the system's working status and parameter information in a timely manner so as to detect abnormal situations in a timely manner. By obtaining status monitoring data, a historical database of the system status can be established to provide a data basis for subsequent analysis and decision-making. By performing time series analysis on the status monitoring data, a time series status monitoring curve of the drive system can be generated to show the dynamic change trend of the system status. Based on the dynamic behavior graph, the drive system time series status monitoring data can be used for abnormal state detection. By identifying patterns and trends that are inconsistent with normal behavior, possible abnormal states can be marked in a timely manner. The dynamic behavior graph can help analysts gain an in-depth understanding of the behavioral characteristics and laws of the drive system, more accurately determine whether an abnormal situation has occurred, and improve the accuracy of abnormal state detection. By clustering analysis, similar types of abnormal states can be classified into the same category, helping analysts to clarify the classification and laws of abnormal states. By tracing the path of the abnormal state, the cause and evolution process of the abnormal state can be deeply understood, providing guidance for subsequent fault diagnosis and problem solving. Analyzing the trends in abnormal status areas can help us understand the development trends and possible evolution directions of abnormal status, and help us take further preventive measures and make decisions. The predictive data generated based on the trend analysis results can provide early warnings of potential abnormal status and take corresponding preventive measures, which helps to avoid equipment failures and production interruptions. The trends and changes of abnormal status can be intuitively displayed through visual views, helping operators to quickly judge the health status of the system and make immediate decisions to ensure the stable operation of the system. Monitoring and decision optimization of abnormal status data based on visual views can improve the accuracy and efficiency of decisions, reduce interference from human factors, and improve the stability and work efficiency of the production line. By building a monitoring decision model, the system can automatically judge abnormal status and make corresponding decisions, reduce dependence on manual operations, and improve the consistency and accuracy of decisions. The decision model based on the recurrent convolutional network can introduce a reinforcement learning mechanism through dilated convolution, continuously optimize the decision-making process, enable the system to have self-learning and adaptability, and improve the effect and performance of abnormal status monitoring.

[0012] Preferably, step S1 includes the following steps:

[0013] Step S11: performing real-time status monitoring on the placement machine drive system to obtain drive system status monitoring data, where the drive system status monitoring data includes drive system operating time data, drive system operating status data, drive system energy consumption data, and drive system fault record data;

[0014] Step S12: performing time series analysis on the drive system state monitoring data to generate drive system time series state monitoring data;

[0015] Step S13: performing a time series evolution analysis on the driving system time series state monitoring data to generate driving system time series state trajectory data;

[0016] Step S14: performing timing curve fitting on the driving system timing state monitoring data to generate a driving system timing state monitoring curve.

[0017] By monitoring and recording the operating time of the drive system, the present invention can understand information such as the equipment's usage, operating cycle, and equipment lifespan. This helps formulate reasonable maintenance plans and promptly carry out equipment repair or replacement to avoid production losses caused by equipment failure and downtime. Monitoring and recording the operating status of the drive system can provide real-time information on the equipment's operating status, including normal operation, standby mode, and failure. By analyzing the operating status data, equipment anomalies or failures can be promptly detected and appropriate measures can be taken to repair and ensure normal operation. By monitoring and recording the energy consumption data of the drive system, the energy efficiency performance of the equipment can be evaluated, the equipment's energy consumption can be understood, and excessive or abnormal energy consumption can be identified. This helps formulate reasonable energy management strategies, improve energy efficiency, and reduce production costs. Recording and analyzing drive system failures can track the frequency, type, and cause of equipment failures. This helps identify potential equipment problems and implement preventive maintenance measures to avoid failures and improve equipment reliability and stability. By performing time series analysis on the monitoring data, changing trends, periodic patterns, and the occurrence of abnormal events in the system status can be revealed. It can help us better understand the operating characteristics of the system, judge the stability and performance of the system, and make corresponding adjustments and improvement measures. By performing evolution analysis on time series data, we can show the changing trajectory of the system state, discover the laws and trends of the system state evolution, understand the dynamic behavior of the system, guide equipment operation optimization, and improve the reliability and performance stability of the system. By performing curve fitting on time series data, we can extract the patterns and trends of the data and display the changes in the system state in the form of curves. We can quickly and intuitively analyze and diagnose the system state, discover system anomalies or potential problems, and take corresponding measures to handle and optimize them.

[0018] Preferably, step S2 includes the following steps:

[0019] Step S21: segmenting the drive system timing state monitoring curve to obtain a drive system timing state monitoring segment set;

[0020] Step S22: extracting features from the drive system timing state monitoring segment set to generate drive system timing state feature data;

[0021] Step S23: performing dynamic behavior analysis on the driving system time series state characteristic data to generate driving system dynamic behavior characteristic data;

[0022] Step S24: performing dynamic behavior evolution processing on the dynamic behavior characteristic data of the drive system to construct a dynamic behavior graph of the drive system;

[0023] Step S25: performing abnormal state detection on the driving system timing state monitoring data to obtain abnormal state data of the driving system;

[0024] Step S26: marking the abnormal state of the driving system dynamic behavior diagram based on the abnormal state data of the driving system to generate abnormal state marking data of the driving system.

[0025] By segmenting the curve into segments, the present invention enables more detailed observation and analysis of the state changes of the drive system under different time periods or events. This provides a deeper understanding of the system's operating process and performance characteristics. If a system anomaly or failure occurs, segmenting the curve can help determine the specific time period or event where the problem occurred, allowing for faster problem location and resolution. Segmenting the curve into segments allows for more efficient management of large amounts of monitoring data. Each segment can be processed, stored, or analyzed independently without causing information confusion or loss. By extracting features, complex time series data can be simplified into more representative feature data, reducing data dimensionality and redundancy for subsequent analysis and processing. The extracted features can describe certain aspects of the system, such as vibration frequency, temperature changes, and current waveforms. These features help understand the system's operating characteristics and abnormal behavior. By extracting features, common patterns or significant changes in the system, such as periodic vibration and high-frequency noise, can be identified. This is very useful for detecting abnormal behavior and predicting failures. By analyzing the changing trends and patterns of feature data, a more accurate description of the system's dynamic behavior can be achieved. This helps understand the system's response speed, stability, and adaptability. Dynamic behavior analysis can help detect abnormal behavior or failures in the system. By monitoring abnormal changes in characteristic data, timely detection and action can be taken to prevent further deterioration of faults. By analyzing the system's dynamic behavior characteristic data, potential improvement points and optimization strategies can be identified. This helps improve system efficiency, reliability, and performance. Dynamic behavior diagrams graphically display the evolution of system behavior, making complex data easier to understand and analyze. Visualization can reveal system trends, periodic behavior, instability, and other phenomena. By observing the evolution of dynamic behavior, patterns or trends within the system can be identified. This helps understand the system's development patterns and predict its future behavior. Dynamic behavior diagrams can be used to detect abnormal behavior or emergencies in the system. By comparing the current behavior with the normal evolution pattern, abnormalities can be promptly detected and addressed. Detecting abnormal conditions can accurately locate and diagnose faults or abnormal conditions in the drive system. This helps provide accurate fault information and enable appropriate maintenance or repair measures. Abnormal condition detection can detect abnormal conditions in the system in advance and provide timely warnings. This helps prevent the escalation of faults and further losses. Abnormal condition data provides a detailed record of system behavior, which can be used for subsequent data analysis, modeling, and optimization. This helps improve system design, maintenance strategies, and performance optimization. By marking abnormal states on the dynamic behavior graph, the location and time of system problems can be intuitively displayed. This helps quickly identify and understand the occurrence and impact of abnormal behavior. The marked abnormal states provide detailed information about the abnormal system behavior, which can be used for further analysis and research on the cause of the abnormality.This helps improve system design and detect the root cause of faults. Abnormal status tag data can be used to generate abnormal reports or notifications, promptly reporting abnormal conditions to relevant personnel. This helps to take emergency measures, optimize maintenance plans, and make decision adjustments.

[0026] Preferably, step S3 includes the following steps:

[0027] Step S31: performing interval discretization processing on the drive system abnormal state mark data to generate drive system abnormal state quantified data;

[0028] Step S32: Perform cluster analysis on the quantified data of abnormal state of the drive system to generate cluster analysis data of abnormal state of the drive system

[0029] Step S33: performing abnormal trajectory analysis on the drive system abnormal state cluster analysis data according to the drive system time series state trajectory data to obtain drive system abnormal state trajectory data;

[0030] Step S34: tracing the abnormal state path of the drive system abnormal state trajectory data to generate the drive system abnormal state path data;

[0031] Step S35: Calculate the abnormal area weight of the drive system abnormal state path data using the drive system abnormal state weight path calculation formula to generate the drive system abnormal state area.

[0032] The present invention converts the original continuous abnormal state value into a discrete data representation through discretization processing, which is more convenient for subsequent data analysis and processing. Discretization processing can suppress the details and noise in the original data, extract the key feature information of the abnormal state, and at the same time reduce the dimension of the data, simplifying the complexity of subsequent analysis. Cluster analysis can help identify and discover similar abnormal state patterns, group and classify abnormal states according to the similarity of the data, so as to better understand and explain the changing laws of abnormal states. Through the results of cluster analysis, the abnormal state quantitative data can be divided into different clusters, which provides a classification basis for subsequent abnormal state analysis and processing, so that corresponding measures can be taken more targetedly. Through abnormal trajectory analysis, the abnormal state can be visualized in time series, so as to more intuitively observe and understand the abnormal state evolution of the drive system at different time points. By observing the abnormal trajectory, we can identify the trends and patterns of abnormal states, such as periodic changes, continuous rise or fall, etc., to provide a basis for subsequent abnormal state prediction and monitoring. By tracing the abnormal state path, we can determine the specific evolution path and process of the abnormal state from the starting point to the end point, which helps to understand the formation mechanism and influencing factors of the abnormal state. The abnormal state path data can reveal the correlation and interaction between different states, help to discover the causal relationship between abnormal states, and further provide a deeper understanding of the abnormal state. Through the abnormal area weight calculation, the abnormal state path can be quantitatively evaluated to determine the degree and importance of the abnormal state, which helps to judge the degree of abnormality. The abnormal area weight calculation can weight the abnormal state path, highlight the important abnormal state areas, provide more precise and accurate abnormal state analysis results, and provide a basis for subsequent decision-making.

[0033] Preferably, the calculation formula of the drive system abnormal state weight path in step S35 is specifically:

[0034]

[0035] Where W is the weight value of the abnormal state path of the driving system, n is the number of nodes in the abnormal state path of the driving system, i is the i-th abnormal state path node of the driving system, d i is the weight of the i-th drive system abnormal state path node in the abnormal state path, j is the abnormal state value of the j-th drive system abnormal state path node, x j is the original state value of the driving system state path, is the original state value of the driving system state path, d nis the abnormal state value of the last drive system abnormal state path node in the path, d1 is the abnormal state value of the first drive system abnormal state path node in the path, f is the number of abnormal states of the drive system, t is the operating time of the drive system, and T is the existence time of the abnormal state of the drive system.

[0036] The present invention The relationship between the weights of nodes in the drive system's abnormal state path and their abnormal state values is summed. By calculating the logarithmic ratio and standard deviation between the node weights and abnormal state values, the importance of the abnormal state path nodes and the degree of dispersion of the abnormal state values can be assessed. A larger logarithmic ratio and a smaller standard deviation indicate that the node contributes significantly to the abnormal state path, making the weight calculation helpful in identifying important nodes in the abnormal state path. Indicates the ratio of the difference between the abnormal state values of the last node and the first node in the path to the number of nodes. By calculating the ratio of the difference between the abnormal state values and the number of nodes, the degree of change in the abnormal state path can be evaluated. A larger difference indicates a larger abnormal state change in the path, so the calculation of the weight is helpful in capturing the abnormal state change of the path. This formula represents the impact of the number of abnormal state events, run time, and abnormal state duration on the weight. By considering the relationship between the number of abnormal state events, run time, and abnormal state duration, the weight calculation can be adjusted to better reflect the actual situation. For example, if the number of abnormal states is high, the run time is long, and the abnormal state duration is short, the weight will increase accordingly, indicating that the path has a greater impact on the abnormal state. The formula calculates the weight of the abnormal state path of the drive system by summing and comparing the weights and abnormal state values of the abnormal state path nodes, and considering the relationship between the number of abnormal states, run time, and abnormal state duration. This calculation is helpful for identifying important nodes in the abnormal state path, capturing abnormal state changes along the path, and adjusting the weight to reflect the impact of the number of abnormal states, run time, and abnormal state duration. This information is important for drive system status analysis and abnormal state management.

[0037] Preferably, step S4 includes the following steps:

[0038] Step S41: performing abnormal state trend analysis on the abnormal state area of the driving system to generate abnormal state trend analysis data of the driving system;

[0039] Step S42: performing trend prediction calculation on the drive system abnormal state trend analysis data using a drive system abnormal state trend prediction calculation formula to generate drive system abnormal state trend prediction data;

[0040] Step S43: Use a deep learning algorithm to visualize the drive system abnormal state trend prediction data to generate a drive system abnormal state trend visualization view.

[0041] By analyzing data from abnormal status areas, the present invention can reveal the changing trends of abnormal conditions, such as whether they are gradually worsening or stabilizing. This helps determine the direction and speed of abnormal conditions. Abnormal status trend analysis can help identify persistent or increasing abnormal conditions, distinguishing temporary abnormalities from potential failure risks. By applying abnormal status trend prediction formulas, the future development trend of drive system abnormal conditions can be predicted. This helps take appropriate measures to prevent potential failures or abnormal conditions in advance. Abnormal status trend prediction data can be used to generate early warning signals or alerts. When abnormal status trends indicate potential problems, relevant personnel can take timely action to avoid production interruptions or equipment damage. By visualizing abnormal status trend prediction data in forms such as charts, curves, or heat maps, the changing trends of abnormal conditions can be intuitively displayed, making them easier to understand and analyze. The abnormal status trend visualization can be updated and displayed in real time, providing real-time monitoring and observation, allowing relevant personnel to quickly understand the latest abnormal status of the drive system. Based on the abnormal status trend visualization, decision makers can more accurately assess the development trend of abnormal conditions and make appropriate decisions and adjustments to optimize monitoring strategies and maintenance plans.

[0042] Preferably, the calculation formula for predicting the abnormal state trend of the driving system in step S42 is specifically:

[0043]

[0044] Where P is the predicted value of the abnormal state trend of the drive system, t1 is the starting time of the abnormal state of the drive system, t2 is the time of the predicted abnormal state trend of the drive system, T is the existence time of the abnormal state of the drive system, f is the number of abnormal states of the drive system, C is the load value of the drive system, ∈ is the trend prediction adjustment factor, d i is the weight of the abnormal state path node of the i-th drive system in the abnormal state path, f is the number of abnormal states of the drive system, and b is the coefficient of the influence of the frequency of abnormal state on trend prediction.

[0045] In this embodiment, by The rate of change of the abnormal state is derived, that is, the trend of the abnormal state. By calculating the rate of change of the abnormal state, the growth rate or reduction rate of the abnormal state can be obtained, thereby predicting the future trend of the abnormal state. It represents the product of the drive system load value and the trend forecast adjustment factor. The drive system load value reflects the current load of the system, and the trend forecast adjustment factor is used to correct the trend forecast. By considering the system load and the adjustment factor, the abnormal state trend can be corrected to make the forecast result more accurate and reliable. This formula represents the impact of the number of abnormal state occurrences and abnormal state path node weights on trend prediction. f represents the number of abnormal state occurrences, which is the number of abnormal state occurrences at the current time. While the value of f may vary at different times, the number of abnormal state occurrences is a constant. For example, if the number of abnormal state occurrences measured at 3:00 PM is 3, the value of f calculated for this calculation is 3. If the number of abnormal state occurrences measured at 4:00 PM is 5, the value of f calculated for this calculation is 5. The number of abnormal state occurrences reflects the frequency of abnormal states, while the abnormal state path node weights represent the importance of different paths. By considering the number of abnormal state occurrences and path node weights, the prediction results can be adjusted to a trend that better reflects actual conditions. This formula provides a prediction of the abnormal state trend of the drive system by taking the derivative and integral of the abnormal state trend and considering the influence of factors such as load, adjustment factor, number of abnormal states, and path node weights. Such predictions are helpful in determining the future direction, speed, and cumulative change of abnormal states, providing important reference for system operation and maintenance and decision-making, and facilitating the implementation of appropriate measures to address and manage abnormal states.

[0046] Preferably, step S5 includes the following steps:

[0047] Step S51: performing knowledge graph reasoning on the drive system abnormal state marked data based on the drive system abnormal state trend visualization view to generate a drive system abnormal state trend knowledge graph;

[0048] Step S52: performing abnormal state optimization analysis on the drive system abnormal state trend knowledge graph to generate abnormal state trend optimization analysis data;

[0049] Step S53: Optimizing the monitoring decision on the abnormal state trend optimization analysis data to construct an abnormal state monitoring optimization decision;

[0050] This invention combines abnormal state data of the drive system with visualizations and, through the reasoning and representation capabilities of a knowledge graph, integrates scattered data into a structured knowledge network. This allows for a more comprehensive description and understanding of relevant information about abnormal states. Knowledge graph reasoning can uncover associations and dependencies between abnormal states, revealing potential causal relationships and influencing factors within the drive system. This helps provide a deeper understanding of the root causes and potential influencing mechanisms of abnormal states. By analyzing the abnormal state trend knowledge graph, optimization objectives for drive system abnormal states can be determined, such as minimizing the duration of abnormal states or minimizing the number of abnormal state occurrences. This helps clarify optimization directions and goals. Based on the abnormal state trend optimization analysis data, corresponding optimization strategies and measures can be formulated to improve the abnormal state trends of the drive system. These can include optimizing maintenance plans, optimizing operating strategies, and upgrading equipment. Based on the abnormal state trend optimization analysis data, abnormal state monitoring strategies can be adjusted and optimized in real time. This helps improve the flexibility and responsiveness of the monitoring system, making monitoring decisions more accurate and effective. By optimizing monitoring decisions, abnormal states can be detected and addressed promptly, preventing potential failures and equipment damage. Optimized monitoring decisions can include setting thresholds, formulating warning rules, adjusting alarm levels, and other measures.

[0051] Preferably, step S6 includes the following steps:

[0052] Step S61: using a recurrent convolutional network to perform convolution preprocessing on the abnormal state monitoring optimization decision to generate a convolution sample set of the abnormal state of the drive system;

[0053] Step S62: performing convolution data segmentation on the drive system abnormal state convolution sample set to generate a drive system abnormal state convolution sequence;

[0054] Step S63: performing dilated convolution on the drive system abnormal state convolution sequence using a dilated convolution algorithm to generate a drive system abnormal state convolution network;

[0055] Step S64: performing spatial pyramid pooling multi-layer sampling on the drive system abnormal state convolution network to generate a drive system abnormal state convolution feature map;

[0056] Step S65: Perform data mining modeling on the convolution feature map of the abnormal state of the driving system to build a driving system monitoring decision model to perform abnormal state monitoring operations.

[0057] The present invention uses convolutional preprocessing in a recurrent convolutional network to extract representative features from raw abnormal state data. These features capture the temporal and spatial information of abnormal states, thereby better describing the characteristics and changes of abnormal states. Convolutional preprocessing can perform data augmentation on abnormal state data, such as translation, scaling, and rotation, to generate more diverse samples. This helps improve the generalization and robustness of the model and better adapt to different abnormal state scenarios. By dividing the abnormal state sample set into convolutional sequences, the abnormal state data can be organized according to specific time windows, forming a continuous sequence representation. This helps capture the temporal evolution of abnormal states and related dynamic characteristics. The generation of convolutional sequences preserves the contextual information of the abnormal state data of the drive system, enabling the model to better understand the relationship between the current state and previous states. This helps improve the accuracy of abnormal state identification and prediction. Using the dilated convolution algorithm, the convolutional network can expand the receptive field without increasing the number of parameters, obtaining a wider range of contextual information. This facilitates the extraction and representation of local features of abnormal state in the drive system. Dilated convolution can reduce the number of parameters in the convolutional network, improving the parameter efficiency and computational efficiency of the model. This is particularly important for large-scale abnormal state datasets and real-time monitoring tasks. Through multi-layer sampling using spatial pyramid pooling, multi-scale and multi-level feature representations can be extracted from the drive system abnormal state convolutional network. This helps comprehensively capture the spatial information and fine-grained features of abnormal states. Spatial pyramid pooling makes the feature map invariant to scale changes, enabling better adaptation to abnormal states at different scales. This helps improve the model's generalization and adaptability. Through data mining modeling, abnormal state patterns and regularities can be learned from the drive system abnormal state convolutional feature map. Based on these learned patterns, effective monitoring decisions can be made, enabling rapid and accurate judgment and response to abnormal states. The drive system monitoring decision model processes abnormal state convolutional feature maps in real time, enabling real-time abnormal state monitoring. The model can make predictions and decisions based on the current feature map, promptly identifying and addressing abnormal states, and improving the system's fault detection and response capabilities. By establishing a monitoring decision model, automated decision-making for abnormal states can be achieved. Based on the learned knowledge and regularities, the model automatically determines the severity and impact of the abnormal state and takes appropriate measures to address it, reducing manual intervention and improving decision consistency and efficiency.

[0058] In this specification, a system for monitoring abnormal conditions of a placement machine drive system is provided, comprising:

[0059] The timing analysis module monitors the status of the placement machine drive system in real time to obtain the drive system status monitoring data; performs timing analysis on the drive system status monitoring data to generate the drive system timing status monitoring curve;

[0060] The abnormal state detection module performs dynamic behavior analysis on the drive system timing state monitoring curve and constructs the drive system dynamic behavior diagram; based on the drive system dynamic behavior diagram, it performs abnormal state detection on the drive system timing state monitoring data and generates drive system abnormal state marking data;

[0061] The path tracing module performs cluster analysis on the drive system abnormal state marking data to generate drive system abnormal state cluster analysis data; and performs abnormal state path tracing on the drive system abnormal state analysis data to generate the drive system abnormal state area;

[0062] The trend analysis module performs abnormal state trend analysis on the abnormal state area of the drive system to generate abnormal state trend prediction data of the drive system; and performs data visualization on the abnormal state trend prediction data of the drive system to generate a visual view of the abnormal state trend of the drive system;

[0063] The decision optimization module optimizes the monitoring decision of the drive system abnormal state data based on the visualization view of the drive system abnormal state trend and generates the abnormal state monitoring optimization decision;

[0064] The convolutional model module uses a recurrent convolutional network to perform dilated convolution on the abnormal state monitoring optimization decision to build a drive system monitoring decision model to perform abnormal state monitoring operations.

[0065] The present invention constructs an abnormal state monitoring system for a SMT drive system. Using a timing analysis module, the system can monitor the drive system's state in real time and acquire drive system state monitoring data. This enables the system to promptly capture and record changes in the drive system's state, enabling real-time monitoring of the system. The abnormal state detection module performs dynamic behavior analysis on the drive system's timing state monitoring curve and constructs a dynamic behavior graph of the drive system. By analyzing the dynamic behavior graph, the system can detect abnormal states in the drive system's timing state monitoring data and generate abnormal state marker data. This facilitates rapid and accurate identification of abnormal states in the drive system. By analyzing cluster analysis data, the system can trace abnormal state paths and identify the specific paths and factors that lead to abnormal states. This facilitates a deeper understanding of the mechanisms that cause abnormal states and provides guidance for problem resolution. By analyzing trend prediction data, the system can predict the development trends of abnormal states and take proactive intervention and remediation measures. This helps avoid potential failures and losses and improves system reliability and stability. The trend analysis module visualizes the abnormal state trend prediction data to generate a visual view of the abnormal state trend. This intuitively displays abnormal state trends and changes, helping users quickly understand and identify the development of abnormal conditions. The decision optimization module optimizes drive system abnormal state monitoring decisions based on abnormal state trend visualization. By analyzing abnormal state data and formulating optimized decisions, the system improves the speed and accuracy of abnormal state response, reducing failure risks and production costs. The convolutional model module utilizes a recurrent convolutional network to perform dilated convolution on abnormal state monitoring optimization decisions, constructing a drive system monitoring decision model. This model automatically learns the characteristics and patterns of abnormal states, enabling automated identification and prediction of abnormal conditions and improving the system's automated monitoring capabilities. The abnormal state monitoring system for the placement machine drive system, through the synergistic effect of modules including real-time monitoring, abnormal state detection, path tracing, trend analysis, data visualization, and decision optimization, provides accurate and reliable drive system status monitoring and abnormal state identification, helping users promptly identify, analyze, and resolve problems, thereby improving production efficiency and quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 A schematic flow chart of a method and system for monitoring abnormal conditions of a placement machine drive system according to the present invention;

[0067] Figure 2 Detailed implementation flow chart of step S1;

[0068] Figure 3 Detailed implementation flow chart of step S2;

[0069] Figure 4Schematic diagram of the detailed implementation steps of step S3. DETAILED DESCRIPTION

[0070] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0071] This application provides a method and system for monitoring the abnormal state of a placement machine drive system. The execution entities of the method and system include, but are not limited to, the following: mechanical equipment, data processing platforms, cloud server nodes, network upload devices, etc. equipped with the system, which can be regarded as general computing nodes of this application. The data processing platform includes, but is not limited to, at least one of an audio and image management system, an information management system, and a cloud data management system.

[0072] See also Figures 1 to 4 The present invention provides a method for monitoring abnormal conditions of a placement machine drive system, the method comprising the following steps:

[0073] Step S1: performing real-time status monitoring on the placement machine drive system to obtain drive system status monitoring data; performing time series analysis on the drive system status monitoring data to generate a drive system time series status monitoring curve;

[0074] Step S2: Performing dynamic behavior analysis on the drive system timing state monitoring curve to construct a drive system dynamic behavior diagram; performing abnormal state detection on the drive system timing state monitoring data based on the drive system dynamic behavior diagram to generate drive system abnormal state marking data;

[0075] Step S3: performing cluster analysis on the drive system abnormal state marking data to generate drive system abnormal state cluster analysis data; performing abnormal state path tracing on the drive system abnormal state cluster analysis data to generate a drive system abnormal state area;

[0076] Step S4: performing abnormal state trend analysis on the abnormal state area of the drive system to generate abnormal state trend prediction data of the drive system; performing data visualization on the abnormal state trend prediction data of the drive system to generate a visual view of the abnormal state trend of the drive system;

[0077] Step S5: Optimizing monitoring decisions for the drive system abnormal state marking data based on the abnormal state trend visualization view of the drive system, and generating an abnormal state monitoring optimization decision;

[0078] Step S6: Use the recurrent convolutional network to perform dilated convolution on the abnormal state monitoring optimization decision to build a drive system monitoring decision model to perform abnormal state monitoring operations.

[0079] By real-time monitoring of the drive system status, the present invention can obtain the system's working status and parameter information in a timely manner so as to detect abnormal situations in a timely manner. By obtaining status monitoring data, a historical database of the system status can be established to provide a data basis for subsequent analysis and decision-making. By performing time series analysis on the status monitoring data, a time series status monitoring curve of the drive system can be generated to show the dynamic change trend of the system status. Based on the dynamic behavior graph, the drive system time series status monitoring data can be used for abnormal state detection. By identifying patterns and trends that are inconsistent with normal behavior, possible abnormal states can be marked in a timely manner. The dynamic behavior graph can help analysts gain an in-depth understanding of the behavioral characteristics and laws of the drive system, more accurately determine whether an abnormal situation has occurred, and improve the accuracy of abnormal state detection. By clustering analysis, similar types of abnormal states can be classified into the same category, helping analysts to clarify the classification and laws of abnormal states. By tracing the path of the abnormal state, the cause and evolution process of the abnormal state can be deeply understood, providing guidance for subsequent fault diagnosis and problem solving. Analyzing the trends in abnormal status areas can help us understand the development trends and possible evolution directions of abnormal status, and help us take further preventive measures and make decisions. The predictive data generated based on the trend analysis results can provide early warnings of potential abnormal status and take corresponding preventive measures, which helps to avoid equipment failures and production interruptions. The trends and changes of abnormal status can be intuitively displayed through visual views, helping operators to quickly judge the health status of the system and make immediate decisions to ensure the stable operation of the system. Monitoring and decision optimization of abnormal status data based on visual views can improve the accuracy and efficiency of decisions, reduce interference from human factors, and improve the stability and work efficiency of the production line. By building a monitoring decision model, the system can automatically judge abnormal status and make corresponding decisions, reduce dependence on manual operations, and improve the consistency and accuracy of decisions. The decision model based on the recurrent convolutional network can introduce a reinforcement learning mechanism through dilated convolution, continuously optimize the decision-making process, enable the system to have self-learning and adaptability, and improve the effect and performance of abnormal status monitoring.

[0080] In the embodiment of the present invention, reference Figure 1 The above is a schematic flow chart of a step of the present invention. In this example, the steps of the abnormal state monitoring method of the electric placement machine drive system include:

[0081] Step S1: performing real-time status monitoring on the placement machine drive system to obtain drive system status monitoring data; performing time series analysis on the drive system status monitoring data to generate a drive system time series status monitoring curve;

[0082] In this embodiment, crawler technology is used to perform real-time status monitoring of the placement machine drive system to obtain drive system status monitoring data and store it in a database or data storage system. The frequency of data acquisition is ensured to be high enough to ensure accurate monitoring of the drive system status. An appropriate time series analysis algorithm is selected to analyze the drive system status monitoring data. Common time series analysis algorithms include moving average, exponential smoothing, and autoregressive models. An appropriate algorithm is selected based on actual conditions and the corresponding parameters are configured. The preprocessed data is analyzed using the selected time series analysis algorithm. Based on the algorithm's requirements, data smoothing, trend decomposition, and periodicity analysis are performed to generate a time series status monitoring curve for the drive system. The results of the time series analysis are converted into curve data and visualized. A charting tool or data visualization software can be used to display the generated time series status monitoring curve in an intuitive manner, making it easier for users to understand and analyze. The generated curve data is updated in real time to the monitoring interface or monitoring system, ensuring that users can monitor changes in the drive system's status in real time. This allows users to promptly detect anomalies or changes and take appropriate measures.

[0083] Step S2: Performing dynamic behavior analysis on the drive system timing state monitoring curve to construct a drive system dynamic behavior diagram; performing abnormal state detection on the drive system timing state monitoring data based on the drive system dynamic behavior diagram to generate drive system abnormal state marking data;

[0084] In this embodiment, the time series state monitoring curve data is preprocessed, including operations such as removing outliers, filling missing values, and smoothing data to ensure data quality and reliability. The overall trend of the curve is observed to determine whether there is a clear upward or downward trend in different time periods. The curve is then tested to see whether it exhibits periodic changes. Periodic analysis methods such as autocorrelation functions and Fourier transforms can be used to find periodic patterns in the curve, calculate the rate of change between adjacent data points, and observe the rate of change of the curve. Sudden and drastic changes may indicate the occurrence of an abnormal state. Based on the analysis results, the state monitoring data of the drive system is mapped into a dynamic behavior graph. The dynamic behavior graph reflects the behavior patterns and change patterns of the system under different states. Based on the characteristics of the dynamic behavior graph and the range of normal operating conditions, a threshold for abnormality detection is set. The time series state monitoring data is compared with the dynamic behavior graph to detect whether there are abnormal conditions that exceed the set threshold. Detected abnormal conditions are marked, which can be done by marking abnormal points on the curve or generating abnormal state marking data to indicate the abnormal time point or interval.

[0085] Step S3: performing cluster analysis on the drive system abnormal state marking data to generate drive system abnormal state cluster analysis data; performing abnormal state path tracing on the drive system abnormal state cluster analysis data to generate a drive system abnormal state area;

[0086] In this embodiment, appropriate features, such as abnormality duration and amplitude, are extracted from the abnormal state marker data. Statistical methods or domain expertise can be used to select features. The extracted features are then normalized so that they are all on the same scale to prevent certain features from overly influencing the clustering results. An appropriate clustering algorithm, such as K-means clustering, hierarchical clustering, or DBSCAN, is then selected for analysis. The selection of a clustering algorithm should be based on the characteristics of the data and the analysis objectives. The standardized abnormal state marker data is clustered using the selected clustering algorithm, grouping similar abnormal state markers into the same cluster. Based on the requirements of the abnormal state analysis, an appropriate path tracing algorithm, such as a depth-first search algorithm or a recursive algorithm, is selected. The path tracing algorithm should be based on the data structure and the analysis objectives. The selected path tracing algorithm is used to identify path relationships between abnormal states in the drive system abnormal state analysis data. For example, the path from the starting abnormal state to the target abnormal state can be obtained by traversing the abnormal state connectivity relationships in the data. Based on the abnormal state path tracing results, similar paths are grouped into the same abnormal state region. Region division can be determined using a clustering algorithm or by setting a threshold.

[0087] Step S4: performing abnormal state trend analysis on the abnormal state area of the drive system to generate abnormal state trend prediction data of the drive system; performing data visualization on the abnormal state trend prediction data of the drive system to generate a visual view of the abnormal state trend of the drive system;

[0088] In this embodiment, statistical analysis is performed on the abnormal state marker data within each abnormal state region. Indicators such as the frequency, duration, and magnitude of abnormalities within each abnormal state region can be calculated, and trend relationships between abnormal states can be explored. Based on the characteristics of the abnormal state and the analysis objectives, an appropriate trend analysis algorithm, such as linear regression, exponential smoothing, or ARIMA, is selected. The choice of trend analysis algorithm depends on the nature of the data and the purpose of the analysis. The selected trend prediction algorithm is used to predict the abnormal state marker data, generating forecasted abnormal state trends for a period of time. The forecast time range can be set as needed, and the layout and style of the visualization view can be designed based on the abnormal state trend prediction data. Line charts, bar charts, area charts, and other formats can be selected to display abnormal state trend changes. Using the selected data visualization tool, the abnormal state trend prediction data can be converted into a visualization based on the designed layout and style. Elements such as labels, legends, and axis scales can be added to enhance the visual effect and information communication.

[0089] Step S5: Optimizing monitoring decisions for the drive system abnormal state marking data based on the abnormal state trend visualization view of the drive system, and generating an abnormal state monitoring optimization decision;

[0090] In this embodiment, the objectives of monitoring decisions are determined, such as reducing the duration of abnormal states, reducing the impact of abnormal states, etc., and warning lines and alarm lines for abnormal states are defined to judge the severity and triggering conditions of abnormal states. The trend changes and periodicity of abnormal states are observed, the trends and characteristics of abnormal states are analyzed, the thresholds and changing patterns of abnormal states are determined, and based on the trends and characteristics of abnormal states, monitoring rules are formulated to identify abnormal states and trigger corresponding alarms. The triggering conditions, judgment rules, and response measures for abnormal states are determined. Existing monitoring decisions are evaluated and analyzed. Based on actual usage and feedback information, possible defects, false alarms, or missed alarms are discovered. The accuracy and efficiency of monitoring decisions are improved by optimizing monitoring rules, adjusting abnormal state thresholds, and improving decision-making processes. The optimized monitoring decisions are recorded in the form of documents or databases, describing the monitoring rules, triggering conditions, and corresponding response measures. This ensures that the optimized decisions for abnormal state monitoring are easy to understand and implement, and can provide accurate guidance for subsequent implementation and maintenance.

[0091] Step S6: Use the recurrent convolutional network to perform dilated convolution on the abnormal state monitoring optimization decision to build a drive system monitoring decision model to perform abnormal state monitoring operations.

[0092] In this embodiment, the collected abnormal state monitoring optimization decision data set is prepared, including input features and corresponding target labels, to ensure the quality and integrity of the data set, perform data cleaning, remove outliers and other preprocessing steps, design the structure of the cyclic convolutional network, including the number of network layers, convolution kernel size, pooling method, etc., according to the characteristics of the data set and the requirements of the problem, select a suitable cyclic convolutional network model, such as LSTM (long short-term memory network) or GRU (gated recurrent unit), etc., divide the data set into training set, validation set and test set, generally use 70% of the data as training set, 10% as validation set, and 20% as test set to ensure that the divided data set can fully represent the feature distribution and sample distribution of the entire data set, use the training set to train the cyclic convolutional network, and update the network through the back propagation algorithm. weight parameters, set appropriate loss functions (such as mean square error, cross entropy, etc.) and optimizers (such as Adam, SGD, etc.) to optimize the performance of the network model, use the validation set to evaluate the trained recurrent convolutional network model, calculate evaluation indicators such as accuracy, precision, recall rate, F1 value, etc., tune and improve the network model based on the evaluation results, such as adjusting network parameters, increasing the number of network layers, etc., use the test set to test the trained and evaluated recurrent convolutional network model, evaluate its performance on unseen samples, analyze the model's performance on the test set, judge the model's generalization ability and accuracy, deploy the trained and tested recurrent convolutional network model to the actual drive system monitoring operation, input the abnormal state monitoring optimization decision into the model according to actual needs, obtain the prediction results for abnormal state monitoring operations.

[0093] In this embodiment, reference Figure 2 The above is a flowchart of the detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include:

[0094] Step S11: performing real-time status monitoring on the placement machine drive system to obtain drive system status monitoring data, where the drive system status monitoring data includes drive system operating time data, drive system operating status data, drive system energy consumption data, and drive system fault record data;

[0095] Step S12: performing time series analysis on the drive system state monitoring data to generate drive system time series state monitoring data;

[0096] Step S13: performing a time series evolution analysis on the driving system time series state monitoring data to generate driving system time series state trajectory data;

[0097] Step S14: performing timing curve fitting on the driving system timing state monitoring data to generate a driving system timing state monitoring curve.

[0098] By monitoring and recording the operating time of the drive system, the present invention can understand information such as the equipment's usage, operating cycle, and equipment lifespan. This helps formulate reasonable maintenance plans and promptly carry out equipment repair or replacement to avoid production losses caused by equipment failure and downtime. Monitoring and recording the operating status of the drive system can provide real-time information on the equipment's operating status, including normal operation, standby mode, and failure. By analyzing the operating status data, equipment anomalies or failures can be promptly detected and appropriate measures can be taken to repair and ensure normal operation. By monitoring and recording the energy consumption data of the drive system, the energy efficiency performance of the equipment can be evaluated, the equipment's energy consumption can be understood, and excessive or abnormal energy consumption can be identified. This helps formulate reasonable energy management strategies, improve energy efficiency, and reduce production costs. Recording and analyzing drive system failures can track the frequency, type, and cause of equipment failures. This helps identify potential equipment problems and implement preventive maintenance measures to avoid failures and improve equipment reliability and stability. By performing time series analysis on the monitoring data, changing trends, periodic patterns, and the occurrence of abnormal events in the system status can be revealed. It can help us better understand the operating characteristics of the system, judge the stability and performance of the system, and make corresponding adjustments and improvement measures. By performing evolution analysis on time series data, we can show the changing trajectory of the system state, discover the laws and trends of the system state evolution, understand the dynamic behavior of the system, guide equipment operation optimization, and improve the reliability and performance stability of the system. By performing curve fitting on time series data, we can extract the patterns and trends of the data and display the changes in the system state in the form of curves. We can quickly and intuitively analyze and diagnose the system state, discover system anomalies or potential problems, and take corresponding measures to handle and optimize them.

[0099] In this embodiment, key points or indicators that need to be monitored are identified, such as the system's operating status, the performance indicators of each module, and the database connection status. Monitoring code is inserted at these key points or monitoring tools are used to collect real-time system data. This data can include system runtime, CPU utilization, memory usage, network request response time, interface call count, etc. The stored monitoring data is analyzed using data analysis tools, visualization tools, or custom scripts. Analysis can include trend analysis, anomaly detection, and correlation analysis of real-time monitoring indicators. An appropriate time series analysis algorithm is selected as needed to analyze the drive system status monitoring data. Common time series analysis algorithms include moving average, exponential smoothing, and autoregressive models. An appropriate algorithm is selected based on the actual situation, and the corresponding parameters are configured. The selected time series analysis algorithm is used to analyze the preprocessed data. Based on the algorithm's requirements, data smoothing, trend decomposition, and periodicity analysis are performed to generate the drive system's time series status monitoring data. An appropriate time series evolution analysis algorithm is selected to analyze the drive system's time series status monitoring data. Common algorithms include cluster analysis and state transition models. Select an appropriate algorithm based on your specific situation and configure the appropriate parameters. Use the selected algorithm to analyze the drive system's time-series state monitoring data and identify the evolving relationships and transition patterns between states. By analyzing information such as state duration and transition frequency, the drive system's time-series state trajectory data is generated. The selected algorithm is then used to perform fitting analysis on the drive system's time-series state monitoring data. Curve fitting allows you to better capture state trends and patterns, generating time-series state monitoring curves for the drive system. These curves can be used to visualize changes in the drive system's state and help identify abnormal conditions or trends.

[0100] In this embodiment, reference Figure 3 The above is a flowchart of the detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include:

[0101] Step S21: segmenting the drive system timing state monitoring curve to obtain a drive system timing state monitoring segment set;

[0102] Step S22: extracting features from the drive system timing state monitoring segment set to generate drive system timing state feature data;

[0103] Step S23: performing dynamic behavior analysis on the driving system time series state characteristic data to generate driving system dynamic behavior characteristic data;

[0104] Step S24: performing dynamic behavior evolution processing on the dynamic behavior characteristic data of the drive system to construct a dynamic behavior graph of the drive system;

[0105] Step S25: performing abnormal state detection on the driving system timing state monitoring data to obtain abnormal state data of the driving system;

[0106] Step S26: marking the abnormal state of the driving system dynamic behavior diagram based on the abnormal state data of the driving system to generate abnormal state marking data of the driving system.

[0107] By segmenting the curve into segments, the present invention enables more detailed observation and analysis of the state changes of the drive system under different time periods or events. This provides a deeper understanding of the system's operating process and performance characteristics. If a system anomaly or failure occurs, segmenting the curve can help determine the specific time period or event where the problem occurred, allowing for faster problem location and resolution. Segmenting the curve into segments allows for more efficient management of large amounts of monitoring data. Each segment can be processed, stored, or analyzed independently without causing information confusion or loss. By extracting features, complex time series data can be simplified into more representative feature data, reducing data dimensionality and redundancy for subsequent analysis and processing. The extracted features can describe certain aspects of the system, such as vibration frequency, temperature changes, and current waveforms. These features help understand the system's operating characteristics and abnormal behavior. By extracting features, common patterns or significant changes in the system, such as periodic vibration and high-frequency noise, can be identified. This is very useful for detecting abnormal behavior and predicting failures. By analyzing the changing trends and patterns of feature data, a more accurate description of the system's dynamic behavior can be achieved. This helps understand the system's response speed, stability, and adaptability. Dynamic behavior analysis can help detect abnormal behavior or failures in the system. By monitoring abnormal changes in characteristic data, timely detection and action can be taken to prevent further deterioration of faults. By analyzing the system's dynamic behavior characteristic data, potential improvement points and optimization strategies can be identified. This helps improve system efficiency, reliability, and performance. Dynamic behavior diagrams graphically display the evolution of system behavior, making complex data easier to understand and analyze. Visualization can reveal system trends, periodic behavior, instability, and other phenomena. By observing the evolution of dynamic behavior, patterns or trends within the system can be identified. This helps understand the system's development patterns and predict its future behavior. Dynamic behavior diagrams can be used to detect abnormal behavior or emergencies in the system. By comparing the current behavior with the normal evolution pattern, abnormalities can be promptly detected and addressed. Detecting abnormal conditions can accurately locate and diagnose faults or abnormal conditions in the drive system. This helps provide accurate fault information and enable appropriate maintenance or repair measures. Abnormal condition detection can detect abnormal conditions in the system in advance and provide timely warnings. This helps prevent the escalation of faults and further losses. Abnormal condition data provides a detailed record of system behavior, which can be used for subsequent data analysis, modeling, and optimization. This helps improve system design, maintenance strategies, and performance optimization. By marking abnormal states on the dynamic behavior graph, the location and time of system problems can be intuitively displayed. This helps quickly identify and understand the occurrence and impact of abnormal behavior. The marked abnormal states provide detailed information about the abnormal system behavior, which can be used for further analysis and research on the cause of the abnormality.This helps improve system design and detect the root cause of faults. Abnormal status tag data can be used to generate abnormal reports or notifications, promptly reporting abnormal conditions to relevant personnel. This helps to take emergency measures, optimize maintenance plans, and make decision adjustments.

[0108] In this embodiment, conditions and parameters for curve segmentation are defined, such as thresholds and window sizes. These parameters are used to determine when to segment the curve into different segments. A curve segmentation algorithm is then applied to segment the drive system timing state monitoring curve according to the set parameters. The curve segmentation algorithm can be based on methods such as thresholds, mutation point detection, and local extreme point detection. After segmentation, multiple drive system timing state monitoring segments are obtained. A selected algorithm is used to extract features from each drive system timing state monitoring segment to obtain corresponding drive system timing state feature data. This feature data describes the statistical, frequency domain, or time domain characteristics of each segment and is used to characterize the characteristics of the segment. An appropriate dynamic behavior analysis algorithm is selected to analyze the drive system timing state feature data. Common methods include cluster analysis and timing pattern recognition. An appropriate algorithm is selected based on the actual situation and the corresponding parameters are configured. The selected algorithm is then used to analyze the drive system timing state feature data to identify dynamic behavior patterns and regularities. Using techniques such as clustering and pattern matching, similar timing state feature data are grouped into the same category, generating dynamic behavior feature data for the drive system. The dynamic behavior feature data of the drive system is processed using the selected algorithm to capture the evolutionary relationship and conversion rules between different features. By establishing a state transition model, mining sequence patterns and other methods, a dynamic behavior graph of the drive system is constructed to represent the relationship between different dynamic behaviors. The drive system time series state monitoring data is analyzed and detected to identify data points of abnormal states. These abnormal state data points may indicate that the system has a fault, anomaly or abnormal operation. According to the abnormal state data of the drive system obtained in step S25, these abnormal states are compared and matched with the dynamic behavior graph of the drive system constructed in step S24. If the abnormal state is associated with a dynamic behavior node or edge, it is marked as an abnormal state. The marked abnormal state data is associated with the dynamic behavior graph of the drive system to generate abnormal state marking data of the drive system. These data can be used for subsequent fault diagnosis, prediction or decision support.

[0109] In this embodiment, reference Figure 4 The above is a flowchart of the detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include:

[0110] Step S31: performing interval discretization processing on the drive system abnormal state mark data to generate drive system abnormal state quantified data;

[0111] Step S32: Perform cluster analysis on the quantified data of abnormal state of the drive system to generate cluster analysis data of abnormal state of the drive system

[0112] Step S33: performing abnormal trajectory analysis on the drive system abnormal state cluster analysis data according to the drive system time series state trajectory data to obtain drive system abnormal state trajectory data;

[0113] Step S34: tracing the abnormal state path of the drive system abnormal state trajectory data to generate the drive system abnormal state path data;

[0114] Step S35: Calculate the abnormal area weight of the drive system abnormal state path data using the drive system abnormal state weight path calculation formula to generate the drive system abnormal state area.

[0115] The present invention converts the original continuous abnormal state value into a discrete data representation through discretization processing, which is more convenient for subsequent data analysis and processing. Discretization processing can suppress the details and noise in the original data, extract the key feature information of the abnormal state, and at the same time reduce the dimension of the data, simplifying the complexity of subsequent analysis. Cluster analysis can help identify and discover similar abnormal state patterns, group and classify abnormal states according to the similarity of the data, so as to better understand and explain the changing laws of abnormal states. Through the results of cluster analysis, the abnormal state quantitative data can be divided into different clusters, which provides a classification basis for subsequent abnormal state analysis and processing, so that corresponding measures can be taken more targetedly. Through abnormal trajectory analysis, the abnormal state can be visualized in time series, so as to more intuitively observe and understand the abnormal state evolution of the drive system at different time points. By observing the abnormal trajectory, we can identify the trends and patterns of abnormal states, such as periodic changes, continuous rise or fall, etc., to provide a basis for subsequent abnormal state prediction and monitoring. By tracing the abnormal state path, we can determine the specific evolution path and process of the abnormal state from the starting point to the end point, which helps to understand the formation mechanism and influencing factors of the abnormal state. The abnormal state path data can reveal the correlation and interaction between different states, help to discover the causal relationship between abnormal states, and further provide a deeper understanding of the abnormal state. Through the abnormal area weight calculation, the abnormal state path can be quantitatively evaluated to determine the degree and importance of the abnormal state, which helps to judge the degree of abnormality. The abnormal area weight calculation can weight the abnormal state path, highlight the important abnormal state areas, provide more precise and accurate abnormal state analysis results, and provide a basis for subsequent decision-making.

[0116] In this embodiment, an appropriate interval division method is determined based on the distribution and characteristics of the abnormal state marker data. For example, equal-width discretization can be used to evenly divide the numerical range into several intervals, or equal-frequency discretization can be used to divide the data into intervals with the same number of records. Each abnormal state marker data is mapped to a corresponding discretized numerical value or category to generate quantitative data on the drive system abnormal state. This quantitative data can better represent the degree or category of the abnormal state. The quantitative data on the drive system abnormal state is input into a clustering algorithm to establish a clustering model and cluster the data. The clustering results are analyzed, and abnormal state patterns in the system are identified based on the characteristics of each cluster and the combination of abnormal states. The clustering results are saved as drive system abnormal state cluster analysis data for subsequent abnormal trajectory analysis and path tracing. The drive system abnormal state analysis data is matched with the time series state trajectory data to identify the time point when the abnormal state occurred and the related time series state. The evolution and change trend of the abnormal state on the time series trajectory are analyzed to explore characteristics such as the duration, periodicity, and frequency of the abnormal state. Extract the time-series trajectory information of abnormal states to generate drive system abnormal state trajectory data for subsequent path tracing and area weight calculation. Based on the evolution characteristics of abnormal states and relevant system knowledge, define rules for abnormal state path tracing. For example, these can include path start and end point conditions, path length limits, and other factors. Based on these rules, find abnormal state paths from the abnormal state trajectory data that meet these conditions. This can be achieved through traversal or recursive algorithms. The found abnormal state paths are recorded to generate drive system abnormal state path data for subsequent abnormal area weight calculation. Based on the specific system conditions and requirements, define a formula for calculating the drive system abnormal state weighted paths. This formula can take into account factors such as the length, frequency, and duration of the abnormal state paths. Calculate the abnormal state path data to determine the weight of each abnormal area. Based on the abnormal area weights, sort the areas from highest to lowest to generate the drive system abnormal state areas. This allows abnormal areas to be prioritized and addressed based on their contribution.

[0117] In this embodiment, the calculation formula of the drive system abnormal state weight path in step S35 is specifically:

[0118]

[0119] Where W is the weight value of the abnormal state path of the driving system, n is the number of nodes in the abnormal state path of the driving system, i is the i-th abnormal state path node of the driving system, d i is the weight of the i-th drive system abnormal state path node in the abnormal state path, j is the abnormal state value of the j-th drive system abnormal state path node, x j is the original state value of the driving system state path, is the original state value of the driving system state path, d n is the abnormal state value of the last drive system abnormal state path node in the path, d1 is the abnormal state value of the first drive system abnormal state path node in the path, f is the number of abnormal states of the drive system, t is the operating time of the drive system, and T is the existence time of the abnormal state of the drive system.

[0120] The present invention The relationship between the weights of nodes in the drive system's abnormal state path and their abnormal state values is summed. By calculating the logarithmic ratio and standard deviation between the node weights and abnormal state values, the importance of the abnormal state path nodes and the degree of dispersion of the abnormal state values can be assessed. A larger logarithmic ratio and a smaller standard deviation indicate that the node contributes significantly to the abnormal state path, making the weight calculation helpful in identifying important nodes in the abnormal state path. Indicates the ratio of the difference between the abnormal state values of the last node and the first node in the path to the number of nodes. By calculating the ratio of the difference between the abnormal state values and the number of nodes, the degree of change in the abnormal state path can be evaluated. A larger difference indicates a larger abnormal state change in the path, so the calculation of the weight is helpful in capturing the abnormal state change of the path. This formula represents the impact of the number of abnormal state events, run time, and abnormal state duration on the weight. By considering the relationship between the number of abnormal state events, run time, and abnormal state duration, the weight calculation can be adjusted to better reflect the actual situation. For example, if the number of abnormal states is high, the run time is long, and the abnormal state duration is short, the weight will increase accordingly, indicating that the path has a greater impact on the abnormal state. The formula calculates the weight of the abnormal state path of the drive system by summing and comparing the weights and abnormal state values of the abnormal state path nodes, and considering the relationship between the number of abnormal states, run time, and abnormal state duration. This calculation is helpful for identifying important nodes in the abnormal state path, capturing abnormal state changes along the path, and adjusting the weight to reflect the impact of the number of abnormal states, run time, and abnormal state duration. This information is important for drive system status analysis and abnormal state management.

[0121] In this embodiment, step S4 includes the following steps:

[0122] Step S41: performing abnormal state trend analysis on the abnormal state area of the driving system to generate abnormal state trend analysis data of the driving system;

[0123] Step S42: performing trend prediction calculation on the drive system abnormal state trend analysis data using a drive system abnormal state trend prediction calculation formula to generate drive system abnormal state trend prediction data;

[0124] Step S43: Use a deep learning algorithm to visualize the drive system abnormal state trend prediction data to generate a drive system abnormal state trend visualization view.

[0125] By analyzing data from abnormal status areas, the present invention can reveal the changing trends of abnormal conditions, such as whether they are gradually worsening or stabilizing. This helps determine the direction and speed of abnormal conditions. Abnormal status trend analysis can help identify persistent or increasing abnormal conditions, distinguishing temporary abnormalities from potential failure risks. By applying abnormal status trend prediction formulas, the future development trend of drive system abnormal conditions can be predicted. This helps take appropriate measures to prevent potential failures or abnormal conditions in advance. Abnormal status trend prediction data can be used to generate early warning signals or alerts. When abnormal status trends indicate potential problems, relevant personnel can take timely action to avoid production interruptions or equipment damage. By visualizing abnormal status trend prediction data in forms such as charts, curves, or heat maps, the changing trends of abnormal conditions can be intuitively displayed, making them easier to understand and analyze. The abnormal status trend visualization can be updated and displayed in real time, providing real-time monitoring and observation, allowing relevant personnel to quickly understand the latest abnormal status of the drive system. Based on the abnormal status trend visualization, decision makers can more accurately assess the development trend of abnormal conditions and make appropriate decisions and adjustments to optimize monitoring strategies and maintenance plans.

[0126] In this embodiment, by analyzing the changing trends of abnormal state data, possible periodic changes, long-term trends, or other specific abnormal state patterns are identified. Based on the analysis results, drive system abnormal state trend analysis data is generated. This can include trend charts, statistical indicators, or other forms of data representation to facilitate subsequent prediction and visualization. The abnormal state trend analysis data is then fed into a prediction calculation formula to perform trend prediction calculations. Based on the requirements of the formula, the predicted trend of the drive system abnormal state over a period of time is calculated. Based on the prediction calculation results, drive system abnormal state trend prediction data is generated. This can include prediction trend charts, prediction indicators, or other forms of data representation to facilitate subsequent visualization and analysis. An appropriate visualization method is selected based on the data characteristics and visualization requirements. This can include line charts, bar charts, area charts, scatter plots, or other more complex visualization techniques. The predicted data is presented using the selected visualization method. Depending on the needs, the predicted trend can be displayed, actual data can be compared with the predicted data, or changes in the abnormal state can be highlighted. The generated abnormal state trend visualization is then verified and adjusted. The prediction data and visualization method are verified to accurately reflect the abnormal state trend, and the visualization is adjusted and optimized as needed. Based on the results of verification and adjustment, a visualization view of the drive system abnormal state trend is generated. This view will intuitively show the trend changes of the drive system abnormal state, helping users understand and analyze the abnormal state situation.

[0127] In this embodiment, the calculation formula for predicting the abnormal state trend of the drive system in step S42 is specifically:

[0128]

[0129] Where P is the predicted value of the abnormal state trend of the drive system, t1 is the starting time of the abnormal state of the drive system, t2 is the time of the predicted abnormal state trend of the drive system, T is the existence time of the abnormal state of the drive system, f is the number of abnormal states of the drive system, C is the load value of the drive system, ∈ is the trend prediction adjustment factor, d i is the weight of the abnormal state path node of the i-th drive system in the abnormal state path, f is the number of abnormal states of the drive system, and b is the coefficient of the influence of the frequency of abnormal state on trend prediction.

[0130] In this embodiment, by The rate of change of the abnormal state is derived, that is, the trend of the abnormal state. By calculating the rate of change of the abnormal state, the growth rate or reduction rate of the abnormal state can be obtained, thereby predicting the future trend of the abnormal state. It represents the product of the drive system load value and the trend forecast adjustment factor. The drive system load value reflects the current load of the system, and the trend forecast adjustment factor is used to correct the trend forecast. By considering the system load and the adjustment factor, the abnormal state trend can be corrected to make the forecast result more accurate and reliable. This formula represents the impact of the number of abnormal state occurrences and abnormal state path node weights on trend prediction. f represents the number of abnormal state occurrences, which is the number of abnormal state occurrences at the current time. While the value of f may vary at different times, the number of abnormal state occurrences is a constant. For example, if the number of abnormal state occurrences measured at 3:00 PM is 3, the value of f calculated for this calculation is 3. If the number of abnormal state occurrences measured at 4:00 PM is 5, the value of f calculated for this calculation is 5. The number of abnormal state occurrences reflects the frequency of abnormal states, while the abnormal state path node weights represent the importance of different paths. By considering the number of abnormal state occurrences and path node weights, the prediction results can be adjusted to a trend that better reflects actual conditions. This formula provides a prediction of the abnormal state trend of the drive system by taking the derivative and integral of the abnormal state trend and considering the influence of factors such as load, adjustment factor, number of abnormal states, and path node weights. Such predictions are helpful in determining the future direction, speed, and cumulative change of abnormal states, providing important reference for system operation and maintenance and decision-making, and facilitating the implementation of appropriate measures to address and manage abnormal states.

[0131] In this embodiment, step S5 includes the following steps:

[0132] Step S51: performing knowledge graph reasoning on the drive system abnormal state marked data based on the drive system abnormal state trend visualization view to generate a drive system abnormal state trend knowledge graph;

[0133] Step S52: performing abnormal state optimization analysis on the drive system abnormal state trend knowledge graph to generate abnormal state trend optimization analysis data;

[0134] Step S53: Optimizing the monitoring decision on the abnormal state trend optimization analysis data to construct an abnormal state monitoring optimization decision;

[0135] This invention combines abnormal state data of the drive system with visualizations and, through the reasoning and representation capabilities of a knowledge graph, integrates scattered data into a structured knowledge network. This allows for a more comprehensive description and understanding of relevant information about abnormal states. Knowledge graph reasoning can uncover associations and dependencies between abnormal states, revealing potential causal relationships and influencing factors within the drive system. This helps provide a deeper understanding of the root causes and potential influencing mechanisms of abnormal states. By analyzing the abnormal state trend knowledge graph, optimization objectives for drive system abnormal states can be determined, such as minimizing the duration of abnormal states or minimizing the number of abnormal state occurrences. This helps clarify optimization directions and goals. Based on the abnormal state trend optimization analysis data, corresponding optimization strategies and measures can be formulated to improve the abnormal state trends of the drive system. These can include optimizing maintenance plans, optimizing operating strategies, and upgrading equipment. Based on the abnormal state trend optimization analysis data, abnormal state monitoring strategies can be adjusted and optimized in real time. This helps improve the flexibility and responsiveness of the monitoring system, making monitoring decisions more accurate and effective. By optimizing monitoring decisions, abnormal states can be detected and addressed promptly, preventing potential failures and equipment damage. Optimized monitoring decisions can include setting thresholds, formulating warning rules, adjusting alarm levels, and other measures.

[0136] In this embodiment, relevant data information is extracted from the visualization of abnormal state trends in the drive system, including the time, location, type, and trend of abnormal events. The extracted abnormal state data is then modeled into a knowledge graph. Using a graph representation method, abnormal events are represented as nodes, and the relationships between nodes are represented as edges. Nodes can include attributes such as abnormality type and location information. Reasoning analysis is performed based on the established knowledge graph of abnormal state trends in the drive system. Inference algorithms are used to discover patterns, associations, and rules based on the relationships and attributes between nodes in the graph. Based on the reasoning results, the information related to abnormal states obtained by reasoning is added to the knowledge graph to form a complete knowledge graph of abnormal state trends in the drive system. Abnormal state analysis is performed based on the knowledge graph of abnormal state trends. The associations, importance, and influencing factors between abnormal events are analyzed. The data obtained from the abnormal state analysis is then modeled. Mathematical models, statistical models, and other methods can be used to represent the relevant data in a form suitable for analysis, such as matrices or vectors. The established model is used to perform abnormal state optimization analysis. Based on the analysis objectives, optimization algorithms and models are applied to obtain optimal results for the abnormal state. Based on the optimization analysis results, abnormal state trend optimization analysis data is generated. This can include data such as optimized values for abnormal conditions and recommended optimization solutions. Abnormal condition trend optimization analysis data is integrated and processed with relevant monitoring data. This includes steps such as data cleaning, preprocessing, and feature engineering. A decision model is established based on the integrated data. Machine learning and optimization algorithms can be used to train the model to optimize abnormal condition monitoring decisions. Monitoring decisions are optimized using the established decision model. The model's predictive and optimization capabilities are used to determine the optimal monitoring strategy, alarm thresholds, and response mechanisms. Based on the optimization results, optimized decisions for abnormal condition monitoring are generated. This includes recommendations for adjusting the monitoring strategy and setting alarm rules.

[0137] In this embodiment, step S6 includes the following steps:

[0138] Step S61: using a recurrent convolutional network to perform convolution preprocessing on the abnormal state monitoring optimization decision to generate a convolution sample set of the abnormal state of the drive system;

[0139] Step S62: performing convolution data segmentation on the drive system abnormal state convolution sample set to generate a drive system abnormal state convolution sequence;

[0140] Step S63: performing dilated convolution on the drive system abnormal state convolution sequence using a dilated convolution algorithm to generate a drive system abnormal state convolution network;

[0141] Step S64: performing spatial pyramid pooling multi-layer sampling on the drive system abnormal state convolution network to generate a drive system abnormal state convolution feature map;

[0142] Step S65: Perform data mining modeling on the convolution feature map of the abnormal state of the driving system to build a driving system monitoring decision model to perform abnormal state monitoring operations.

[0143] The present invention uses convolutional preprocessing in a recurrent convolutional network to extract representative features from raw abnormal state data. These features capture the temporal and spatial information of abnormal states, thereby better describing the characteristics and changes of abnormal states. Convolutional preprocessing can perform data augmentation on abnormal state data, such as translation, scaling, and rotation, to generate more diverse samples. This helps improve the generalization and robustness of the model and better adapt to different abnormal state scenarios. By dividing the abnormal state sample set into convolutional sequences, the abnormal state data can be organized according to specific time windows, forming a continuous sequence representation. This helps capture the temporal evolution of abnormal states and related dynamic characteristics. The generation of convolutional sequences preserves the contextual information of the abnormal state data of the drive system, enabling the model to better understand the relationship between the current state and previous states. This helps improve the accuracy of abnormal state identification and prediction. Using the dilated convolution algorithm, the convolutional network can expand the receptive field without increasing the number of parameters, obtaining a wider range of contextual information. This facilitates the extraction and representation of local features of abnormal state in the drive system. Dilated convolution can reduce the number of parameters in the convolutional network, improving the parameter efficiency and computational efficiency of the model. This is particularly important for large-scale abnormal state datasets and real-time monitoring tasks. Through multi-layer sampling using spatial pyramid pooling, multi-scale and multi-level feature representations can be extracted from the drive system abnormal state convolutional network. This helps comprehensively capture the spatial information and fine-grained features of abnormal states. Spatial pyramid pooling makes the feature map invariant to scale changes, enabling better adaptation to abnormal states at different scales. This helps improve the model's generalization and adaptability. Through data mining modeling, abnormal state patterns and regularities can be learned from the drive system abnormal state convolutional feature map. Based on these learned patterns, effective monitoring decisions can be made, enabling rapid and accurate judgment and response to abnormal states. The drive system monitoring decision model processes abnormal state convolutional feature maps in real time, enabling real-time abnormal state monitoring. The model can make predictions and decisions based on the current feature map, promptly identifying and addressing abnormal states, and improving the system's fault detection and response capabilities. By establishing a monitoring decision model, automated decision-making for abnormal states can be achieved. Based on the learned knowledge and regularities, the model automatically determines the severity and impact of the abnormal state and takes appropriate measures to address it, reducing manual intervention and improving decision consistency and efficiency.

[0144] In this embodiment, the collected monitoring data is preprocessed, including data cleaning, outlier processing, and missing value filling. To ensure data quality and integrity, the time series data is converted into a format suitable for processing by a recurrent convolutional network. A sliding window technique can be used to slice the time series data into fixed-length subsequences. The convolution sample set of the drive system abnormal state is sliced to generate a convolution sequence. The sliding window technique can be used to slice the data in the sample set into fixed-length sequences, which are then used as input to generate a convolution sequence of the drive system abnormal state. Each sequence represents an abnormal state feature over a period of time. The convolution sequence of the drive system abnormal state is processed using a dilated convolution algorithm. Dilated convolution can expand the receptive field and capture a wider range of contextual information. The abnormal state sequence processed by dilated convolution is input into a convolutional neural network to construct a convolutional network for the drive system abnormal state. The convolutional network can include multiple convolutional and pooling layers, and spatial pyramid pooling is performed on the output of the drive system abnormal state convolution network. Spatial pyramid pooling can effectively extract features at different scales. Features from multiple pyramid pooling layers are sampled and combined to generate a convolution feature map of the drive system abnormal state. The feature graph represents abnormal state characteristics at different levels and scales. Data mining and modeling are performed using the convolutional feature graph data of the drive system abnormal state. Machine learning algorithms or deep learning methods can be used to construct a monitoring decision model. Based on the data mining modeling results, a drive system monitoring decision model is constructed. This model can be used to classify, predict, or determine abnormal states based on a given abnormal state feature graph. The constructed drive system monitoring decision model is then used in actual abnormal state monitoring operations. By inputting abnormal state feature graph data, decision predictions are made and abnormal state monitoring and identification are performed. Based on the decision results, appropriate measures are taken to address the abnormal state.

[0145] In this specification, a system for monitoring abnormal conditions of a placement machine drive system is provided, comprising:

[0146] The timing analysis module monitors the status of the placement machine drive system in real time to obtain the drive system status monitoring data; performs timing analysis on the drive system status monitoring data to generate the drive system timing status monitoring curve;

[0147] The abnormal state detection module performs dynamic behavior analysis on the drive system timing state monitoring curve and constructs the drive system dynamic behavior diagram; based on the drive system dynamic behavior diagram, it performs abnormal state detection on the drive system timing state monitoring data and generates drive system abnormal state marking data;

[0148] The path tracing module performs cluster analysis on the drive system abnormal state marking data to generate drive system abnormal state cluster analysis data; and performs abnormal state path tracing on the drive system abnormal state analysis data to generate the drive system abnormal state area;

[0149] The trend analysis module performs abnormal state trend analysis on the abnormal state area of the drive system to generate abnormal state trend prediction data of the drive system; and performs data visualization on the abnormal state trend prediction data of the drive system to generate a visual view of the abnormal state trend of the drive system;

[0150] The decision optimization module optimizes the monitoring decision of the drive system abnormal state data based on the visualization view of the drive system abnormal state trend and generates the abnormal state monitoring optimization decision;

[0151] The convolutional model module uses a recurrent convolutional network to perform dilated convolution on the abnormal state monitoring optimization decision to build a drive system monitoring decision model to perform abnormal state monitoring operations.

[0152] The present invention constructs an abnormal state monitoring system for a SMT drive system. Using a timing analysis module, the system can monitor the drive system's state in real time and acquire drive system state monitoring data. This enables the system to promptly capture and record changes in the drive system's state, enabling real-time monitoring of the system. The abnormal state detection module performs dynamic behavior analysis on the drive system's timing state monitoring curve and constructs a dynamic behavior graph of the drive system. By analyzing the dynamic behavior graph, the system can detect abnormal states in the drive system's timing state monitoring data and generate abnormal state marker data. This facilitates rapid and accurate identification of abnormal states in the drive system. By analyzing cluster analysis data, the system can trace abnormal state paths and identify the specific paths and factors that lead to abnormal states. This facilitates a deeper understanding of the mechanisms that cause abnormal states and provides guidance for problem resolution. By analyzing trend prediction data, the system can predict the development trends of abnormal states and take proactive intervention and remediation measures. This helps avoid potential failures and losses and improves system reliability and stability. The trend analysis module visualizes the abnormal state trend prediction data to generate a visual view of the abnormal state trend. This intuitively displays abnormal state trends and changes, helping users quickly understand and identify the development of abnormal conditions. The decision optimization module optimizes drive system abnormal state monitoring decisions based on abnormal state trend visualization. By analyzing abnormal state data and formulating optimized decisions, the system improves the speed and accuracy of abnormal state response, reducing failure risks and production costs. The convolutional model module utilizes a recurrent convolutional network to perform dilated convolution on abnormal state monitoring optimization decisions, constructing a drive system monitoring decision model. This model automatically learns the characteristics and patterns of abnormal states, enabling automated identification and prediction of abnormal conditions and improving the system's automated monitoring capabilities. The abnormal state monitoring system for the placement machine drive system, through the synergistic effect of modules including real-time monitoring, abnormal state detection, path tracing, trend analysis, data visualization, and decision optimization, provides accurate and reliable drive system status monitoring and abnormal state identification, helping users promptly identify, analyze, and resolve problems, thereby improving production efficiency and quality.

[0153] The present invention is therefore intended to be illustrative and non-restrictive in all respects, with the scope of the invention being defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the application documents are intended to be embraced therein.

[0154] It should be understood that although the terms "first," "second," and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used solely to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0155] The foregoing description is intended only to provide specific embodiments of the present invention, intended to enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A method for monitoring abnormal state of a placement machine drive system, characterized in that: The following steps are involved: Step S1: Performing real-time status monitoring on the placement machine drive system to obtain drive system status monitoring data; Perform time series analysis on the drive system status monitoring data to generate the drive system time series status monitoring curve; Step S2: Performing dynamic behavior analysis on the drive system timing state monitoring curve to construct a drive system dynamic behavior diagram; performing abnormal state detection on the drive system timing state monitoring data based on the drive system dynamic behavior diagram to generate drive system abnormal state marking data; Step S3: performing cluster analysis on the drive system abnormal state marking data to generate drive system abnormal state cluster analysis data; Perform abnormal state path tracing on the drive system abnormal state cluster analysis data to generate the drive system abnormal state area; Step S4: performing abnormal state trend analysis on the abnormal state area of the drive system to generate abnormal state trend prediction data of the drive system; Perform data visualization on the abnormal state trend prediction data of the drive system to generate a visual view of the abnormal state trend of the drive system; Step S5: Optimizing monitoring decisions for the drive system abnormal state marking data based on the abnormal state trend visualization view of the drive system, and generating an abnormal state monitoring optimization decision; Step S6: Use the recurrent convolutional network to perform dilated convolution on the abnormal state monitoring optimization decision to build a drive system monitoring decision model to perform abnormal state monitoring operations.

2. The method according to claim 1, characterized in that The specific steps of step S1 are: Step S11: performing real-time status monitoring on the placement machine drive system to obtain drive system status monitoring data, where the drive system status monitoring data includes drive system operating time data, drive system operating status data, drive system energy consumption data, and drive system fault record data; Step S12: performing time series analysis on the drive system state monitoring data to generate drive system time series state monitoring data; Step S13: performing a time series evolution analysis on the driving system time series state monitoring data to generate driving system time series state trajectory data; Step S14: performing timing curve fitting on the driving system timing state monitoring data to generate a driving system timing state monitoring curve.

3. The method according to claim 1, characterized in that The specific steps of step S2 are: Step S21: segmenting the drive system timing state monitoring curve to obtain a drive system timing state monitoring segment set; Step S22: extracting features from the drive system timing state monitoring segment set to generate drive system timing state feature data; Step S23: performing dynamic behavior analysis on the driving system time series state characteristic data to generate driving system dynamic behavior characteristic data; Step S24: performing dynamic behavior evolution processing on the dynamic behavior characteristic data of the drive system to construct a dynamic behavior graph of the drive system; Step S25: performing abnormal state detection on the driving system timing state monitoring data to obtain abnormal state data of the driving system; Step S26: marking the abnormal state of the driving system dynamic behavior diagram based on the abnormal state data of the driving system to generate abnormal state marking data of the driving system.

4. The method according to claim 1, wherein The specific steps of step S3 are: Step S31: performing interval discretization processing on the drive system abnormal state mark data to generate drive system abnormal state quantified data; Step S32: performing cluster analysis on the quantified data of abnormal state of the driving system to generate cluster analysis data of abnormal state of the driving system; Step S33: performing abnormal trajectory analysis on the drive system abnormal state cluster analysis data according to the drive system time series state trajectory data to obtain drive system abnormal state trajectory data; Step S34: tracing the abnormal state path of the drive system abnormal state trajectory data to generate the drive system abnormal state path data; Step S35: Calculate the abnormal area weight of the drive system abnormal state path data using the drive system abnormal state weight path calculation formula to generate the drive system abnormal state area.

5. The method according to claim 4, characterized in that The specific calculation formula of the drive system abnormal state weight path in step S35 is: Where W is the weight value of the abnormal state path of the driving system, n is the number of nodes in the abnormal state path of the driving system, i is the i-th abnormal state path node of the driving system, d i is the weight of the i-th drive system abnormal state path node in the abnormal state path, j is the abnormal state value of the j-th drive system abnormal state path node, x j is the original state value of the driving system state path, is the original state value of the driving system state path, d n is the abnormal state value of the last drive system abnormal state path node in the path, d1 is the abnormal state value of the first drive system abnormal state path node in the path, f is the number of abnormal states of the drive system, t is the operating time of the drive system, and T is the existence time of the abnormal state of the drive system.

6. The method according to claim 1, characterized in that The specific steps of step S4 are: Step S41: performing abnormal state trend analysis on the abnormal state area of the driving system to generate abnormal state trend analysis data of the driving system; Step S42: performing trend prediction calculation on the drive system abnormal state trend analysis data using a drive system abnormal state trend prediction calculation formula to generate drive system abnormal state trend prediction data; Step S43: Use a deep learning algorithm to visualize the drive system abnormal state trend prediction data to generate a drive system abnormal state trend visualization view.

7. The method according to claim 6, characterized in that The calculation formula for predicting the abnormal state trend of the drive system in step S42 is specifically: Where P is the predicted value of the abnormal state trend of the drive system, t1 is the starting time of the abnormal state of the drive system, t2 is the time of the predicted abnormal state trend of the drive system, T is the existence time of the abnormal state of the drive system, f is the number of abnormal states of the drive system, C is the load value of the drive system, ∈ is the trend prediction adjustment factor, d i is the weight of the abnormal state path node of the i-th drive system in the abnormal state path, f is the number of abnormal states of the drive system, and b is the coefficient of the influence of the frequency of abnormal state on trend prediction.

8. The method according to claim 1, characterized in that The specific steps of step S5 are: Step S51: performing knowledge graph reasoning on the drive system abnormal state marked data based on the drive system abnormal state trend visualization view to generate a drive system abnormal state trend knowledge graph; Step S52: performing abnormal state optimization analysis on the drive system abnormal state trend knowledge graph to generate abnormal state trend optimization analysis data; Step S53: Optimizing the monitoring decision on the abnormal state trend optimization analysis data to construct an abnormal state monitoring optimization decision.

9. The method according to claim 1, characterized in that The specific steps of step S43 are: Step S61: using a recurrent convolutional network to perform convolution preprocessing on the abnormal state monitoring optimization decision to generate a convolution sample set of the abnormal state of the drive system; Step S62: performing convolution data segmentation on the drive system abnormal state convolution sample set to generate a drive system abnormal state convolution sequence; Step S63: performing dilated convolution on the drive system abnormal state convolution sequence using a dilated convolution algorithm to generate a drive system abnormal state convolution network; Step S64: performing spatial pyramid pooling multi-layer sampling on the drive system abnormal state convolution network to generate a drive system abnormal state convolution feature map; Step S65: Perform data mining modeling on the convolution feature map of the abnormal state of the driving system to build a driving system monitoring decision model to perform abnormal state monitoring operations.

10. An abnormal state monitoring system for a placement machine drive system, characterized in that: The method for monitoring the abnormal state of the placement machine drive system according to claim 1 comprises: The timing analysis module monitors the status of the placement machine drive system in real time to obtain the drive system status monitoring data; performs timing analysis on the drive system status monitoring data to generate the drive system timing status monitoring curve; The abnormal state detection module performs dynamic behavior analysis on the drive system timing state monitoring curve and constructs the drive system dynamic behavior diagram; based on the drive system dynamic behavior diagram, it performs abnormal state detection on the drive system timing state monitoring data and generates drive system abnormal state marking data; The path tracing module performs cluster analysis on the drive system abnormal state marking data to generate drive system abnormal state cluster analysis data; and performs abnormal state path tracing on the drive system abnormal state analysis data to generate the drive system abnormal state area; The trend analysis module performs abnormal state trend analysis on the abnormal state area of the drive system to generate abnormal state trend prediction data of the drive system; and performs data visualization on the abnormal state trend prediction data of the drive system to generate a visual view of the abnormal state trend of the drive system; The decision optimization module optimizes the monitoring decision of the drive system abnormal state data based on the visualization view of the drive system abnormal state trend and generates the abnormal state monitoring optimization decision; The convolutional model module uses a recurrent convolutional network to perform dilated convolution on the abnormal state monitoring optimization decision to build a drive system monitoring decision model to perform abnormal state monitoring operations.

Images