A working state monitoring method and system for a high-speed laser film-making and striping machine

Abstract

The present application relates to the technical field of state monitoring, and provides a method and system for monitoring the working state of a high-speed laser film-making and stripping machine. The method includes: receiving predetermined parameters and control schemes of the high-speed laser film-making and stripping machine; disassembling component features to obtain control decisions of each component; extracting the control decision of the first component, predicting its health status, and obtaining a health status matrix; controlling the equipment, and monitoring in real time through a sensor network to obtain the status information of the first component; predicting risks based on the health status matrix and status information, and generating a risk coefficient; if the risk coefficient is greater than a predetermined value, generating an early warning signal to warn the equipment. The method solves the technical problem that there is a lack of effective means to monitor the working state of each component in real time and predict potential failure risks during the operation of the high-speed laser film-making and stripping machine, resulting in untimely maintenance of the high-speed laser film-making and stripping machine.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, specifically to the field of state monitoring technology, and in particular to a working state monitoring method and system for a high-speed laser film-making and striping machine. Background Art

[0002] With the rapid development and technological innovation of modern manufacturing, high-speed laser film-making and slitting machines, as high-precision and high-efficiency production equipment, play an indispensable role in many fields such as electronics, automobiles, aerospace, etc. This type of machine uses laser technology to achieve precise cutting and slitting of materials, greatly improving production efficiency and product quality. However, with the continuous expansion of production scale and the increasing complexity of processing technology, the working status monitoring and management of high-speed laser film-making and slitting machines face unprecedented challenges. Traditional working status monitoring methods are difficult to detect and solve potential problems in equipment operation in a timely manner, and may also cause equipment failures due to untimely monitoring, affecting production progress and product quality. In addition, traditional methods lack in-depth mining and analysis of equipment operation data, and cannot accurately predict the trend of equipment status changes, let alone provide strong support for preventive maintenance and optimized scheduling of equipment. Summary of the invention

[0003] The present application provides a working status monitoring method and system for a high-speed laser filming and striping machine, aiming to solve the technical problem that during the operation of the high-speed laser filming and striping machine, there is a lack of effective means to monitor the working status of each component in real time and predict potential failure risks, resulting in untimely maintenance of the high-speed laser filming and striping machine.

[0004] In view of the above problems, the present application provides a method and system for monitoring the working status of a high-speed laser film-making and slitting machine.

[0005] The first aspect disclosed in the present application provides a method for monitoring the working status of a high-speed laser film-making and striping machine, the method comprising: receiving predetermined working parameters of the high-speed laser film-making and striping machine, obtaining a predetermined control scheme for the device and predetermined working environment information for the device; performing component feature disassembly according to the predetermined control scheme for the device, and obtaining predetermined control decisions for each component of the high-speed laser film-making and striping machine; extracting a first component predetermined control decision corresponding to a first component of the high-speed laser film-making and striping machine according to the predetermined control decisions for each component; and monitoring the working status health of the first component according to the predetermined working environment information for the device and the predetermined control decision for the first component. Prediction, obtain the first component health status prediction matrix; control the high-speed laser filming and striping integrated machine based on the predetermined control decisions of each component, and monitor the high-speed laser filming and striping integrated machine in real time through the sensor monitoring network to obtain the working status monitoring information of the first component; based on the first component health status prediction matrix, perform risk prediction on the first component according to the first component working status monitoring information, and generate a first component state risk coefficient; if the first component state risk coefficient is greater than / equal to the predetermined state risk coefficient, generate a first risk warning signal, and perform an abnormal warning on the high-speed laser filming and striping integrated machine according to the first risk warning signal.

[0006] Another aspect disclosed in the present application provides a working status monitoring system for a high-speed laser film-making and striping machine, the system comprising: a working parameter receiving module, the working parameter receiving module is used to receive the predetermined working parameters of the high-speed laser film-making and striping machine, and obtain the predetermined control scheme and the predetermined working environment information of the equipment; a component feature disassembly module, the component feature disassembly module is used to disassemble the component features according to the predetermined control scheme of the equipment, and obtain the predetermined control decisions of each component of the high-speed laser film-making and striping machine; a control decision extraction module, the control decision extraction module is used to extract the first component predetermined control decision corresponding to the first component of the high-speed laser film-making and striping machine according to the predetermined control decisions of each component; a health prediction module, the health prediction module is used to determine the health status of the first component according to the predetermined working environment information of the equipment and the predetermined control decision of the first component. The first component performs a working status health prediction to obtain a first component health status prediction matrix; a real-time monitoring module, the real-time monitoring module is used to control the high-speed laser filming and striping integrated machine based on the predetermined control decisions of each component, and to perform real-time monitoring of the high-speed laser filming and striping integrated machine through a sensor monitoring network to obtain the working status monitoring information of the first component; a risk prediction module, the risk prediction module is used to perform risk prediction on the first component based on the first component health status prediction matrix and the working status monitoring information of the first component to generate a first component state risk coefficient; an abnormal warning module, the abnormal warning module is used to generate a first risk warning signal if the state risk coefficient of the first component is greater than / equal to a predetermined state risk coefficient, and to perform an abnormal warning on the high-speed laser filming and striping integrated machine according to the first risk warning signal.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] The above-mentioned working state monitoring method of a high-speed laser film-making and striping machine receives and determines the predetermined working parameters, predetermined control schemes and predetermined working environment information of the high-speed laser film-making and striping machine. This information is the basis for ensuring that the equipment can work efficiently and stably as expected. Subsequently, according to these predetermined control schemes, the specific control decisions of each component are disassembled. The purpose of this is to manage the operating state of each component more accurately, because the overall performance of the equipment often depends on the coordinated work of its various components. Afterwards, according to the predetermined control decisions of each component, the predetermined control decision of the first component is extracted, and the health state of the component is predicted in combination with the working environment information to obtain a health state prediction matrix to describe the possible operating state of the first component in the future. Through the sensor monitoring network throughout the equipment, the working state information of all components including the first component is obtained in real time. This information is combined with the previous health state prediction matrix to predict the risk of the first component and calculate the state risk coefficient. If the state risk coefficient reaches or exceeds the predetermined state risk coefficient, a risk warning signal is generated to remind maintenance personnel to pay attention to possible abnormal conditions of the first component, avoid equipment failure, and ensure production continuity and product quality.

[0009] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0011] Figure 1 The figure is a flow chart of a method for monitoring the working status of a high-speed laser film-making and slitting machine in one embodiment.

[0012] Figure 2 The following is an architecture diagram of a working status monitoring system for a high-speed laser film-making and slitting machine in one embodiment.

[0013] Explanation of the accompanying drawings: working parameter receiving module 1, component feature disassembly module 2, control decision extraction module 3, health prediction module 4, real-time monitoring module 5, risk prediction module 6, abnormal warning module 7. DETAILED DESCRIPTION

[0014] The embodiments of the present application provide a method and system for monitoring the working status of a high-speed laser filming and striping machine, thereby solving the technical problem that during the operation of the high-speed laser filming and striping machine, there is a lack of effective means to monitor the working status of each component in real time and predict potential failure risks, resulting in untimely maintenance of the high-speed laser filming and striping machine.

[0015] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0016] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or inherent to these processes, methods, products or devices.

[0017] Embodiment 1, as Figure 1 As shown, the present application provides a method for monitoring the working status of a high-speed laser film-making and striping machine, the method comprising:

[0018] Receive the predetermined working parameters of the high-speed laser film-making and slitting machine, and obtain the predetermined control plan of the equipment and the predetermined working environment information of the equipment.

[0019] In an embodiment of the present application, when the high-speed laser film-making and striping machine is ready to run, the system terminal will receive a series of predetermined operating parameters of the device. These parameters are the benchmarks for the device to perform tasks according to specific requirements, such as laser power, cutting speed, material thickness, etc. Based on these predetermined operating parameters, the system terminal obtains the device's predetermined control scheme for the device. This scheme specifies how the device should adjust the status of each component during operation to ensure that the overall performance is optimized. At the same time, the predetermined working environment information of the device is also obtained, including external factors such as temperature, humidity, vibration, etc. that may affect the operation of the device. By understanding this information, the device can be appropriately adjusted or configured in advance to cope with possible environmental challenges.

[0020] The component characteristics are disassembled according to the predetermined control scheme of the equipment to obtain the predetermined control decisions of each component of the high-speed laser film-making and slitting machine.

[0021] In one embodiment, after obtaining the predetermined control scheme of the high-speed laser film-making and striping machine, the system terminal performs component feature disassembly on the scheme, that is, decomposing the complex control process of the entire device into specific control tasks of each component, such as the laser, cutting head, transmission system, etc. Through component feature disassembly, the system terminal can understand the role of each component in the operation of the device and how to accurately control them according to the predetermined control scheme. After completing the component feature disassembly, the system terminal generates a series of predetermined control decisions for each component. These decisions are the key to the smooth operation of the device according to the predetermined control scheme. They ensure that each component in the device can work together to realize the overall function of the device.

[0022] According to the predetermined control decisions of each component, a first component predetermined control decision corresponding to the first component of the high-speed laser film-making and slitting machine is extracted.

[0023] In one embodiment, the system terminal randomly extracts a component from the components of the high-speed laser film-making and slitting machine as the first component, and uses the corresponding predetermined control decision as the first component predetermined control decision for subsequent health prediction of the working status of the first component.

[0024] The working state health prediction of the first component is performed according to the predetermined working environment information of the equipment and the predetermined control decision of the first component to obtain a first component health state prediction matrix.

[0025] In one embodiment, the system terminal performs a health prediction on the working state of the first component based on the predetermined working environment conditions of the device and the preset control strategy of the first component. This prediction process is intended to evaluate in advance whether the first component will be in good operating condition in the future. The system terminal generates a health status prediction matrix for the first component through steps such as registration recognition and concentrated interval recognition. This matrix contains information about the health status of the first component in different time periods or under different working conditions in the future, which helps to discover potential problems in advance and optimize the control strategy to ensure the overall performance and stability of the device.

[0026] Furthermore, the present application provides a method for performing working state health prediction on the first component according to the predetermined working environment information of the device and the predetermined control decision of the first component to obtain a first component health state prediction matrix, including:

[0027] According to the high-speed laser film-making and striping integrated machine, health status monitoring records are retrieved to obtain a first health status monitoring record set corresponding to the first component; the first health status monitoring record set is aligned and identified according to the predetermined working environment information of the equipment and the predetermined control decision of the first component to obtain a first health status prediction alignment space; data cleaning and classification are performed according to the first health status prediction alignment space to obtain multiple health status prediction alignment domains.

[0028] Preferably, the system terminal retrieves the health status monitoring records related to the first component according to the operation history of the high-speed laser film-making and striping machine to form a first health status monitoring record set. This first health status monitoring record set reflects the working status of the first component in different time periods and working environments. Subsequently, the system terminal uses the predetermined working environment information of the equipment and the predetermined control decision of the first component as a reference to align and identify the first health status monitoring record set. This process is to construct a first health status prediction alignment space that meets the predetermined requirements by performing weighted alignment calculation on the health status monitoring records in the first health status monitoring record set. This space is a framework for analyzing and comparing the health status data of the first component under the same benchmark. Afterwards, in the first health status prediction alignment space, the system terminal cleans the internal working status monitoring data to remove noise and irrelevant data, and checks whether there are missing values, records the number and location of missing values, and then uses the mean of the five adjacent data to supplement the missing values. After the data cleaning is completed, the system terminal determines the neighborhood size ε and the minimum number of points MinPts as the DBSCAN algorithm input parameters according to actual needs and data characteristics. Then, the distance between each two working state monitoring data after cleaning is calculated according to the Euclidean distance, and then the configured DBSCAN algorithm is used to divide these working state monitoring data into different clusters to obtain multiple health state prediction registration domains. Each registration domain represents a different health state mode that the first component may present under specific working conditions or time periods. This helps to more accurately predict and diagnose the future health status of the first component, providing strong support for maintenance and optimization.

[0029] Centralized interval recognition is performed according to the multiple health status prediction and registration domains to obtain multiple health prediction and registration recognition results; and the first component health status prediction matrix is ​​generated according to the multiple health prediction and registration recognition results.

[0030] Preferably, after obtaining multiple health status prediction registration domains, the system terminal uses the working status monitoring data in each registration domain as a data point, and applies the K-means clustering algorithm to perform cluster analysis on all working status monitoring data. Among them, the cluster number K value is selected based on the elbow rule. After the clustering is completed, each data point will be assigned to a cluster center. For each cluster center, the corresponding prediction value range, that is, the concentrated interval, is calculated, and then the frequency of each concentrated interval is determined, that is, the proportion of the working status monitoring data in the interval in all working status monitoring data, and the weight of the interval is determined according to the frequency of the interval. The concentrated interval that appears at a high frequency is given a higher weight. Subsequently, the recognition results of each registration domain are integrated, and the concentrated interval of each working status monitoring data is summarized to form multiple health prediction registration recognition results. In the summary process, the system terminal performs weighted calculation on the summarized results by the Bayesian weighted average method to ensure the accuracy of the comprehensive results. Afterwards, based on these health prediction registration recognition results, the system terminal generates a health status prediction matrix of the first component. This matrix is ​​a two-dimensional array, which contains the health status prediction information of the first component at different times or conditions. Each row of the matrix represents a specific time point or condition, while each column represents different health status prediction information. By looking at this matrix, you can quickly understand the health status change trend of the first component and the health problems that may occur under different conditions.

[0031] Further, the present application provides registering and identifying the first health status monitoring record set according to the predetermined working environment information of the device and the predetermined control decision of the first component to obtain a first health status prediction registration space, including:

[0032] Traverse the first health status monitoring record set and extract the first health status monitoring record, wherein the first health status monitoring record includes the first sample working environment information, the first sample control decision and the first healthy sample working status monitoring information corresponding to the first component; perform a weighted registration analysis on the first health status monitoring record according to the predetermined working environment information of the equipment and the predetermined control decision of the first component to generate a first sample comprehensive registration coefficient.

[0033] Optionally, the system terminal traverses the first health status monitoring record set and extracts the first health status monitoring record therefrom. The first health status monitoring record records in detail the first sample working environment information, the first sample control decision and the first healthy sample working status monitoring information corresponding to the first component. Subsequently, the system terminal uses the predetermined working environment information of the equipment and the predetermined control decision of the first component as a benchmark, and performs registration and identification on the first health status monitoring record according to the pre-built environment registration and identification channel and the control registration and identification channel to obtain the corresponding registration coefficient. Afterwards, the obtained registration coefficient is weightedly calculated based on the weighted registration weight condition to generate the first sample comprehensive registration coefficient. This first sample comprehensive registration coefficient reflects the degree of closeness between the actual working environment, control decision and working status in the record and the predetermined ideal state, thereby helping to evaluate the health status of the component.

[0034] Determine whether the first sample comprehensive registration coefficient is greater than / equal to the predetermined comprehensive registration coefficient; if the first sample comprehensive registration coefficient is greater than / equal to the predetermined comprehensive registration coefficient, add the first healthy sample working status monitoring information to the first health status prediction registration space.

[0035] Optionally, after calculating the first sample comprehensive registration coefficient, the system terminal determines whether the first sample comprehensive registration coefficient is greater than or equal to a predetermined comprehensive registration coefficient. This predetermined comprehensive registration coefficient is a threshold set based on historical experience and expert advice, and is used to measure the degree of compliance between the health status monitoring record and the standards required by the health status prediction registration space. If the first sample comprehensive registration coefficient is greater than or equal to the predetermined comprehensive registration coefficient, it means that the working status monitoring information reflected by the first health status monitoring record matches the requirements of the health status prediction registration space. Therefore, the system terminal adds the first health sample working status monitoring information of the first health status monitoring record to the first health status prediction registration space. The purpose of this is to accumulate more high-quality data in order to more accurately predict and evaluate the health status of components.

[0036] Further, the present application provides performing weighted registration analysis on the first health status monitoring record according to the predetermined working environment information of the device and the predetermined control decision of the first component to generate a first sample comprehensive registration coefficient, including:

[0037] Based on the twin neural network, the environment registration and recognition channel and the control registration and recognition channel are constructed.

[0038] Optionally, in order to perform accurate weighted registration analysis, the system terminal constructs an environmental registration recognition channel and a control registration recognition channel based on the architecture of the twin neural network, including the number of neurons in the input layer, hidden layer and output layer of each self-network, loss function, etc. The core functions of these two channels are to use the characteristics of the twin neural network to identify the similarity between input data, but they differ in training data and target applications. For the environmental registration recognition channel, the system terminal divides the sample device predetermined working environment information, sample working environment information, and sample environment registration coefficient to obtain a training set and a test set. The training set is then divided into multiple small batches, each of which contains multiple pairs of sample device predetermined working environment information, sample working environment information and their corresponding sample environment registration coefficients. For each small batch, the system terminal inputs the paired input samples into the two sub-networks of the environmental registration recognition channel for forward propagation calculation. In the output layer, the Euclidean distance is used to calculate the distance between the embedded vectors output by the two sub-networks, and the predicted value of the environmental registration coefficient is calculated based on the distance. Subsequently, the difference between the predicted environment registration coefficient and the sample environment registration coefficient is calculated using the contrast loss function to obtain the loss value. According to the loss value, the gradient of each parameter in the environment registration recognition channel is calculated by the back propagation algorithm. Then the optimization algorithm SGD is used to update the parameters of the environment registration recognition channel according to the gradient to minimize the loss function. Repeat the above steps until the training round reaches the preset upper limit. After that, the trained environment registration recognition channel is evaluated using the test set, and the environment registration recognition accuracy, recall rate and other indicators of the environment registration recognition channel on unseen data are calculated, and the performance of the environment registration recognition channel under different working environment conditions is analyzed to identify possible deviations and deficiencies. If the evaluation result does not meet the actual needs, the environment registration recognition channel is optimized and adjusted, including adjusting hyperparameters such as learning rate, batch size, loss function weight, and introducing regularization technology to prevent overfitting. Otherwise, the current environment registration recognition channel is output. For the control registration recognition channel, the system terminal uses the same training method as above to train the final control registration recognition channel according to the sample component predetermined control decision, sample control decision, and sample control registration coefficient.

[0039] The predetermined working environment information of the device and the first sample working environment information are input into the environmental registration identification channel to generate a first sample environmental registration coefficient; based on the predetermined control decision of the first component and the first sample control decision, according to the control registration identification channel, a first sample control registration coefficient is obtained; based on the weighted registration weight condition, the first sample environmental registration coefficient and the first sample control registration coefficient are weightedly calculated to obtain the first sample comprehensive registration coefficient.

[0040] Optionally, after constructing the environment registration recognition channel and the control registration recognition channel, the system terminal takes the predetermined working environment information of the device and the first sample working environment information as input and inputs them into the environment registration recognition channel. The environment registration recognition channel compares and analyzes the two inputs based on the feature representation and similarity measurement method learned during the training process, and outputs the first sample environment registration coefficient, which quantifies the degree of registration between the first sample working environment and the predetermined working environment. Similarly, the system terminal inputs the first component predetermined control decision and the first sample control decision into the control registration recognition channel to generate the first sample control registration coefficient. After obtaining the first sample environment registration coefficient and the first sample control registration coefficient, the system terminal calculates the first sample comprehensive registration coefficient according to the weighted registration weight condition. Among them, the weighted registration weight condition specifies two weights, which correspond to the importance of the first environment registration coefficient and the first control registration coefficient in the calculation of the comprehensive registration coefficient, and the two weights are determined based on historical experience and expert advice. The system terminal calculates the first sample comprehensive registration coefficient by multiplying the two registration coefficients by their corresponding weights and summing them. This coefficient comprehensively considers the alignment of both the working environment and the control decision, providing a more comprehensive basis for subsequent decision-making or evaluation.

[0041] The high-speed laser film-making and slitting machine is controlled based on the predetermined control decisions of each component, and the high-speed laser film-making and slitting machine is monitored in real time through a sensor monitoring network to obtain the working status monitoring information of the first component.

[0042] In one embodiment, the system terminal guides and controls the operation of the high-speed laser film-making and striping machine according to the predetermined control decisions of each component. Each component in the high-speed laser film-making and striping machine will follow the decision content to perform its specific function to ensure that the entire device can operate in a coordinated and efficient manner. In order to grasp the operating status of the high-speed laser film-making and striping machine in real time and discover potential problems in a timely manner, the system terminal collects various real-time data such as temperature, pressure, vibration, speed, etc. based on the deployed sensor monitoring network. The data collected by these sensors will be summarized by the system terminal to generate the first component working status monitoring information, which reflects in detail the current working status, performance parameters and possible abnormal conditions of the high-speed laser film-making and striping machine.

[0043] Based on the health status prediction matrix of the first component and according to the working status monitoring information of the first component, risk prediction is performed on the first component to generate a first component status risk coefficient.

[0044] In one embodiment, the system terminal organizes and transforms the working status monitoring information of the first component to obtain the first component status monitoring matrix corresponding to the working status monitoring information of the first component. Subsequently, the first component status monitoring matrix and the first component health status prediction matrix are evaluated for deviation depth, and the deviation depth evaluation result is compared with the predetermined requirements. If the deviation depth evaluation result meets the predetermined requirements, the system terminal performs risk prediction on the deviation analysis results of the first component status monitoring matrix and the first component health status prediction matrix based on the first component risk prediction channel to generate a first component status risk coefficient. This first component status risk coefficient is a quantitative indicator that reflects the health status risk level of the first component at present and in the future. If the risk coefficient is high, it means that the first component has a higher risk of failure. On the contrary, it indicates that the first component is currently in a relatively healthy state.

[0045] Further, the present application provides a method for performing risk prediction on the first component based on the health status prediction matrix of the first component and according to the working status monitoring information of the first component to generate a first component status risk coefficient, including:

[0046] Organizing the working status monitoring information of the first component and establishing a first component status monitoring matrix; performing deviation analysis based on the first component health status prediction matrix and the first component status monitoring matrix to obtain a first component status health deviation analysis result; performing deviation depth evaluation based on the first component status health deviation analysis result to obtain a first component health deviation depth coefficient;

[0047] Preferably, the system terminal comprehensively organizes the obtained first component working state monitoring information, which includes the measured values ​​of multiple parameters such as temperature, pressure, vibration frequency, etc. at different time points. By arranging these information in the same format and order as the first component health state prediction matrix, a first component state monitoring matrix is ​​established. This matrix is ​​a data set that reflects the working state of the first component in real time, and provides a basis for subsequent analysis. Subsequently, the first component health state prediction matrix and the first component state monitoring matrix are subjected to deviation analysis to find the difference or deviation between the current monitoring state and the predicted health state. In the process of deviation analysis, the system terminal first aligns the time node in the first component health state prediction matrix with the time node in the first component state monitoring matrix to ensure that the two are compared in the same time frame, and then ensures that the parameter items in the prediction matrix and the monitoring matrix correspond one to one, so as to perform direct comparative analysis later. Subsequently, for each parameter item, the absolute difference between the monitoring value and the predicted value, that is, the deviation degree, is calculated, and each calculated deviation degree is compared with the deviation threshold value to determine which parameter monitoring values ​​exceed the threshold range of the predicted value, and obtain the first component state health deviation analysis result, which reveals the mismatch between the current health state of the first component and the predicted health state. Afterwards, the system terminal traverses the obtained health deviation analysis results of the first component state. In each traversal, the traversed deviation is compared with the depth deviation threshold, and the deviation greater than or equal to the depth deviation threshold is determined as the depth deviation. Among them, the deviation threshold and the depth deviation threshold are set according to the characteristics and operating requirements of the first component, and are used to measure the degree of deviation. Then, the system terminal assigns a weight to each deep deviation based on the importance of the indicators represented by these deep deviations, combined with historical experience and expert advice. Finally, each deep deviation is multiplied by the corresponding weight, and the results are accumulated to calculate the health deviation depth coefficient of the first component. This coefficient reflects the degree to which the health state of the first component deviates from the normal level. The higher the coefficient, the deeper the deviation and the worse the health state of the component.

[0048] Determine whether the health deviation depth coefficient of the first component is greater than / equal to a predetermined health deviation depth coefficient; if the health deviation depth coefficient of the first component is greater than / equal to the predetermined health deviation depth coefficient, activate a pre-constructed first component risk prediction channel; based on the health deviation analysis result of the first component state, obtain the first component state risk coefficient according to the first component risk prediction channel.

[0049] Preferably, after obtaining the health deviation depth coefficient of the first component, the system terminal compares the health deviation depth coefficient of the first component with the predetermined health deviation depth coefficient to determine whether the health status of the first component has reached a level that requires special attention. If the health deviation depth coefficient of the first component is greater than or equal to the predetermined health deviation depth coefficient, it means that the health status of the first component has or is about to exceed the acceptable range, and there is a high risk. At this time, the system terminal activates a pre-designed risk prediction channel. This channel contains a series of preset models for analyzing and predicting the risks that the first component may face. After the risk prediction channel is activated, the system terminal calculates a specific risk coefficient, namely the first component state risk coefficient, based on the health deviation analysis results of the first component state through the risk prediction channel. This risk coefficient quantifies the degree of risk that may be brought about by the current state of the first component, and provides a basis for subsequent decision-making and response.

[0050] Further, the present application provides a method for obtaining a risk coefficient of the first component state based on the first component state health deviation analysis result and according to the first component risk prediction channel, including:

[0051] The first component risk prediction channel includes K risk prediction models corresponding to the first component, where K is a positive integer greater than 1.

[0052] Optionally, in order to calculate the risk coefficient of the first component state, the system terminal constructs a first component risk prediction channel, which includes K risk prediction models corresponding to the first component, where K is a positive integer greater than 1. In the process of constructing the first component risk prediction channel, the system terminal first obtains K groups of sample state health deviation analysis results and K groups of sample risk prediction coefficients, and each sample risk prediction coefficient corresponds to a group of sample state health deviation analysis results. Subsequently, the system terminal pre-sets a neural network framework, including setting the number of neurons in the input layer, hidden layer, and output layer, setting ReLU as the activation function, setting the mean square error as the loss function, etc. After the neural network framework is set, the system terminal randomly extracts a group of sample state health deviation analysis results and a corresponding group of sample risk prediction coefficients, and divides these sample state health deviation analysis results and the corresponding sample risk prediction coefficients to obtain a training set and a validation set, and then uses random numbers to initialize the neural network framework, that is, assigning a small random value to the weight and bias. After that, the system terminal divides the training set into multiple small batches. In each iterative training, the system terminal inputs the data of the current small batch into the neural network, calculates it layer by layer through the network layer until it reaches the output layer, and calculates the risk prediction coefficient at the output layer, and then compares it with the sample risk prediction coefficient, and calculates the loss value through the mean square error. Then, the loss value is used to calculate the gradient of the network parameters, that is, the partial derivative of the loss value with respect to each parameter. This step involves the application of the chain rule to backpropagate the gradient from the output layer to the input layer. Then, the optimization algorithm SGD and the calculated gradient are used to update the network parameters to minimize the loss function, that is, adjust the parameters to make the prediction of the neural network on the training data more accurate. During the training process, in addition to using the training data to update the parameters of the neural network, the performance of the neural network is also evaluated on the validation set. If the performance of the neural network on the validation set does not improve significantly or decreases in multiple consecutive iterations, the training is stopped to avoid overfitting. After the training stops, the system terminal outputs the current neural network as a risk prediction model. Repeat the above training process, and use the remaining K-1 groups of sample state health deviation analysis results and K-1 groups of sample risk prediction coefficients to build K-1 risk prediction models. Finally, the system terminal connects the K risk prediction models in parallel to build the first component risk prediction channel.

[0053] Input the health deviation analysis results of the first component state into the K risk prediction models to obtain K risk prediction coefficients; perform proportion calculation according to the K risk prediction accuracy parameters corresponding to the K risk prediction models to obtain K risk prediction incentive coefficients; perform weighted calculation on the K risk prediction coefficients according to the K risk prediction incentive coefficients to obtain the first component state risk coefficient.

[0054] Optionally, after constructing the first component risk prediction channel, the system terminal inputs the first component state health deviation analysis result into the first component risk prediction channel, and the first component risk prediction channel generates K risk prediction coefficients based on the internal K risk prediction models. These coefficients represent the first component risk prediction channel's judgment of the degree of risk that may exist in the current state of the component. Subsequently, the system terminal sums the risk prediction accuracy parameters of each model, and then uses the risk prediction accuracy parameters of each model and the summed result to calculate the proportion, and obtains the risk prediction incentive coefficient of each model. Among them, the risk prediction accuracy parameter of each model refers to the accuracy index obtained by testing the model using data that is not used for training and verification. Afterwards, the system terminal uses the calculated risk prediction incentive coefficient of each model as the weight of the model, multiplies it with the risk prediction coefficient corresponding to the model, and then sums the K products to obtain the first component state risk coefficient. This risk coefficient combines the prediction results of multiple models and reflects the overall risk level that the first component may face in its current state.

[0055] If the first component state risk coefficient is greater than / equal to a predetermined state risk coefficient, a first risk warning signal is generated, and an abnormal warning is issued to the high-speed laser film-making and slitting machine according to the first risk warning signal.

[0056] In one embodiment, if the calculated state risk coefficient of the first component is greater than or equal to the predetermined state risk coefficient, it means that the first component has hidden dangers or is in a high-risk state. At this time, the system terminal will immediately generate a first risk warning signal. The function of this signal is to issue a clear warning to the maintenance personnel, indicating that the first component of the high-speed laser filming and slitting machine is currently in a high-risk state and may have hidden dangers of failure or performance degradation.

[0057] Furthermore, the present application provides an abnormal environment early warning, including:

[0058] According to the sensor monitoring network, the real-time working environment information corresponding to the high-speed laser film-making and slitting integrated machine is obtained; based on the predetermined working environment information of the equipment, the real-time working environment information is detected for abnormality to obtain abnormal working environment information; according to the abnormal working environment information, an abnormal level evaluation is performed to generate an abnormal environment level; based on the abnormal environment level, an abnormal environment warning is issued to the high-speed laser film-making and slitting integrated machine.

[0059] Preferably, during the operation of the high-speed laser film-making and striping machine, the system terminal obtains the working environment information of the device in real time through the sensor monitoring network deployed, including parameters such as temperature, humidity, and light intensity. This information is crucial to ensure the normal operation of the high-speed laser film-making and striping machine, because any abnormality in environmental parameters may affect the performance and life of the high-speed laser film-making and striping machine. Subsequently, the system terminal compares the real-time acquired working environment information with the predetermined working environment information of the device. Through comparison and analysis, it is identified which environmental parameters have an absolute difference greater than or equal to the environmental deviation threshold with the predetermined environmental parameters, thereby determining the presence of abnormal working environment information. Once abnormal working environment information is detected, the system terminal assigns a weight to each parameter according to the importance of the environmental parameters recorded in the abnormal working environment information, and then uses these weights to perform weighted summation on the absolute difference to obtain a comprehensive abnormal index. Afterwards, the position of this comprehensive abnormal index in the preset level interval is determined to generate the corresponding abnormal environment level. Among them, the preset level interval is set according to historical data, equipment specifications and safety standards. Then, based on this abnormal environment level, the system terminal will issue an abnormal environment warning to the maintenance personnel. The warning information will clearly indicate the specific situation, level and possible consequences of the current environmental abnormality, so that maintenance personnel can respond quickly and take necessary measures to adjust or improve the working environment to ensure the stable operation of the high-speed laser filming and slitting machine.

[0060] In summary, the embodiments of the present application have at least the following technical effects:

[0061] The embodiment of the present application first receives the working parameters and control scheme of the device, and obtains relevant working environment information. Based on this information, the various components of the device are analyzed and decomposed, and the control decision of each component is extracted. Subsequently, the health status of the component is predicted according to the working environment of the device and the control decision of the component, and a health status prediction matrix is ​​generated. This matrix is ​​used to judge the health status of each component. In the actual operation process, the working status of the device is monitored in real time using the sensor monitoring network. According to the real-time monitoring data and the health status prediction matrix, risk assessment is performed to generate the risk coefficient of the component. If the risk coefficient exceeds a predetermined threshold, an early warning signal is generated to indicate that the device may be abnormal. In order to generate the health status prediction matrix, relevant data is extracted from the historical monitoring records, and data cleaning and classification are performed. Then, the concentrated interval of the health status is identified through cluster analysis, and these results are integrated into the health status prediction matrix. In the registration and identification process, the health monitoring records are traversed, sample data is extracted and weighted analysis is performed, and a comprehensive registration coefficient is generated. If the registration coefficient meets the predetermined standard, these data are used for health status prediction. Afterwards, by collating and analyzing the real-time monitoring data, comparing it with the prediction matrix, evaluating the deviation, calculating the risk coefficient, and conducting risk prediction and early warning, the normal operation and maintenance of the equipment are ensured. These technical effects jointly solve the technical problem that there is a lack of effective means to monitor the working status of each component in real time and predict the potential failure risk during the operation of the high-speed laser film-making and striping machine, which leads to the untimely maintenance of the high-speed laser film-making and striping machine. It realizes the health status prediction and risk warning of the key components of the high-speed laser film-making and striping machine, improves the stability and reliability of the operation of the high-speed laser film-making and striping machine, and reduces the failure downtime of the high-speed laser film-making and striping machine.

[0062] Embodiment 2 is based on the same inventive concept as the working state monitoring method of a high-speed laser film-making and slitting machine in the above embodiment. Figure 2As shown, the present application provides a working status monitoring system for a high-speed laser film-making and striping machine, and the system includes: a working parameter receiving module 1: the working parameter receiving module 1 is used to receive the predetermined working parameters of the high-speed laser film-making and striping machine, and obtain the predetermined control scheme of the equipment and the predetermined working environment information of the equipment; a component feature disassembly module 2: the component feature disassembly module 2 is used to disassemble the component features according to the predetermined control scheme of the equipment, and obtain the predetermined control decisions of each component of the high-speed laser film-making and striping machine; a control decision extraction module 3: the control decision extraction module 3 is used to extract the first component predetermined control decision corresponding to the first component of the high-speed laser film-making and striping machine according to the predetermined control decisions of each component; a health prediction module 4: the health prediction module 4 is used to predict the health of the first component according to the predetermined working environment information of the equipment and the predetermined control decision of the first component The first component performs a working status health prediction to obtain a first component health status prediction matrix; real-time monitoring module 5: the real-time monitoring module 5 is used to control the high-speed laser film-making and striping integrated machine based on the predetermined control decisions of each component, and to perform real-time monitoring of the high-speed laser film-making and striping integrated machine through a sensor monitoring network to obtain the working status monitoring information of the first component; risk prediction module 6: the risk prediction module 6 is used to perform risk prediction on the first component based on the first component health status prediction matrix and the first component working status monitoring information, and generate a first component state risk coefficient; abnormal warning module 7: the abnormal warning module 7 is used to generate a first risk warning signal if the first component state risk coefficient is greater than / equal to the predetermined state risk coefficient, and perform an abnormal warning on the high-speed laser film-making and striping integrated machine according to the first risk warning signal.

[0063] Furthermore, the health prediction module 4 is also used to perform the following method:

[0064] According to the high-speed laser film-making and striping integrated machine, health status monitoring records are retrieved to obtain a first health status monitoring record set corresponding to the first component; the first health status monitoring record set is registered and identified according to the predetermined working environment information of the equipment and the predetermined control decision of the first component to obtain a first health status prediction registration space; data cleaning and classification are performed according to the first health status prediction registration space to obtain multiple health status prediction registration domains; concentrated interval identification is performed according to the multiple health status prediction registration domains to obtain multiple health prediction registration identification results; based on the multiple health prediction registration identification results, a health status prediction matrix of the first component is generated.

[0065] Furthermore, the health prediction module 4 is also used to perform the following method:

[0066] Traverse the first health status monitoring record set and extract the first health status monitoring record, wherein the first health status monitoring record includes the first sample working environment information, the first sample control decision and the first healthy sample working status monitoring information corresponding to the first component; perform weighted registration analysis on the first health status monitoring record according to the predetermined working environment information of the equipment and the predetermined control decision of the first component to generate a first sample comprehensive registration coefficient; determine whether the first sample comprehensive registration coefficient is greater than / equal to the predetermined comprehensive registration coefficient; if the first sample comprehensive registration coefficient is greater than / equal to the predetermined comprehensive registration coefficient, add the first healthy sample working status monitoring information to the first health status prediction registration space.

[0067] Furthermore, the health prediction module 4 is also used to perform the following method:

[0068] According to the twin neural network, an environmental registration recognition channel and a control registration recognition channel are constructed; the predetermined working environment information of the device and the first sample working environment information are input into the environmental registration recognition channel to generate a first sample environmental registration coefficient; based on the first component predetermined control decision and the first sample control decision, according to the control registration recognition channel, a first sample control registration coefficient is obtained; based on the weighted registration weight condition, the first sample environmental registration coefficient and the first sample control registration coefficient are weightedly calculated to obtain the first sample comprehensive registration coefficient.

[0069] Furthermore, the risk prediction module 6 is also used to perform the following method:

[0070] Organize the working status monitoring information of the first component and establish a first component status monitoring matrix; perform deviation analysis based on the first component health status prediction matrix and the first component status monitoring matrix to obtain a first component status health deviation analysis result; perform deviation depth evaluation based on the first component status health deviation analysis result to obtain a first component health deviation depth coefficient; determine whether the first component health deviation depth coefficient is greater than / equal to a predetermined health deviation depth coefficient; if the first component health deviation depth coefficient is greater than / equal to the predetermined health deviation depth coefficient, activate a pre-constructed first component risk prediction channel; based on the first component status health deviation analysis result, obtain the first component status risk coefficient according to the first component risk prediction channel.

[0071] Furthermore, the risk prediction module 6 is also used to perform the following method:

[0072] The risk prediction channel of the first component includes K risk prediction models corresponding to the first component, wherein K is a positive integer greater than 1; the health deviation analysis results of the first component state are input into the K risk prediction models to obtain K risk prediction coefficients; a proportion calculation is performed based on K risk prediction accuracy parameters corresponding to the K risk prediction models to obtain K risk prediction incentive coefficients; a weighted calculation is performed on the K risk prediction coefficients based on the K risk prediction incentive coefficients to obtain the first component state risk coefficient.

[0073] Furthermore, the abnormal warning module 7 is also used to execute the following method:

[0074] According to the sensor monitoring network, the real-time working environment information corresponding to the high-speed laser film-making and slitting integrated machine is obtained; based on the predetermined working environment information of the equipment, the real-time working environment information is detected for abnormality to obtain abnormal working environment information; according to the abnormal working environment information, an abnormal level evaluation is performed to generate an abnormal environment level; based on the abnormal environment level, an abnormal environment warning is issued to the high-speed laser film-making and slitting integrated machine.

[0075] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0076] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0077] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A method for monitoring the working status of a high-speed laser film-making and slitting machine, characterized in that: The method comprises: Receive the predetermined working parameters of the high-speed laser film-making and slitting machine, and obtain the predetermined control plan of the equipment and the predetermined working environment information of the equipment; Disassemble the components according to the predetermined control scheme of the equipment to obtain the predetermined control decision of each component of the high-speed laser film-making and slitting machine; Extracting a first component predetermined control decision corresponding to a first component of the high-speed laser film-making and slitting machine according to the predetermined control decisions of each component; Performing working state health prediction on the first component according to the predetermined working environment information of the device and the predetermined control decision of the first component to obtain a first component health state prediction matrix; The high-speed laser film-making and slitting machine is controlled based on the predetermined control decisions of each component, and the high-speed laser film-making and slitting machine is monitored in real time through a sensor monitoring network to obtain working status monitoring information of the first component; Based on the health status prediction matrix of the first component, performing risk prediction on the first component according to the working status monitoring information of the first component, and generating a first component status risk coefficient; If the state risk coefficient of the first component is greater than / equal to the predetermined state risk coefficient, a first risk warning signal is generated, and an abnormal warning is issued to the high-speed laser film-making and slitting machine according to the first risk warning signal; Among them, performing working state health prediction on the first component according to the predetermined working environment information of the device and the predetermined control decision of the first component to obtain the first component health state prediction matrix includes: Performing a health status monitoring record search according to the high-speed laser film-making and striping integrated machine to obtain a first health status monitoring record set corresponding to the first component; Performing registration and identification on the first health status monitoring record set according to the predetermined working environment information of the device and the predetermined control decision of the first component to obtain a first health status prediction registration space; Performing data cleaning and classification according to the first health state prediction registration space to obtain multiple health state prediction registration domains; Performing concentrated interval recognition according to the multiple health status prediction and registration domains to obtain multiple health prediction and registration recognition results; generating a health status prediction matrix of the first component according to the plurality of health prediction registration and identification results; The first health status monitoring record set is registered and identified according to the predetermined working environment information of the device and the predetermined control decision of the first component to obtain a first health status prediction registration space, including: Traversing the first health status monitoring record set, extracting a first health status monitoring record, wherein the first health status monitoring record includes first sample working environment information, first sample control decision and first healthy sample working status monitoring information corresponding to the first component; According to the predetermined working environment information of the device and the predetermined control decision of the first component, performing a weighted registration analysis on the first health status monitoring record to generate a first sample comprehensive registration coefficient; Determining whether the first sample comprehensive registration coefficient is greater than / equal to a predetermined comprehensive registration coefficient; If the first sample comprehensive registration coefficient is greater than / equal to the predetermined comprehensive registration coefficient, adding the first healthy sample working state monitoring information to the first health state prediction registration space; According to the predetermined working environment information of the device and the predetermined control decision of the first component, a weighted registration analysis is performed on the first health status monitoring record to generate a first sample comprehensive registration coefficient, including: According to the twin neural network, the environment registration recognition channel and the control registration recognition channel are constructed; Inputting the predetermined working environment information of the device and the first sample working environment information into the environment registration recognition channel to generate a first sample environment registration coefficient; Based on the first component predetermined control decision and the first sample control decision, obtaining a first sample control registration coefficient according to the control registration identification channel; Performing weighted calculation on the first sample environment registration coefficient and the first sample control registration coefficient based on the weighted registration weight condition to obtain the first sample comprehensive registration coefficient; Wherein, based on the health status prediction matrix of the first component, risk prediction is performed on the first component according to the working status monitoring information of the first component to generate a first component status risk coefficient, including: Arrange the working status monitoring information of the first component and establish a first component status monitoring matrix; Perform deviation analysis according to the first component health status prediction matrix and the first component status monitoring matrix to obtain a first component status health deviation analysis result; Performing a deviation depth evaluation according to the health deviation analysis result of the first component state to obtain a health deviation depth coefficient of the first component; Determining whether the healthy deviation depth coefficient of the first component is greater than / equal to a predetermined healthy deviation depth coefficient; If the health deviation depth coefficient of the first component is greater than / equal to the predetermined health deviation depth coefficient, activating a pre-constructed first component risk prediction channel; Based on the health deviation analysis result of the first component state and according to the first component risk prediction channel, the first component state risk coefficient is obtained.

2. A method for monitoring the working status of a high-speed laser film-making and slitting machine as claimed in claim 1, characterized in that: Based on the first component state health deviation analysis result and according to the first component risk prediction channel, obtaining the first component state risk coefficient includes: The first component risk prediction channel includes K risk prediction models corresponding to the first component, where K is a positive integer greater than 1; Inputting the first component state health deviation analysis result into the K risk prediction models to obtain K risk prediction coefficients; Performing a proportion calculation based on the K risk prediction accuracy parameters corresponding to the K risk prediction models to obtain K risk prediction incentive coefficients; The K risk prediction coefficients are weightedly calculated according to the K risk prediction incentive coefficients to obtain the first component state risk coefficient.

3. A method for monitoring the working status of a high-speed laser film-making and slitting machine as claimed in claim 1, characterized in that: The method comprises: According to the sensor monitoring network, real-time working environment information corresponding to the high-speed laser film-making and slitting machine is obtained; Performing abnormality detection on the real-time working environment information based on the predetermined working environment information of the device to obtain abnormal working environment information; Performing an abnormality level evaluation according to the abnormal working environment information to generate an abnormal environment level; Based on the abnormal environment level, an abnormal environment warning is issued to the high-speed laser film-making and slitting machine.

4. A working status monitoring system for a high-speed laser film-making and striping machine, characterized in that: The steps for implementing the working status monitoring method of a high-speed laser film-making and slitting machine as described in any one of claims 1 to 3 include: Working parameter receiving module: receiving the predetermined working parameters of the high-speed laser film-making and slitting machine, and obtaining the predetermined control scheme of the equipment and the predetermined working environment information of the equipment; Component feature disassembly module: disassembly of component features according to the predetermined control scheme of the equipment to obtain the predetermined control decision of each component of the high-speed laser film-making and slitting machine; A control decision extraction module: extracting a first component predetermined control decision corresponding to a first component of the high-speed laser film-making and slitting integrated machine according to the predetermined control decisions of each component; Health prediction module: performs working state health prediction on the first component according to the predetermined working environment information of the device and the predetermined control decision of the first component, and obtains a first component health state prediction matrix; Real-time monitoring module: controlling the high-speed laser film-making and slitting integrated machine based on the predetermined control decisions of each component, and monitoring the high-speed laser film-making and slitting integrated machine in real time through a sensor monitoring network to obtain the working status monitoring information of the first component; Risk prediction module: based on the health status prediction matrix of the first component and according to the working status monitoring information of the first component, risk prediction is performed on the first component to generate a first component status risk coefficient; Abnormal warning module: if the state risk coefficient of the first component is greater than / equal to the predetermined state risk coefficient, a first risk warning signal is generated, and an abnormal warning is issued to the high-speed laser film-making and slitting machine according to the first risk warning signal.

Images