Intelligent fault diagnosis method, system and device and storage medium

Abstract

The invention provides an intelligent fault diagnosis method, system and device and a storage medium, and relates to the technical field of industrial automation, and the method comprises the steps: obtaining the operation parameters of a plurality of devices in an industrial automation production line, and enabling the devices to be sequentially connected according to a production technology process; inputting the operation parameters of each device into a preset fault detection model to generate a first diagnosis result of each device; obtaining a topological relation between the devices, establishing a device association diagram in combination with a production process flow and the topological relation, and generating association parameters between the devices according to the device association diagram; generating a second diagnosis result of each piece of equipment according to the associated parameters; and combining the first diagnosis result and the second diagnosis result to generate a final fault diagnosis result of each device. The technical effect of the application is that accurate and comprehensive fault diagnosis of the equipment is realized.

Description

Technical Field

[0001] The present application relates to the field of industrial automation technology, and specifically to an intelligent fault diagnosis method, system, device and storage medium. Background Art

[0002] In industrial automation production lines, equipment failures can lead to reduced production efficiency, lower product quality, and even serious consequences such as production line shutdown. Therefore, timely and accurate diagnosis of equipment failures is of great significance to ensure the stable operation of the production line.

[0003] At present, fault diagnosis of industrial automation production lines mainly adopts the single-device diagnosis method, that is, by collecting the operating parameters of a single device, the fault detection model is used to diagnose the fault of the device. However, in the actual production process, there are complex associations between the devices on the production line, and the failure of one device may affect the operating status of other devices associated with it. The traditional single-device diagnosis method ignores the correlation between devices and cannot effectively identify the fault propagation phenomenon caused by the correlation between devices, thus affecting the accuracy of fault diagnosis. Summary of the invention

[0004] The present application provides an intelligent fault diagnosis method, system, device and storage medium for realizing accurate and comprehensive fault diagnosis of a device.

[0005] In a first aspect, the present application provides an intelligent fault diagnosis method, the method comprising: obtaining operating parameters of multiple devices in an industrial automation production line, wherein each of the devices is connected in sequence according to a production process; inputting the operating parameters of each of the devices into a preset fault detection model to generate a first diagnosis result for each of the devices; obtaining a topological relationship between each of the devices, establishing a device association diagram in combination with the production process and the topological relationship, and generating association parameters between each of the devices based on the device association diagram; generating a second diagnosis result for each of the devices based on the association parameters; and generating a final fault diagnosis result for each of the devices by combining the first diagnosis result and the second diagnosis result.

[0006] By adopting the above technical solution, the limitations of a single diagnostic method are effectively overcome by using both the preset fault detection model and the equipment association analysis. On the one hand, by analyzing the operating parameters of each device through the preset fault detection model, the abnormal state of a single device can be quickly identified; on the other hand, by establishing an equipment association diagram that reflects the physical connection and process association between devices, and analyzing based on the association parameters, the propagation path of the fault can be effectively tracked. The fusion of these two diagnostic results not only ensures the accurate identification of single equipment faults, but also provides insight into the fault propagation relationship between devices, significantly improving the accuracy of fault diagnosis.

[0007] Optionally, the method also includes: obtaining historical operation data of each of the devices, the historical operation data including device operation parameters and corresponding fault labels; performing feature extraction on the historical operation data to obtain a fault feature vector; dividing the fault feature vector into a training data set and a verification data set according to a preset ratio; training the training data set to obtain an initial fault detection model; inputting the verification data set into the initial fault detection model for verification to obtain a model verification result; adjusting the initial fault detection model according to the model verification result to obtain a preset fault detection model.

[0008] By adopting the above technical solution, the original operation data is first converted into fault feature vectors through feature extraction, which improves the data characterization ability; then the model training and verification are carried out by separating the training set and the verification set to avoid the overfitting problem; finally, the model is optimized and adjusted based on the verification results to ensure that the preset fault detection model has good generalization ability. This data-driven model construction method makes full use of the experience of historical fault cases, enabling the model to accurately identify abnormal situations similar to historical fault patterns, thereby improving the reliability of fault detection.

[0009] Optionally, the combination of the production process and the topological relationship to establish an equipment association diagram includes: obtaining the process sequence of each device in the production process; determining the material transfer method between each device according to the topological relationship; using each device as a node and the material transfer method as a connecting line to generate an initial association diagram, and adding an information flow identifier to the initial association diagram, wherein the information flow identifier represents the data transmission relationship between each device; marking the process parameters of each device in the initial association diagram; and generating an equipment association diagram based on the process sequence, the material transfer method, the information flow identifier and the process parameters.

[0010] By adopting the above technical solution, the mutual influence relationship between devices can be clearly seen by visually integrating information such as process sequence, material transfer, information flow and process parameters. This association diagram not only reflects the physical connection relationship between devices, but also contains data transmission paths and key process parameter information, providing a complete basis for subsequent fault propagation analysis. For example, when a device is abnormal, the affected downstream equipment can be quickly tracked through the association diagram, and the spread and impact of the fault can be accurately judged in combination with the changes in process parameters.

[0011] Optionally, generating association parameters between the devices according to the device association diagram includes: determining the upstream and downstream relationship between the devices according to the information flow direction identifier in the device association diagram, defining the device located upstream as the source device, and defining the device located downstream as the target device; obtaining the fault type of the source device, and determining the fault propagation parameter corresponding to the fault type according to a preset fault propagation rule; calculating the propagation coefficient of the fault propagation parameter in combination with the material transfer method and process parameters between the source device and the target device; multiplying the fault propagation parameter with the corresponding propagation coefficient to obtain the degree of association between the source device and the target device; traversing all device pairs in the device association diagram, obtaining the degree of association between each pair of the devices, and generating a degree of association matrix; processing the degree of association matrix according to a preset threshold value to generate association parameters between the devices.

[0012] By adopting the above technical solution, the solution realizes the quantitative evaluation of the impact of fault propagation between devices by establishing a systematic correlation parameter calculation method. By defining the relationship between devices as source devices and target devices, and combining fault propagation rules, material transfer methods and process parameters to calculate the propagation coefficient, the evaluation of the fault impact is more accurate. In particular, through the construction of the correlation matrix, the mutual influence between all pairs of devices is quantified, and through the screening of preset thresholds, the correlation relationships with significant influence are retained. This quantitative analysis method based on multiple factors can accurately identify the critical path and impact of fault propagation. For example, in an injection molding production line, when the hopper dryer has a temperature abnormality, this method can accurately calculate the impact on the downstream injection molding machine, making the fault warning more timely and accurate.

[0013] Optionally, generating a second diagnostic result for each of the devices based on the association parameters includes: inputting the association parameters into a preset device association model to obtain a device fault impact matrix; calculating the fault association strength between each of the devices based on the device fault impact matrix; comparing the fault association strength with a preset fault propagation threshold, if the fault association strength is greater than the preset fault propagation threshold, determining that a fault association exists between the corresponding devices; generating a second diagnostic result for each of the devices based on the fault association and in combination with preset fault diagnosis rules.

[0014] By adopting the above technical solution, the fault propagation relationship between devices can be intuitively displayed by converting the correlation parameters into the device fault impact matrix; and based on the comparison of the fault correlation strength with the preset threshold, the weakly correlated interference items are effectively filtered out, and the strongly correlated fault propagation links are highlighted. This fault correlation determination method based on quantitative analysis, combined with the preset fault diagnosis rules, can accurately identify the source device of the fault and the affected device.

[0015] Optionally, the combination of the first diagnostic result and the second diagnostic result to generate a final fault diagnosis result for each of the devices includes: obtaining a first fault probability of each of the devices in the first diagnostic result, and a second fault probability of each of the devices in the second diagnostic result; weighted summing the first probability and the second probability according to a preset weight coefficient to obtain a comprehensive fault probability of each of the devices; comparing the comprehensive fault probability with a preset fault judgment threshold, and if there is a first device whose comprehensive fault probability is not less than the preset fault judgment threshold, determining that the first device has a fault; if there is a second device whose comprehensive fault probability is less than the preset fault judgment threshold, determining that the second device does not have a fault.

[0016] By adopting the above technical solution, the first fault probability and the second fault probability are weighted by the preset weight coefficient, which fully considers the advantages and reliability of the two diagnostic methods; and the comparison mechanism based on the preset fault judgment threshold provides an objective quantitative standard for fault judgment. This fusion diagnosis method not only retains the sensitivity of the preset fault detection model to single device anomalies, but also takes into account the recognition ability of device association analysis to fault propagation, which significantly improves the reliability of the diagnostic results.

[0017] Optionally, after generating the final fault diagnosis result of each of the devices, it also includes: generating a fault repair instruction according to the final fault diagnosis result of each of the devices, and sending the fault repair instruction to the terminal device of the maintenance personnel so that the maintenance personnel can perform maintenance.

[0018] By adopting the above technical solution, the final fault diagnosis result is automatically converted into a fault repair instruction and directly pushed to the maintenance personnel's terminal device, which significantly shortens the time interval from fault discovery to maintenance response and improves maintenance efficiency.

[0019] In a second aspect, the present application provides an intelligent fault diagnosis system, which includes: an acquisition module, a first diagnosis module, a generation module, a second diagnosis module and a combination module; wherein the acquisition module is used to acquire the operating parameters of multiple devices in an industrial automation production line, wherein each of the devices is connected in sequence according to a production process; the first diagnosis module is used to input the operating parameters of each of the devices into a preset fault detection model to generate a first diagnosis result for each of the devices; the generation module is used to acquire the topological relationship between each of the devices, establish a device association diagram in combination with the production process and the topological relationship, and generate association parameters between each of the devices based on the device association diagram; the second diagnosis module is used to generate a second diagnosis result for each of the devices based on the association parameters; the combination module is used to combine the first diagnosis result and the second diagnosis result to generate a final fault diagnosis result for each of the devices.

[0020] In the third aspect, the present application provides an electronic device, adopting the following technical solution: including a processor, a memory, a user interface and a network interface, the memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes a computer program such as any one of the above-mentioned intelligent fault diagnosis methods.

[0021] In a fourth aspect, the present application provides a computer-readable storage medium, which adopts the following technical solution: storing a computer program that can be loaded by a processor and execute any of the above-mentioned intelligent fault diagnosis methods.

[0022] In summary, the present application includes at least one of the following beneficial technical effects: 1. By using both the preset fault detection model and equipment association analysis, the limitations of a single diagnostic method are effectively overcome. On the one hand, by analyzing the operating parameters of each device through the preset fault detection model, the abnormal state of a single device can be quickly identified; on the other hand, by establishing an equipment association diagram that reflects the physical connection and process association between devices, and analyzing based on the association parameters, the propagation path of the fault can be effectively tracked. The fusion of these two diagnostic results not only ensures the accurate identification of single device faults, but also provides insight into the fault propagation relationship between devices, significantly improving the accuracy of fault diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a flowchart of an intelligent fault diagnosis method provided by an embodiment of the present application; Figure 2 It is a structural diagram of an intelligent fault diagnosis system provided in an embodiment of the present application; Figure 3It is a structural schematic diagram of an electronic device provided in an embodiment of the present application.

[0024] Description of reference numerals: 1000, electronic device; 1001, processor; 1002, communication bus; 1003, user interface; 1004, network interface; 1005, memory. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0026] In the description of the embodiments of the present application, words such as "illustrative", "for example" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "illustrative", "for example" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "illustrative", "for example" or "for example" is intended to present related concepts in a concrete way.

[0027] As the degree of industrial automation continues to increase, the correlation between equipment in modern production lines is becoming stronger and stronger, and the complexity of equipment fault diagnosis is also increasing. At present, the fault diagnosis of industrial automation production lines mainly has the following problems: Traditional fault diagnosis methods mainly rely on parameter monitoring and threshold judgment of a single device. Although this method is simple to operate, it cannot effectively identify faults caused by correlations between devices, resulting in insufficient accuracy of diagnostic results. For example, a minor anomaly in an upstream device may cause a failure in a downstream device, but diagnosing only the downstream device often fails to find the root cause of the failure.

[0028] Although some existing intelligent fault diagnosis systems use machine learning and other technologies to identify faults, they mainly focus on the operating status of a single device and ignore the correlation between devices. This makes it difficult to accurately determine the propagation path and impact range of faults in complex production processes.

[0029] Equipment failures in production lines have propagation and chain reaction characteristics. When a key device fails, it may affect the operating status of adjacent equipment through material transfer or process parameter changes. Existing technologies lack systematic analysis and modeling of this fault propagation mechanism.

[0030] A large amount of equipment operation data and fault records are accumulated during industrial production, but these historical data have not been fully utilized. Existing fault diagnosis methods often lack in-depth mining of historical data and cannot effectively extract fault characteristics and laws.

[0031] Due to the lack of unified fault diagnosis standards and models, it is difficult to effectively integrate the fault diagnosis results of different devices, which affects the overall diagnosis effect. At the same time, the diagnostic results are not interpretable enough, making it difficult to provide clear guidance for equipment maintenance.

[0032] Therefore, there is an urgent need for an intelligent fault diagnosis method that can comprehensively consider the independent operation status of equipment and the correlation between equipment, so as to improve the accuracy and efficiency of fault diagnosis in industrial automation production lines.

[0033] Figure 1 FIG. 1 is a flow chart of an intelligent fault diagnosis method provided by an embodiment of the present application. Figure 1 As shown, the method includes S101-S105: S101, obtaining operating parameters of multiple devices in an industrial automation production line, wherein each device is connected in sequence according to a production process flow.

[0034] In the specific implementation process, it is first necessary to obtain the operating parameters of multiple devices in the industrial automation production line. Since modern industrial automation production lines usually contain multiple interrelated devices, these devices are connected in sequence according to the requirements of the production process to form a complete production line. In order to realize intelligent fault diagnosis of the production line, the system collects the operating parameters of the equipment in real time through sensors deployed on each device. These operating parameters include but are not limited to physical quantities such as temperature, pressure, vibration, speed, current, voltage, etc. of the equipment, as well as process parameters such as product quality parameters and equipment status codes.

[0035] For example, for an injection molding machine, its operating parameters include mold temperature, injection pressure, holding time, etc.; for a conveyor belt, its operating parameters include transmission speed, motor current, tension, etc.; for a packaging machine, its operating parameters include sealing temperature, packaging speed, air pressure, etc. The system adopts a distributed data acquisition architecture to transmit the data collected by each sensor to the central control system through industrial Ethernet or fieldbus. To ensure the accuracy and real-time nature of the data, the system uses high-precision sensors for parameter acquisition, and the sampling frequency is adaptively adjusted according to the changing characteristics of different parameters, usually at the millisecond level. At the same time, the system pre-processes the collected raw data, including outlier filtering, data standardization and other operations to ensure the reliability of subsequent analysis. By establishing an equipment operating parameter database, the system can grasp the operating status of each equipment on the production line in real time, providing a data basis for subsequent fault diagnosis.

[0036] This parameter collection method based on the complete production line can not only reflect the operating status of a single device, but also reflect the correlation between devices, which helps to discover fault problems caused by mutual influence between devices.

[0037] S102, inputting the operating parameters of each device into a preset fault detection model to generate a first diagnosis result of each device.

[0038] After obtaining the equipment operating parameters, it is necessary to perform a preliminary diagnosis on the operating status of each device. Due to the variety of equipment types and complex failure modes in industrial automation production lines, it is difficult to accurately identify potential faults simply by relying on fixed threshold judgment or manual experience. Therefore, this embodiment uses a preset fault detection model to analyze the equipment operating parameters and generate a first diagnosis result for each device.

[0039] Specifically, the preset fault detection model adopts a deep learning network structure, including a feature extraction layer, a fault feature learning layer, and a fault classification layer. The feature extraction layer uses a multi-layer convolutional neural network to automatically extract time domain features and frequency domain features from the original operating parameters; the fault feature learning layer uses a long short-term memory network (LSTM) to capture the time dependency of the parameter sequence; the fault classification layer outputs the probability distribution of the fault type through a fully connected layer and a softmax function.

[0040] In actual applications, the system first divides the input operating parameters into time windows and standardizes the data to ensure that the data format meets the model input requirements. Subsequently, the processed data is calculated through each level of the model in turn. For example, for the operating parameters of the injection molding machine, the model can simultaneously analyze multi-dimensional features such as the mold temperature curve and the injection pressure change trend, and identify potential faults such as mold temperature out of control and injection pressure abnormalities. The output results of the model include fault type judgment and corresponding fault probability, which constitute the first diagnostic result. Through the analysis of the preset fault detection model, the system can quickly and accurately identify abnormal conditions at the level of individual equipment, providing a basis for subsequent correlation analysis.

[0041] The method for constructing a preset fault detection model specifically includes: Since the equipment failure modes in industrial automation production lines are complex and diverse, it is difficult to build accurate fault diagnosis rules relying solely on expert experience. Therefore, a data-driven approach is adopted to achieve fault detection by training a deep learning model with historical data.

[0042] The system first obtains the historical operating data of each device from the database of the equipment management system. These data include the operating parameters of the equipment under different working conditions, such as time series data of physical quantities such as temperature, pressure, vibration, and fault tags confirmed by maintenance records.

[0043] For example, for injection molding machines, historical data contains complete parameter records under normal operation and various fault conditions, and each record is marked with a specific fault type or normal state label. In order to extract fault features in the data, the system adopts a multi-scale feature extraction method, including time domain features (statistics such as mean, variance, peak value, etc.), frequency domain features (spectral analysis, wavelet transform coefficients, etc.), and time-frequency domain features. These features are organized into feature vectors of fixed dimensions, each of which corresponds to a complete operating state sample. Considering the need for generalization performance evaluation of the model, the system randomly divides the feature vectors into training data sets and validation data sets according to a preset ratio of 8:2, while ensuring that the distribution of samples of different fault types in the two data sets is basically the same. The initial fault detection model is trained using the training data set, and the model adopts a deep learning network structure, including a feature extraction layer, a fault feature learning layer, and a fault classification layer.

[0044] During the training process, the batch gradient descent algorithm is used to optimize the model parameters, and the cross entropy loss function is used to measure the difference between the model prediction results and the true labels. After the training is completed, the validation data set is input into the model for testing to obtain the model validation results, including performance indicators such as accuracy, precision, and recall. Based on the validation results, the system optimizes the model structure and parameters, such as increasing or decreasing the number of network layers, adjusting the learning rate, adding regularization terms, etc., and repeatedly trains and validates until the model performance meets the preset indicator requirements, and finally obtains the preset fault detection model. This model training method based on historical data can make full use of equipment operation experience to build a fault detection model with strong generalization ability, providing reliable technical support for subsequent real-time fault diagnosis.

[0045] S103, obtaining the topological relationship between the devices, establishing a device association diagram in combination with the production process and the topological relationship, and generating association parameters between the devices according to the device association diagram.

[0046] First, the system obtains the physical connection relationship between each device through the communication topology of the industrial fieldbus or industrial Ethernet, and reads the production process information from the production management system. Since equipment failures in industrial automation production lines often have propagation and chain reaction characteristics, it is difficult to accurately locate the source of the failure and predict the scope of the failure by only considering the operating status of a single device. Therefore, the system constructs an equipment association diagram based on the obtained topological relationship and production process.

[0047] In the equipment association diagram, each device is represented as a node, and the material transfer channels between devices are represented as directed connecting lines. For example, in an injection molding production line, the injection molding machine, conveyor belt, testing equipment, etc. are connected according to the process sequence, and the materials flow in a specific direction. The system marks the material transfer method of each connecting line in the association diagram, such as mechanical transmission, pneumatic transmission or hydraulic transmission, and adds information flow direction identification to indicate the data communication relationship between devices. At the same time, the key process parameters of the equipment, such as temperature, pressure, speed, etc., are marked at each node. Based on the complete equipment association diagram, the system further analyzes the association parameters between devices.

[0048] First, determine the upstream and downstream relationship of the equipment, define the upstream equipment as the source equipment, and the downstream equipment as the target equipment. Then, according to the pre-defined fault propagation rules, analyze the propagation characteristics of different types of faults, such as mechanical vibration faults propagating through the connection structure, temperature anomalies propagating through heat conduction, etc. Combined with the material transfer mode and process parameters, the system calculates the fault propagation coefficient, which reflects the attenuation degree of the fault propagating from the source equipment to the target equipment. By multiplying the fault propagation parameter with the propagation coefficient, the correlation between the equipment is obtained, and finally a complete correlation matrix is ​​generated. The system applies a preset threshold to filter the correlation matrix to obtain the final equipment-to-equipment correlation parameter, which is a numerical indicator used to quantify the degree of influence of a device's fault or abnormality on the operating status of other equipment. For example, if the value of the correlation parameter of equipment A to equipment B is 0.8, it means that the fault of equipment A will be transmitted to equipment B with an influence of 80%. This analysis method based on the equipment association diagram can effectively capture the fault propagation law between equipment and provide reliable data support for subsequent associated fault diagnosis.

[0049] Based on the above embodiment, as an optional implementation, in S103, combining the production process flow and the topological relationship, establishing the equipment association diagram specifically includes S301-S305: S301, obtaining the process sequence of each device in the production process.

[0050] S302: Determine the material transfer method between the devices according to the topological relationship.

[0051] S303, taking each device as a node and the material transfer method as a connecting line, an initial association diagram is generated, and an information flow mark is added to the initial association diagram, where the information flow mark represents the data transmission relationship between each device.

[0052] S304, marking the process parameters of each device in the initial association diagram.

[0053] S305, generating an equipment association diagram according to the process sequence, material transfer method, information flow identification and process parameters.

[0054] In industrial automation production lines, there are complex process connections and information interaction relationships between equipment. In order to accurately describe and analyze these associations, a complete equipment association diagram needs to be constructed. First, the system obtains detailed information on the production process from the production management system and determines the process sequence of each equipment in the entire production process. For example, in an injection molding production line, the process sequence may be: raw material conveying equipment → drying equipment → injection molding machine → conveyor belt → testing equipment → packaging equipment. This sequence relationship reflects the basic process of product processing. On the basis of determining the process sequence, the system analyzes the material transfer method between each device according to the physical layout and connection method of the equipment. Material transfer methods include mechanical transmission (such as conveyor belts, robotic arms), pneumatic transmission (such as pneumatic conveying pipelines), manual transmission, etc. For example, a vacuum conveying system may be used to transfer raw materials from drying equipment to injection molding machines, while finished products are automatically transferred from injection molding machines to conveyor belts through robotic arms.

[0055] Based on the information about the process sequence and the material transfer method, the system starts to build an initial association diagram. In this diagram, each device is represented as a node, and the material transfer method is represented by different types of connecting lines. In order to more completely describe the association relationship between devices, the system adds information flow identifiers to the initial association diagram. These identifiers reflect the data transmission and control signal transmission relationship between devices. For example, an injection molding machine sends operating parameters such as temperature and pressure to the central control system, and receives material status information from the upstream drying equipment. These information flows usually do not completely overlap with the material flows, because some devices may only have information exchange but no material transfer.

[0056] After adding the information flow identifier, the system will mark the process parameters for each equipment node in the association diagram. These process parameters include the key operating parameters of the equipment (such as temperature, pressure, speed, etc.) and process control parameters (such as processing time, process accuracy requirements, etc.). For example, the injection molding machine node will mark parameters such as mold temperature range, injection pressure setting value, and holding time. These parameter information is of great significance for understanding the equipment operation status and fault diagnosis.

[0057] Finally, the system integrates all information such as process sequence, material transfer method, information flow identification and process parameters to generate the final equipment association diagram. This association diagram not only shows the physical connection and logical relationship between the equipment, but also contains detailed process parameters and information interaction details. For example, through the association diagram, you can clearly see the material flow direction, control signal transmission path, and process parameter requirements of each key node between the injection molding machine and its adjacent equipment. This multi-dimensional association diagram can provide comprehensive reference information for fault diagnosis and equipment maintenance, helping technicians better understand the mutual influence relationship between equipment.

[0058] Based on the above embodiment, as an optional implementation, in S103, generating association parameters between devices according to the device association diagram specifically includes S401-S406: S401, according to the information flow direction identifier in the device association diagram, determine the upstream and downstream relationship between the devices, define the upstream device as the source device, and define the downstream device as the target device.

[0059] S402, obtaining the fault type of the source device, and determining a fault propagation parameter corresponding to the fault type according to a preset fault propagation rule.

[0060] S403, calculating a propagation coefficient of a fault propagation parameter in combination with a material transfer mode and process parameters between the source device and the target device.

[0061] S404: multiply the fault propagation parameter by the corresponding propagation coefficient to obtain the correlation between the source device and the target device.

[0062] S405, traverse all device pairs in the device association graph, obtain the corresponding association degrees between each device, and generate an association degree matrix.

[0063] S406, processing the correlation matrix according to a preset threshold value to generate correlation parameters between the devices.

[0064] In the intelligent fault diagnosis system, in order to accurately quantify the fault impact relationship between devices, it is necessary to generate scientific and reasonable association parameters based on the device association diagram. The system first determines the upstream and downstream relationship between devices based on the information flow direction identifier in the association diagram, defines the source end of the information flow as the source device, and the receiving end as the target device.

[0065] For example, in an injection molding production line, the drying equipment is used as the source equipment, and the downstream injection molding machine is used as the target equipment. After determining the equipment relationship, the system obtains the possible fault types of the source equipment from the fault knowledge base, and determines the fault propagation parameters corresponding to each fault type according to the preset fault propagation rules. The fault propagation rules are established based on expert experience and historical data, and describe the propagation characteristics of different types of faults. For example, the temperature control failure of the drying equipment may cause abnormal moisture content of the material, and its fault propagation parameter may be set to 0.8, indicating that this type of fault has strong propagation.

[0066] In order to more accurately describe the propagation characteristics of faults in the actual process environment, the system needs to calculate the propagation coefficient based on the material transfer method and process parameters between the source device and the target device. Different material transfer methods have different propagation characteristics. For example, the propagation coefficient of direct mechanical transmission is larger, while the propagation coefficient through the cache device is smaller. At the same time, the matching degree of process parameters will also affect the propagation coefficient. For example, when the process parameter requirements of the source device and the target device are closely related, the propagation coefficient is larger. By establishing a propagation coefficient calculation model, the system can calculate the corresponding propagation coefficient according to the specific conditions of the material transfer method and process parameters. For example, in the case where the drying equipment and the injection molding machine are connected by a vacuum conveying system, considering the short conveying distance and the close process requirements, the propagation coefficient may be calculated as 0.9.

[0067] By multiplying the fault propagation parameter with the propagation coefficient, the system obtains the correlation between the source device and the target device. This correlation comprehensively reflects the physical connection characteristics and process correlation between the devices. By traversing all possible pairs of devices in the device association diagram, the system calculates the correlation between each pair of devices and organizes these correlations into a correlation matrix. In a system with n devices, the correlation matrix is ​​an n×n square matrix, where the matrix element aij represents the degree of influence of device i on device j. For example, in a simple system containing drying equipment, injection molding machines, and conveyor belts, the correlation matrix may be as follows: [1.0 0.72 0.22] [0.15 1.0 0.63] [0.08 0.27 1.0] Finally, the system processes the correlation matrix according to the preset threshold and sets the correlation below the threshold to 0 to reduce the computational complexity and highlight the important correlation relationship. The processed correlation matrix is ​​the final correlation parameter, which will be used to analyze the impact of fault propagation between devices in the subsequent fault diagnosis process.

[0068] S104: Generate a second diagnosis result for each device according to the associated parameters.

[0069] After obtaining the correlation parameters between devices, the system needs to further analyze the correlation of device failures and generate a second diagnostic result that takes into account the mutual influence between devices. Since device failures in industrial automation production lines often do not occur in isolation, the abnormality of a single device may affect the operating status of other devices through process flow or physical connection, so the operating status of the equipment needs to be re-evaluated based on the correlation parameters.

[0070] The system first converts the associated parameters into a weight matrix, where each element in the matrix represents the degree of impact of the source device failure on the target device. For example, in an injection molding production line, abnormal temperature control of the injection molding machine will directly affect the product quality on the downstream conveyor belt. This impact relationship is quantified by the size of the associated parameters. The system uses a fault propagation model based on a graph neural network, taking the initial state information and associated parameters of the device as input to simulate the propagation process of the fault in the device network.

[0071] Specifically, the model first initializes the state vector of each device node based on the first diagnostic result of each device node; then, through multiple rounds of message passing mechanisms, each node is able to aggregate information from adjacent nodes. In the message passing process, the importance of information is determined by the association parameter, and a larger association parameter means a stronger information transmission intensity. After multiple rounds of iterative calculations, the model outputs the updated state of each device node. This state information reflects the fault state after considering the correlation between devices. The system converts the updated state information into a specific fault type and fault probability through a decoder network, thereby generating a second diagnostic result. Compared with the first diagnostic result, the second diagnostic result not only takes into account the operating parameters of the device itself, but also includes the impact information from related devices, which can more accurately reflect the actual operating status of the device.

[0072] For example, when the failure probability of a device is found to be significantly higher than the first diagnosis result, it may indicate that the failure is caused by abnormal propagation of upstream devices. This diagnostic method based on correlation analysis can help maintenance personnel better understand the propagation path of the fault and take more targeted maintenance measures.

[0073] Based on the above embodiment, as an optional implementation, in S104, generating the second diagnosis result of each device according to the associated parameters specifically includes S501-S504: S501, inputting association parameters into a preset equipment association model to obtain an equipment failure impact matrix.

[0074] S502: Calculate the fault correlation strength between the devices according to the device fault impact matrix.

[0075] S503, comparing the fault association strength with a preset fault propagation threshold, if the fault association strength is greater than the preset fault propagation threshold, determining that a fault association exists in the corresponding device.

[0076] S504: Generate a second diagnosis result for each device according to the fault association and preset fault diagnosis rules.

[0077] In the fault diagnosis process of industrial automation production lines, due to the complex associations between devices, the failure of a single device may trigger a chain reaction, causing abnormalities in related devices. In order to accurately identify this fault propagation phenomenon, the system needs to perform in-depth analysis based on association parameters.

[0078] First, the system inputs the aforementioned generated association parameters into the preset device association model. This model uses a graph neural network structure, which can handle complex topological relationships between devices and consider the timing characteristics of fault propagation. Through model calculation, the system obtains the device fault impact matrix, which not only includes the direct fault propagation impact, but also considers the cumulative effect of the indirect propagation path.

[0079] For example, in an injection molding production line, the failure of the drying equipment may affect the injection molding machine through material quality, and then affect the subsequent product quality inspection equipment. This chain propagation effect will be reflected in the fault impact matrix.

[0080] Based on the fault impact matrix, the system uses a weighted network analysis method to calculate the fault correlation strength between each device. The calculation of fault correlation strength takes into account multiple factors, including direct impact strength, number of indirect impact paths, propagation attenuation coefficient, etc. The system uses an improved PageRank algorithm to iteratively calculate the fault impact to obtain a more accurate correlation strength value.

[0081] For example, for the fault correlation strength between the drying equipment and the injection molding machine, not only the direct correlation parameters between them should be considered, but also the indirect impact path formed through other equipment. If a propagation path passes through multiple intermediate devices, its impact strength will decay as the propagation distance increases.

[0082] The system compares the calculated fault correlation strength with the preset fault propagation threshold. This threshold is determined based on historical fault data and expert experience and is used to screen out fault correlation relationships with significant impact. When the fault correlation strength between a pair of devices exceeds the preset threshold, the system determines that there is a significant fault correlation between the two devices.

[0083] For example, if the fault correlation strength between the drying equipment and the injection molding machine is 0.85, and the preset threshold is 0.7, the system will determine that there is a fault correlation between the two devices, and it is necessary to focus on the mutual influence between them during fault diagnosis.

[0084] Finally, the system generates the second diagnosis result for each device based on the determined fault association and the preset fault diagnosis rules. The fault diagnosis rules include the propagation mode of equipment faults, typical fault characteristics, and corresponding processing strategies. When generating the second diagnosis result, the system not only considers the fault status of the equipment itself, but also the fault impact of related equipment. For example, when an injection molding machine is abnormal, the system will trace the status of the upstream drying equipment and analyze whether there is a fault propagation, so as to provide a more accurate fault diagnosis result.

[0085] The preset equipment association model is a mathematical model used to analyze and predict the fault propagation relationship between industrial equipment. Its construction process mainly includes the following key steps: First, based on the process flow and equipment layout of the production line, a network of physical connections between devices is established. This network reflects the direct connection between devices. For example, in an injection molding production line, the drying equipment is connected to the injection molding machine through a material conveying system, and the injection molding machine is connected to the subsequent cooling system through a conveyor belt. This physical connection constitutes the basic channel for fault propagation.

[0086] Secondly, collect and organize equipment operation data and historical fault records. Operation data includes process parameters (such as temperature, pressure, speed, etc.), status information (such as startup time, load rate, etc.) and quality inspection data of each equipment. Historical fault records include information such as fault type, occurrence time, duration, and impact range. These data are important basis for building models.

[0087] Next, analyze the process correlation between equipment. Process correlation is reflected in two aspects: parameter transfer and quality impact. For example, the temperature and humidity of the drying equipment directly affect the material quality, which in turn affects the molding quality of the injection molding machine. By analyzing the correlation and change rules of process parameters, the strength of process correlation between equipment can be determined.

[0088] Then, a time series correlation model is established. This model considers the time characteristics of fault propagation, including propagation delay and cumulative effect. For example, temperature anomalies in drying equipment may cause defects in injection molded products only after a certain period of time, and this effect may intensify over time. By analyzing the time series characteristics of faults in historical data, a correlation model that reflects time dependence can be established.

[0089] Finally, the above analysis results are integrated to build a comprehensive correlation model. This model unifies physical connections, process correlations, and timing characteristics to form a complete fault propagation prediction framework. The input of the model includes the current status of the equipment, operating parameters, and historical data, and the output is the prediction of the impact of faults between devices.

[0090] The validation and optimization of the model is an ongoing process. By comparing the model prediction results with the actual failure cases, the model parameters and structure are continuously adjusted to improve the prediction accuracy. For example, if the model predicts that the failure of the drying equipment will affect the quality of the injection molding machine within 2 hours, and the actual situation is 3 hours, the timing parameters need to be adjusted accordingly.

[0091] Taking the injection molding production line of a smart manufacturing factory as an example, the construction and application process of the preset equipment association model is explained in detail: the production line mainly includes the following core equipment: hopper dryer, injection molding machine, mold temperature controller, manipulator, conveyor belt and online detection equipment. Based on the production process, these equipment form a closely related production system.

[0092] First, the basic model is established by analyzing the physical connection relationship of the equipment. For example, the hopper dryer conveys the dried plastic particles to the injection molding machine hopper through the vacuum conveying system; the injection molding machine is connected to the mold temperature controller through the water circuit and the electronic control system; the molded product is taken out by the robot and placed on the conveyor belt; and finally, it is conveyed to the online detection equipment through the conveyor belt for quality inspection.

[0093] Secondly, collect the key operating parameters of each device. Taking the hopper dryer as an example, the main monitoring parameters include: drying temperature (155±5℃), dew point temperature (-40±5℃), heating power (45kW), material residence time (4 hours), etc. The injection molding machine monitors parameters such as injection pressure (85MPa), holding time (5s), mold temperature (85℃), cooling time (25s), etc. These parameters constitute the basic data of the equipment's operating status.

[0094] By analyzing six months of production data, a typical case of fault propagation was found: when the heating system of the hopper dryer fluctuated, causing the drying temperature to drop to 140°C, the products produced by the injection molding machine began to have surface defects about 2 hours later. Specifically, the moisture content of the material increased (from 0.02% to 0.05%), silver threads appeared during the injection molding process (due to the high moisture content of the material), watermarks appeared on the surface of the product, and the defective product detection rate of the online inspection equipment increased from 0.5% to 2.8%.

[0095] Based on this type of historical cases, the process correlation strength index between equipment was established: hopper dryer → injection molding machine: correlation strength 0.85; injection molding machine → robot: correlation strength 0.65; injection molding machine → online detection: correlation strength 0.78; mold temperature controller → injection molding machine: correlation strength 0.72.

[0096] In actual applications, the model successfully predicted many fault propagation events. For example, when the dew point temperature of the hopper dryer was detected to be abnormal (increased to -30°C), the model immediately calculated the fault propagation risk: it predicted that the quality of the injection molded product would be affected in 2-3 hours, the degree of impact was expected to cause the defective rate to increase by more than 2%, and it was recommended to adjust the injection molding machine process parameters or perform preventive maintenance.

[0097] Based on the early warning information, the maintenance personnel promptly checked the refrigeration system of the dryer and found that there was a refrigerant leakage problem in the dryer.

[0098] S105, combining the first diagnosis result and the second diagnosis result to generate a final fault diagnosis result of each device.

[0099] After obtaining the first diagnostic result that reflects the independent operation status of the equipment and the second diagnostic result that considers the correlation of the equipment, the system needs to fuse these two types of diagnostic information to generate a more comprehensive and accurate final fault diagnosis result. Since the first diagnostic result is mainly based on the operating parameters of the equipment itself, and the second diagnostic result reflects the mutual influence between the equipment, these two types of results have their own characteristics and complement each other, so a reasonable fusion strategy needs to be adopted.

[0100] The system first builds a multi-source information fusion model based on the attention mechanism, which can adaptively adjust the weights of the two types of diagnosis results. Specifically, for each device, the system calculates the confidence score of the first diagnosis result and the second diagnosis result, and the confidence score is determined according to the probability distribution characteristics of the diagnosis result and the historical diagnosis accuracy.

[0101] For example, when the first diagnostic result shows that a certain injection molding machine has a temperature anomaly with a failure probability of 0.85, and the second diagnostic result shows that the temperature anomaly may be caused by upstream equipment and the failure probability is 0.65, the system will weighted average the failure probability according to the confidence of the two diagnostic results. The fusion model dynamically calculates the importance weights of the two types of diagnostic results through the attention layer. The calculation of the weights takes into account the consistency of the diagnostic results, the operating environment of the equipment, and historical diagnostic experience. When there are significant differences between the two types of diagnostic results, the system will further analyze the causes of the differences, such as checking whether the fault propagation path is blocked. After determining the weights, the system weightedly fuses the two types of diagnostic results to generate the final fault type judgment and fault probability.

[0102] The final fault diagnosis results include not only the fault type and probability information, but also the possible propagation path and impact range of the fault. For example, the system may output: "There is an abnormality in the temperature control system of the injection molding machine (fault probability 0.78), which may be caused by a fault in the feeding system and may affect the subsequent molding quality." This fusion diagnosis method can effectively combine the advantages of independent operation status of equipment and associated impact analysis, provide more accurate and instructive diagnosis results, and help maintenance personnel quickly locate the source of the fault and take targeted maintenance measures.

[0103] Based on the above embodiment, as an optional implementation, in S105, combining the first diagnosis result and the second diagnosis result to generate the final fault diagnosis result of each device specifically includes S601-S604: S601, obtaining a first failure probability of each device in a first diagnosis result and a second failure probability of each device in a second diagnosis result.

[0104] S602: Perform a weighted summation of the first probability and the second probability according to a preset weight coefficient to obtain a comprehensive failure probability of each device.

[0105] S603, comparing the comprehensive failure probability with a preset failure judgment threshold, if the comprehensive failure probability of the first device is not less than the preset failure judgment threshold, determining that the first device has a failure.

[0106] S604: If the comprehensive probability of the existence of a second device failure is less than a preset failure determination threshold, it is determined that the second device does not fail.

[0107] In the specific implementation, first obtain the fault probabilities obtained by the two diagnostic methods. Taking the hopper dryer as an example, the first diagnostic result based on the preset model shows that its fault probability is 0.82 (first fault probability), while the second diagnostic result based on equipment association analysis shows that its fault probability is 0.75 (second fault probability). Similarly, obtain the two types of fault probability values ​​for other equipment such as injection molding machines and mold temperature controllers.

[0108] Considering that the preset model has a strong mathematical basis and historical data support, and equipment association analysis can better reflect the real-time production status, they are given weight coefficients of 0.6 and 0.4 respectively when calculating the comprehensive failure probability.

[0109] For example, for the hopper dryer, the comprehensive failure probability is calculated as: 0.82×0.6+0.75×0.4=0.792. This weighted calculation method not only retains the theoretical advantages of the preset model, but also takes into account the real-time nature of equipment association analysis.

[0110] In order to accurately determine whether the equipment is faulty, the system sets a preset fault determination threshold of 0.75. When the comprehensive fault probability of a certain equipment is not less than the threshold, the equipment is deemed to be faulty. In the above case, the comprehensive fault probability of the hopper dryer is 0.792, which is greater than the preset threshold of 0.75, so it is determined to be faulty. On the contrary, if the comprehensive fault probability of the mold temperature controller is 0.45, which is less than the preset threshold, it is determined to be in normal operation.

[0111] After generating the final fault diagnosis results for each device, it also includes: According to the final fault diagnosis result of each device, a fault repair instruction is generated and sent to the terminal device of the maintenance personnel so that the maintenance personnel can perform maintenance.

[0112] In order to achieve rapid response and efficient handling of faults, the system automatically generates targeted maintenance instructions based on the final fault diagnosis results. Taking the hopper dryer fault as an example, the system integrates information such as fault type, fault location, and recommended maintenance plan into standardized maintenance instructions, which are directly pushed to the mobile terminal of the nearest maintenance personnel through the factory MES system. The maintenance instructions include specific fault parameters (such as dew point temperature -30℃ beyond the normal range), possible fault causes (refrigeration system leakage) and recommended maintenance steps.

[0113] Based on the above method, the present application also discloses an intelligent fault diagnosis system, such as Figure 2 As shown, Figure 2 is a structural diagram of an intelligent fault diagnosis system provided in an embodiment of the present application, the system comprises: an acquisition module, a first diagnosis module, a generation module, a second diagnosis module and a combination module; wherein, An acquisition module is used to acquire the operating parameters of multiple devices in an industrial automation production line, wherein the devices are connected in sequence according to a production process flow; a first diagnosis module is used to input the operating parameters of each device into a preset fault detection model to generate a first diagnosis result for each device; a generation module is used to acquire the topological relationship between the devices, establish a device association diagram in combination with the production process flow and the topological relationship, and generate association parameters between the devices based on the device association diagram; a second diagnosis module is used to generate a second diagnosis result for each device based on the association parameters; and a combination module is used to combine the first diagnosis result and the second diagnosis result to generate a final fault diagnosis result for each device.

[0114] It should be noted that: when the device provided in the above embodiment realizes its function, only the division of the above functional modules is used as an example. In actual application, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0115] See also Figure 3 , is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 3 As shown, the electronic device 1000 may include: at least one processor 1001 , at least one network interface 1004 , a user interface 1003 , a memory 1005 , and at least one communication bus 1002 .

[0116] The communication bus 1002 is used to realize the connection and communication between these components.

[0117] The user interface 1003 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 1003 may also include a standard wired interface and a wireless interface.

[0118] The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0119] Among them, the processor 1001 may include one or more processing cores. The processor 1001 uses various interfaces and lines to connect various parts in the entire server, and executes various functions of the server and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 1005, and calling data stored in the memory 1005. Optionally, the processor 1001 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 1001 can integrate one or a combination of a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU) and a modem. Among them, the CPU mainly processes the operating system, user interface and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 1001, and it can be implemented separately through a chip.

[0120] Among them, the memory 1005 may include a random access memory (Random Access Memory, RAM) and may also include a read-only memory (Read-Only Memory). Optionally, the memory 1005 includes a non-transitory computer-readable storage medium. The memory 1005 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 1005 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 1005 may optionally also be at least one storage device located away from the aforementioned processor 1001. As Figure 3 As shown, the memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and an application program of an intelligent fault diagnosis method.

[0121] exist Figure 3In the electronic device 1000 shown, the user interface 1003 is mainly used to provide an input interface for the user and obtain data input by the user; and the processor 1001 can be used to call an application program storing an intelligent fault diagnosis method in the memory 1005. When executed by one or more processors, the electronic device executes one or more methods described in the above embodiments.

[0122] An electronic device readable storage medium stores instructions, which, when executed by one or more processors, enable the electronic device to execute one or more of the methods described in the above embodiments.

[0123] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the order of the actions described, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for the present application.

[0124] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0125] In the several embodiments provided in the present application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of the units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0126] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0127] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0128] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned memory includes: various media that can store program codes, such as USB flash drives, mobile hard drives, magnetic disks or optical disks.

[0129] The above is only an exemplary embodiment of the present disclosure, and the scope of the present disclosure cannot be limited thereto. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the specification and practicing the disclosure here, those skilled in the art will easily think of other embodiments of the present disclosure. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the technical field not recorded in the present disclosure. The description and examples are only regarded as exemplary, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. An intelligent fault diagnosis method, characterized in that: The method comprises: Obtaining operating parameters of a plurality of devices in an industrial automation production line, wherein each of the devices is connected in sequence according to a production process flow; Inputting the operating parameters of each of the devices into a preset fault detection model to generate a first diagnosis result of each of the devices; Acquire the topological relationship between the devices, establish a device association diagram in combination with the production process and the topological relationship, and generate association parameters between the devices according to the device association diagram; generating a second diagnosis result for each of the devices according to the associated parameters; The first diagnostic result and the second diagnostic result are combined to generate a final fault diagnosis result of each of the devices.

2. The intelligent fault diagnosis method according to claim 1, characterized in that: The method further comprises: Acquire historical operation data of each of the devices, wherein the historical operation data includes device operation parameters and corresponding fault labels; Extracting features from the historical operation data to obtain a fault feature vector; Dividing the fault feature vector into a training data set and a verification data set according to a preset ratio; Training the training data set to obtain an initial fault detection model; Inputting the verification data set into the initial fault detection model for verification, and obtaining a model verification result; The initial fault detection model is adjusted according to the model verification result to obtain a preset fault detection model.

3. The intelligent fault diagnosis method according to claim 1, characterized in that: The step of combining the production process flow and the topological relationship to establish a device association diagram includes: Obtaining the process sequence of each of the equipment in the production process; Determine the material transfer mode between the devices according to the topological relationship; Taking each of the devices as a node and the material transfer method as a connecting line, an initial association diagram is generated, and an information flow direction mark is added to the initial association diagram, wherein the information flow direction mark represents the data transmission relationship between each of the devices; Marking the process parameters of each of the devices in the initial association diagram; An equipment association diagram is generated according to the process sequence, the material transfer method, the information flow identification and the process parameters.

4. The intelligent fault diagnosis method according to claim 3, characterized in that: Generating association parameters between the devices according to the device association diagram includes: According to the information flow direction identifier in the device association diagram, determining the upstream and downstream relationship between the devices, defining the upstream device as a source device, and defining the downstream device as a target device; Obtaining a fault type of the source device, and determining a fault propagation parameter corresponding to the fault type according to a preset fault propagation rule; Calculate the propagation coefficient of the fault propagation parameter in combination with the material transfer mode and process parameters between the source device and the target device; Multiplying the fault propagation parameter by a corresponding propagation coefficient to obtain a correlation between the source device and the target device; Traversing all device pairs in the device association graph, obtaining the corresponding association degrees between each pair of the devices, and generating an association degree matrix; The association matrix is ​​processed according to a preset threshold to generate association parameters between the devices.

5. The intelligent fault diagnosis method according to claim 1, characterized in that: Generating a second diagnosis result of each of the devices according to the associated parameters includes: Inputting the association parameters into a preset equipment association model to obtain an equipment failure impact matrix; Calculating the fault correlation strength between the devices according to the device fault impact matrix; Comparing the fault association strength with a preset fault propagation threshold, if the fault association strength is greater than the preset fault propagation threshold, determining that a fault association exists in the corresponding device; According to the fault association and in combination with preset fault diagnosis rules, a second diagnosis result of each of the devices is generated.

6. The intelligent fault diagnosis method according to claim 1, characterized in that: The combining the first diagnosis result and the second diagnosis result to generate a final fault diagnosis result of each of the devices includes: Obtaining a first failure probability of each of the devices in the first diagnosis result, and a second failure probability of each of the devices in the second diagnosis result; According to a preset weight coefficient, the first probability and the second probability are weighted and summed to obtain a comprehensive failure probability of each of the devices; Comparing the comprehensive failure probability with a preset failure determination threshold, if the comprehensive failure probability of the first device is not less than the preset failure determination threshold, determining that the first device has a failure; If the comprehensive failure probability of the second device is less than the preset failure determination threshold, it is determined that the second device does not have a failure.

7. The intelligent fault diagnosis method according to claim 1, characterized in that: After generating the final fault diagnosis result of each device, the method further includes: A fault repair instruction is generated according to the final fault diagnosis result of each of the devices, and the fault repair instruction is sent to the terminal device of the maintenance personnel so that the maintenance personnel can perform maintenance.

8. An intelligent fault diagnosis system, characterized in that: The system comprises: an acquisition module, a first diagnosis module, a generation module, a second diagnosis module and a combination module; wherein, The acquisition module is used to acquire the operating parameters of multiple devices in the industrial automation production line, wherein each of the devices is connected in sequence according to the production process; The first diagnostic module is used to input the operating parameters of each of the devices into a preset fault detection model to generate a first diagnostic result for each of the devices; The generating module is used to obtain the topological relationship between the devices, establish a device association diagram in combination with the production process and the topological relationship, and generate association parameters between the devices according to the device association diagram; The second diagnosis module is used to generate a second diagnosis result of each of the devices according to the associated parameters; The combining module is used to combine the first diagnostic result and the second diagnostic result to generate a final fault diagnosis result of each of the devices.

9. An electronic device, characterized in that: It includes a processor, a memory, a user interface and a network interface, the memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: A computer program is stored which can be loaded by a processor and execute the method according to any one of claims 1 to 7.

Images