Fault reason studying, judging and analyzing system based on machine learning

Abstract

The invention discloses a fault reason research and judgment analysis system based on machine learning, and belongs to the technical field of power system fault diagnosis. The system comprises a data acquisition module, a data preprocessing module, a feature engineering module, a machine learning research and judgment core, a fault reason judgment module and a user interface and interaction module. According to the method, the real-time early warning of equipment tripping is realized by integrating and analyzing massive equipment operation data and constructing a model by using a time sequence data analysis method. The system further adopts advanced machine learning algorithms such as a convolutional neural network, association rule learning, hierarchical clustering and the like, and combines professional knowledge and expert experience to deeply analyze fault recording data so as to identify, study and judge specific reasons of equipment tripping.

Description

Technical Field

[0001] The present invention relates to the technical field of power system fault diagnosis, and in particular to a fault cause determination and analysis system based on machine learning. Background Art

[0002] In the field of power systems, the stable operation of substation equipment is crucial to ensuring the reliability and safety of the power grid. However, traditional substation equipment fault diagnosis methods mainly rely on manual experience and simple data processing technology, which have problems such as low efficiency, lack of real-time performance, and limited accuracy when processing large-scale, heterogeneous data and complex fault modes.

[0003] With the development of big data and machine learning technologies, it has become possible to use these advanced technologies to conduct in-depth analysis and intelligent diagnosis of substation equipment operation data. Although some studies have applied machine learning algorithms to power transformer fault diagnosis, most of them focus on fault classification, and the accurate judgment and location of fault causes are still insufficient.

[0004] Therefore, there is an urgent need to develop a fault cause analysis system based on machine learning to accurately determine and locate the cause of the fault. Summary of the invention

[0005] In order to overcome the shortcomings of the background technology, the present invention proposes a fault cause analysis system based on machine learning, which aims to realize the rapid and accurate diagnosis of substation tripping faults. The system can automatically mine fault information from massive equipment operation data, quickly and accurately diagnose the cause of the fault, and provide corresponding processing suggestions, thereby improving the automation and intelligence level of fault processing, thereby reducing the impact of faults on power grid operation and improving the service quality and efficiency of power supply companies.

[0006] The technical solution adopted by the present invention is: A fault cause analysis system based on machine learning, comprising a data acquisition module, a data preprocessing module, a feature engineering module, a machine learning analysis core, a fault cause judgment module and a user interface and interaction module; the data acquisition module is connected to the substation equipment, the data preprocessing module is connected to the data acquisition module, the feature engineering module is connected to the data preprocessing module, the machine learning analysis core is connected to the feature engineering module, the fault cause judgment module is connected to the machine learning analysis core, and the user interface and interaction module is connected to the fault cause judgment module; The data acquisition module collects the SOE alarm information, telesignaling and telemetry data, operation ticket and work ticket data of the substation equipment in real time; The data preprocessing module cleans, time synchronizes and normalizes the collected data; The feature engineering module extracts time series data features and performs feature selection; The machine learning analysis core uses algorithms to analyze large amounts of data and obtain fault prediction results based on the collected data; The fault cause judgment module is based on the prediction results of the machine learning analysis core. The fault cause judgment module analyzes and judges the fault cause to obtain the fault analysis results and risk assessment information; The user interface and interaction module displays the fault analysis results and risk assessment information of the fault cause judgment module to realize user interactive operation.

[0007] Furthermore, the data acquisition module uses the following formula to perform time synchronization processing on the data during the data acquisition process: , in, is the timestamp after synchronization; is the timestamp of the original data; is the time calibration offset.

[0008] Furthermore, the data preprocessing module uses the Z-Score standardization method to normalize the data and eliminate the dimension and magnitude differences between different data sources; The Z-Score standardized calculation formula is as follows: , in, It is the standardized data; is the original data point; is the mean of the data set; is the standard deviation.

[0009] Furthermore, the feature engineering module uses time series data analysis methods to extract key features from SOE alarm information, telesignaling and telemetry, operation tickets and work tickets; uses semantic recognition algorithms to identify key information in substation SOE alarms and builds a fault signal classification and recognition model; and uses machine learning algorithms, including convolutional neural networks (CNNs), association rule learning and hierarchical clustering, to train feature analysis models for equipment tripping fault causes.

[0010] Furthermore, the feature engineering module uses a convolutional neural network (CNN) to automatically extract features of device video images for feature extraction, and uses an activation function (ReLU) to enhance the nonlinear expression capability of features; The activation function is as follows: , in, For input data; Represents the activation function.

[0011] Furthermore, the feature engineering module selects the features most relevant to the fault cause through a mutual information method and calculates the mutual information between the features and the fault type; The calculation formula of the mutual information method is as follows: , in, It is the mutual information between feature set A and fault type B; A represents feature set; B represents fault type; is the joint probability distribution; is the marginal probability distribution of feature A; The marginal probability distribution of feature B.

[0012] Furthermore, the machine learning analysis core includes: a convolutional neural network unit for extracting features from video image data, an association rule learning unit for mining potential association rules in data, and a hierarchical clustering algorithm unit for classifying device status; The machine learning analysis core conducts intelligent analysis on the causes of substation equipment tripping through comprehensive online monitoring, preventive testing, status evaluation, family defects and environmental disaster data.

[0013] Furthermore, the machine learning analysis core uses a machine learning algorithm to analyze the cause of the fault; First, the model is built and the long short-term memory network (LSTM) is used for time series prediction to screen the list of faulty devices that trip. The screening calculation formula is as follows: , in, is the hidden state at time step t; is the input; is the cell state at the previous time step; is the hidden state at the previous time step; is a function of the LSTM unit; are model parameters; Second, a random forest classifier is used to classify the possible fault types: The classification calculation formula is as follows: , in, is the final classification result; is the classification result of the t-th tree; T is the total number of trees.

[0014] Further, the fault cause determination module includes a fault signal classification and recognition model, and the fault signal classification and recognition model recognizes key alarm information; The fault cause judgment module divides faults into three categories according to risk factors: Low-risk failures: have little impact on system operation and only require recording and monitoring; Medium-risk failures: require immediate response, some system adjustments or temporary measures; High-risk failures: pose a serious threat to system security and require immediate action to avoid potential serious consequences; The fault cause judgment module divides faults into three categories according to their severity: Minor fault: does not affect the normal operation of the equipment, but may reduce efficiency or performance; General failure: requires certain maintenance or adjustment, which will temporarily affect the operation of the equipment; Serious failure: Causes the equipment to stop functioning completely and requires urgent repair or replacement of parts.

[0015] Furthermore, in the process of displaying the fault analysis results and risk assessment information of the fault cause determination module by the user interface and the interaction module: data visualization interaction, using data visualization technology to display the fault analysis results and risk assessment, and using a heat map to represent the equipment status; , in, is the heat map in coordinates The color value at It is at coordinates The function value at; N is the total number of function values; It means to sum all function values ​​from 1 to N.

[0016] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: 1. Accurate fault cause analysis Through rich data collection and in-depth feature engineering processing, the system can extract multi-dimensional data features generated by substation equipment during operation, thereby accurately determining the cause of the fault. Compared with the simple fault classification method, this provides a more detailed fault analysis.

[0017] 2. Multi-level fault assessment The system can not only identify the type of fault, but also conduct a multi-level assessment of the fault based on risk factor and severity. This hierarchical management method enables operators to quickly determine the degree of harm caused by the fault, so as to take appropriate emergency measures and minimize risks.

[0018] 3. Intelligent prediction and prevention By integrating multiple machine learning algorithms (such as convolutional neural networks, association rule learning, and hierarchical clustering), the system is able to predict failures and identify potential failure modes, so that preventive measures can be taken in advance to reduce product and equipment downtime.

[0019] 4. Real-time monitoring and feedback The real-time data collection and analysis module in the system enables operators to obtain equipment status and fault diagnosis results in the first place, dynamically monitor the equipment operation status, and respond to sudden faults in a timely manner.

[0020] 5. Improve decision support Combined with insights from data analysis, the system can provide data-driven decision support for equipment maintenance and management, helping managers to make more scientific equipment operation and maintenance plans and improve equipment reliability and safety.

[0021] 6. User-friendly interactive interface Through the user interface and interactive module, operators can intuitively obtain fault analysis and risk assessment information, assist in daily operation and maintenance work, and improve response efficiency.

[0022] In summary, the machine learning-based fault cause analysis system integrates and analyzes massive equipment operation data, uses time series data analysis methods, builds models, and realizes real-time warning of equipment tripping. The system further uses advanced machine learning algorithms such as convolutional neural networks, association rule learning, and hierarchical clustering, combined with professional knowledge and expert experience, to conduct in-depth analysis of fault recording data to identify and determine the specific causes of equipment tripping, provide more comprehensive and accurate fault cause analysis and location, and provide effective support for the safe operation and management of power transformers. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0024] Figure 1 It is a structural block diagram of the fault cause analysis system based on machine learning of the present invention; Figure 2 This is a flow chart of the fault cause analysis system based on machine learning of the present invention. DETAILED DESCRIPTION

[0025] The following will be combined with the drawings of the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without creative work are within the scope of protection of the present invention.

[0026] This embodiment provides a fault cause analysis system based on machine learning, such as Figure 1 As shown, the fault cause analysis system based on machine learning includes a data acquisition module, a data preprocessing module, a feature engineering module, a machine learning analysis core, a fault cause judgment module and a user interface and interaction module.

[0027] Among them, the data acquisition module is connected to the substation equipment, the data preprocessing module is connected to the data acquisition module, the feature engineering module is connected to the data preprocessing module, the machine learning analysis core is connected to the feature engineering module, the fault cause judgment module is connected to the machine learning analysis core, and the user interface and interaction module are connected to the fault cause judgment module.

[0028] The functions of each module of the fault cause analysis system based on machine learning are as follows: Data acquisition module: The data acquisition module is connected to the substation equipment; the data acquisition module is used to collect SOE alarm information, telesignaling and telemetry data, operation ticket and work ticket data of the substation equipment in real time.

[0029] Data preprocessing module: The data preprocessing module is connected to the data acquisition module; the data preprocessing module is used to clean, time synchronize and normalize the collected data.

[0030] Feature engineering module: The feature engineering module is connected to the data preprocessing module; the feature engineering module is used to extract time series data features from the preprocessed collected data and perform feature selection.

[0031] Machine learning analysis core: The machine learning analysis core is connected to the feature engineering module; the machine learning analysis core uses algorithms to analyze large amounts of data to complete intelligent recognition patterns; and obtains fault prediction results based on the collected data.

[0032] The core of machine learning analysis includes: Convolutional neural network unit: used to extract features from video image data; Association rule learning unit: used to mine potential association rules in data; Hierarchical clustering algorithm unit: used to classify device status; Through comprehensive online monitoring, preventive testing, status evaluation, family defects and environmental disaster data, intelligent analysis of the reasons for substation equipment tripping can be achieved.

[0033] Fault cause judgment module: The fault cause judgment module is connected to the machine learning analysis core; based on the prediction results of the machine learning analysis core, the fault cause judgment module analyzes and judges the cause of the fault to obtain the fault analysis results and risk assessment information.

[0034] The fault cause judgment module divides faults into three categories according to risk factors: Low-risk failures: have little impact on system operation and only require recording and monitoring; Medium-risk failures: require immediate response, some system adjustments or temporary measures; High-risk failures: pose a serious threat to system security and require immediate action to avoid potential serious consequences.

[0035] The fault cause judgment module divides faults into three categories according to their severity: Minor fault: does not affect the normal operation of the equipment, but may reduce efficiency or performance; General failure: requires certain maintenance or adjustment, which will temporarily affect the operation of the equipment; Serious failure: Causes the equipment to stop functioning completely and requires urgent repair or replacement of parts.

[0036] User interface and interaction module: The user interface and interaction module is connected to the fault cause judgment module; the user interface and interaction module displays the fault analysis results and risk assessment information of the fault cause judgment module to realize user interactive operation.

[0037] This machine learning-based fault cause analysis system integrates and analyzes massive equipment operation data, uses time series data analysis methods to build models, and achieves real-time warning of equipment tripping. The system further uses advanced machine learning algorithms such as convolutional neural networks, association rule learning, and hierarchical clustering, combined with professional knowledge and expert experience, to conduct in-depth analysis of fault recording data to identify and determine the specific causes of equipment tripping, provide more comprehensive and accurate fault cause analysis and location, and provide effective support for the safe operation and management of power transformers.

[0038] like Figure 2 As shown in the figure, the working principle of the fault cause analysis system based on machine learning is as follows: First, click Start through the user interface and interactive module, and the machine learning-based fault cause analysis system will start working; Step 1: Real-time data collection of substation equipment: The data collection module collects the SOE alarm information, telemetering data, operation ticket and work ticket data of the substation equipment in real time; the data collection module uses the following formula to perform time synchronization processing on the data: , in, is the timestamp after synchronization; is the timestamp of the original data; is the time calibration offset.

[0039] By collecting substation equipment data in real time through step one, the time consistency of the alarm information and telemetry data of the substation equipment can be ensured, so that subsequent analysis is based on accurate time series data, reducing analysis errors caused by time deviation.

[0040] Step 2: Data preprocessing: Data preprocessing includes data cleaning and normalization; Data cleaning: The data preprocessing module cleans the incoming data, removes invalid or erroneous records, and uses time synchronization technology to ensure the consistency of the data in the time dimension; performs data preprocessing, including denoising, outlier removal, and data standardization; Normalization: The Z-Score standardization method was used to normalize the data to eliminate the dimension and magnitude differences between different data sources; The Z-Score standardized calculation formula is as follows: , in, It is the standardized data; is the original data point; is the mean of the data set; is the standard deviation.

[0041] Data preprocessing in step 2 can clean invalid and erroneous data records and improve data quality. Denoising and outlier removal ensure that the data used in model training is more credible, and Z-Score standardization eliminates the dimensional differences between data from different sources, making the influence of each feature in model training more balanced.

[0042] Step 3: Extract and select features: Feature Engineering: The feature engineering module uses time series data analysis methods to extract key features from multi-dimensional data such as SOE alarm information, telesignaling and telemetry, operation tickets and work tickets. The semantic recognition algorithm is used to identify key information in the SOE alarm of substation equipment, such as opening and closing, reclosing and protection actions, and to build a fault signal classification and recognition model.

[0043] Feature engineering uses time series analysis methods to extract key features, allowing the model to more effectively capture important information in the data. Key information is extracted through semantic recognition algorithms to improve the accuracy and speed of fault identification.

[0044] Through machine learning algorithms, including convolutional neural network (CNN), association rule learning and hierarchical clustering, a feature analysis model of the cause of equipment tripping fault is trained.

[0045] Feature extraction uses convolutional neural network (CNN) to automatically extract features of device video images, and uses activation function ReLU (Rectified Linear Unit) to enhance the nonlinear expression ability of features: , in, For input data; Represents the activation function.

[0046] The convolutional neural network (CNN) can automatically extract high-level features from device video images, enhancing the model's ability to identify complex fault modes. The use of the ReLU activation function further enhances the nonlinear ability of feature modeling.

[0047] Mutual Information Method: The most relevant features to the fault cause are selected through the mutual information method, and the mutual information between the features and the fault type is calculated; The calculation formula of the mutual information method is as follows: , in, It is the mutual information between feature set A and fault type B; A represents feature set; B represents fault type; is the joint probability distribution; is the marginal probability distribution of feature A; The marginal probability distribution of feature B.

[0048] By calculating the mutual information between features and fault types, the features most relevant to the fault causes are screened out, reducing the computational complexity and improving the accuracy and efficiency of the model.

[0049] Step 4: Use machine learning algorithms to determine the cause of the fault: Machine Learning Models: After the equipment trips, the machine learning analysis core and fault cause judgment module use fault recording data, combined with online monitoring, preventive testing, status evaluation and other data, and use fault cause analysis technology based on machine learning algorithms to conduct a comprehensive analysis.

[0050] Through the feature analysis model, the cause of equipment tripping can be intelligently determined, and the equipment status and operation risks can be evaluated.

[0051] Use machine learning algorithms to determine the cause of failure: First, the model is built and the long short-term memory network (LSTM) is used for time series prediction to screen the list of faulty devices that trip. The screening calculation formula is as follows: , in, is the hidden state at time step t; is the input; is the cell state at the previous time step; is the hidden state at the previous time step; is a function of the LSTM unit; are model parameters.

[0052] Second, a random forest classifier is used to classify the possible fault types: The classification calculation formula is as follows: , in, is the final classification result; is the classification result of the t-th tree; T is the total number of trees.

[0053] Step 4 uses machine learning algorithms to analyze the cause of the fault. It combines a variety of data display and analysis technologies and uses machine learning models, such as LSTM and random forest, to automatically analyze the cause of the fault. This comprehensive analysis can quickly identify faulty equipment, improve response speed, and reduce human errors.

[0054] In addition, the long short-term memory network (LSTM) performs well in processing time series data and can capture the time dependency in the data; the random forest improves the robustness of classification and reduces the risk of overfitting by integrating multiple decision trees.

[0055] Furthermore, the fault cause judgment module includes a fault signal classification and recognition model, which is based on the output fault type, by defining a risk factor R and a severity assessment factor S, and mapping them in combination with expert knowledge and historical data; The calculation formula of risk factor R is as follows: ; In the formula, is the risk factor; is a function; The fault cause judgment module divides faults into three categories based on the risk factor R. The rules are as follows: Low-risk faults: , which has little impact on system operation and only needs to be recorded and monitored; Medium-risk failures: , requiring immediate response, requiring partial system adjustments or temporary measures; High-risk failures: , which poses a serious threat to system security and requires immediate action to avoid potential serious consequences.

[0056] The calculation formula of severity assessment coefficient S is as follows: ; In the formula, is the severity assessment coefficient; is a function; The fault cause judgment module divides faults into three categories according to severity based on the severity evaluation function S. The rules are as follows: Minor faults: , does not affect the normal operation of the equipment, but may reduce efficiency or performance; General faults: , which requires certain maintenance or adjustment, and will temporarily affect the operation of the equipment; Serious failure: , causing the equipment to completely stop running and requiring emergency repair or replacement of components. The first threshold and the second threshold are set according to actual needs.

[0057] Step 5: Data visualization interaction: Visualization: Use advanced data visualization technologies, such as SVG, Canvas, WebGL, etc., to achieve BI data visualization and display fault analysis results and risk assessment information.

[0058] Through a variety of chart components, such as pie charts, line charts, heat maps, etc., it provides users with intuitive fault analysis and decision support.

[0059] Data visualization interaction uses data visualization technology to display fault analysis results and risk assessment, and uses heat maps to represent equipment status: , in, is the heat map in coordinates The color value at It is at coordinates The function value at; N is the total number of function values; It means to sum all function values ​​from 1 to N.

[0060] Visual presentation uses advanced data visualization technologies, such as SVG and Canvas, to enable users to understand fault analysis and evaluation information in an intuitive way, supporting better decision-making. Visualization tools such as heat maps can quickly display equipment status for easy monitoring and management.

[0061] User Interaction: Design user interaction logic to allow users to input specific parameters through the interface, and the system will update the analysis results in real time based on the input parameters.

[0062] User interaction allows users to input parameters according to their own needs, and update and feedback fault analysis results in real time, which enhances the flexibility and operability of the system and improves the user experience.

[0063] In summary, through the data processing and analysis of the above steps, the overall system can more effectively perform fault identification, cause analysis and risk assessment, providing strong guarantees for the safe operation of substation equipment.

[0064] The above formulas are all dimensionless and numerical calculations. The formula is a formula obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formula are set by technicians in this field according to actual conditions. The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods.

[0065] Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and describes these embodiments in detail in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A fault cause analysis system based on machine learning, characterized by: It includes a data acquisition module, a data preprocessing module, a feature engineering module, a machine learning analysis core, a fault cause judgment module and a user interface and interaction module; the data acquisition module is connected to the substation equipment, the data preprocessing module is connected to the data acquisition module, the feature engineering module is connected to the data preprocessing module, the machine learning analysis core is connected to the feature engineering module, the fault cause judgment module is connected to the machine learning analysis core, and the user interface and interaction module is connected to the fault cause judgment module; The data acquisition module collects the SOE alarm information, telesignaling and telemetry data, operation ticket and work ticket data of the substation equipment in real time; The data preprocessing module cleans, time synchronizes and normalizes the collected data; The feature engineering module extracts time series data features and performs feature selection; The machine learning analysis core uses algorithms to analyze large amounts of data and obtain fault prediction results based on the collected data; The fault cause judgment module is based on the prediction results of the machine learning analysis core. The fault cause judgment module analyzes and judges the fault cause to obtain the fault analysis results and risk assessment information; The user interface and interaction module displays the fault analysis results and risk assessment information of the fault cause judgment module to realize user interactive operation.

2. The fault cause analysis system based on machine learning according to claim 1 is characterized in that: The data acquisition module uses the following formula to perform time synchronization processing on the data during the data acquisition process: ， in, is the timestamp after synchronization; is the timestamp of the original data; is the time calibration offset.

3. The fault cause analysis system based on machine learning according to claim 1 is characterized in that: The data preprocessing module uses the Z-Score standardization method to normalize the data to eliminate the dimension and magnitude differences between different data sources; The Z-Score standardized calculation formula is as follows: ， in, It is the standardized data; is the original data point; is the mean of the data set; is the standard deviation.

4. The fault cause analysis system based on machine learning according to claim 1 is characterized in that: The feature engineering module uses time series data analysis methods to extract key features from SOE alarm information, telesignaling and telemetry, operation tickets and work tickets; The semantic recognition algorithm is used to identify the key information in the SOE alarm of the substation equipment, and a fault signal classification and recognition model is constructed. The feature analysis model of the cause of the equipment tripping fault is trained through machine learning algorithms, including convolutional neural network CNN, association rule learning and hierarchical clustering.

5. The fault cause analysis system based on machine learning according to claim 4 is characterized in that: The feature engineering module uses a convolutional neural network (CNN) to automatically extract features of device video images and uses an activation function (ReLU) to enhance the nonlinear expression capability of features. The activation function is as follows: ， in, For input data; Represents the activation function.

6. The fault cause analysis system based on machine learning according to claim 4 is characterized in that: The feature engineering module selects the features most relevant to the fault cause through the mutual information method and calculates the mutual information between the features and the fault type; The calculation formula of the mutual information method is as follows: ， in, It is the mutual information between feature set A and fault type B; A represents feature set; B represents fault type; is the joint probability distribution; is the marginal probability distribution of feature A; The marginal probability distribution of feature B.

7. The fault cause analysis system based on machine learning according to claim 1 is characterized in that: The machine learning analysis core includes: a convolutional neural network unit for extracting features from video image data, an association rule learning unit for mining potential association rules in data, and a hierarchical clustering algorithm unit for classifying device status; The machine learning analysis core conducts intelligent analysis on the causes of substation equipment tripping through comprehensive online monitoring, preventive testing, status evaluation, family defects and environmental disaster data.

8. The fault cause analysis system based on machine learning according to claim 7 is characterized in that: The machine learning analysis core uses a machine learning algorithm to analyze the cause of the fault; First, the model is constructed and the long short-term memory network (LSTM) is used for time series prediction to screen the list of faulty devices that trip. The screening calculation formula is as follows: ， Among them, is the hidden state at time step t; is the input; is the cell state at the previous time step; is the hidden state at the previous time step; is a function of the LSTM unit; are model parameters; Second, a random forest classifier is used to classify the possible fault types: The classification calculation formula is as follows: ， in, is the final classification result; is the classification result of the t-th tree; T is the total number of trees.

9. The fault cause analysis system based on machine learning according to claim 1 is characterized in that: The fault cause determination module includes a fault signal classification and recognition model, which identifies key alarm information; The fault cause judgment module divides faults into three categories according to risk factors: Low-risk failures: have little impact on system operation and only require recording and monitoring; Medium-risk failures: require immediate response, some system adjustments or temporary measures; High-risk failures: pose a serious threat to system security and require immediate action to avoid potential serious consequences; The fault cause judgment module divides faults into three categories according to their severity: Minor fault: does not affect the normal operation of the equipment, but may reduce efficiency or performance; General failure: requires certain maintenance or adjustment, which will temporarily affect the operation of the equipment; Serious failure: Causes the equipment to stop functioning completely and requires urgent repair or replacement of parts.

10. The fault cause analysis system based on machine learning according to claim 1 is characterized in that: The user interface and the interaction module display the fault analysis results and risk assessment information of the fault cause judgment module in the process of: data visualization interaction, using data visualization technology to display the fault analysis results and risk assessment, and using a heat map to represent the equipment status; ， in, is the heat map in coordinates The color value at It is at coordinates The function value at; N is the total number of function values; It means to sum all function values ​​from 1 to N.

Images