Bolt type capacitor fault detection and repair method and device, and medium

Abstract

The invention relates to the technical field of fault detection and repair of a bolt type capacitor. The bolt-type capacitor fault detection and repair method comprises the following steps: acquiring health state information of a bolt-type capacitor; based on the health state information, a hierarchical fault detection mechanism is adopted, faults of the bolt type capacitor are divided into multiple levels, the multiple levels comprise a slight fault, a serious fault and an emergency fault, and a corresponding detection strategy is implemented for each level; if the fault type is a slight fault or a serious fault, historical fault information is acquired, the development trend of the fault is analyzed based on the historical fault information and the fault type, and fault development early warning information is generated; and determining a risk type according to the fault development early warning information, obtaining a repair strategy based on the risk type, and executing repair operation by the automatic repair system according to the repair strategy. According to the invention, the fault detection accuracy is improved, and the maintenance efficiency and safety of the capacitor are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of fault detection and repair of bolt-type capacitors, and in particular to a method, device and medium for fault detection and repair of bolt-type capacitors. Background Art

[0002] Bolt-on capacitors are widely used in the fields of electricity, industrial control, etc. As important power storage equipment, they play the role of stable output and efficient transmission of electric energy. However, during long-term use, bolt-on capacitors are prone to failures due to overload, environmental changes, aging and other factors. Common types of failures include capacitance drop, short circuit, open circuit and other problems, which seriously affect the normal operation of the equipment and may cause safety accidents.

[0003] Traditional methods for detecting faults in bolted capacitors rely mainly on manual inspection and regular maintenance. These methods rely on manual judgment and are unable to monitor the health status of the capacitor in real time, making it easy to overlook potential failure risks. At the same time, existing fault diagnosis methods lack accurate classification of fault types, resulting in poor accuracy and timeliness of diagnostic results, and fail to provide effective automatic repair solutions. The lag in fault diagnosis and repair often prevents bolted capacitors from taking effective repair measures in a timely manner when a fault occurs, thus affecting the long-term stability and safety of the equipment. Summary of the invention

[0004] In order to improve the accuracy of fault detection and effectively enhance the maintenance efficiency and safety of capacitors, the present application provides a method, device and medium for detecting and repairing bolt-type capacitor faults.

[0005] The above-mentioned invention objective of the present application is achieved through the following technical solutions: Get health status information of bolt-on capacitors; Based on the health status information, a hierarchical fault detection mechanism is adopted to classify the faults of the bolt-type capacitor into multiple levels, wherein the multiple levels include minor faults, serious faults and emergency faults, and a corresponding detection strategy is implemented for each level; If the fault type is a minor fault or a major fault, historical fault information is obtained, and based on the historical fault information and the fault type, a fault development trend is analyzed to generate fault development warning information; According to the fault development warning information, the risk type is determined, based on the risk type, a repair strategy is obtained, and according to the repair strategy, the automated repair system performs a repair operation.

[0006] By adopting the above technical solution, the health status information of the capacitor is acquired and analyzed in real time, combined with the hierarchical fault detection mechanism, different levels of faults, such as minor faults, serious faults and emergency faults, can be accurately identified and distinguished. For each type of fault, a corresponding detection strategy is adopted to ensure the accuracy and timeliness of fault judgment. Based on historical fault information, the fault development trend is predicted using methods such as regression analysis and time series analysis, which enables the system to identify the potential aggravation or expansion of the fault in advance and provide data support for subsequent repair measures. This not only helps to discover and solve problems in a timely manner, but also avoids equipment damage or shutdown caused by untimely handling. By analyzing the fault development warning information, corresponding repair strategies are formulated according to risk types (such as low risk, medium risk and high risk), ensuring the system's rapid response to different risks.

[0007] In a preferred example, the present application may be further configured as follows: the health status information includes: Acquiring voltage information, current information, temperature information, vibration information and environmental information of the bolt-type capacitor; The voltage information, the current information, the temperature information, the vibration information and the environmental information are integrated through a Kalman filter algorithm to obtain the health status information.

[0008] By adopting the above technical solution, by obtaining voltage, current, temperature, vibration and environmental information, and fusing this information with the Kalman filter algorithm, the noise in various signals is effectively removed, ensuring the accuracy and reliability of health status information. The comprehensive analysis of different sensor data can more comprehensively reflect the working status of the capacitor, thereby improving the accuracy of fault diagnosis.

[0009] In a preferred example, the present application may be further configured as follows: based on the health status information, a hierarchical fault detection mechanism is adopted to classify the fault of the bolt-type capacitor into multiple levels, wherein the multiple levels include minor faults, serious faults and emergency faults, including: Comparing the health status information with preset health status thresholds, wherein the health status thresholds include a normal working range, a minor fault threshold, a major fault threshold, and an emergency fault threshold; If the health status information does not exceed the normal working range, it is determined to be in a normal working state and fault detection is stopped; If the health status information exceeds the normal working range, the hierarchical fault detection mechanism is adopted; If the health status information exceeds the normal working range but does not exceed the minor fault threshold, it is determined to be the minor fault; if the health status information exceeds the minor fault threshold but does not exceed the serious fault threshold, it is determined to be the serious fault; if the health status information exceeds the serious fault threshold, it is determined to be the emergency fault.

[0010] By adopting the above technical solution, the faults of bolt-type capacitors are divided into multiple levels (minor faults, serious faults and emergency faults), and different detection strategies can be adopted according to the different fault degrees. This grading method can more accurately identify the severity of the fault, and perform appropriate detection and repair measures according to its urgency to avoid misjudgment and missed judgment. By comparing the health status information with the preset health status threshold, the working status of the capacitor can be judged in real time. When the health status information exceeds the normal working range, the system can quickly trigger the graded fault detection mechanism, discover potential problems in time and take corresponding measures to reduce the risk of failure and improve the operating stability of the equipment.

[0011] In a preferred example, the present application may be further configured as follows: if the fault type is a minor fault or a serious fault, historical fault information is obtained, and based on the historical fault information and the fault type, the fault development trend is analyzed to generate fault development warning information, including: If the fault type is the minor fault, then based on the historical fault information, a linear regression analysis method is used to predict the future development trend of the minor fault, and generate corresponding fault development warning information; If the fault type is the severe fault, then based on the historical fault information, a time series analysis method is used to predict the future aggravation trend of the severe fault, and generate corresponding fault development warning information.

[0012] By adopting the above technical solution, the linear regression analysis method is used to predict minor faults, which can accurately predict their future development trends based on historical fault information, thereby timely discovering the potential risk of aggravation of minor faults. Similarly, the time series analysis method is used to predict serious faults, which can analyze the aggravation trend of faults.

[0013] In a preferred example, the present application may be further configured as follows: if the fault type is the minor fault, a linear regression analysis method is used based on the historical fault information to predict the future development trend of the minor fault, and generate corresponding fault development warning information, including: Extract key feature information from the historical fault information, wherein the key feature information includes the fault occurrence time and the fault frequency, and use the key feature information to establish a linear regression model, wherein the fault occurrence time is used as an independent variable and the fault frequency is used as a dependent variable, and a linear relationship of the fault development trend is fitted to obtain a trained linear regression model; The trained linear regression model is applied to the future development trend of the minor fault to obtain a prediction result. Based on the prediction result, the probability value of future fault occurrence is calculated, and the probability value is compared with the preset warning threshold. If the probability value exceeds the preset warning threshold, the corresponding fault development warning information is generated.

[0014] By adopting the above technical solution, by extracting key feature information from historical fault information (such as fault occurrence time and frequency), and using the linear regression model to establish a mathematical relationship of fault development trends, the historical pattern of faults and future development paths can be accurately understood. By using the trained linear regression model for fault prediction, the probability of future faults can be calculated. By comparing with the preset warning threshold, if the prediction result exceeds the threshold, fault development warning information will be generated in time. This enables maintenance personnel to take effective preventive measures before the problem worsens, avoiding more serious problems caused by minor faults not being handled in time.

[0015] In a preferred example, the present application may be further configured as follows: if the fault type is the severe fault, based on the historical fault information, a time series analysis method is used to predict the future aggravation trend of the severe fault, and corresponding fault development warning information is generated, including: Acquire time series feature information from the historical fault information, the time series feature data including the time series of fault occurrence, the fault severity and the fault frequency, use the time series feature information data to input into the LSTM model for training, and obtain the trained LSTM model through the long-term dependency between the time series of fault occurrence and the fault severity and the fault frequency; The trained LSTM model is applied to the prediction of the serious fault, and the probability of the serious fault occurring at a predetermined time point is calculated. If the probability exceeds a preset fault aggravation warning threshold, the fault development warning information is triggered.

[0016] By adopting the above technical solution, the trained LSTM model can predict the probability of serious faults at a predetermined time point in the future based on historical fault information. When the probability of a fault predicted by the LSTM model exceeds the preset fault aggravation warning threshold, the system will promptly trigger the fault development warning information. This real-time warning can help the operation and maintenance team take emergency measures before the fault further deteriorates, avoid the spread of equipment faults, reduce equipment downtime, and ensure the stable operation of the system.

[0017] In a preferred example, the present application may be further configured as follows: determining the risk type according to the fault development warning information, obtaining a repair strategy based on the risk type, and performing a repair operation by the automated repair system according to the repair strategy, including: Analyze the severity of the fault according to the fault development warning information, and classify the severity into low risk, medium risk and high risk; If the risk type is the low risk, a regular inspection and monitoring strategy is formulated; if the risk type is the medium risk, a preventive repair strategy is formulated; if the risk type is the high risk, an emergency repair strategy is formulated.

[0018] By adopting the above technical solution, according to the fault development warning information, the severity of the fault can be accurately analyzed and divided into three levels: low risk, medium risk and high risk. Through this classification, the operation and maintenance personnel can clearly understand the potential threat level of each fault, so as to allocate resources more reasonably and take corresponding treatment measures. According to the fault development warning information, the severity of the fault can be accurately analyzed and divided into three levels: low risk, medium risk and high risk. Through this classification, the operation and maintenance personnel can clearly understand the potential threat level of each fault, so as to allocate resources more reasonably and take corresponding treatment measures. For low-risk faults, regular inspection and monitoring strategies can be automatically formulated. By regularly monitoring the equipment status, the system can timely detect and solve potential problems before the fault develops further, thereby ensuring that the equipment operates stably for a long time and avoiding greater impact of the fault on the system. For medium-risk faults, preventive repair strategies will be automatically started. This means that the operation and maintenance team can take measures in advance according to the trend and development of the fault, such as replacing vulnerable parts, adjusting the operating process or performing local repairs, to reduce the possibility of further deterioration of the fault, thereby avoiding greater losses caused by equipment failure. For high-risk faults, emergency repair strategies are triggered. This strategy will prioritize maintenance resources to ensure that problems are resolved quickly, avoid excessive equipment downtime, and prevent failures from causing serious impacts on the production process, equipment safety, or personnel safety. By responding quickly, the system can minimize the risk of accidents.

[0019] The second object of the invention is achieved by the following technical solutions: A bolt-type capacitor fault detection and repair device, the bolt-type capacitor fault detection and repair device comprising: A health status information acquisition module, used to acquire health status information of bolt-type capacitors; A hierarchical fault detection module, configured to classify the faults of the bolt-type capacitor into multiple levels based on the health status information and using a hierarchical fault detection mechanism, wherein the multiple levels include minor faults, serious faults and emergency faults, and implement a corresponding detection strategy for each level; A fault development trend analysis module is used to obtain historical fault information if the fault type is a minor fault or a major fault, analyze the fault development trend based on the historical fault information and the fault type, and generate fault development warning information; The automatic repair execution module is used to determine the risk type according to the fault development warning information, obtain the repair strategy based on the risk type, and according to the repair strategy, the automatic repair system executes the repair operation.

[0020] By adopting the above technical solution, the health status information of the capacitor is acquired and analyzed in real time, combined with the hierarchical fault detection mechanism, different levels of faults, such as minor faults, serious faults and emergency faults, can be accurately identified and distinguished. For each type of fault, a corresponding detection strategy is adopted to ensure the accuracy and timeliness of fault judgment. Based on historical fault information, the fault development trend is predicted using methods such as regression analysis and time series analysis, which enables the system to identify the potential aggravation or expansion of the fault in advance and provide data support for subsequent repair measures. This not only helps to discover and solve problems in a timely manner, but also avoids equipment damage or shutdown caused by untimely handling. By analyzing the fault development warning information, corresponding repair strategies are formulated according to risk types (such as low risk, medium risk and high risk), ensuring the system's rapid response to different risks.

[0021] The third objective of the present application is achieved through the following technical solutions: A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned bolt-type capacitor fault detection and repair method when executing the computer program.

[0022] The fourth objective of the present application is achieved through the following technical solutions: A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned bolt-type capacitor fault detection and repair method.

[0023] In summary, the present application includes at least one of the following beneficial technical effects: 1. By acquiring and analyzing the health status information of the capacitor in real time, combined with the hierarchical fault detection mechanism, it can accurately identify and distinguish different levels of faults, such as minor faults, serious faults and emergency faults. For each type of fault, a corresponding detection strategy is adopted to ensure the accuracy and timeliness of fault judgment. Based on historical fault information, regression analysis and timing analysis are used to predict the development trend of faults, which enables the system to identify the potential aggravation or expansion of faults in advance and provide data support for subsequent repair measures. This not only helps to discover and solve problems in a timely manner, but also avoids equipment damage or shutdown due to untimely handling. By analyzing the early warning information of fault development, corresponding repair strategies are formulated according to the risk type (such as low risk, medium risk and high risk), ensuring the system's rapid response to different risks; 2. Through the trained LSTM model, the probability of serious faults occurring at a predetermined time point in the future can be predicted based on historical fault information. When the probability of a fault predicted by the LSTM model exceeds the preset fault aggravation warning threshold, the system will promptly trigger the fault development warning information. This real-time warning can help the operation and maintenance team take emergency measures before the fault further deteriorates, avoid the spread of equipment faults, reduce equipment downtime, and ensure the stable operation of the system; 3. By extracting key feature information from historical fault information (such as fault occurrence time and frequency) and using a linear regression model to establish a mathematical relationship for fault development trends, we can accurately understand the historical patterns and future development paths of faults. By using a trained linear regression model for fault prediction, we can calculate the probability of future faults. By comparing with the preset warning threshold, if the prediction result exceeds the threshold, a fault development warning information will be generated in a timely manner. This enables maintenance personnel to take effective preventive measures before the problem worsens, avoiding more serious problems caused by minor faults not being handled in time. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a structural schematic diagram of a method for detecting and repairing a bolt-type capacitor fault in an embodiment of the application; Figure 2 It is a flowchart for implementing step S10 in the method for detecting and repairing a bolt-type capacitor fault in one embodiment of the present application; Figure 3 It is a flowchart for implementing step S20 in the method for detecting and repairing a bolt-type capacitor fault in one embodiment of the present application; Figure 4 It is a flowchart for implementing step S30 in the method for detecting and repairing a bolt-type capacitor fault in one embodiment of the present application; Figure 5 It is a flowchart for implementing step S301 in the method for detecting and repairing a bolt-type capacitor fault in one embodiment of the present application; Figure 6 It is a flowchart for implementing step S302 in the method for detecting and repairing a bolt-type capacitor fault in one embodiment of the present application; Figure 7 It is a flowchart for implementing step S40 in the method for detecting and repairing a bolt-type capacitor fault in one embodiment of the present application; Figure 8 This is a principle block diagram of a bolt-type capacitor fault detection and repair device in one embodiment of the present application; Fig. 9 It is a schematic diagram of a device in an embodiment of the present application. DETAILED DESCRIPTION

[0025] The present application is further described in detail below in conjunction with the accompanying drawings.

[0026] In one embodiment, if Figure 1 As shown, the present application discloses a method for detecting and repairing a bolt-type capacitor fault, which specifically includes the following steps: S10: Obtain health status information of the bolt-type capacitor.

[0027] Specifically, the health status information of bolt-type capacitors is collected through multiple sensors, covering parameters such as capacitor voltage, current, temperature, vibration, and environmental factors. This information is monitored in real time through different measurement technologies and sensor configurations to ensure that the state of the capacitor under different working conditions can be fully reflected. Voltage information reflects the voltage stability of the capacitor, current information reflects the load of the capacitor, temperature information helps to determine whether the capacitor is overheated, vibration information can reveal whether the capacitor has mechanical damage, and environmental information includes humidity, air pressure and other external factors that may affect the performance of the capacitor. After all this information is collected by the data acquisition module, it is processed using fusion algorithms such as Kalman filtering to eliminate noise interference and finally obtain accurate health status information.

[0028] S20: Based on the health status information, a hierarchical fault detection mechanism is adopted to classify the faults of the bolt-type capacitor into multiple levels, including minor faults, serious faults and emergency faults, and a corresponding detection strategy is implemented for each level.

[0029] Specifically, the acquired health status information will be compared with the preset health status thresholds, which represent different fault levels, such as "normal operating range", "minor fault threshold", "serious fault threshold" and "urgent fault threshold". These thresholds are set based on the normal operating range of the bolt-type capacitor and historical fault data experience. When the health status information (such as voltage, current, temperature, etc.) is within the normal operating range, it means that the capacitor is in normal working condition, and the system will suspend fault detection and allow the device to continue to operate normally; when the health status information exceeds the normal operating range but is still below the minor fault threshold, the system will determine it as a minor fault and continue to monitor, and may issue a reminder, but will not immediately start the repair operation; when the health status information exceeds the minor fault threshold but does not exceed the major fault threshold, the system will start a detailed fault detection program, record data and prepare preventive repair measures; when the health status information exceeds the major fault threshold, the system will consider that the device has an urgent fault, immediately start the emergency repair program or send an alarm signal, and the system may shut down until the fault is repaired.

[0030] S30: If the fault type is a minor fault or a major fault, historical fault information is obtained, and based on the historical fault information and the fault type, the development trend of the fault is analyzed to generate fault development warning information.

[0031] Specifically, if the fault type is detected as a minor fault or a major fault, the system will first extract relevant data from the stored historical fault information, including the time when the fault occurred, the frequency of the fault, the evolution of the fault type, the conditions under which the fault occurred, and the past repairs. This historical fault information is used to understand the past trends of faults and provide a basis for future fault predictions. Based on this historical information and the current fault type, the system will apply data analysis methods (such as linear regression analysis or time series analysis) to analyze the changing trends of faults. For minor faults, the relationship between the frequency and time of faults may be analyzed through a linear regression model to predict the frequency and time of future faults; for major faults, time series analysis methods (such as LSTM models) may be used to analyze the historical patterns of faults, combined with the current fault characteristics, to predict whether the fault will intensify or undergo more serious changes. Finally, through these trend analyses, the system will generate fault development warning information.

[0032] Furthermore, the system is used to monitor, analyze and repair the failure of bolt-type capacitors. This system can collect and process the health status information of bolt-type capacitors in real time, and evaluate the failure type, development trend and risk through intelligent algorithms, and finally generate corresponding repair strategies and automatically perform repair operations. S40: Determine the risk type according to the fault development warning information, obtain a repair strategy based on the risk type, and execute the repair operation according to the repair strategy by the automated repair system.

[0033] Specifically, according to the generated fault development warning information, the system will first analyze the key features in the warning information, such as the probability of fault occurrence, fault severity, development trend, etc., to determine the risk type of the fault. Risk types are usually divided into three categories: low risk, medium risk and high risk. For example, if the warning information shows that the probability of fault occurrence is low and the impact is small, it is judged as low risk. If the predicted probability of fault occurrence is high and may cause a large impact, it is judged as high risk. In this way, the system can accurately divide the risk type according to the potential impact and probability of occurrence of the fault. After determining the risk type, the system will formulate corresponding repair strategies for different risk types based on preset rules and strategies. For low-risk faults, the system may recommend regular inspection and monitoring, and non-emergency maintenance of the equipment. For medium-risk faults, the system may recommend preventive repairs or component replacement. For high-risk faults, the system will automatically generate emergency repair strategies and prioritize repair tasks to ensure that the fault hidden dangers are eliminated as soon as possible. According to the determined repair strategy, the automated repair system will perform specific repair operations, including starting the repair equipment, executing the repair program, or repairing the fault through manual intervention.

[0034] In one embodiment, if Figure 2 As shown, in step S10, the health status information includes: S101: Obtain voltage information, current information, temperature information, vibration information, and environmental information of the bolt-type capacitor.

[0035] Specifically, voltage information is obtained through voltage sensors installed at both ends of the capacitor. The sensor measures the working voltage of the capacitor in real time and transmits the data to the control system; current information is obtained through current sensors, which are used to monitor the intensity and fluctuation of current and transmit real-time current data to the system; temperature information is obtained by temperature sensors, which are usually installed on or near the surface of the capacitor and are used to monitor the working temperature of the capacitor. Real-time collection is performed through thermocouples or RTDs (resistance temperature detectors); vibration information is obtained through vibration sensors, which are installed on the capacitor structure and evaluate the operating status of the capacitor by detecting the frequency and amplitude of mechanical vibration; environmental information is obtained through environmental sensors, which are usually used to monitor surrounding environmental parameters such as temperature and humidity.

[0036] S102: The voltage information, current information, temperature information, vibration information and environmental information are integrated through the Kalman filter algorithm to obtain health status information.

[0037] Specifically, each sensor information (voltage, current, temperature, vibration, environment) is preliminarily estimated, and the error of each sensor data is quantified. The error size is usually represented by the noise covariance matrix of the sensor. Then, the Kalman filter calculates a weighted average based on the measured value and error size of each sensor data through a recursive formula. Sensors with higher accuracy (smaller errors) will receive higher weights, thus occupying a larger proportion in the fused health status information. The Kalman filter will dynamically adjust the weights and update the impact of each sensor information on the health status estimation in real time, and finally obtain a fused health status information.

[0038] Furthermore, it is assumed that the following information is obtained at a certain time point: voltage information V k , current information I k , temperature information T k , vibration information Vb k , environment information Env k In order to compare and fuse data from different sensors under the same standard, each data is first standardized to ensure that they are in the same dimension. For example, the minimum-maximum standardization method can be used to map each sensor data value to the [0,1] interval. The calculation method is as follows: , where A, B, C, D, and E refer to the voltage information after standard processing, the current information after standard processing, the temperature information after standard processing, the vibration information after standard processing, and the environmental information after standard processing, respectively. Next, by setting the weight of each data w 1, w 2 , w 3， w 4， w 5 According to the influence of different sensors on health status assessment, the weighted average is calculated and the standardized data is fused into a comprehensive health status information H k : In this process, the weight values ​​w1, w2, w3, w4, and w5 represent the importance of different sensors to the final health status information. These weight values ​​can be adjusted according to actual conditions and experience to ensure that the impact of sensors such as voltage, temperature, vibration, and environment is appropriately weighted. Then, in order to ensure the stability of the health status information and avoid abnormal fluctuations in a short period of time from adversely affecting the overall evaluation, the calculated health status information is smoothed, usually using a sliding average method or a weighted sliding average method. Assuming that a window size N is selected, H is smoothed by the following formula k To smooth: , where Hsmooth It is the fused health status information.

[0039] In one embodiment, if Figure 3 As shown, in step S20, based on the health status information, a hierarchical fault detection mechanism is adopted to classify the faults of the bolt-type capacitor into multiple levels, including minor faults, serious faults and emergency faults, including: S201: Compare the health status information with a preset health status threshold, where the health status threshold includes a normal operating range, a minor fault threshold, a major fault threshold, and an emergency fault threshold.

[0040] Specifically, the health status information is compared with multiple preset thresholds to determine the operating status of the capacitor. The preset health status thresholds are set based on historical data and equipment performance standards, and they reflect the performance boundaries of the capacitor in different working stages or fault states. Specifically, first, the acquired health status information is usually a multidimensional data, including multiple indicators such as voltage, current, temperature, vibration, etc., which comprehensively reflect the current operating status of the capacitor. Then, these health status information will be compared with the preset thresholds one by one. For example, if an indicator such as current or temperature exceeds the upper or lower limit of the normal working range, it can be judged as a sign of fault, and further compared with the thresholds of minor faults, major faults and emergency faults. For each indicator, first check whether it is within the normal working range. If it is within the range, the capacitor is considered to be in a normal state; if it exceeds the normal working range but does not reach the minor fault threshold, it enters a minor fault state; if it exceeds the minor fault threshold but does not reach the major fault threshold, it enters a major fault state; if it exceeds the major fault threshold, it is determined to be an emergency fault state.

[0041] S202: If the health status information does not exceed the normal working range, it is determined to be in a normal working state and fault detection is stopped.

[0042] Specifically, if all health status information is within the normal working range, it means that the bolt-type capacitor has no signs of abnormal fault. At this time, the system will calibrate the current state of the capacitor as "normal working state" and further decide to stop the fault detection process. The specific operation process is to first check whether the various sensor data (such as voltage, current, temperature, vibration, etc.) are within the preset normal working range. If their values ​​do not exceed the normal range, the system will automatically determine that the capacitor is in a normal state. In this case, the fault detection process will no longer continue because there is no need for further fault diagnosis and repair operations. At this time, the monitoring system can enter a sleep or standby state and continue to observe the operating status of the capacitor.

[0043] S203: If the health status information exceeds the normal working range, a hierarchical fault detection mechanism is adopted.

[0044] Specifically, if certain parameters of the health status information (such as voltage, current, temperature, etc.) exceed the preset normal working range, it is necessary to enable the hierarchical fault detection mechanism. This means that the system will adopt different fault detection strategies according to the severity of the fault. First, the system will analyze the parameters that exceed the normal working range, determine the degree of the excess range, and then determine whether it is a minor fault, a major fault, or an emergency fault. If an indicator is slightly beyond the normal range, the system may determine it as a minor fault and take simpler detection measures; if the degree of excess range is large, it may be determined as a major fault or an emergency fault, and the system will start the hierarchical fault detection mechanism.

[0045] S204: If the health status information exceeds the normal operating range but does not exceed the minor fault threshold, it is determined to be a minor fault; if the health status information exceeds the minor fault threshold but does not exceed the major fault threshold, it is determined to be a major fault; if the health status information exceeds the major fault threshold, it is determined to be an emergency fault.

[0046] Specifically, the level of the fault is determined step by step based on the comparison results of the health status information with each fault threshold. First, if a certain health status information (such as voltage, current, temperature, etc.) exceeds the normal working range but is still within the minor fault threshold, the system will determine the fault as a minor fault. At this time, the system will record the abnormal situation, but will not take complex repair measures immediately. Instead, it will enter a light monitoring mode to track the change trend of the parameter. Next, if the health status information further exceeds the minor fault threshold but has not yet reached the major fault threshold, the system will determine the fault as a major fault. At this time, the system will trigger more sensitive detection measures, which may require regular inspections or the initiation of preventive repair operations. Finally, if the health status information exceeds the major fault threshold, the system will determine it as an emergency fault.

[0047] In one embodiment, as shown in step 4, in step S30, if the fault type is a minor fault or a major fault, historical fault information is obtained, and based on the historical fault information and the fault type, the development trend of the fault is analyzed to generate fault development warning information, including: S301: If the fault type is a minor fault, a linear regression analysis method is used based on historical fault information to predict the future development trend of the minor fault and generate corresponding fault development warning information.

[0048] Specifically, information related to minor faults is extracted from historical fault information, mainly including the time, frequency, and possible triggering conditions of the fault. This information will be used as input to analyze the trend of faults. Through the linear regression method, the time of fault occurrence is used as the independent variable, and the frequency or severity of the fault is used as the dependent variable to calculate the relationship between the two. Specifically, by calculating the relationship between the frequency or number of faults and time over a period of time, a simple linear relationship is obtained. Based on this relationship, the trend of minor faults in a certain period of time in the future can be inferred, such as whether the fault frequency may increase or whether it may stabilize at the current level. Then, the predicted results are compared with the preset warning threshold. If the probability of the fault is higher than the threshold, the system will generate a warning message.

[0049] S302: If the fault type is a serious fault, based on historical fault information, a time series analysis method is used to predict the future aggravation trend of the serious fault, and generate corresponding fault development warning information.

[0050] Specifically, firstly, relevant time series information is extracted from the historical fault information, such as the time point when the fault occurred, the frequency of the fault, the severity of each fault, etc. Then, based on this information, the time series analysis method is used to predict the future fault development trend. Specifically, the time series analysis method includes calculating the time interval between faults and the changing trend of the fault severity. For example, by calculating the time difference between each fault occurrence, it is determined whether there is a trend of aggravation. If the frequency of the fault gradually decreases or the severity gradually increases, it means that the fault may be aggravating. By counting the occurrence of historical faults, the trend warning information of the possible aggravation of the fault in the future period is generated. If the severity of the warning information exceeds the predetermined threshold, the corresponding warning information is generated.

[0051] In one embodiment, as shown in step 5, in step S301, if the fault type is a minor fault, a linear regression analysis method is used based on historical fault information to predict the future development trend of the minor fault and generate corresponding fault development warning information, including: S3011: Extract key feature information from historical fault information, including fault occurrence time and fault frequency. Use the key feature information to establish a linear regression model, with the fault occurrence time as the independent variable and the fault frequency as the dependent variable. Fit the linear relationship of the fault development trend to obtain a trained linear regression model.

[0052] Specifically, the occurrence time and frequency of each fault are extracted from the historical fault information. The occurrence time of each fault refers to the specific moment when each fault occurs, which is usually recorded in the form of a timestamp, while the fault frequency indicates the number of times a certain type of fault occurs in a unit time. For example, fault records can be collected in the past few months or years, and each record point corresponds to the specific time of a fault and the number of times the fault type occurs in that time period. Through these data, a time series can be obtained, in which each data point of the time series contains time information and the corresponding fault frequency. Next, the fault occurrence time is used as the independent variable (x) and the fault frequency is used as the dependent variable (y) to establish a linear regression model. The purpose of linear regression is to find the most suitable relationship between data points by fitting a straight line, that is, to find the best fitting line by minimizing the sum of squares of the error between the predicted value and the actual value. In this process, the least squares method can be used to solve the regression coefficient. The least squares method calculates the vertical distance from all data points to the fitting line and minimizes the sum of squares of these distances to obtain a linear relationship that minimizes the error. The equation of this line is: , where m is the slope, which indicates the rate of change of the fault frequency with the time of fault occurrence, and b is the intercept, which indicates the initial fault frequency at time 0. During the fitting process, these two parameters are adjusted to minimize the error between the actual fault data point and the regression line. After the fitting is completed, the obtained linear regression model can predict the future fault frequency.

[0053] S3012: Apply the trained linear regression model to the future development trend of minor faults to obtain prediction results. Based on the prediction results, calculate the probability value of future faults, compare the probability value with the preset warning threshold, and generate corresponding fault development warning information if the probability value exceeds the preset warning threshold.

[0054] Specifically, the trained linear regression model is used to calculate based on the predicted future fault occurrence time (independent variable) and the historical fault frequency (dependent variable). The form of the linear regression model is: y=mx+b, where y is the fault frequency, x is the time (i.e., the moment when the fault occurs), m is the slope of the regression model, and b is the intercept. By substituting the fault occurrence time as input into the regression model, the predicted fault frequency at that moment is obtained. In other words, based on the time when the fault occurs (for example, predicting the next hour, the next quarter, etc.), the model outputs a fault frequency value. Assume that through statistical analysis of historical fault data, it is concluded that the probability of a fault occurrence can be proportional to the fault frequency (or has other specific relationships). For example, in historical data, when the fault frequency is higher, the probability of a fault occurrence is also higher, then the predicted fault frequency value can be used to calculate the probability of a fault occurrence at that moment. Calculate the probability of a fault: By using historical data to establish a mathematical model of the probability of a fault and the fault frequency, assuming that the relationship between the fault probability (P) and the fault frequency (f) is linear, it is expressed as: , where α and β are coefficients obtained by fitting historical data. In this way, after obtaining the predicted fault frequency, the probability of fault occurrence can be calculated by the above formula. After obtaining the probability value of fault occurrence, it is compared with the pre-set warning threshold. For example, assuming that a warning threshold P is set thresh , when the probability of failure exceeds the threshold, it indicates that the failure risk is high and an early warning needs to be issued. （f） Greater than the preset threshold P thresh , it means that the possibility of fault occurrence is high, thus generating corresponding fault development warning information.

[0055] In one embodiment, as shown in step 6, in step S302, if the fault type is a serious fault, a time series analysis method is used based on historical fault information to predict the future aggravation trend of the serious fault, and generate corresponding fault development warning information, including: S3021: Obtain time series feature information from historical fault information. The time series feature data includes the time series of fault occurrence, fault severity, and fault frequency. The time series feature information data is input into the LSTM model for training. The trained LSTM model is obtained through the long-term dependency between the time series of fault occurrence and the fault severity and fault frequency.

[0056] Specifically, the time series of fault occurrence, fault severity and fault frequency are extracted from the historical fault information. The time series refers to the actual time of each fault occurrence, which can usually be expressed in timestamps or continuous time units (such as hours, days, etc.). The severity of the fault refers to the impact of each fault event, which can be quantified in a numerical way, such as 1 for a minor fault, 2 for a moderate fault, and 3 for a serious fault; the frequency of the fault refers to the number or frequency of the fault occurring within a certain period of time. This information is combined into a time series feature data set, in which each data point contains features such as time, fault severity and frequency, forming a time series containing multiple historical data points. Then, these time series feature data are input into the LSTM (Long Short-Term Memory) model for training. LSTM is a special recursive neural network (RNN) that can effectively capture long-term dependencies in long time series when processing time series data. The LSTM model controls the storage and forgetting of information through internal gating mechanisms (such as input gate, forget gate, and output gate), thereby overcoming the gradient vanishing problem that occurs when traditional RNN models are trained in long sequences. It can remember information over a long period of time and is suitable for predicting time series data such as faults. During the training process, the LSTM model automatically learns to extract the complex relationship between the time, severity, and frequency of faults from the time series, and finds an optimal combination of model parameters by continuously adjusting model parameters and optimizing the loss function. Specifically, the LSTM model calculates the state of each time point based on the input historical fault data, and recursively combines the current fault information with the previous historical information to predict future fault development trends. After training, the LSTM model can capture the long-term dependencies between the various features in the time series and generate a trained LSTM model.

[0057] S3022: Apply the trained LSTM model to the prediction of severe faults, calculate the probability of severe faults occurring at a predetermined time point, and trigger fault development warning information if the probability exceeds the preset fault aggravation warning threshold.

[0058] Specifically, the key time series feature information (including fault occurrence time, fault severity, and fault occurrence frequency, etc.) in the historical fault information is input into the pre-trained LSTM model. LSTM (Long Short-Term Memory Network) can effectively capture the long-term dependencies in time series information, that is, considering the impact of past faults, predict the possibility of future faults. After the information is input, the LSTM model will gradually process the input information through multiple hidden layers, and perform feature learning based on the time series of historical faults and the corresponding fault severity information. Specifically, LSTM gradually adjusts the network weights to convert historical fault information into a high-dimensional feature representation to predict the fault state at a certain point in the future. The trained LSTM model can output a probability value of a serious fault at a certain point in the future. This probability value reflects the possibility of a serious fault occurring in a given time range. Then, the calculated probability value will be compared with the preset fault aggravation warning threshold. The preset warning threshold can be obtained through empirical information or historical fault mode analysis, and is a standard value that characterizes the risk of fault occurrence. If the calculated probability of fault occurrence is greater than the set threshold, the fault development warning information is triggered.

[0059] In one embodiment, as shown in step 7, in step S40, the risk type is determined according to the fault development warning information, and a repair strategy is obtained based on the risk type. According to the repair strategy, the automated repair system performs a repair operation, including: S401: Analyze the severity of the fault based on the fault development warning information and classify the severity into low risk, medium risk and high risk.

[0060] Specifically, the classification is based on several key features of the fault, including the frequency of occurrence of the fault, the scope of impact, the development trend, and the predicted time. First, the following data is extracted from the fault development warning information: The frequency of fault occurrence refers to the frequency of occurrence of the fault within a certain time range. The higher the frequency, the more likely the fault has been transformed from an occasional problem to a systemic problem. The scope of impact refers to the degree of impact of the fault on the system, and the assessment of whether the fault is limited to a local component or has expanded to multiple components. The development trend of the fault refers to the analysis of the evolution trend of the fault by comparing historical data with current data. If the trend shows that the fault is intensifying, it means that the risk is high. The predicted time refers to the assessment of the risk based on the predicted time of the fault. The closer the time to the fault occurs, the higher the urgency. Based on these key features, the faults are divided into three categories: low risk, medium risk, and high risk by setting thresholds. The specific classification rules are as follows: Low risk, if the fault frequency is low (for example, only 1-2 times a month), the scope of impact is limited to individual components, the fault development trend is relatively stable, and the time to the next fault is far (for example, more than a few months), then it is judged as low risk. Medium risk: if the frequency of failure is moderate (e.g., 1-2 times per week), and the scope of impact begins to expand to multiple components, but the overall impact is small, and the failure development trend shows signs of aggravation (e.g., the frequency of the last two failures is higher than the previous one), it is judged as medium risk. High risk: if the frequency of failure is high (e.g., multiple times per day), the scope of impact is wide, and the failure development trend is significantly aggravated (e.g., a large negative impact or the speed of failure diffusion is accelerated), and the predicted time for failure is short (e.g., within a few hours or a day), it is judged as high risk.

[0061] S402: If the risk type is low risk, a regular inspection and monitoring strategy is formulated; if the risk type is medium risk, a preventive repair strategy is formulated; if the risk type is high risk, an emergency repair strategy is formulated.

[0062] Specifically, for low-risk situations, regular inspections and monitoring are carried out. The specific operation is to set a schedule for regular inspections, such as monthly or quarterly inspections, focusing on the operating status of the equipment, possible hidden dangers, and checking whether the relevant parameters are within the normal range; at the same time, set up monitoring equipment to collect the status data of the equipment in real time, including voltage, current, temperature, vibration and other information, and conduct automatic monitoring to ensure that warnings are received at the early stage of the problem. For medium-risk situations, preventive repairs are implemented. According to historical fault information, hidden dangers that may develop into serious problems are analyzed, and measures are taken in advance to repair them to prevent further aggravation of the fault. At this time, partial replacement, adjustment or local repair may be carried out according to the health status of the equipment; regular system inspections are carried out, the test results of each time are recorded, and the problems are handled in stages to avoid greater losses due to ignoring early symptoms. For high-risk situations, emergency repair measures must be taken immediately. At this time, a quick response is required to shut down the equipment, troubleshoot and repair the problem. Immediate action should be taken to carry out repairs, replace defective parts, adjust operating status, etc., to ensure that the equipment is quickly restored to a normal working state; if it cannot be repaired on site, spare equipment must be quickly mobilized to replace it to minimize equipment downtime.

[0063] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0064] In one embodiment, a bolt-type capacitor fault detection and repair device is provided, and the bolt-type capacitor fault detection and repair device corresponds to the bolt-type capacitor fault detection and repair method in the above embodiment. Figure 8 As shown, the bolt-type capacitor fault detection and repair device includes a health status information acquisition module, a graded fault detection module, a fault development trend analysis module and an automatic repair execution module. The detailed description of each functional module is as follows: A health status information acquisition module, used to acquire health status information of bolt-type capacitors; A hierarchical fault detection module is used to classify the faults of bolt-type capacitors into multiple levels based on health status information and adopt a hierarchical fault detection mechanism. The multiple levels include minor faults, serious faults and emergency faults, and implement corresponding detection strategies for each level; The fault development trend analysis module is used to obtain historical fault information if the fault type is a minor fault or a major fault, analyze the fault development trend based on the historical fault information and the fault type, and generate fault development warning information; The automatic repair execution module is used to determine the risk type according to the fault development warning information, obtain the repair strategy based on the risk type, and the automatic repair system executes the repair operation according to the repair strategy.

[0065] Optionally, the health status information acquisition module includes: An information acquisition submodule is used to acquire voltage information, current information, temperature information, vibration information and environmental information of the bolt-type capacitor; The fusion calculation submodule is used to fuse voltage information, current information, temperature information, vibration information and environmental information through the Kalman filter algorithm to obtain health status information.

[0066] Optionally, the hierarchical fault detection module includes: The health status information comparison submodule is used to compare the health status information with the preset health status thresholds, which include the normal working range, the minor fault threshold, the major fault threshold and the emergency fault threshold; A normal working state determination submodule is used to determine that the normal working state is in normal working state and stop fault detection if the health state information does not exceed the normal working range; A hierarchical fault detection submodule is used to adopt a hierarchical fault detection mechanism if the health status information exceeds the normal working range; The fault type determination submodule is used to determine that if the health status information exceeds the normal operating range but does not exceed the minor fault threshold, it is a minor fault; if the health status information exceeds the minor fault threshold but does not exceed the major fault threshold, it is determined to be a major fault; if the health status information exceeds the major fault threshold, it is determined to be an emergency fault.

[0067] Optional, fault development trend analysis module includes: A minor fault trend prediction submodule is used to predict the future development trend of a minor fault based on historical fault information and generate corresponding fault development warning information if the fault type is a minor fault. The severe fault aggravation prediction submodule is used to predict the future aggravation trend of severe faults based on historical fault information and generate corresponding fault development warning information if the fault type is severe fault.

[0068] Optional, minor fault trend prediction submodule includes: The historical fault feature extraction and linear regression modeling unit is used to extract key feature information from the historical fault information. The key feature information includes the fault occurrence time and the fault frequency. The linear regression model is established using the key feature information, in which the fault occurrence time is used as the independent variable and the fault frequency is used as the dependent variable. The linear relationship of the fault development trend is fitted to obtain a trained linear regression model. The minor fault prediction and warning unit is used to apply the trained linear regression model to the future development trend of minor faults to obtain the prediction results. According to the prediction results, the probability value of future faults is calculated and the probability value is compared with the preset warning threshold. If the probability value exceeds the preset warning threshold, the corresponding fault development warning information is generated.

[0069] Optionally, the severe fault aggravation prediction submodule includes: The time series feature extraction and LSTM training unit is used to obtain time series feature information from historical fault information. The time series feature data includes the time series of fault occurrence, fault severity and fault frequency. The time series feature information is input into the LSTM model for training. The trained LSTM model is obtained through the long-term dependence between the time series of fault occurrence and the fault severity and fault frequency. The severe fault prediction and warning unit is used to apply the trained LSTM model to the prediction of severe faults, calculate the probability of severe faults occurring at a predetermined time point, and trigger the fault development warning information if the probability exceeds the preset fault aggravation warning threshold.

[0070] Optionally, the automatic repair execution module includes: The fault severity analysis submodule is used to analyze the severity of the fault according to the fault development warning information and classify the severity into low risk, medium risk and high risk; The risk response strategy formulation submodule is used to formulate a regular inspection and monitoring strategy if the risk type is low risk, a preventive repair strategy if the risk type is medium risk, and an emergency repair strategy if the risk type is high risk.

[0071] The specific definition of the bolt-type capacitor fault detection and repair device can be found in the above definition of the bolt-type capacitor fault detection and repair method, which will not be repeated here. Each module in the above-mentioned bolt-type capacitor fault detection and repair device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0072] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Fig. 9As shown. The computer device includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a bolt-type capacitor fault detection and repair method is implemented.

[0073] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program: Get health status information of bolt-on capacitors; Based on the health status information, a hierarchical fault detection mechanism is adopted to classify the faults of bolt-type capacitors into multiple levels, including minor faults, serious faults and emergency faults, and a corresponding detection strategy is implemented for each level; If the fault type is a minor fault or a major fault, historical fault information is obtained, and based on the historical fault information and the fault type, the fault development trend is analyzed to generate fault development warning information; According to the fault development warning information, the risk type is determined, and based on the risk type, the repair strategy is obtained. According to the repair strategy, the automated repair system performs the repair operation.

[0074] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: Get health status information of bolt-on capacitors; Based on the health status information, a hierarchical fault detection mechanism is adopted to classify the faults of bolt-type capacitors into multiple levels, including minor faults, serious faults and emergency faults, and a corresponding detection strategy is implemented for each level; If the fault type is a minor fault or a major fault, historical fault information is obtained, and based on the historical fault information and the fault type, the fault development trend is analyzed to generate fault development warning information; According to the fault development warning information, the risk type is determined, and based on the risk type, the repair strategy is obtained. According to the repair strategy, the automated repair system performs the repair operation.

[0075] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0076] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0077] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for detecting and repairing a bolt-type capacitor fault, characterized in that: The bolt-type capacitor fault detection and repair method comprises: Get health status information of bolt-on capacitors; Based on the health status information, a hierarchical fault detection mechanism is adopted to classify the faults of the bolt-type capacitor into multiple levels, wherein the multiple levels include minor faults, serious faults and emergency faults, and a corresponding detection strategy is implemented for each level; If the fault type is a minor fault or a major fault, historical fault information is obtained, and based on the historical fault information and the fault type, a fault development trend is analyzed to generate fault development warning information; According to the fault development warning information, the risk type is determined, based on the risk type, a repair strategy is obtained, and according to the repair strategy, the automated repair system performs a repair operation.

2. The bolt-type capacitor fault detection and repair method according to claim 1, characterized in that: The health status information includes: Acquiring voltage information, current information, temperature information, vibration information and environmental information of the bolt-type capacitor; The voltage information, the current information, the temperature information, the vibration information and the environmental information are integrated through a Kalman filter algorithm to obtain the health status information.

3. The bolt-type capacitor fault detection and repair method according to claim 1, characterized in that: Based on the health status information, a hierarchical fault detection mechanism is adopted to classify the faults of the bolt-type capacitor into multiple levels, wherein the multiple levels include minor faults, serious faults and emergency faults, including: Comparing the health status information with preset health status thresholds, wherein the health status thresholds include a normal working range, a minor fault threshold, a major fault threshold, and an emergency fault threshold; If the health status information does not exceed the normal working range, it is determined to be in a normal working state and fault detection is stopped; If the health status information exceeds the normal working range, the hierarchical fault detection mechanism is adopted; If the health status information exceeds the normal working range but does not exceed the minor fault threshold, it is determined to be the minor fault; if the health status information exceeds the minor fault threshold but does not exceed the serious fault threshold, it is determined to be the serious fault; if the health status information exceeds the serious fault threshold, it is determined to be the emergency fault.

4. The method for detecting and repairing a bolt-type capacitor fault according to claim 1, characterized in that: If the fault type is a minor fault or a major fault, historical fault information is obtained, and based on the historical fault information and the fault type, the fault development trend is analyzed to generate fault development warning information, including: If the fault type is the minor fault, then based on the historical fault information, a linear regression analysis method is used to predict the future development trend of the minor fault, and generate corresponding fault development warning information; If the fault type is the severe fault, then based on the historical fault information, a time series analysis method is used to predict the future aggravation trend of the severe fault, and generate corresponding fault development warning information.

5. The method for detecting and repairing a bolt-type capacitor fault according to claim 4, characterized in that: If the fault type is the minor fault, a linear regression analysis method is used based on the historical fault information to predict the future development trend of the minor fault and generate corresponding fault development warning information, including: Extract key feature information from the historical fault information, wherein the key feature information includes the fault occurrence time and the fault frequency, and use the key feature information to establish a linear regression model, wherein the fault occurrence time is used as an independent variable and the fault frequency is used as a dependent variable, and a linear relationship of the fault development trend is fitted to obtain a trained linear regression model; The trained linear regression model is applied to the future development trend of the minor fault to obtain a prediction result. Based on the prediction result, the probability value of future fault occurrence is calculated, and the probability value is compared with the preset warning threshold. If the probability value exceeds the preset warning threshold, the corresponding fault development warning information is generated.

6. The method for detecting and repairing a bolt-type capacitor fault according to claim 4, characterized in that: If the fault type is a serious fault, based on the historical fault information, a time series analysis method is used to predict the future aggravation trend of the serious fault, and corresponding fault development warning information is generated, including: Acquire time series feature information from the historical fault information, the time series feature data including the time series of fault occurrence, the fault severity and the fault frequency, use the time series feature information data to input into the LSTM model for training, and obtain the trained LSTM model through the long-term dependency between the time series of fault occurrence and the fault severity and the fault frequency; The trained LSTM model is applied to the prediction of the serious fault, and the probability of the serious fault occurring at a predetermined time point is calculated. If the probability exceeds a preset fault aggravation warning threshold, the fault development warning information is triggered.

7. The method for detecting and repairing a bolt-type capacitor fault according to claim 1, characterized in that: The step of determining the risk type according to the fault development warning information, obtaining a repair strategy based on the risk type, and executing a repair operation by the automated repair system according to the repair strategy includes: Analyze the severity of the fault according to the fault development warning information, and classify the severity into low risk, medium risk and high risk; If the risk type is the low risk, a regular inspection and monitoring strategy is formulated; if the risk type is the medium risk, a preventive repair strategy is formulated; if the risk type is the high risk, an emergency repair strategy is formulated.

8. A bolt-type capacitor fault detection and repair device, characterized in that: The bolt-type capacitor fault detection and repair device comprises: A health status information acquisition module, used to acquire health status information of bolt-type capacitors; A hierarchical fault detection module, configured to classify the faults of the bolt-type capacitor into multiple levels based on the health status information and using a hierarchical fault detection mechanism, wherein the multiple levels include minor faults, serious faults and emergency faults, and implement a corresponding detection strategy for each level; A fault development trend analysis module is used to obtain historical fault information if the fault type is a minor fault or a major fault, analyze the fault development trend based on the historical fault information and the fault type, and generate fault development warning information; The automatic repair execution module is used to determine the risk type according to the fault development warning information, obtain the repair strategy based on the risk type, and according to the repair strategy, the automatic repair system executes the repair operation.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the bolt-type capacitor fault detection and repair method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the bolt-type capacitor fault detection and repair method according to any one of claims 1 to 7 are implemented.

Images