Intelligent fault diagnosis method, device and equipment for circuit breaker and medium

Abstract

The method is mainly applied to the technical field of power system fault diagnosis. The invention discloses an intelligent fault diagnosis method, device and equipment for a circuit breaker and a medium, and the method comprises the steps: updating a preset fault recognition model according to a historical data set, so as to enhance the incidence relation between data features and fault features; acquiring real-time operation data of the target circuit breaker, and performing feature extraction on the real-time operation data to obtain a plurality of data features; key data features are screened out from the multiple data features, and fault features corresponding to the key data features are determined based on the association relationship between the enhanced data features and the fault features; outputting a recognition result corresponding to the fault feature through the updated fault recognition model, wherein the recognition result comprises each fault type and a confidence coefficient corresponding to each fault type; and when the confidence coefficient of any fault type is greater than a preset threshold value corresponding to the fault type, sending out an alarm notification. According to the invention, the precision and efficiency of circuit breaker fault detection can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power system fault diagnosis, and particularly to an intelligent fault diagnosis method, device, equipment and medium for a circuit breaker. Background Art

[0002] A circuit breaker is a key device for protecting and controlling circuits in a power system, and its normal operation is crucial for the stability and safety of the power system. However, during long-term operation, a circuit breaker may fail due to mechanical wear, electrical aging, environmental factors (such as humidity, temperature), etc. Common fault types include mechanical jamming, contact wear, insulation aging, operating mechanism failure, etc. If these faults are not identified and processed in a timely manner, they may lead to power system accidents, causing serious economic losses and social impacts. In recent years, the popularization of intelligent sensors and Internet of Things technology has promoted the development of circuit breaker fault diagnosis methods, but the existing technologies still have the disadvantages of low diagnostic accuracy and calculation efficiency. Summary of the Invention

[0003] The present invention provides an intelligent fault diagnosis method, device, equipment and medium for a circuit breaker, which can improve the accuracy and efficiency of detecting faults of circuit breakers in a power system.

[0004] The present invention provides an intelligent fault diagnosis method for a circuit breaker, and the method includes: Updating a preset fault recognition model according to a historical data set to enhance the correlation between data features and fault features; Obtaining real-time operation data of a target circuit breaker, and performing feature extraction on the real-time operation data to obtain a plurality of data features; Screening out key data features from the plurality of data features, and determining the fault features corresponding to the key data features based on the enhanced correlation between data features and fault features; Outputting an identification result corresponding to the fault features through the updated fault recognition model, where the identification result includes each fault type and the confidence level corresponding to each fault type; When the confidence level of any one of the fault types is greater than a preset threshold corresponding to the fault type, an alarm notification is sent.

[0005] Optionally, the intelligent fault diagnosis method for a circuit breaker further includes steps of constructing the fault recognition model and training the fault recognition model; The constructing the fault recognition model and training the fault recognition model includes: Obtaining a plurality of target parameters, and classifying each target parameter into a corresponding feature parameter group; Encoding each feature parameter group to generate a label corresponding to each feature parameter group; Generate the historical data set based on each of the feature parameter groups and each of the labels; After constructing a model based on a random forest classifier and training the model according to the historical data set, generate a preliminary fault identification model; By traversing the parameter combinations of the preliminary fault identification model, determine the parameter combination with the best parameter configuration, where the parameter combination includes the number of decision trees, depth, and splitting threshold; Adjust the number of decision trees, depth, and splitting threshold of the preliminary fault identification model according to the parameter combination with the best parameter configuration, and use the adjusted fault identification model as the optimized fault identification model.

[0006] Optionally, the intelligent fault diagnosis method for the circuit breaker includes: Extract features from the data in the historical data set to obtain a plurality of historical data features; Sort each of the historical data features based on a preset priority; Filter each sorted historical data feature to obtain target features; Collect data for the target circuit breaker to obtain real-time operating data related to the target features.

[0007] Optionally, the step of filtering each sorted historical data feature to obtain target features includes: Generate a feature quantity list, where the feature quantity list includes a plurality of random quantity values; Select a current random quantity value from the feature quantity list; Use the trained filtering model to filter out the selected features from the plurality of historical data features and classify them into a selected feature subset, where the number of features in the selected feature subset is equal to the current random quantity value; Based on the selected feature subset, determine the performance evaluation value of the filtering model and use it as the performance evaluation value corresponding to the current random quantity value; Return to execute the step of selecting a current random quantity value from the feature quantity list until each random quantity value in the feature quantity list is traversed, and obtain the performance evaluation value corresponding to each random quantity value; Filter out the optimal performance evaluation value from all the performance evaluation values, and use the random quantity value corresponding to the optimal performance evaluation value as the optimal feature quantity value; Filter each sorted historical data feature according to the optimal feature quantity value to obtain target features.

[0008] Optionally, the intelligent fault diagnosis method for the circuit breaker further includes: Obtain historical operation data within a preset time period, and filter the historical operation data to obtain a mechanical vibration signal group, an electrical waveform signal group, and a temperature and humidity sensor signal group; Through data synchronization and alignment processing of the mechanical vibration signal group, the electrical waveform signal group, and the temperature and humidity sensor signal group, to determine a unified time axis for the mechanical vibration signal group, the electrical waveform signal group, and the temperature and humidity sensor signal group; Based on the unified time axis, combine the mechanical vibration signal group, the electrical waveform signal group, and the temperature and humidity sensor signal group into the historical data set.

[0009] Optionally, the process of determining a unified time axis for the mechanical vibration signal group, the electrical waveform signal group, and the temperature and humidity sensor signal group by performing data synchronization and alignment processing on the mechanical vibration signal group, the electrical waveform signal group, and the temperature and humidity sensor signal group includes: Obtain a first signal and a second signal, where the first signal is a signal of the mechanical vibration signal group, and the second signal is a signal of the electrical waveform signal group; Generate a cost matrix based on the Euclidean distance between the first signal and the second signal; By determining the minimum cumulative cost path of the cost matrix, and determining the horizontal offset range and vertical offset range of the minimum cumulative cost path on the time axis according to a preset constraint radius, to determine the unified time axis of the first signal and the second signal.

[0010] Optionally, the process of determining a unified time axis for the mechanical vibration signal group, the electrical waveform signal group, and the temperature and humidity sensor signal group by performing data synchronization and alignment processing on the mechanical vibration signal group, the electrical waveform signal group, and the temperature and humidity sensor signal group includes: Within the temperature and humidity sensor signal group, generate a continuous signal by means of spline interpolation, where the time stamp of the continuous signal is the same as the time stamp of the signal of the electrical waveform signal group, and the interpolation formula used for the spline interpolation is:

[0011] where a represents the signal value and t represents the time.

[0012] The present invention also provides a circuit breaker intelligent fault diagnosis device, and the device includes: A model training module, configured to update a preset fault recognition model according to a historical data set to enhance the correlation between data features and fault features; A feature extraction module, configured to obtain real-time operation data of a target circuit breaker, and perform feature extraction on the real-time operation data to obtain a plurality of data features; A screening module, configured to screen out key data features from the plurality of data features, and determine a fault feature corresponding to the key data feature based on an association relationship between the enhanced data feature and the fault feature; An output module, configured to output an identification result corresponding to the fault feature through an updated fault identification model, where the identification result includes each fault type and a confidence level corresponding to each fault type; An alarm module, configured to send an alarm notification when the confidence level of any one of the fault types is greater than a preset threshold corresponding to the fault type.

[0013] The present invention further provides an electronic device, where the electronic device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the intelligent fault diagnosis method for a circuit breaker described in any one of the above is implemented.

[0014] The present invention further provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the intelligent fault diagnosis method for a circuit breaker described in any one of the above is implemented.

[0015] The present invention has at least the following beneficial effects: In the technical solution of the present invention, first, a historical data set is used to update a fault identification model, strengthening the association relationship between data features and fault features, so that the model can more accurately identify fault patterns; second, real-time operation data of a target circuit breaker is obtained and features are extracted, key data features are screened out, reducing interference from redundant information, and further improving diagnostic accuracy. Based on the updated model, the fault type and the confidence level are output, and an alarm is issued when the confidence level is higher than the threshold, ensuring that faults can be detected in time. Generally speaking, the technical solution of the present invention optimizes the model in combination with historical data, accurately extracts key features in real time and quickly judges faults, significantly improving the accuracy and efficiency of circuit breaker fault detection, and providing a strong guarantee for the stable operation of the power system. Description of the Drawings

[0016] The drawings are used to provide a further understanding of the technical solution of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the technical solution of the present invention, and do not constitute a limitation to the technical solution of the present invention.

[0017] Figure 1 is a flowchart of the steps of the intelligent fault diagnosis method for a circuit breaker in this embodiment; Figure 2 is a flowchart of the steps of constructing and training a fault identification model in the intelligent fault diagnosis method for a circuit breaker in this embodiment; Figure 3 It is another step flowchart of the intelligent fault diagnosis method for the circuit breaker in this embodiment; Figure 4 It is the step flowchart of step S303 in the intelligent fault diagnosis method for the circuit breaker in this embodiment; Figure 5 It is the step flowchart of generating the historical data set in the intelligent fault diagnosis method for the circuit breaker in this embodiment; Figure 6 It is the structural schematic diagram of the intelligent fault diagnosis device for the circuit breaker in this embodiment; Figure 7 It is the structural schematic diagram of an electronic device. Detailed implementation manners

[0018] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0019] Please refer to Figure 1 , Figure 1 It is the step flowchart of the intelligent fault diagnosis method for the circuit breaker in this embodiment.

[0020] This embodiment provides an intelligent fault diagnosis method for a circuit breaker, and the method includes: S101. Update the preset fault recognition model according to the historical data set to enhance the correlation between the data features and the fault features.

[0021] S102. Obtain the real-time operation data of the target circuit breaker, and perform feature extraction on the real-time operation data to obtain multiple data features.

[0022] S103. Screen out the key data features from the multiple data features, and determine the fault features corresponding to the key data features based on the enhanced correlation between the data features and the fault features.

[0023] S104. Output the recognition result corresponding to the fault feature through the updated fault recognition model, and the recognition result includes each fault type and the confidence level corresponding to each fault type.

[0024] S105. When the confidence level of any fault type is greater than the preset threshold corresponding to the fault type, send an alarm notification.

[0025] In some embodiments, the historical data set includes a training set, a validation set, and a test set. Specifically, the operation data of a 10 kV circuit breaker in a substation over 3 years is collected, including 12,000 samples in the normal state and 4 types of fault samples. The data is divided into a 70% training set, a 15% validation set, and a 15% test set to ensure class balance.

[0026] In step S104 of some embodiments, by inputting the values of key data features into the updated fault identification model, key data features such as current features and switching time features are standardized by Z-score to avoid the influence of dimensional differences on model training.

[0027] In step S105 of some embodiments, the fault identification results are output through a visualization interface and an alarm interface. The model parameters are updated monthly based on new data to adapt to the aging state of the device, and the prediction results and confidence levels are recorded for subsequent analysis and optimization.

[0028] It can be understood that the current peak value and the opening and closing time are key data features, which are directly related to the mechanical action of the circuit breaker and the arc energy. The humidity has low importance, and it may be considered to be excluded in subsequent feature screening due to the small difference in sample environments.

[0029] Please refer to Figure 2 , Figure 2 is the flowchart of the steps for constructing and training a fault identification model in the intelligent fault diagnosis method of the circuit breaker in this embodiment.

[0030] In some embodiments, an intelligent fault diagnosis method for a circuit breaker further includes steps of constructing a fault identification model and training the fault identification model.

[0031] The specific implementation manners of the steps of constructing a fault identification model and training the fault identification model are as follows: S201. Obtain a plurality of target parameters and classify each target parameter into a corresponding feature parameter group.

[0032] S202. Encode each feature parameter group to generate a label corresponding to each feature parameter group.

[0033] S203. Generate a historical data set based on each feature parameter group and each label.

[0034] S204. Based on a random forest classifier, construct a model and perform model training according to the historical data set to generate a preliminary fault identification model.

[0035] S205. By traversing the parameter combinations of the preliminary fault identification model, determine the parameter combination with the best parameter configuration, where the parameter combination includes the number of decision trees, the depth, and the splitting threshold.

[0036] S206. Adjust the number of decision trees, depth, and splitting threshold of the preliminary fault identification model according to the parameter combination configured with the optimal parameters, and use the adjusted fault identification model as the optimized fault identification model.

[0037] It can be understood that in this embodiment, the fault identification model is constructed and optimized through a series of fine steps, significantly improving the accuracy and reliability of fault diagnosis. First, by obtaining and classifying the target parameters, a clear structured data foundation is provided for the model input. Then, the feature parameter groups are encoded and labels are generated, further standardizing the data format and providing high-quality input for model training. Based on the generated historical data set, a preliminary model is constructed using a random forest classifier, and by traversing the parameter combinations, the optimal parameter configuration is accurately determined, including key parameters such as the number of decision trees, depth, and splitting threshold. Finally, the preliminary model is adjusted according to the optimal parameter configuration to generate the optimized fault identification model. This process not only makes full use of the historical data, but also greatly improves the model performance through parameter optimization, enabling it to more accurately identify faults and providing an efficient and reliable solution for related fault diagnosis tasks.

[0038] Specifically, the initialization parameters of the random forest classifier include: n_estimators = 100, random_state = 42, and the Gini impurity is used as the node splitting criterion. The parameter search grid is defined as: n_estimators ∈ {50, 100, 200}, max_depth ∈ {None, 10, 20}, min_samples_split ∈ {2, 5, 10}.

[0039] In step S202 of some embodiments, the label can be 0, 1, 2, or 3, where 0 represents normal; 1 represents mechanical jamming; 2 represents contact wear; 3 represents insulation aging.

[0040] In step S206 of some embodiments, the generalization ability of the model on unseen data is ensured through 5-fold cross-validation (5-FoldCV). The cross-validation score (such as 0.91) is more reliable than a single split of the training set / test set. The hyperparameters are optimized through grid search to improve the generalization ability.

[0041] Specifically, the number of decision trees refers to n_estimators. The more trees there are, the more stable it is, but the computational cost increases; the depth refers to max_depth, that is, the maximum depth of a single tree, which is used to control model overfitting; the splitting threshold refers to max_features, that is, the maximum number of features considered when each tree splits (default sqrt).

[0042] It can be understood that the way to adjust the number, depth and splitting threshold of decision trees of the preliminary fault identification model is through optimizing the training of the preliminary fault identification model. Specifically, a feature importance evaluation method based on Gini impurity is adopted to dynamically select the top N high-importance features for node splitting; during the training process, the generation of multiple decision trees is accelerated through a parallel computing framework, and the maximum depth of each tree is set not to exceed a preset threshold to control the complexity of the model; after each round of iteration, the out-of-bag error estimation is used to verify the model performance, and the sample weight distribution is adjusted based on the verification results to optimize the subsequent training process.

[0043] Please refer to Figure 3 , Figure 3 This is another flowchart of the circuit breaker intelligent fault diagnosis method of this embodiment.

[0044] In some embodiments, a circuit breaker intelligent fault diagnosis method includes: S301: Extract features from the historical data set to obtain multiple historical data features.

[0045] S302: Sort each historical data feature based on a preset priority.

[0046] S303: Screen each sorted historical data feature to obtain a target feature.

[0047] S304: Collect data from the target circuit breaker to obtain real-time operating data related to the target characteristics.

[0048] It can be understood that this embodiment significantly improves the accuracy and efficiency of circuit breaker fault diagnosis by optimizing the data processing process. First, feature extraction is performed on the historical data set to mine multiple key historical data features, laying the foundation for subsequent analysis. Next, these features are sorted based on preset priorities. This process helps to screen out the features that are most valuable for fault diagnosis and avoids the interference of redundant information. Through further screening, target features are obtained. These features are strictly screened and the most relevant data for fault diagnosis. Finally, real-time operating data related to the target features is collected for the target circuit breaker to ensure that the collected data is highly matched with the fault diagnosis requirements, further improving the pertinence and accuracy of the diagnosis. Overall, this solution optimizes the feature selection and data collection links through a refined data processing process, allowing fault diagnosis to focus more accurately on key information, thereby effectively improving diagnostic efficiency and reliability, and providing strong support for the rapid and accurate identification of circuit breaker faults.

[0049] In some embodiments, the way to collect data from the circuit breaker is to install sensors at key positions of the circuit breaker and connect them to the data acquisition system, ensuring that the sensors work synchronously with the timer, performing multiple opening and closing operations and collecting the opening and closing times (ms), peak current (A), vibration spectrum energy (J), temperature change rate (℃ / s), and humidity (%).

[0050] In step S301 of some embodiments, the features extracted from the historical data set are: vibration features, electrical features, and environmental features. The vibration features can be obtained by extracting the spectrum energy (J) of the signal in the 0-500Hz frequency band through FFT; the electrical features can be obtained by extracting the arc energy E arc , current harmonic distortion rate (THD); the environmental features can be obtained by extracting the temperature change rate (△T / △t) and temperature and humidity fluctuation gradient. It can be understood that the vibration spectrum energy reflects the looseness or wear of mechanical components, the arc energy is directly related to contact ablation, and the temperature change rate captures the risk of local overheating.

[0051] In a specific embodiment, the arc energy is calculated based on the integration of the current waveform:

[0052] where : contact resistance (fitted according to historical data, default value 0.05Ω). The arc energy is used as an electrical feature and input into the model to reflect the degree of contact ablation.

[0053] Please refer to Figure 4 , Figure 4 which is the flowchart of step S303 in the intelligent fault diagnosis method of the circuit breaker in this embodiment.

[0054] In some embodiments, step S303 includes: S401. Generate a feature quantity list, and the feature quantity list includes multiple random quantity values.

[0055] S402. Select a current random quantity value from the feature quantity list.

[0056] S403. Use the trained screening model to screen out the selected features from multiple historical data features and classify them into the selected feature subset, where the number of features in the selected feature subset is equal to the current random quantity value.

[0057] S404. Based on the selected feature subset, determine the performance evaluation value of the screening model and use it as the performance evaluation value corresponding to the current random quantity value.

[0058] S405. Return to execute step S402 until each random quantity value in the feature quantity list is traversed, and obtain the performance evaluation value corresponding to each random quantity value.

[0059] S406. Select the optimal performance evaluation value from all the performance evaluation values, and use the random quantity value corresponding to the optimal performance evaluation value as the optimal feature quantity value.

[0060] S407. Screen each sorted historical data feature according to the optimal feature quantity value to obtain the target feature.

[0061] In some embodiments, traverse each candidate feature quantity in the feature quantity list and perform the following operations: Based on the Recursive Feature Elimination (RFE) algorithm, with the classifier as the base estimator, configure the RFE algorithm to select the features of the current candidate feature quantity; use the training data set to fit the RFE model to obtain the selected feature subset; perform cross-validation on the selected feature subset and calculate the average value of the model performance metrics; if the current model performance metric exceeds the best performance metric, update the best performance metric and the best feature quantity.

[0062] Optionally, the base estimator parameters of the RFE algorithm include at least one of the number of trees, the maximum depth, and the splitting criterion.

[0063] Optionally, K-fold cross-validation is used for cross-validation, where K is an integer greater than 1, and the model performance metric is one of accuracy, F1 score, or AUC value. The specific steps of cross-validation include: Divide the training data set into K subsets; sequentially use each subset as the validation set and the remaining subsets as the training subsets, train the model using the selected feature subset and verify; summarize the K verification results and calculate the average value of the model performance metrics.

[0064] Optionally, the feature quantity list is generated in the following way: according to the feature dimension of the input data, generate candidate feature quantities within a predetermined range at a fixed step length, or receive a user-defined set of candidate feature quantities.

[0065] Please refer to Figure 5 , Figure 5 which is the flowchart of the steps for generating the historical data set in the circuit breaker intelligent fault diagnosis method of this embodiment.

[0066] In some embodiments, a circuit breaker intelligent fault diagnosis method further includes: S501. Obtain the historical operation data within a preset time period, and screen the historical operation data to obtain a mechanical vibration signal group, an electrical waveform signal group, and a temperature and humidity sensor signal group.

[0067] S502. Perform data synchronization and alignment processing on the mechanical vibration signal group, electrical waveform signal group, and temperature and humidity sensor signal group to determine the unified time axis of the mechanical vibration signal group, electrical waveform signal group, and temperature and humidity sensor signal group.

[0068] S503. Based on the unified time axis, merge the mechanical vibration signal group, electrical waveform signal group, and temperature and humidity sensor signal group into a historical data set.

[0069] It can be understood that in this embodiment, through multi-dimensional data fusion and processing, the comprehensiveness and accuracy of circuit breaker fault diagnosis are significantly improved. First, the scheme obtains historical operation data within a preset time period and filters it into a mechanical vibration signal group, an electrical waveform signal group, and a temperature and humidity sensor signal group, covering the key physical quantities of circuit breaker operation. Then, perform data synchronization and alignment processing on these different types of signal groups to determine the unified time axis, solve the problem of inconsistent time bases for multi-source data, and ensure the consistency and comparability of each signal group in time. Finally, based on the unified time axis, merge the three groups of signals into a complete historical data set, providing a rich, accurate, and consistent data basis for the training and analysis of subsequent fault diagnosis models. This multi-dimensional data fusion method can more comprehensively reflect the operating state of the circuit breaker, thereby improving the accuracy and reliability of fault diagnosis and providing a strong guarantee for the stable operation of the power system.

[0070] Optionally, the steps of obtaining historical operation data include: Load the circuit breaker data set and read the original data file containing fault type labels; Detect and process data missing values, and perform data cleaning by deleting data records containing missing values; Divide the data set into a feature matrix and a label vector, where the label is the fault type column; Split the data into a training set and a test set; Perform standardization processing on continuous features, and use a standardization scaler to normalize the specified continuous features of the training set and the test set.

[0071] In some embodiments, the specific method for processing missing values is: count the number of missing values for each feature. If the proportion of missing values exceeds the preset threshold, delete the corresponding feature; otherwise, delete the sample rows containing missing values.

[0072] In some embodiments, the specific implementation manner of performing data synchronization and alignment processing on the mechanical vibration signal group, electrical waveform signal group, and temperature and humidity sensor signal group to determine the unified time axis of the mechanical vibration signal group, electrical waveform signal group, and temperature and humidity sensor signal group includes: Obtain a first signal and a second signal, where the first signal is a signal of a mechanical vibration signal group, and the second signal is a signal of an electrical waveform signal group; generate a cost matrix based on the Euclidean distance between the first signal and the second signal; determine the minimum cumulative cost path of the cost matrix, and determine the horizontal offset range and the vertical offset range of the minimum cumulative cost path on the time axis according to a preset constraint radius, so as to determine the unified time axis of the first signal and the second signal.

[0073] Specifically, if the sampling rate of the mechanical vibration signal is 1 kHz and the sampling rate of the electrical waveform signal is 10 kHz. The multi-source heterogeneous data synchronization and alignment method is adopted to align the mechanical vibration signal and the electrical waveform signal through dynamic time warping (DTW). Non-uniform sampling alignment is performed on the two signals: calculate the cost matrix of the two signals (Cost Matrix), and define the distance function as the Euclidean distance. The minimum cumulative cost path is found through the dynamic programming algorithm to align the time axis. Set the window constraint radius = 5% to limit the horizontal and vertical offset ranges of the alignment path and prevent excessive distortion of the timing relationship.

[0074] In some embodiments, the specific implementation manner of determining the unified time axis of the mechanical vibration signal group, the electrical waveform signal group, and the temperature and humidity sensor signal group by performing data synchronization and alignment processing on the mechanical vibration signal group, the electrical waveform signal group, and the temperature and humidity sensor signal group includes: Within the temperature and humidity sensor signal group, a continuous signal is generated by spline interpolation, where the time stamp of the continuous signal is the same as the time stamp of the signal of the electrical waveform signal group, and the interpolation formula used for the spline interpolation method is:

[0075] where a represents the signal value and t represents the time.

[0076] For example, the sampling rate of the electrical waveform signal is 10 kHz, and the sampling rate of the temperature and humidity sensor signal is 1 Hz. A continuous signal with the same time stamp as the electrical data (10 kHz) is generated by the cubic spline interpolation method.

[0077] In some embodiments, a circuit breaker intelligent fault diagnosis method further includes: Implement model lightweighting by reducing the number of trees or restricting the model depth, and optimize the model performance through hyperparameter tuning and cross-validation; deploy the lightweighted model to an embedded device, adopt multi-threading or coroutine asynchronous processing for data acquisition and model prediction tasks, and perform batch prediction after accumulating multiple groups of data to reduce the I / O overhead.

[0078] Optionally, the specific implementation of cross-validation and parameter optimization is as follows: Define a parameter search grid that contains combinations of candidate values for the number of decision trees, maximum depth, and minimum sample split threshold; perform hyperparameter optimization on the training set through the grid search algorithm, and select the parameter combination with the optimal model performance. Specifically, the grid search adopts K-fold cross-validation with K = 5, and uses the classification accuracy as the model performance evaluation index.

[0079] In any of the above embodiments, first, use the historical data set to update the fault identification model to strengthen the correlation between data features and fault features, so that the model can more accurately identify fault patterns; second, obtain the real-time operation data of the target circuit breaker in real time and extract features, screen out key data features, reduce the interference of redundant information, and further improve the diagnostic accuracy. Based on the updated model, output the fault type and confidence level. When the confidence level is higher than the threshold, an alarm is issued to ensure that the fault can be detected in time. Overall, this technical solution optimizes the model by combining historical data, accurately extracts key features in real time and quickly judges faults, significantly improving the accuracy and efficiency of circuit breaker fault detection, and providing a strong guarantee for the stable operation of the power system.

[0080] Please refer to Figure 6 , Figure 6 which is the structural schematic diagram of the circuit breaker intelligent fault diagnosis device of this embodiment.

[0081] This embodiment also provides a circuit breaker intelligent fault diagnosis device, including: A model training module 601, configured to update a preset fault identification model according to a historical data set to enhance the correlation between data features and fault features.

[0082] A feature extraction module 602, configured to obtain real-time operation data of a target circuit breaker and perform feature extraction on the real-time operation data to obtain a plurality of data features.

[0083] A screening module 603, configured to screen out key data features from the plurality of data features and determine fault features corresponding to the key data features based on the enhanced correlation between data features and fault features.

[0084] An output module 604, configured to output an identification result corresponding to a fault feature through the updated fault identification model, where the identification result includes each fault type and the confidence level corresponding to each fault type.

[0085] An alarm module 605, configured to issue an alarm notification when the confidence level of any fault type is greater than a preset threshold corresponding to the fault type.

[0086] Those of ordinary skill in the art can understand that all or some of the steps and devices disclosed in the above methods can be implemented as software, firmware, hardware, and their appropriate combinations. Some physical components or all physical components can be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include a computer storage medium (or non-transitory medium) and a communication medium (or transitory medium). As is well known to those of ordinary skill in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassette, tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. It is well known to those of ordinary skill in the art that a communication medium typically contains computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transmission mechanism, and can include any information delivery medium.

[0087] It can be understood that the content in the above method embodiments is applicable to the device embodiments of the present application. The functions specifically implemented by the device embodiments of the present application are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those of the above method embodiments.

[0088] The embodiments of the present application further provide an electronic device, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the intelligent fault diagnosis method of the circuit breaker as described in any one of the above.

[0089] Reference Figure 7 , Figure 7 illustrates the hardware structure of an electronic device according to another embodiment. The electronic device includes: A processor 701, which can be implemented in a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, etc., and is used to execute relevant programs to implement the technical solutions provided by the embodiments of the present application; The memory 702 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM), etc. The memory 702 can store an operating device and other application programs. When implementing the technical solutions provided in the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 702, and the processor 701 is used to call and execute the intelligent fault diagnosis method of the circuit breaker in the embodiments of this application; The input / output interface 703 is used to implement information input and output; The communication interface 704 is used to implement communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or through wireless means (such as mobile network, WIFI, Bluetooth, etc.); The bus 705 transmits information between various components of the device (such as the processor 701, the memory 702, the input / output interface 703, and the communication interface 704); Among them, the processor 701, the memory 702, the input / output interface 703, and the communication interface 704 achieve communication connections with each other inside the device through the bus 705.

[0090] It can be understood that the content in the above method embodiments is applicable to the embodiments of this electronic device. The functions specifically implemented in the embodiments of this electronic device are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0091] The embodiments of this application also provide a computer-readable storage medium, in which there is a program executable by a processor. When the program executable by the processor is executed by the processor, it is used to implement the intelligent fault diagnosis method of the circuit breaker as described in any one of the above specific embodiments.

[0092] The embodiments of this application also disclose a computer program product, including a computer program or computer instructions. The computer program or computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer program or computer instructions from the computer-readable storage medium, and the processor executes the computer program or computer instructions, so that the computer device executes the intelligent fault diagnosis method of the circuit breaker as described in any of the previous embodiments.

[0093] It can be understood that the content in the above method embodiments is applicable to the embodiments of this storage medium. The functions specifically implemented in the embodiments of this storage medium are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0094] In the description of the present application and the above-mentioned accompanying drawings, the terms "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments of the present application described herein can be implemented in an order different from those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product or device comprising a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices. It should be understood that in the present application, "at least one (item)" means one or more, and "a plurality" means two or more.

[0095] In several embodiments provided by the present application, it should be understood that the disclosed apparatus, device and method can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be an indirect coupling or communication connection through some interfaces, devices or units, and can be in the form of electricity, machinery or others.

[0096] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0097] In addition, the functional units in each embodiment of the present application can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0098] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0099] Although the description of this application has been quite detailed and has particularly described several of the described embodiments, it is not intended to be limited to any of these details or embodiments or any particular embodiment, but rather should be regarded as effectively covering the intended scope of this application by considering the prior art to provide a broad interpretation of these claims. In addition, the application has been described above with embodiments foreseeable by the inventors for the purpose of providing a useful description, and non-substantive changes to this application that are not currently foreseeable may still represent equivalent changes to this application.

Claims

1. An intelligent fault diagnosis method for a circuit breaker, characterized in that, The method includes: Updating a preset fault identification model according to a historical data set to enhance the correlation between data features and fault features; Obtaining real-time operation data of a target circuit breaker, and extracting features from the real-time operation data to obtain a plurality of data features; Screening out key data features from the plurality of data features, and determining the fault features corresponding to the key data features based on the enhanced correlation between data features and fault features; Outputting an identification result corresponding to the fault features through the updated fault identification model, where the identification result includes each fault type and the confidence level corresponding to each fault type; When the confidence level of any one of the fault types is greater than a preset threshold corresponding to the fault type, an alarm notification is sent.

2. The intelligent fault diagnosis method for a circuit breaker according to claim 1, wherein The method further includes steps of constructing the fault identification model and training the fault identification model; The constructing the fault identification model and training the fault identification model includes: Obtaining a plurality of target parameters, and classifying each target parameter into a corresponding feature parameter group; Encoding each feature parameter group to generate a label corresponding to each feature parameter group; Generating the historical data set based on each feature parameter group and each label; Generating a preliminary fault identification model after constructing a model based on a random forest classifier and training the model according to the historical data set; Determining a parameter combination with an optimal parameter configuration by traversing the parameter combinations of the preliminary fault identification model, where the parameter combination includes the number of decision trees, depth, and splitting threshold; Adjusting the number of decision trees, depth, and splitting threshold of the preliminary fault identification model according to the parameter combination with the optimal parameter configuration, and using the adjusted fault identification model as the optimized fault identification model.

3. A method for intelligent fault diagnosis of a circuit breaker according to claim 1, characterized in that, The method includes: Extracting features from the data of the historical data set to obtain a plurality of historical data features; Sorting each historical data feature based on a preset priority; Screening each sorted historical data feature to obtain target features; Collecting data of the target circuit breaker to obtain real-time operation data related to the target features.

4. The intelligent fault diagnosis method of a circuit breaker according to claim 3, characterized in that The screening each sorted historical data feature to obtain target features includes: Generating a feature quantity list, where the feature quantity list includes a plurality of random quantity values; Selecting a current random quantity value from the feature quantity list; Using the trained screening model to screen out selected features from the plurality of historical data features and classify them into a selected feature subset, where the number of features in the selected feature subset is equal to the current random quantity value; Determining a performance evaluation value of the screening model based on the selected feature subset and using it as the performance evaluation value corresponding to the current random quantity value; Returning to execute the step of selecting a current random quantity value from the feature quantity list until each random quantity value in the feature quantity list is traversed, and obtaining the performance evaluation value corresponding to each random quantity value; Select the optimal performance evaluation value from all the performance evaluation values, and use the corresponding random quantity value of the optimal performance evaluation value as the optimal feature quantity value; Screen each sorted historical data feature according to the optimal feature quantity value to obtain the target feature.

5. A method for intelligent fault diagnosis of a circuit breaker according to claim 1, characterized in that, The method further includes: Obtain historical operation data within a preset time period, and screen the historical operation data to obtain a mechanical vibration signal group, an electrical waveform signal group, and a temperature and humidity sensor signal group; Perform data synchronization and alignment processing on the mechanical vibration signal group, the electrical waveform signal group, and the temperature and humidity sensor signal group to determine the unified time axis of the mechanical vibration signal group, the electrical waveform signal group, and the temperature and humidity sensor signal group; Based on the unified time axis, merge the mechanical vibration signal group, the electrical waveform signal group, and the temperature and humidity sensor signal group into the historical data set.

6. The intelligent fault diagnosis method for a circuit breaker according to claim 5, wherein The performing data synchronization and alignment processing on the mechanical vibration signal group, the electrical waveform signal group, and the temperature and humidity sensor signal group to determine the unified time axis of the mechanical vibration signal group, the electrical waveform signal group, and the temperature and humidity sensor signal group includes: Obtain a first signal and a second signal, where the first signal is the signal of the mechanical vibration signal group and the second signal is the signal of the electrical waveform signal group; Generate a cost matrix based on the Euclidean distance between the first signal and the second signal; Determine the minimum cumulative cost path of the cost matrix, and determine the horizontal offset range and vertical offset range of the minimum cumulative cost path on the time axis according to a preset constraint radius to determine the unified time axis of the first signal and the second signal.

7. A method for intelligent fault diagnosis of a circuit breaker according to claim 5, characterized in that, The performing data synchronization and alignment processing on the mechanical vibration signal group, the electrical waveform signal group, and the temperature and humidity sensor signal group to determine the unified time axis of the mechanical vibration signal group, the electrical waveform signal group, and the temperature and humidity sensor signal group includes: Within the temperature and humidity sensor signal group, generate a continuous signal by means of spline interpolation, where the time stamp of the continuous signal is the same as the time stamp of the signal of the electrical waveform signal group, and the interpolation formula used for the spline interpolation is: where a represents the signal value and t represents the time.

8. An intelligent fault diagnosis device for a circuit breaker, characterized in that, The device includes: A model training module, configured to update a preset fault identification model according to a historical data set to enhance the correlation between data features and fault features; A feature extraction module, configured to obtain real-time operation data of a target circuit breaker, and perform feature extraction on the real-time operation data to obtain a plurality of data features; A screening module, configured to screen out key data features from the plurality of data features, and determine the fault features corresponding to the key data features based on the enhanced correlation between data features and fault features; An output module, configured to output an identification result corresponding to the fault feature through the updated fault identification model, where the identification result includes each fault type and the confidence level corresponding to each fault type; An alarm module, configured to send an alarm notification when the confidence level of any one of the fault types is greater than a preset threshold corresponding to the fault type.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the circuit breaker intelligent fault diagnosis method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the circuit breaker intelligent fault diagnosis method according to any one of claims 1 to 7.

Images