SF6 circuit breaker fault prediction method and system considering breaking arc effect
By establishing a model relating cumulative arc energy to SF6 gas decomposition product content, and using LSTM and SSA-DBN models, optimizing data, and combining it with the ratio judgment method, the problem of inaccurate SF6 circuit breaker fault prediction under the influence of arc interruption was solved, and accurate diagnosis and prediction of circuit breaker faults were achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- STATE GRID LIAONING ELECTRIC POWER CO LTD
- Filing Date
- 2023-01-31
- Publication Date
- 2026-06-23
Smart Images
Figure CN116106735B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of power equipment condition prediction and fault diagnosis technology, specifically relating to a fault prediction method and system for SF6 circuit breakers that takes into account the effect of interrupting electric arc. Background Technology
[0002] With the construction and development of the power industry, ensuring the safe and stable operation of the power grid is of paramount importance, and the reliable operation of power equipment plays a decisive role. Circuit breakers, as key equipment in the power grid, play a crucial role in controlling and protecting the power system, and timely and accurate assessment of the current and future operating status of the power grid is of great significance.
[0003] In recent years, SF6 has been increasingly widely used in high-voltage switchgear due to its stable chemical properties and superior arc-extinguishing and insulation performance. Meanwhile, further developments in the SF6 gas decomposition mechanism, detection technology, and testing instruments have made gas analysis technology a key technology for predicting and assessing the internal condition of SF6 circuit breakers.
[0004] Current gas analysis techniques mostly rely on calculating SO2, H2S, SOF2, and S2OF. 10 The changes in the content, percentage, and ratio of SF6 gas decomposition products such as CO2 and CF4 are used to predict the short-term operating status of circuit breakers. However, the on-site detection conditions for SF6 gas decomposition products are complex, and it is impossible to fully detect all gas decomposition products, which limits the current gas analysis technology. In addition, when the circuit breaker is performing a normal breaking operation, the breaking arc will affect the decomposition of SF6 gas, causing some gas decomposition products to accumulate inside the equipment. Current gas analysis technology cannot account for the impact of accumulated arc energy on the content of SF6 gas decomposition products, thus limiting its ability to predict faults in SF6 circuit breakers. Summary of the Invention
[0005] The technical problem to be solved by the present invention is to provide a fault prediction method and system for SF6 circuit breakers that takes into account the effect of arc breaking in order to address the shortcomings of the prior art. This method and system solves the technical problem that the existing technology cannot diagnose the fault status of circuit breakers based on the content of characteristic gases, and enables the prediction of the operating status of circuit breakers and timely and accurate judgment of the internal condition of equipment, thus providing technical support for condition-based maintenance based on fault diagnosis.
[0006] The present invention adopts the following technical solution:
[0007] A fault prediction method for SF6 circuit breakers that takes into account the effect of arc interruption includes the following steps:
[0008] S1. Establish the circuit breaker historical interruption dataset, historical operation dataset, and historical fault dataset respectively;
[0009] S2. Based on the circuit breaker historical interruption dataset obtained in step S1, build a model of the relationship between cumulative arc energy and SF6 gas decomposition product content. Based on the model of the relationship between cumulative arc energy and SF6 gas decomposition product content, obtain the SF6 gas decomposition product content corresponding to different interruption times and interruption currents.
[0010] S3. Based on the historical operation dataset obtained in step S1, if the circuit breaker has been interrupted during the historical operation period, combine the cumulative arc energy-SF6 gas decomposition product content relationship model obtained in step S2 to eliminate the influence of the circuit breaker interruption arc on the generation of SF6 gas decomposition products in the fault state, optimize the operation data based on the piecewise linear interpolation method, update the historical operation dataset obtained in step S1, and use it as the historical operation data to be predicted.
[0011] S4. Based on the historical operating data obtained in step S3, build an LSTM-based prediction model for the content of SF6 gas decomposition products and obtain the prediction results of the content of SF6 gas decomposition products.
[0012] S5. Based on the historical fault dataset obtained in step S1, build an SF6 circuit breaker fault diagnosis model based on the ratio judgment method and the SSA-DBN classification method. Use the prediction results of the SF6 gas decomposition product content obtained in step S4 as input to obtain the fault diagnosis results and realize the fault prediction of SF6 circuit breakers.
[0013] Specifically, in step S1, the arc data and SF6 gas decomposition product content after the circuit breaker is opened are collected to establish a historical circuit breaker opening dataset; the SF6 gas decomposition product content during circuit breaker operation is collected to establish a historical operation dataset for predicting SF6 gas decomposition product content; and the SF6 gas decomposition product content after a circuit breaker fault is collected to establish a historical fault dataset for SF6 circuit breaker fault prediction. SF6 gas decomposition products include SO2, H2S, SOF2, and S2OF. 10 The substances involved include HF, H2O, CO, CO2, and CF4. Fault types include partial / corona discharge faults, spark discharge faults, arc discharge faults, and local overheating faults.
[0014] Specifically, in step S2, the relationship model between cumulative arc energy and SF6 gas decomposition product content is as follows:
[0015] φ gas =θ0+θ1I+θ2Times+θ3I 2 +θ4ITimes+θ5Times 2
[0016] Where I is the breaking current, Times is the number of breaking cycles, and θ0, θ1, K, θ5 are the model fitting coefficients.
[0017] Specifically, in step S3, the historical operational data to be predicted is as follows:
[0018] S301. Collect historical data on the content of SF6 gas decomposition products and establish an initial historical operation dataset;
[0019] S302. Determine whether the circuit breaker has ever broken during the historical operation period. If it has, proceed to step S303; otherwise, proceed to step S304.
[0020] S303. Input the number of interruptions and the interruption current into the cumulative arc energy-SF6 gas decomposition product content relationship model to obtain the change in SF6 gas decomposition products caused by the interruption arc.
[0021] S304. Calculate the content of SF6 gas decomposition products when the arc is not interrupted, and obtain the historical operation dataset that eliminates the influence of the arc interruption.
[0022] S305. Optimize the historical running dataset in step S304 using piecewise linear interpolation;
[0023] S306. Establish the final historical data set of SF6 gas decomposition product content.
[0024] Specifically, in step S4, the LSTM-based model for predicting the content of SF6 gas decomposition products is built as follows:
[0025] S401. Take the final obtained historical data set of SF6 gas decomposition product content as input and process the data into array form;
[0026] S402. Convert the time series data into a supervised data sequence, and perform a difference operation on the input data sequence, i.e., subtract two adjacent values.
[0027] S403. Divide the data sequence after the difference operation in step S402 into a training set and a test set;
[0028] S404. Readjust the data, scaling the data values to between (-1, 1) using a scaler;
[0029] S405, Configure the LSTM neural network;
[0030] S406. Use the training set obtained in step S403 to train the LSTM neural network obtained in step S405 to obtain a prediction model of SF6 gas decomposition product content based on LSTM.
[0031] S407. Set the prediction time scale, and use the network model trained in step S406 to predict the test set obtained in step S403 to obtain the prediction results of the SF6 gas decomposition product content.
[0032] Specifically, in step S5, the fault diagnosis of SF6 circuit breakers based on the ratio judgment method and the SSA-DBN classification method is as follows:
[0033] S501. Use the predicted content of SF6 gas decomposition products as input;
[0034] S502. Determine whether the content of each SF6 gas decomposition product is within the normal standard range. If yes, proceed to step S503; otherwise, proceed to step S504.
[0035] The contents of S503 and SF6 gas decomposition products are normal, the circuit breaker is in normal condition, and outputs fault prediction or fault diagnosis results.
[0036] S504, when SO2, H2S, SOF2, S2OF 10 If all four SF6 gas decomposition products are detected, proceed to step S505; otherwise, proceed to step S506.
[0037] S505. The ratio judgment method based on the generation mechanism of SF6 gas decomposition products is selected to diagnose the fault type of the circuit breaker.
[0038] S506. Select the SSA-DBN probabilistic generation model based on deep learning to diagnose the fault types of the circuit breaker.
[0039] S507. Output the fault type diagnosis result of the circuit breaker in step S505 or step S506.
[0040] Furthermore, in step S505, the ratio judgment method is specifically as follows:
[0041] Utilizing the SO2, H2S, SOF2, and S2OF decomposition products of SF6 gas, which best reflect the fault state. 10 Based on the effects of electrical fault discharge intensity and local overheating fault temperature on the four gases, establish the SO2 / H2S and SOF2 / S2OF ratios. 10 The relationship between ratios and specific fault types:
[0042] When the SF6 gas decomposition product content of the circuit breaker is abnormal, if 1 ≤ SO2 / H2S ≤ 2, it is diagnosed as a high-energy arc discharge fault; if 2 ≤ SO2 / H2S ≤ 6, it is diagnosed as a low-energy arc discharge fault; if 5 ≤ SOF2 / S2OF 10If ≤24, the diagnosis is partial discharge / corona discharge fault; if 7≤SO2 / H2S≤24, the diagnosis is local overheating fault.
[0043] Furthermore, in step S506, the specific steps for building the deep learning-based SSA-DBN probabilistic generation model are as follows:
[0044] Using the historical fault dataset of S5061 and SF6 gas decomposition product content as input, the data is normalized to the (0,1) interval, and the training set and test set are divided proportionally.
[0045] The S5062 and SSA algorithms are used for population initialization, determining the DBN network structure, and setting the initial model parameters.
[0046] S5063. Set the ratio of discoverers to followers in the SSA algorithm;
[0047] S5064. Convert the position parameters of each sparrow in the SSA algorithm into weights and biases in the DBN model, and start optimizing the parameters of the DBN model.
[0048] S5065. Select the minimum error rate of the DBN model as the fitness function;
[0049] S5066. When the population in the SSA algorithm discovers food, the positions of the discoverer and the followers are updated sequentially, and a warning person is randomly selected and its position is updated.
[0050] S5067. Calculate and sort the fitness function values based on the updated position information in step S5066.
[0051] S5068. If the optimal fitness function value obtained after sorting satisfies both individual optimum and global optimum, proceed to step S5069; otherwise, proceed to step S5063.
[0052] S5069. Obtain the weights and biases of the DBN model after optimization by the SSA algorithm;
[0053] S50610. Based on the weights and biases in step S5069, perform training to obtain the optimal SSA-DBN classification method model.
[0054] Secondly, embodiments of the present invention provide an SF6 circuit breaker fault prediction system that takes into account the effect of arc breaking, comprising:
[0055] The data module establishes a historical circuit breaker interruption dataset, a historical operation dataset, and a historical fault dataset, respectively.
[0056] The first module builds a model of the relationship between cumulative arc energy and SF6 gas decomposition product content based on the historical interruption dataset of the circuit breaker obtained from the data module. Based on the model of the relationship between cumulative arc energy and SF6 gas decomposition product content, the content of SF6 gas decomposition products corresponding to different interruption times and interruption currents is obtained.
[0057] The update module, based on the historical operation dataset obtained from the data module, if the circuit breaker has been interrupted during the historical operation period, combines the cumulative arc energy-SF6 gas decomposition product content relationship model obtained from the first construction module to eliminate the influence of the circuit breaker interruption arc on the SF6 gas decomposition products generated by the fault state, optimizes the operation data based on the piecewise linear interpolation method, and updates the historical operation dataset obtained from the data module as the historical operation data to be predicted.
[0058] The second module builds a prediction model for the content of SF6 gas decomposition products based on LSTM, using historical operating data obtained from the update module, and obtains the prediction results for the content of SF6 gas decomposition products.
[0059] The prediction module, based on the historical fault dataset obtained from the data module, builds an SF6 circuit breaker fault diagnosis model using the ratio judgment method and the SSA-DBN classification method. The prediction results of the SF6 gas decomposition product content obtained from the second construction module are used as input to obtain the fault diagnosis results, thereby realizing the fault prediction of SF6 circuit breakers.
[0060] Compared with the prior art, the present invention has at least the following beneficial effects:
[0061] This invention discloses a fault prediction method for SF6 circuit breakers that considers the effect of interrupting electric arcs. First, the SF6 gas decomposition product content data of the circuit breaker is processed and optimized to obtain a continuous time-series historical operation dataset. Then, the incremental impact of different interrupting conditions on SF6 gas decomposition products is obtained through a cumulative arc energy-SF6 gas decomposition product content relationship model. After eliminating the effect of interrupting electric arcs, the predicted SF6 gas decomposition product content is obtained based on the SF6 gas decomposition product content prediction model. Furthermore, a circuit breaker fault diagnosis model is used to achieve fault prediction for the SF6 circuit breaker. This invention is based on the SF6 decomposition mechanism, utilizes the SF6 gas decomposition product content to achieve circuit breaker fault prediction, and uses the fault prediction results to achieve real-time analysis and auxiliary decision-making for the circuit breaker's operating status. To eliminate the impact of circuit breaker interruption arcs on the content of SF6 gas decomposition products generated under fault conditions, the relationship between the number of interruptions, interruption current, and SF6 gas decomposition product content was explored. After eliminating the impact of interruption arcs, a historical operational dataset of SF6 gas decomposition product content was established, and an SF6 gas decomposition product prediction model was built to predict the SF6 gas decomposition product content. Using the prediction results of SF6 gas decomposition product content, a fault diagnosis model for SF6 circuit breakers was implemented to diagnose circuit breaker faults.
[0062] Furthermore, the decomposition products of SF6 gas include SO2, H2S, SOF2, and S2OF. 10 HF, H2O, CO, CO2, CF4; Fault types include partial / corona discharge faults, spark discharge faults, arc discharge faults, and local overheating faults.
[0063] Furthermore, establishing a model relating cumulative arc energy to SF6 gas decomposition product content aims to fully consider the impact of the interrupting arc when using SF6 decomposition products to determine the circuit breaker fault status. The principle is as follows: when a circuit breaker interrupts, the arc energy triggers SF6 gas decomposition. With the increase in the number of interruptions and changes in the interrupting current, the content of some SF6 gas decomposition products continuously increases. Using this model, the content of SF6 gas decomposition products generated by interruptions within the circuit breaker is calculated based on the number of interruptions and the interrupting current. Then, if an interruption occurred during the monitoring period of the initial historical operating dataset, the change in SF6 gas decomposition product content caused by the cumulative interrupting arc is subtracted from the SF6 gas decomposition product content in the dataset to obtain a historical operating dataset that eliminates the impact of the interrupting arc. Additionally, due to the complexity of on-site conditions, there are data gaps in the actual measured SF6 gas decomposition products; therefore, a piecewise linear interpolation method is used to optimize the historical operating dataset to eliminate the impact of the interrupting arc. Piecewise linear interpolation has the advantages of low computational complexity and small interpolation error while ensuring data continuity. When applied in prediction models, it exhibits good convergence and high stability.
[0064] Furthermore, the content of SF6 gas decomposition products in the circuit breaker is predicted to obtain the changing trend of SF6 gas decomposition product content in advance. The prediction results are used as input to the circuit breaker fault diagnosis model to obtain the circuit breaker fault prediction results. If the prediction results indicate that the circuit breaker will fail, corresponding measures are taken to avoid the fault based on the predicted fault type. LSTM is a neural network used to process sequential data. The network mainly includes three parts: a forgetting stage, a selective memory stage, and an output stage. Each stage uses a Sigmoid layer to convert the current input into a value between 0 and 1, implementing a gating method. Among them, the forgetting gate (forgetting stage) determines what information to discard from the cell state, i.e., forgetting unimportant information and remembering important information; the input gate (selective memory stage) determines what information to store in the cell state, i.e., remembering more important information and less unimportant information; the output gate (output stage) determines what information to update in the cell state, and which information will be used as the output of the current state.
[0065] Furthermore, a fault diagnosis model for SF6 circuit breakers based on the ratio judgment method and the SSA-DBN classification method is constructed. The monitoring frequency of SF6 gas decomposition product content is low; commonly used electrochemical sensors in field monitoring can only detect small amounts of gaseous decomposition products such as SO2, H2S, and CO. Moreover, due to the complexity of factors involved in SF6 decomposition, actual data is scattered. Given these limitations in detection frequency, detection methods, and current data, the aforementioned model can provide solutions to these limitations, maximizing the accuracy of circuit breaker fault diagnosis.
[0066] Furthermore, when the content of SF6 gas decomposition products is abnormal, if the content values of each gas are relatively comprehensive, the ratio judgment method is used to determine the fault type of the circuit breaker. The ratio judgment method is a fault judgment method based on the generation mechanism of SF6 gas decomposition products. It uses Weibull, normal, and log-normal distributions to fit the data characteristic parameters under each fault type condition, performs parameter estimation and error analysis, and finally determines the characteristic gas ratios under each fault state. For example, when 1 ≤ SO2 / H2S ≤ 2, it is diagnosed as a high-energy arc discharge fault; if 2 ≤ SO2 / H2S ≤ 6, it is diagnosed as a low-energy arc discharge fault; if 5 ≤ SOF2 / S2OF... 10 If the concentration of SO2 / H2S is ≤24, the fault is diagnosed as partial discharge / corona discharge; if 7≤SO2 / H2S≤24, the fault is diagnosed as local overheating. When the concentration of SF6 gas decomposition products is abnormal, if there is limited data for each gas concentration, a classification method based on SSA-DBN is used. SSA-DBN is a probabilistic generative model based on deep learning. Leveraging its ability to mine sample features and its advantage of hierarchical processing of sample feature information, it determines the probability of each fault type occurring in the circuit breaker when the concentration of SF6 gas decomposition products is abnormal. DBN consists of multiple Restricted Boltzmann Machine (RBM) layers stacked together. Its core idea is to acquire data features through a layer-by-layer unsupervised learning mechanism, and then establish a correspondence between data features and target output through supervised learning. Furthermore, the SSA algorithm optimizes the weights and biases of the DBN network through processes such as initializing the population size, selecting the fitness function, finding the individual optimum and the global optimum, updating position information, and calculating the fitness function value, further improving the convergence speed and recognition accuracy of the DBN network.
[0067] It is understandable that the beneficial effects of the second aspect mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here.
[0068] In summary, the present invention has the following advantages:
[0069] (1) The SF6 decomposition products are used to diagnose the fault status of the circuit breaker. The measurement of the SF6 decomposition products does not require disassembling the circuit breaker or power outage operation.
[0070] (2) Using LSTM as the basic model for predicting SF6 decomposition products can provide a high accuracy in predicting the content of SF6 decomposition products while performing rapid calculations.
[0071] (3) By integrating DBN and SSA algorithms, a deep learning-based SSA-DBN probability generation model was built, which improved the accuracy of circuit breaker fault diagnosis and was able to obtain the probability of occurrence of each fault type.
[0072] (4) The impact of the arc breaking action on the SF6 decomposition products generated when the circuit breaker is faulty is considered, which improves the accuracy of using SF6 decomposition products to diagnose or predict the fault status of the circuit breaker.
[0073] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description
[0074] Figure 1 This is a flowchart illustrating the entire process of the method of the present invention;
[0075] Figure 2 This is a schematic diagram showing the fitting results of the model relating cumulative arc energy to SF6 gas decomposition product content in this invention.
[0076] Figure 3 A flowchart for establishing a historical operational dataset of SF6 gas decomposition product content in this invention;
[0077] Figure 4 A flowchart illustrating the construction of an LSTM-based model for predicting the content of SF6 gas decomposition products in this invention.
[0078] Figure 5 This is a flowchart of the SF6 circuit breaker fault diagnosis method based on the ratio judgment method and SSA-DBN classification method of the present invention.
[0079] Figure 6 The flowchart for building the DBN model in this invention;
[0080] Figure 7 The flowchart for constructing the SSA-DBN classification method model for this invention is shown below;
[0081] Figure 8 A schematic diagram of experimental data on the SF6 decomposition products of a certain circuit breaker;
[0082] Figure 9 A schematic diagram showing the SO2 content after considering the effect of arc interruption;
[0083] Figure 10 This is a schematic diagram of a model for predicting the content of SF6 decomposition products. Detailed Implementation
[0084] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0085] In the description of this invention, it should be understood that the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.
[0086] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.
[0087] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. Additionally, the character " / " in this document generally indicates that the preceding and following objects have an "or" relationship.
[0088] It should be understood that although terms such as first, second, third, etc., may be used in the embodiments of the present invention to describe the preset range, these preset ranges should not be limited to these terms. These terms are only used to distinguish the preset ranges from one another. For example, without departing from the scope of the embodiments of the present invention, the first preset range may also be referred to as the second preset range, and similarly, the second preset range may also be referred to as the first preset range.
[0089] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."
[0090] The accompanying drawings illustrate various structural schematic diagrams according to embodiments disclosed in this invention. These drawings are not to scale, and some details have been enlarged for clarity, and some details may have been omitted. The shapes of the various regions and layers shown in the drawings, as well as their relative sizes and positional relationships, are merely exemplary and may deviate from reality due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art can design regions / layers with different shapes, sizes, and relative positions as needed.
[0091] This invention provides a fault prediction method for SF6 circuit breakers that considers the effect of arc breaking. During circuit breaker operation, arc breaking, electrical faults, and local overheating faults can all cause different types and degrees of chemical reactions in the decomposition products of SF6 gas. By optimizing historical operating data, field detection data, and experimental data, a cumulative arc energy-SF6 gas decomposition product content relationship model, an SF6 gas decomposition product content prediction model, and an SF6 circuit breaker fault diagnosis model are established. First, the cumulative arc energy-SF6 gas decomposition product content relationship model is used to obtain the incremental impact of different number of interruptions and interruption currents on SF6 gas decomposition products. After eliminating the influence of arc breaking, the SF6 gas decomposition product content prediction model is used to obtain the predicted result of SF6 gas decomposition product content. Finally, the SF6 circuit breaker fault diagnosis model is used to achieve fault prediction for the SF6 circuit breaker.
[0092] Based on the content of SF6 gas decomposition products, this invention enables fault prediction and diagnosis of SF6 circuit breaker status, timely and accurate judgment of the internal condition of circuit breaker equipment, real-time analysis and auxiliary decision-making of circuit breaker operation risk level, and provides technical support for condition-based maintenance based on fault diagnosis.
[0093] Please see Figure 1 This invention discloses a fault prediction method for SF6 circuit breakers that considers the effect of arc breaking. Based on the generation mechanism of SF6 gas decomposition products, it first collects and processes data on the content of SF6 gas decomposition products, the number of breaks, and the breaking current during circuit breaker operation. Then, based on whether the circuit breaker has broken and the breaking conditions, it eliminates the influence of arc breaking on the content of SF6 gas decomposition products generated inside the circuit breaker, obtaining a continuous time series of SF6 gas decomposition products to be predicted. A prediction model for the content of SF6 gas decomposition products is established based on LSTM (Long Short-Term Memory) network to predict the content of SF6 gas decomposition products. Finally, the prediction results are input into the circuit breaker fault diagnosis model to obtain the diagnosis results, thus realizing fault prediction for SF6 circuit breakers. The specific steps are as follows:
[0094] S1. Collect data on the content of SF6 gas decomposition products, number of interruptions, and interruption current during the operation of the circuit breaker or when a fault occurs, to obtain historical operation dataset and historical fault dataset; process the historical operation dataset and historical fault dataset using interpolation and eigenvalue normalization methods, and determine whether the circuit breaker has interrupted during the data collection period. If it has, proceed to step S2; otherwise, proceed to step S3.
[0095] The gaseous decomposition products include SO2, H2S, SOF2, and S2OF. 10 , HF, H2O, CO, CO2, CF4.
[0096] S2. Based on the cumulative arc energy-SF6 gas decomposition product content relationship model, obtain the SF6 gas decomposition product content corresponding to different number of interruptions and interruption currents, and eliminate the influence of the interruption arc on the change of SF6 gas decomposition product content.
[0097] Please see Figure 2 The cumulative arc energy-SF6 gas decomposition product content relationship model uses the number of interruptions and interruption current as independent variables and the SF6 gas decomposition product content as dependent variable, and performs multivariate polynomial fitting. When the order of the polynomial is 2, the cumulative arc energy-SF6 gas decomposition product content relationship model is shown in Equation 1:
[0098] φ gas =θ0+θ1I+θ2Times+θ3I 2 +θ4ITimes+θ5Times 2 (1)
[0099] Where I is the breaking current, Times is the number of breaking cycles, and θ0, θ1, K, θ5 are the model fitting coefficients.
[0100] To verify the rationality of the model relating cumulative arc energy to SF6 gas decomposition product content, the following was conducted: Figure 2 The experimental data were fitted and calculated, and the fitting results are as follows: Figure 2 As shown in the figure, the green origin represents the true value, and the orange triangle represents the fitted value. It can be seen that the true value and the fitted value almost overlap, and the model's fit R2 value is 1, indicating that the model can be used to describe the relationship between the number of interruptions, the interruption current, and the SO2 content. The established relationship between the cumulative arc energy and the SF6 gas decomposition product content is shown in Equation 2.
[0101] φ SO2 =-58739.597-4.592I+687.133Times-94.127I 2 +270.101Times - 2.858Times 2 (2).
[0102] S3. Establish a historical operational dataset of SF6 gas decomposition product content as historical operational data to be predicted;
[0103] Please see Figure 3 To establish a historical operational dataset of SF6 gas decomposition product content, the specific steps are as follows:
[0104] S301. Collect historical data on the content of SF6 gas decomposition products and establish an initial historical operation dataset;
[0105] S302. Determine whether the circuit breaker has ever broken during the historical operation period. If it has, proceed to step S303; otherwise, proceed to step S304.
[0106] S303. Input the number of interruptions and the interruption current into the cumulative arc energy-SF6 gas decomposition product content relationship model expressed by formula (2) to obtain the change in SF6 gas decomposition products caused by the interruption arc.
[0107] S304. Calculate the content of SF6 gas decomposition products when the arc is not interrupted, and obtain the historical operation dataset that eliminates the influence of the arc interruption.
[0108] S305. Optimize the historical running dataset in step S304 using piecewise linear interpolation;
[0109] S306. Establish the final historical data set of SF6 gas decomposition product content.
[0110] S4. Using the historical data to be predicted in step S3 as input, establish an LSTM-based prediction model for the content of SF6 gas decomposition products and obtain the prediction results of the content of SF6 gas decomposition products.
[0111] Please see Figure 4 The following steps are taken to build an LSTM-based model for predicting the content of SF6 gas decomposition products:
[0112] S401. Take the final obtained historical data set of SF6 gas decomposition product content as input and process the data into array form;
[0113] S402. Convert the time series data into a supervised data sequence, and perform a difference operation on the input data sequence, i.e., subtract two adjacent values.
[0114] S403. Divide the training set and the test set into an 8:2 ratio;
[0115] S404. Readjust the data, scaling the data values to between (-1, 1) using a scaler;
[0116] S405. Configure the LSTM neural network: Set the parameters of each network, such as the number of neuron cores, the number of iterations, the optimizer parameters, etc.
[0117] S406. Perform training to obtain a prediction model for the content of SF6 gas decomposition products based on LSTM.
[0118] S407. Set the prediction time scale, use the trained network model to predict the test data, and obtain the prediction results of the SF6 gas decomposition product content.
[0119] S5. Using the predicted results of the SF6 gas decomposition product content in step S4 as input, and based on the ratio judgment method and SSA-DBN classification method, the SF6 circuit breaker fault diagnosis is realized.
[0120] Please see Figure 5 The specific steps for fault diagnosis of SF6 circuit breakers based on the ratio judgment method and SSA-DBN classification method are as follows:
[0121] S501. Use the predicted content of SF6 gas decomposition products as input;
[0122] S502. Determine whether the content of each SF6 gas decomposition product is within the normal standard range. If yes, proceed to step S503; otherwise, proceed to step S504.
[0123] The contents of S503 and SF6 gas decomposition products are normal. At this time, the circuit breaker is in normal condition and outputs fault prediction or fault diagnosis results.
[0124] S504, judge SO2, H2S, SOF2, S2OF 10 If all four SF6 gas decomposition products are detected, proceed to step S505; otherwise, proceed to step S506.
[0125] S505. The ratio judgment method based on the generation mechanism of SF6 gas decomposition products is selected to diagnose the fault type of the circuit breaker.
[0126] The ratio judgment method is as follows:
[0127] Utilizing the SO2, H2S, SOF2, and S2OF decomposition products of SF6 gas, which best reflect the fault state. 10 Based on the effects of electrical fault discharge intensity and local overheating fault temperature on the four gases, establish the SO2 / H2S and SOF2 / S2OF ratios. 10 Correspondence between ratio and specific fault type: When the content of SF6 gas decomposition products in the circuit breaker is abnormal, if 1≤SO2 / H2S≤2, it is diagnosed as a high-energy arc discharge fault; if 2≤SO2 / H2S≤6, it is diagnosed as a low-energy arc discharge fault; if 5≤SOF2 / S2OF 10 If ≤24, the diagnosis is partial discharge / corona discharge fault; if 7≤SO2 / H2S≤24, the diagnosis is local overheating fault.
[0128] S506. Select the SSA-DBN probabilistic generation model based on deep learning to diagnose the fault types of the circuit breaker.
[0129] Please see Figure 6 and Figure 7The specific steps for building a deep learning-based SSA-DBN probabilistic generation model are as follows:
[0130] Using historical fault datasets of S5061 and SF6 gas decomposition product content as input, the data is normalized to the (0,1) interval, and the training set and test set are divided in an 8:2 ratio.
[0131] Population initialization is performed in the S5062 and SSA algorithms, and the DBN network structure is built at the same time. Parameters such as weights and biases of visible and hidden layers, number of iterations, momentum parameter, learning rate, number of neurons, and number of network layers are set.
[0132] S5063. Set the ratio of discoverers to followers in the SSA algorithm. If the fitness function value changes in step S5068, adjust the ratio of discoverers to followers according to the fitness function value.
[0133] S5064. Convert the position parameters of each sparrow in the SSA algorithm into weights and biases in the DBN model, and start optimizing the parameters of the DBN model.
[0134] S5065. Select the minimum error rate of the DBN model as the fitness function;
[0135] In the S5066 and SSA algorithms, when the population discovers food, the positions of the discoverer and the followers are updated sequentially, and a warning animal is randomly selected and its position is updated.
[0136] S5067. Calculate and sort the fitness function values based on the updated position information in step S5066.
[0137] S5068. Determine whether the optimal fitness function value obtained after sorting satisfies the individual optimum and the global optimum. If it does, proceed to step S5069; otherwise, proceed to step S5063.
[0138] S5069. Obtain the weights and biases of the DBN model after optimization by the SSA algorithm;
[0139] S50610. Based on the weights and biases in step S5069, perform training to obtain the optimal SSA-DBN classification method model.
[0140] S507. Output the fault type diagnosis result of the circuit breaker in step S505 or step S506.
[0141] S6. By predicting the content of SF6 gas decomposition products in step S4 and SF6 circuit breaker fault diagnosis in step S5, the final fault prediction result is obtained.
[0142] In another embodiment of the present invention, an SF6 circuit breaker fault prediction system considering the effect of arc breaking is provided. This system can be used to implement the above-mentioned SF6 circuit breaker fault prediction method considering the effect of arc breaking. Specifically, the SF6 circuit breaker fault prediction system considering the effect of arc breaking includes a data module, a first construction module, an update module, a second construction module, and a prediction module.
[0143] Among them, the data module establishes the circuit breaker historical interruption dataset, historical operation dataset, and historical fault dataset, respectively;
[0144] The first module builds a model of the relationship between cumulative arc energy and SF6 gas decomposition product content based on the historical interruption dataset of the circuit breaker obtained from the data module. Based on the model of the relationship between cumulative arc energy and SF6 gas decomposition product content, the content of SF6 gas decomposition products corresponding to different interruption times and interruption currents is obtained.
[0145] The update module, based on the historical operation dataset obtained from the data module, if the circuit breaker has been interrupted during the historical operation period, combines the cumulative arc energy-SF6 gas decomposition product content relationship model obtained from the first construction module to eliminate the influence of the circuit breaker interruption arc on the SF6 gas decomposition products generated by the fault state, optimizes the operation data based on the piecewise linear interpolation method, and updates the historical operation dataset obtained from the data module as the historical operation data to be predicted.
[0146] The second module builds a prediction model for the content of SF6 gas decomposition products based on LSTM, using historical operating data obtained from the update module, and obtains the prediction results for the content of SF6 gas decomposition products.
[0147] The prediction module, based on the historical fault dataset obtained from the data module, builds an SF6 circuit breaker fault diagnosis model using the ratio judgment method and the SSA-DBN classification method. The prediction results of the SF6 gas decomposition product content obtained from the second construction module are used as input to obtain the fault diagnosis results, thereby realizing the fault prediction of SF6 circuit breakers.
[0148] In another embodiment of the present invention, a terminal device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to achieve a corresponding method flow or corresponding function. The processor described in this embodiment can be used in the operation of an SF6 circuit breaker fault prediction method considering the effect of arc interruption, including:
[0149] Historical interruption datasets, historical operation datasets, and historical fault datasets for circuit breakers were established. Based on the historical interruption dataset, a model was constructed to determine the relationship between cumulative arc energy and SF6 gas decomposition product content. This model yielded the SF6 gas decomposition product content corresponding to different interruption counts and interruption currents. Based on the historical operation dataset, if the circuit breaker had interrupted during historical operation, the influence of the interruption arc on SF6 gas decomposition product generation during fault conditions was eliminated using the cumulative arc energy-SF6 gas decomposition product content model. The operation data was optimized using piecewise linear interpolation, and the historical operation dataset was updated as the historical operation data to be predicted. Based on the historical operation data, an LSTM-based SF6 gas decomposition product content prediction model was built to obtain the predicted SF6 gas decomposition product content. Based on the historical fault dataset, an SF6 circuit breaker fault diagnosis model was built using the ratio judgment method and the SSA-DBN classification method. The predicted SF6 gas decomposition product content was used as input to obtain the fault diagnosis result, thus achieving fault prediction for SF6 circuit breakers.
[0150] In another embodiment of the present invention, a storage medium is provided, specifically a computer-readable storage medium (memory). This computer-readable storage medium is a memory device in a terminal device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and extended storage media supported by the terminal device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device.
[0151] One or more instructions stored in a computer-readable storage medium can be loaded and executed by a processor to implement the corresponding steps of the SF6 circuit breaker fault prediction method considering the effect of arc interruption in the above embodiments; one or more instructions in the computer-readable storage medium are loaded and executed by the processor to perform the following steps:
[0152] Historical interruption datasets, historical operation datasets, and historical fault datasets for circuit breakers were established. Based on the historical interruption dataset, a model was constructed to determine the relationship between cumulative arc energy and SF6 gas decomposition product content. This model yielded the SF6 gas decomposition product content corresponding to different interruption counts and interruption currents. Based on the historical operation dataset, if the circuit breaker had interrupted during historical operation, the influence of the interruption arc on SF6 gas decomposition product generation during fault conditions was eliminated using the cumulative arc energy-SF6 gas decomposition product content model. The operation data was optimized using piecewise linear interpolation, and the historical operation dataset was updated as the historical operation data to be predicted. Based on the historical operation data, an LSTM-based SF6 gas decomposition product content prediction model was built to obtain the predicted SF6 gas decomposition product content. Based on the historical fault dataset, an SF6 circuit breaker fault diagnosis model was built using the ratio judgment method and the SSA-DBN classification method. The predicted SF6 gas decomposition product content was used as input to obtain the fault diagnosis result, thus achieving fault prediction for SF6 circuit breakers.
[0153] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.
[0154] The invention will be explained using experimental data from a certain circuit breaker as an example. Figure 8 The image shows the SF6 decomposition product content record obtained from a certain experimental monitoring of the circuit breaker. This record is the historical operation dataset of the circuit breaker.
[0155] According to the SF6 circuit breaker fault prediction method proposed in this invention, the first step is to check whether the circuit breaker has experienced any interruption in the recent period of this experimental monitoring. The results show that the circuit breaker experienced an interruption within one week before the monitoring of SF6 decomposition products, but the SF6 decomposition products after the interruption were not measured. Therefore, in order to eliminate the influence of normal circuit breaker interruption on fault diagnosis, the content of SF6 decomposition products generated by this interruption is calculated based on the historical interruption records of the circuit breaker, that is, the cumulative arc energy of the circuit breaker-SF6 gas decomposition product content relationship model is established as shown in Equation (3). It should be noted that only SO2, H2S and other gases were monitored in the historical interruption records of the circuit breaker, and SOF2, SO2F2, SOF4 were not generated. Therefore, Equation (3) is the relationship model between SO2 and interruption current and interruption number.
[0156] φ SO2 =-58739.597-4.592I+687.133Times-94.127I 2 +270.101Times - 2.858Times 2 (3)
[0157] The SO2 content generated during this interruption was predicted using equation (3), and the result was 0.345 μL / L. This led to the optimization of the SO2 content in the historical operation dataset, such as... Figure 9 As shown.
[0158] Using the optimized historical operating dataset as the prediction object, an SF6 decomposition product prediction model for this circuit breaker is established based on the LSTM model, such as... Figure 10As shown in the figure, the established prediction model has a good fit, with fitting accuracies of 99.691%, 96.215%, 97.582%, and 99.089% for the historical operating data of SO2, SOF2, SO2F2, and SOF4, respectively. Therefore, LSTM can be used as the basic model for predicting SF6 decomposition products of circuit breakers. The predicted SF6 decomposition product content of this circuit breaker for the next 5 hours is shown in the table.
[0159] Table 1. Predicted results of SF6 decomposition product content
[0160] Prediction time h <![CDATA[SO2μL / L]]> <![CDATA[SOF2μL / L]]> <![CDATA[SO2F2μL / L]]> <![CDATA[SOF4μL / L]]> 101 3.556 36.645 13.348 15.966 102 3.563 37.188 13.435 16.023 103 3.569 37.611 13.517 16.060 104 3.573 37.912 13.589 16.090 105 3.576 38.125 13.650 16.113
[0161] Based on historical fault datasets, an SF6 circuit breaker fault diagnosis model based on SSA-DBN was established. The historical fault dataset was divided into training and test sets in an 8:2 ratio. After optimizing the DBN model parameters using SSA, the DBN model parameters were: 16 hidden layers, a learning rate of 0.1 rbm, a ReLU activation function, and a decay term of 0.000037. Results showed that the diagnostic accuracy on the training set was 97.312%, and the diagnostic accuracy on the test set was 95.161%. The established SSA-DBN fault diagnosis model can effectively detect circuit breaker faults using SF6 decomposition products. Therefore, the SF6 decomposition product content prediction results in Table 1 were used for diagnosis. The diagnosis results are shown in Table 2. The circuit breaker has the highest probability of partial discharge faults within the next 5 hours, followed by arc discharge faults.
[0162] Table 2 Fault prediction results of a certain circuit breaker
[0163]
[0164] In summary, this invention provides a fault prediction method and system for SF6 circuit breakers that considers the effect of arc interruption. When predicting the content of SF6 gas decomposition products, it takes into account the influence of arc interruption on the changes in SF6 gas decomposition product content. On the one hand, based on the cumulative arc energy-SF6 gas decomposition product content relationship model, it obtains the content of SF6 gas decomposition products under different interruption times and interruption currents. On the other hand, after eliminating the effect of arc interruption, it achieves prediction of SF6 gas decomposition product content only under fault conditions, thus providing a more comprehensive understanding of the future fault development trend of the circuit breaker. When predicting circuit breaker faults, it considers the limitations of the complex generation mechanism of SF6 gas decomposition products, long detection cycle, limited types of detected gases, and discrete detection results in actual situations. It introduces a fault diagnosis method combining the ratio judgment method and the SSA-DBN classification method, realizing fault prediction for SF6 circuit breakers and providing guidance for the operation and maintenance of circuit breakers.
[0165] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to 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. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.
[0166] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.
[0167] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.
[0168] In the embodiments provided by this invention, it should be understood that the disclosed devices / terminals and methods can be implemented in other ways. For example, the device / terminal embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.
[0169] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0170] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.
[0171] If the integrated module / unit is implemented as 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, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.
[0172] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0173] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0174] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0175] The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solution based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.
Claims
1. A fault prediction method for SF6 circuit breakers considering the effect of arc interruption, characterized in that, Includes the following steps: S1. Establish the circuit breaker historical interruption dataset, historical operation dataset, and historical fault dataset respectively; S2. Based on the circuit breaker historical interruption dataset obtained in step S1, build a model of the relationship between cumulative arc energy and SF6 gas decomposition product content. Based on the model of the relationship between cumulative arc energy and SF6 gas decomposition product content, obtain the SF6 gas decomposition product content corresponding to different interruption times and interruption currents. S3. Based on the historical operation dataset obtained in step S1, if the circuit breaker has been interrupted during the historical operation period, combine the cumulative arc energy-SF6 gas decomposition product content relationship model obtained in step S2 to eliminate the influence of the circuit breaker interruption arc on the generation of SF6 gas decomposition products in the fault state, optimize the operation data based on the piecewise linear interpolation method, update the historical operation dataset obtained in step S1, and use it as the historical operation data to be predicted. S4. Based on the historical operating data obtained in step S3, build an LSTM-based prediction model for the content of SF6 gas decomposition products and obtain the prediction results of the content of SF6 gas decomposition products. S5. Based on the historical fault dataset obtained in step S1, build an SF6 circuit breaker fault diagnosis model based on the ratio judgment method and the SSA-DBN classification method. Use the prediction results of the SF6 gas decomposition product content obtained in step S4 as input to obtain the fault diagnosis results and realize the fault prediction of SF6 circuit breakers.
2. The SF6 circuit breaker fault prediction method considering the effect of arc interruption according to claim 1, characterized in that, In step S1, the arc data and SF6 gas decomposition product content after the circuit breaker is opened are collected to establish a historical circuit breaker opening dataset; the SF6 gas decomposition product content during circuit breaker operation is collected to establish a historical operation dataset for predicting SF6 gas decomposition product content; the SF6 gas decomposition product content after a circuit breaker fault is collected to establish a historical fault dataset for SF6 circuit breaker fault prediction. SF6 gas decomposition products include SO2, H2S, SOF2, and S2OF. 10 The substances involved include HF, H2O, CO, CO2, and CF4. Fault types include partial / corona discharge faults, spark discharge faults, arc discharge faults, and local overheating faults.
3. The SF6 circuit breaker fault prediction method considering the effect of arc interruption according to claim 1, characterized in that, In step S2, the specific model for the relationship between cumulative arc energy and SF6 gas decomposition product content is as follows: f gas =θ0+θ1I+θ2Times+θ3I 2 +θ4ITimes+θ5Times 2 Where I is the breaking current, Times is the number of breaking cycles, and θ0, θ1, K, θ5 are the model fitting coefficients.
4. The SF6 circuit breaker fault prediction method considering the effect of arc interruption according to claim 1, characterized in that, In step S3, the historical operational data to be predicted is as follows: S301. Collect historical data on the content of SF6 gas decomposition products and establish an initial historical operation dataset; S302. Determine whether the circuit breaker has ever broken during the historical operation period. If it has, proceed to step S303; otherwise, proceed to step S304. S303. Input the number of interruptions and the interruption current into the cumulative arc energy-SF6 gas decomposition product content relationship model to obtain the change in SF6 gas decomposition products caused by the interruption arc. S304. Calculate the content of SF6 gas decomposition products when the arc is not interrupted, and obtain the historical operation dataset that eliminates the influence of the arc interruption. S305. Optimize the historical running dataset in step S304 using piecewise linear interpolation; S306. Establish the final historical data set of SF6 gas decomposition product content.
5. The SF6 circuit breaker fault prediction method considering the effect of arc interruption according to claim 1, characterized in that, In step S4, the LSTM-based model for predicting the content of SF6 gas decomposition products is constructed as follows: S401. Take the final obtained historical data set of SF6 gas decomposition product content as input and process the data into array form; S402. Convert the time series data into a supervised data sequence, and perform a difference operation on the input data sequence, i.e., subtract two adjacent values. S403. Divide the data sequence after the difference operation in step S402 into a training set and a test set; S404. Readjust the data, scaling the data values to between (-1, 1) using a scaler; S405, Configure the LSTM neural network; S406. Use the training set obtained in step S403 to train the LSTM neural network obtained in step S405 to obtain a prediction model of SF6 gas decomposition product content based on LSTM. S407. Set the prediction time scale, and use the network model trained in step S406 to predict the test set obtained in step S403 to obtain the prediction results of the SF6 gas decomposition product content.
6. The SF6 circuit breaker fault prediction method considering the effect of arc interruption according to claim 1, characterized in that, In step S5, the fault diagnosis of SF6 circuit breakers based on the ratio judgment method and the SSA-DBN classification method is as follows: S501. Use the predicted content of SF6 gas decomposition products as input; S502. Determine whether the content of each SF6 gas decomposition product is within the normal standard range. If yes, proceed to step S503; otherwise, proceed to step S504. The contents of S503 and SF6 gas decomposition products are normal, the circuit breaker is in normal condition, and outputs fault prediction or fault diagnosis results. S504, when SO2, H2S, SOF2, S2OF 10 If all four SF6 gas decomposition products are detected, proceed to step S505; otherwise, proceed to step S506. S505. The ratio judgment method based on the generation mechanism of SF6 gas decomposition products is selected to diagnose the fault type of the circuit breaker. S506. Select the SSA-DBN probabilistic generation model based on deep learning to diagnose the fault types of the circuit breaker. S507. Output the fault type diagnosis result of the circuit breaker in step S505 or step S506.
7. The SF6 circuit breaker fault prediction method considering the effect of arc interruption according to claim 6, characterized in that, In step S505, the ratio judgment method is specifically as follows: Utilizing the SO2, H2S, SOF2, and S2OF decomposition products of SF6 gas, which best reflect the fault state. 10 Based on the effects of electrical fault discharge intensity and local overheating fault temperature on the four gases, establish the SO2 / H2S and SOF2 / S2OF ratios. 10 Correspondence between ratio and specific fault type: When the SF6 gas decomposition product content of the circuit breaker is abnormal, if 1 ≤ SO2 / H2S ≤ 2, it is diagnosed as a high-energy arc discharge fault; if 2 ≤ SO2 / H2S ≤ 6, it is diagnosed as a low-energy arc discharge fault; if 5 ≤ SOF2 / S2OF 10 If ≤24, the diagnosis is partial discharge / corona discharge fault; if 7≤SO2 / H2S≤24, the diagnosis is local overheating fault.
8. The SF6 circuit breaker fault prediction method considering the effect of arc interruption according to claim 6, characterized in that, In step S506, the specific steps for building the deep learning-based SSA-DBN probabilistic generation model are as follows: Using the historical fault dataset of S5061 and SF6 gas decomposition product content as input, the data is normalized to the (0,1) interval, and the training set and test set are divided proportionally. The S5062 and SSA algorithms are used for population initialization, determining the DBN network structure, and setting the initial model parameters. S5063. Set the ratio of discoverers to followers in the SSA algorithm; S5064. Convert the position parameters of each sparrow in the SSA algorithm into weights and biases in the DBN model, and start optimizing the parameters of the DBN model. S5065. Select the minimum error rate of the DBN model as the fitness function; S5066. When the population in the SSA algorithm discovers food, the positions of the discoverer and the followers are updated sequentially, and a warning person is randomly selected and its position is updated. S5067. Calculate and sort the fitness function values based on the updated position information in step S5066. S5068. If the optimal fitness function value obtained after sorting satisfies both individual optimum and global optimum, proceed to step S5069; otherwise, proceed to step S5063. S5069. Obtain the weights and biases of the DBN model after optimization by the SSA algorithm; S50610. Based on the weights and biases in step S5069, perform training to obtain the optimal SSA-DBN classification method model.
9. A fault prediction system for SF6 circuit breakers that considers the effect of arc breaking, characterized in that, include: The data module establishes a historical circuit breaker interruption dataset, a historical operation dataset, and a historical fault dataset, respectively. The first module builds a model of the relationship between cumulative arc energy and SF6 gas decomposition product content based on the historical interruption dataset of the circuit breaker obtained from the data module. Based on the model of the relationship between cumulative arc energy and SF6 gas decomposition product content, the content of SF6 gas decomposition products corresponding to different interruption times and interruption currents is obtained. The update module, based on the historical operation dataset obtained from the data module, if the circuit breaker has been interrupted during the historical operation period, combines the cumulative arc energy-SF6 gas decomposition product content relationship model obtained from the first construction module to eliminate the influence of the circuit breaker interruption arc on the SF6 gas decomposition products generated by the fault state, optimizes the operation data based on the piecewise linear interpolation method, and updates the historical operation dataset obtained from the data module as the historical operation data to be predicted. The second module builds a prediction model for the content of SF6 gas decomposition products based on LSTM, using historical operating data obtained from the update module, and obtains the prediction results for the content of SF6 gas decomposition products. The prediction module, based on the historical fault dataset obtained from the data module, builds an SF6 circuit breaker fault diagnosis model using the ratio judgment method and the SSA-DBN classification method. The prediction results of the SF6 gas decomposition product content obtained from the second construction module are used as input to obtain the fault diagnosis results, thereby realizing the fault prediction of SF6 circuit breakers.