An online diagnosis method and device for capacitor breakdown of capacitive voltage transformer

Abstract

The present invention discloses an online diagnosis method and device for capacitor breakdown of a capacitor voltage transformer, by constructing a historical capacitor breakdown data set of the capacitor voltage transformer; extracting from the historical capacitor breakdown data set and constructing a first capacitor breakdown data set based on the inter-group amplitude ratio parameter and the zero-sequence voltage direction variation error characteristic parameter; optimizing the first RBF model based on the Optics algorithm to obtain a second RBF model, and training the second RBF model based on the first capacitor breakdown data set to obtain an optimal RBF model; inputting the capacitor breakdown data of the capacitor voltage transformer to be tested in real time into the optimal RBF model, so that the optimal RBF model outputs the capacitor breakdown online diagnosis result of the capacitor voltage transformer to be tested. The technical solution of the present invention can realize real-time and accurate monitoring of the capacitor breakdown fault of the capacitor voltage transformer, and improve the accuracy of the capacitor breakdown diagnosis result.

Description

Technical Field

[0001] The present invention relates to the technical field of capacitor breakdown diagnosis, and in particular to an online diagnosis method and device for capacitor breakdown of a capacitive voltage transformer. Background Art

[0002] In the prior art, capacitor breakdown fault identification of capacitor elements in capacitive voltage transformers is generally based on massive secondary voltage data of the capacitive voltage transformer, using data mining technology to achieve real-time monitoring of capacitor breakdown, and using secondary voltage changes to identify capacitor breakdown faults of capacitor elements of the capacitive voltage transformer. However, changes in the secondary voltage of the capacitive voltage transformer may be caused by capacitor breakdown or by circuit fluctuations or other interferences. For example, the power-on of parallel reactors and parallel capacitor groups will cause sudden changes in voltage. Therefore, it is not enough to use only secondary voltage changes for capacitor breakdown monitoring. In addition, human experience judgment is required throughout the entire process of capacitor breakdown monitoring, which causes the judgment result to be affected by human subjectivity, resulting in low accuracy of the judgment result. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide an online diagnosis method and device for capacitor breakdown of a capacitive voltage transformer, which can realize real-time and accurate monitoring of the capacitor breakdown fault of the capacitive voltage transformer and improve the accuracy of the capacitor breakdown diagnosis result.

[0004] In order to solve the above technical problems, the present invention provides an online diagnosis method for capacitor breakdown of a capacitive voltage transformer, comprising:

[0005] Obtain historical capacitance breakdown data of capacitor voltage transformers in various capacitance breakdown states, and construct a historical capacitance breakdown data set;

[0006] Extracting an inter-group amplitude ratio parameter and a zero-sequence voltage direction variation error characteristic parameter from the historical capacitor breakdown data set, and constructing a first capacitor breakdown data set based on the inter-group amplitude ratio parameter and the zero-sequence voltage direction variation error characteristic parameter;

[0007] Optimizing the first RBF model based on the Optics algorithm to obtain a second RBF model, and training the second RBF model based on the first capacitor breakdown data set to obtain an optimal RBF model;

[0008] Capacitor breakdown data of the capacitive voltage transformer to be tested is acquired in real time, and the capacitor breakdown data is input into an optimal RBF model, so that the optimal RBF model outputs an online diagnosis result of the capacitor breakdown of the capacitive voltage transformer to be tested.

[0009] In one possible implementation, historical capacitor breakdown data of a capacitive voltage transformer in each capacitor breakdown state, wherein each capacitor breakdown state includes a capacitor non-breakdown state, a high-voltage capacitor breakdown state, two high-voltage capacitor breakdown states, a medium-voltage capacitor breakdown state and two medium-voltage capacitor breakdown states, and the historical capacitor breakdown data includes historical secondary voltage amplitudes and historical phases at each moment.

[0010] In a possible implementation, extracting the inter-group amplitude ratio parameter from the historical capacitance breakdown data set specifically includes:

[0011] Dividing the historical capacitor breakdown data set into a training set and a test set according to a preset ratio;

[0012] A feature vector based on the inter-group amplitude ratio parameter is constructed, and the inter-group amplitude ratio parameter is extracted from the historical capacitor breakdown data set to obtain a first inter-group amplitude ratio parameter of the training set and a second inter-group amplitude ratio parameter of the test set, respectively. The feature vector based on the inter-group amplitude ratio parameter is constructed as follows:

[0013]

[0014] Among them, A zj represents the amplitude ratio parameter between groups, n represents the number of capacitive voltage transformers, V Ai ,V Bi ,V Ci Represents the three-phase secondary voltage amplitude of the i-th capacitor voltage transformer.

[0015] In a possible implementation, extracting the zero-sequence voltage direction variation error characteristic parameter from the historical capacitor breakdown data set specifically includes:

[0016] Respectively obtain the zero-sequence voltage direction change corresponding to the voltage breakdown change moment of each capacitive voltage transformer in the training set and the test set, and simultaneously obtain the standard zero-sequence voltage direction change;

[0017] The error between the zero-sequence voltage direction variation corresponding to each capacitive voltage transformer and the standard zero-sequence voltage direction variation is calculated to obtain a first zero-sequence voltage direction variation error characteristic parameter of the training set and a second zero-sequence voltage direction variation error characteristic parameter of the test set, wherein the zero-sequence voltage direction variation error characteristic parameter is as follows:

[0018] B zj =(ε1…ε n );

[0019] Among them, B zjIt represents the error characteristic parameter of the zero-sequence voltage direction change, n represents the number of capacitive voltage transformers, ε i is the zero-sequence voltage direction change error of the i-th capacitive voltage transformer.

[0020] In a possible implementation, constructing a first capacitor breakdown data set based on the inter-group amplitude ratio parameter and the zero-sequence voltage direction variation error characteristic parameter specifically includes:

[0021] Acquire and obtain a first training set based on the first inter-group amplitude ratio parameter and the first zero-sequence voltage direction change amount error characteristic parameter;

[0022] Acquire and obtain a first test set based on the second inter-group amplitude ratio parameter and the second zero-sequence voltage direction change amount error characteristic parameter;

[0023] The first capacitance breakdown data set is constructed based on the first training set and the first test set.

[0024] In a possible implementation manner, after constructing the first capacitance breakdown data set, the method further includes:

[0025] Perform z-score normalization processing on the first training set to obtain a first normalized training set after z-score normalization processing, wherein the z-score normalization formula is as follows:

[0026]

[0027] Among them, μ is the characteristic mean of this dimension, and σ is the characteristic standard deviation of this dimension;

[0028] updating the first training set according to the first normalized training set to obtain a second training set, wherein the second training set includes a first feature vector set and a first category set;

[0029] Based on the feature mean and the feature standard deviation, performing z-score normalization processing on the first test set to obtain a first normalized test set after the z-score normalization processing;

[0030] According to the first normalized test set, the first test set is updated to obtain a second test set, wherein the second test set includes a second feature vector set and a second category set.

[0031] In a possible implementation, optimizing the first RBF model based on the Optics algorithm to obtain a second RBF model, and training the second RBF model based on the first capacitor breakdown data set to obtain an optimal RBF model, specifically includes:

[0032] Performing nonlinear change processing on the hidden layer of the first RBF model based on a radial basis function, wherein the radial basis function is a Gaussian function;

[0033] Acquire the center of the Gaussian function based on the Optics algorithm, so as to optimize the first RBF model based on the center of the Gaussian function to obtain a second RBF model;

[0034] Assigning data to multiple hyperparameters of the second RBF model and the Optics algorithm;

[0035] Acquire the second training set, and input the second training set into the second RBF model to obtain a third RBF model;

[0036] Acquire the second test set, and input the second test set into the third RBF model, so that the third RBF model outputs a prediction classification result of the second test set, and generates a prediction classification result set;

[0037] Comparing the prediction classification result set with the second category set to obtain the accuracy of the third RBF model, and determining whether the accuracy of the third RBF model reaches a preset accuracy, if so, retaining the third RBF model and its corresponding accuracy, and if not, not retaining the third RBF model and its corresponding accuracy;

[0038] Determine whether the number of the third RBF models currently obtained reaches a preset threshold number. If not, return to the step of "assigning data to multiple hyperparameters of the second RBF model and the Optics algorithm". If yes, proceed to the next step.

[0039] The multiple third RBF models retained are compared in terms of accuracy to obtain the third RBF model corresponding to the highest accuracy, and the third RBF model with the highest accuracy is set as the optimal RBF model.

[0040] The embodiment of the present invention further provides an online diagnosis device for capacitor breakdown of a capacitive voltage transformer, comprising: a data set construction module, a first capacitor breakdown data set construction module, an RBF model optimization module and a capacitor breakdown online diagnosis module;

[0041] The data set construction module is used to construct a historical capacitance breakdown data set based on the historical capacitance breakdown data of the capacitive voltage transformer in each capacitance breakdown state;

[0042] The first capacitor breakdown data set construction module is used to extract the inter-group amplitude ratio parameter and the zero-sequence voltage direction change amount error characteristic parameter from the historical capacitor breakdown data set, and construct the first capacitor breakdown data set based on the inter-group amplitude ratio parameter and the zero-sequence voltage direction change amount error characteristic parameter;

[0043] The RBF model optimization module is used to optimize the first RBF model based on the Optics algorithm to obtain a second RBF model, and perform model training on the second RBF model based on the first capacitor breakdown data set to obtain an optimal RBF model;

[0044] The capacitor breakdown online diagnosis module is used to obtain the capacitor breakdown data of the capacitor voltage transformer to be tested in real time, and input the capacitor breakdown data into the optimal RBF model so that the optimal RBF model outputs the capacitor breakdown online diagnosis result of the capacitor voltage transformer to be tested.

[0045] In one possible implementation, the data set constructs historical capacitor breakdown data of a capacitive voltage transformer in each capacitor breakdown state in the module, wherein each capacitor breakdown state includes a capacitor non-breakdown state, a high-voltage capacitor breakdown state, two high-voltage capacitor breakdown states, a medium-voltage capacitor breakdown state, and two medium-voltage capacitor breakdown states, and the historical capacitor breakdown data includes historical secondary voltage amplitudes and historical phases at each moment.

[0046] In a possible implementation, the first capacitance breakdown data set construction module is used to extract the inter-group amplitude ratio parameter from the historical capacitance breakdown data set, specifically including:

[0047] Dividing the historical capacitor breakdown data set into a training set and a test set according to a preset ratio;

[0048] A feature vector based on the inter-group amplitude ratio parameter is constructed, and the inter-group amplitude ratio parameter is extracted from the historical capacitor breakdown data set to obtain a first inter-group amplitude ratio parameter of the training set and a second inter-group amplitude ratio parameter of the test set, respectively. The feature vector based on the inter-group amplitude ratio parameter is constructed as follows:

[0049]

[0050] Among them, A zj represents the amplitude ratio parameter between groups, n represents the number of capacitive voltage transformers, V Ai ,V Bi ,V Ci Represents the three-phase secondary voltage amplitude of the i-th capacitor voltage transformer.

[0051] In a possible implementation, the first capacitor breakdown data set construction module is used to extract the zero-sequence voltage direction change amount error characteristic parameter from the historical capacitor breakdown data set, specifically including:

[0052] Respectively obtain the zero-sequence voltage direction change corresponding to the voltage breakdown change moment of each capacitive voltage transformer in the training set and the test set, and simultaneously obtain the standard zero-sequence voltage direction change;

[0053] The error between the zero-sequence voltage direction variation corresponding to each capacitive voltage transformer and the standard zero-sequence voltage direction variation is calculated to obtain a first zero-sequence voltage direction variation error characteristic parameter of the training set and a second zero-sequence voltage direction variation error characteristic parameter of the test set, wherein the zero-sequence voltage direction variation error characteristic parameter is as follows:

[0054] B zj =(ε1…ε n );

[0055] Among them, B zj It represents the error characteristic parameter of the zero-sequence voltage direction change, n represents the number of capacitive voltage transformers, ε i is the zero-sequence voltage direction change error of the i-th capacitive voltage transformer.

[0056] In a possible implementation, the first capacitor breakdown data set construction module is used to construct a first capacitor breakdown data set based on the inter-group amplitude ratio parameter and the zero-sequence voltage direction change amount error characteristic parameter, specifically including:

[0057] Acquire and obtain the first training set based on the first inter-group amplitude ratio parameter and the first zero-sequence voltage direction change amount error characteristic parameter;

[0058] Acquire and obtain a first test set based on the second inter-group amplitude ratio parameter and the second zero-sequence voltage direction change amount error characteristic parameter;

[0059] The first capacitance breakdown data set is constructed based on the first training set and the first test set.

[0060] In a possible implementation, an online diagnostic device for capacitor breakdown of a capacitive voltage transformer provided by an embodiment of the present invention further includes: a first capacitor breakdown data set processing module;

[0061] The first capacitor breakdown data set processing module is used to perform z-score normalization processing on the first training set to obtain a first normalized training set after the z-score normalization processing, wherein the z-score normalization formula is as follows:

[0062]

[0063] Among them, μ is the characteristic mean of this dimension, and σ is the characteristic standard deviation of this dimension;

[0064] The first capacitor breakdown data set processing module is used to update the first training set according to the first normalized training set to obtain a second training set, wherein the second training set includes a first feature vector set and a first category set;

[0065] The first capacitor breakdown data set processing module is used to perform z-score normalization processing on the first test set based on the feature mean and the feature standard deviation to obtain a first normalized test set after the z-score normalization processing;

[0066] The first capacitance breakdown data set processing module is used to update the first test set according to the first normalized test set to obtain a second test set, wherein the second test set includes a second feature vector set and a second category set.

[0067] In a possible implementation, the RBF model optimization module is used to optimize the first RBF model based on the Optics algorithm to obtain a second RBF model, and perform model training on the second RBF model based on the first capacitor breakdown data set to obtain an optimal RBF model, specifically including:

[0068] Performing nonlinear change processing on the hidden layer of the first RBF model based on a radial basis function, wherein the radial basis function is a Gaussian function;

[0069] Acquire the center of the Gaussian function based on the Optics algorithm, so as to optimize the first RBF model based on the center of the Gaussian function to obtain a second RBF model;

[0070] Assigning data to multiple hyperparameters of the second RBF model and the Optics algorithm;

[0071] Acquire the second training set, and input the second training set into the second RBF model to obtain a third RBF model;

[0072] Acquire the second test set, and input the second test set into the third RBF model, so that the third RBF model outputs a prediction classification result of the second test set, and generates a prediction classification result set;

[0073] Comparing the prediction classification result set with the second category set to obtain the accuracy of the third RBF model, and determining whether the accuracy of the third RBF model reaches a preset accuracy, if so, retaining the third RBF model and its corresponding accuracy, and if not, not retaining the third RBF model and its corresponding accuracy;

[0074] Determine whether the number of the third RBF models currently obtained reaches a preset threshold number. If not, return to the step of "assigning data to multiple hyperparameters of the second RBF model and the Optics algorithm". If yes, proceed to the next step.

[0075] The multiple third RBF models retained are compared in terms of accuracy to obtain the third RBF model corresponding to the highest accuracy, and the third RBF model with the highest accuracy is set as the optimal RBF model.

[0076] An embodiment of the present invention also provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, an online diagnosis method for capacitor breakdown of a capacitive voltage transformer as described in any one of the above items is implemented.

[0077] An embodiment of the present invention also provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute an online diagnosis method for capacitor breakdown of a capacitive voltage transformer as described in any one of the above.

[0078] The online diagnosis method and device for capacitor breakdown of a capacitive voltage transformer according to the embodiment of the present invention have the following beneficial effects compared with the prior art:

[0079] By acquiring the historical capacitor breakdown data of the capacitor voltage transformer in each capacitor breakdown state, a historical capacitor breakdown data set is constructed, and a first capacitor breakdown data set is constructed based on the amplitude ratio parameter between groups and the error characteristic parameter of the zero-sequence voltage direction change amount extracted from the historical capacitor breakdown data set; the first RBF model is optimized based on the Optics algorithm to obtain a second RBF model, and the second RBF model is trained based on the first capacitor breakdown data set to obtain an optimal RBF model; the capacitor breakdown data of the capacitor voltage transformer to be tested obtained in real time is input into the optimal RBF model, so that the optimal RBF model outputs the capacitor breakdown online diagnosis result of the capacitor voltage transformer to be tested. Compared with the prior art, the technical solution of the present invention performs online diagnosis of the capacitor breakdown fault of the capacitor voltage transformer based on the optimal RBF model, without human intervention, reducing the error caused by human factors, realizing real-time and accurate monitoring of the capacitor breakdown fault of the capacitor voltage transformer, and improving the accuracy of the capacitor breakdown diagnosis result. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Figure 1 It is a flow chart of an embodiment of an online diagnosis method for capacitor breakdown of a capacitive voltage transformer provided by the present invention;

[0081] Figure 2 It is a structural schematic diagram of an embodiment of an online diagnostic device for capacitor breakdown of a capacitive voltage transformer provided by the present invention;

[0082] Figure 3 is a schematic diagram of an equivalent circuit model of a capacitive voltage transformer according to an embodiment of the present invention;

[0083] Figure 4 is a vector diagram of three-phase voltages under an ideal state of an embodiment provided by the present invention;

[0084] Figure 5 It is a schematic diagram of the corresponding relationship between the single-phase voltage amplitude and the zero-sequence voltage direction of an embodiment provided by the present invention;

[0085] Figure 6 It is a vector diagram of three-phase voltage when the B phase amplitude is abnormal according to an embodiment of the present invention;

[0086] Figure 7 It is a schematic diagram of a circuit model of a capacitive voltage transformer according to an embodiment of the present invention. DETAILED DESCRIPTION

[0087] The following will be combined with the accompanying drawings in the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0088] Example 1

[0089] See also Figure 1 , Figure 1 FIG. 1 is a flow chart of an embodiment of an online diagnosis method for capacitor breakdown of a capacitive voltage transformer provided by the present invention. Figure 1 As shown, the method includes steps 101 to 104, which are specifically as follows:

[0090] Step 101: Acquire historical capacitance breakdown data of capacitor voltage transformers in various capacitance breakdown states, and construct a historical capacitance breakdown data set.

[0091] In one embodiment, the typical category of the capacitor breakdown state of the capacitor voltage transformer is defined as L={L normal ,L 1H ,L 2H ,L 1M ,L 2M}, where L normal Indicates that the capacitor is not broken down, L 1G Indicates a high voltage capacitor breakdown, L 2H Indicates that two high-voltage capacitors are broken down, L 1M Indicates a medium voltage capacitor breakdown, L 2M Indicates that two medium voltage capacitors are broken down. Other capacitor breakdown states are analyzed based on these typical categories. Therefore, in this embodiment, only the above five capacitor states are considered.

[0092] In one embodiment, historical capacitor breakdown data of a capacitive voltage transformer in each capacitor breakdown state is obtained, and a historical capacitor breakdown data set is constructed. Specifically, historical capacitor breakdown data at each moment corresponding to the capacitive voltage transformer in each capacitor breakdown state on the same bus is obtained, and a historical capacitor breakdown data set D is constructed, wherein each capacitor breakdown state includes a capacitor non-breakdown state, a high-voltage capacitor breakdown state, two high-voltage capacitor breakdown states, a medium-voltage capacitor breakdown state, and two medium-voltage capacitor breakdown states, and the historical capacitor breakdown data includes historical secondary voltage amplitudes and historical phases at each moment.

[0093] Step 102: extracting the inter-group amplitude ratio parameter and the zero-sequence voltage direction variation error characteristic parameter from the historical capacitor breakdown data set, and constructing a first capacitor breakdown data set based on the inter-group amplitude ratio parameter and the zero-sequence voltage direction variation error characteristic parameter.

[0094] In one embodiment, a parameter A based on the amplitude ratio between groups is constructed. zj The feature vector of is as follows:

[0095]

[0096] Among them, A zj represents the amplitude ratio parameter between groups, n represents the number of capacitive voltage transformers, V Ai ,V Bi ,V Ci Represents the three-phase secondary voltage amplitude of the i-th capacitor voltage transformer.

[0097] In one embodiment, the reason for using the inter-group amplitude ratio parameter to determine the capacitor breakdown is: the dielectric loss of the high-voltage capacitor and the medium-voltage capacitor is equivalent to a parallel resistance, and a T-type equivalent circuit is used to construct an equivalent circuit model of a capacitive voltage transformer taking into account the internal insulation parameters. The schematic diagram of the equivalent circuit model of the capacitive voltage transformer is as follows: Figure 3 As shown, Figure 3 In, X CH and X C< are the equivalent capacitive reactances of the high voltage divider capacitor and the medium voltage divider capacitor, R H and R M are the equivalent resistances of the dielectric losses of the two, X L For compensation reactor, X T1 , X >2 , R1, R2 are equivalent parameters of the intermediate transformer, Z m is the magnetizing reactance of the intermediate transformer, Z L is the equivalent secondary load converted to the primary side of the intermediate transformer. m and Z L s is regarded as infinite, that is, the output port of the capacitive voltage divider is regarded as open circuit. At this time, the secondary output voltage of the capacitive voltage transformer taking into account the internal insulation parameters is:

[0098]

[0099] in, They are the primary voltage and secondary output voltage of the capacitive voltage transformer respectively, and K is the transformation ratio of the intermediate transformer.

[0100] Converting the dielectric loss equivalent resistance into the dielectric loss factor, the expression of the output voltage of the capacitive voltage transformer can be obtained as follows:

[0101]

[0102] When the capacity and dielectric loss of the voltage divider capacitor of the capacitive voltage transformer change abnormally, the additional ratio difference caused by the change of insulation parameters in the capacitive voltage transformer is:

[0103]

[0104] Among them, C A =C H (C H +C M ) -1 , C Hi , C Mi tanσ Hi tanσ Mi These are all internal insulation parameters that have undergone abnormal changes. Therefore, capacitor breakdown will cause the ratio difference to change.

[0105] According to the capacitance voltage transformer measurement error ratio relationship When two selected groups of capacitor voltage transformers with the same voltage and the same machine model are located on the same bus or two parallel buses, the primary voltage amplitudes corresponding to the two capacitor voltage transformers can be considered to be equal:

[0106]

[0107] It can be approximately considered that the secondary voltage amplitude ratio of the capacitor voltage transformer is only related to the ratio difference of the capacitor voltage transformer, that is, the change of the secondary voltage amplitude ratio can be used to reflect the state of capacitor breakdown.

[0108] In one embodiment, the zero-sequence voltage direction variation error characteristic parameter B is constructed zj , where B zj =(ε1…ε n ).

[0109] Specifically, the time point t at which the capacitor breaks down is obtained i , take the direction of the vector synthesis of the zero-sequence voltage at m moments before the capacitor breaks down The direction of the vector synthesis of the zero-sequence voltage at m moments after the capacitor breaks down The direction change of the synthesized zero-sequence voltage vector sum is Collect the historical data of ΔDir corresponding to the occurrence of voltage breakdown changes. Each corresponding vector change amount and the standard zero-sequence voltage direction change amount are ε = ΔDir - ΔDir(U0). In this embodiment, the box plot method is also used to obtain the appropriate error range of ε. When ε is within this error range, it is considered that the change in the zero-sequence voltage direction is caused by capacitor breakdown. Q3 + 1.5IQR < ε < Q1 - 1.5IQR, where Q3 is the upper quartile, Q1 is the lower quartile, and IQR = Q3 - Q1.

[0110] In one embodiment, the reason for using the error characteristic parameter of the zero-sequence voltage direction change amount to judge capacitor breakdown is that in the ideal state, the zero-sequence voltage is the vector sum of one-third of the three-phase voltages The magnitudes of the three-phase voltages are equal, and they are spaced 120° apart from each other. The magnitude of the zero-sequence voltage is 0, and the direction is 0°, as Figure 4 shown. Figure 4 is the vector schematic diagram of the three-phase voltages in the ideal state provided by the present invention. In the ideal state, when the magnitude of the single-phase voltage changes, the corresponding zero-sequence direction will change and the corresponding will occur, as Figure 5 shown. Figure 5 is the schematic diagram of the correspondence between the magnitude of the single-phase voltage and the zero-sequence voltage direction, and it is considered that Figure 5 the change amount of the zero-sequence voltage direction in is the standard zero-sequence voltage direction change amount. When the magnitude of phase B increases, the zero-sequence direction will change by 120°, as Figure 6 shown. Figure 6 is the vector schematic diagram of the three-phase voltages when the magnitude of phase B is abnormal.

[0111] If at time t i when the three-phase state is normal, its zero-sequence voltage direction is Dir(U 0,i ), and when the magnitude of the single-phase voltage changes at time t i+1 , its zero-sequence voltage direction is Dir(U 0,i+1 ), then the corresponding change amount of the zero-sequence voltage direction is ΔDir(U 0,i ) = Dir(U 0,i+1 ) - Dir(U 0,i ).

[0112] Figure 7 is the schematic diagram of the circuit model of the capacitive voltage transformer, as Figure 7 shown. The primary voltage is U1, and the secondary voltage is U2. Then the corresponding relationship between the two is When the high-voltage capacitor breaks down, When the medium-voltage capacitor breaks down, That is, when the high voltage breaks down, the secondary voltage amplitude of the corresponding phase decreases, and when the low voltage breaks down, the secondary voltage amplitude of the corresponding phase increases. That is, the corresponding relationship between the single-phase voltage amplitude and the zero-sequence voltage direction is also applicable to the relationship between the amplitude change caused by voltage breakdown and the zero-sequence voltage direction change.

[0113] In one embodiment, the historical capacitor breakdown data set D is divided into a training set and a test set according to a preset ratio, wherein the preset ratio is 7:3. Specifically, the number of samples in the historical capacitor breakdown data set D is set to n D , then the number of samples in the training set is n training =0.7n D , the number of samples in the test set is n test =0.3n D , randomly extract samples without replacement from the historical capacitor breakdown data set D to the training sample set D training In total, n training The remaining samples in the historical capacitor breakdown dataset D constitute the test set D test .

[0114] In one embodiment, the inter-group amplitude ratio parameter of the historical capacitor breakdown data set is extracted to obtain a first inter-group amplitude ratio parameter of the training set and a second inter-group amplitude ratio parameter of the test set.

[0115] In one embodiment, the zero-sequence voltage direction change corresponding to each capacitor voltage transformer in the training set and the test set at the time when the voltage breakdown change occurs is obtained respectively, and the standard zero-sequence voltage direction change is obtained at the same time; the error between the zero-sequence voltage direction change corresponding to each capacitor voltage transformer and the standard zero-sequence voltage direction change is calculated to obtain the first zero-sequence voltage direction change error characteristic parameter of the training set and the second zero-sequence voltage direction change error characteristic parameter of the test set.

[0116] In one embodiment, a first training set is obtained based on the first inter-group amplitude ratio parameter and the first zero-sequence voltage direction change error characteristic parameter; a first test set is obtained based on the second inter-group amplitude ratio parameter and the second zero-sequence voltage direction change error characteristic parameter; and the first capacitor breakdown data set is constructed based on the first training set and the first test set.

[0117] In one embodiment, z-score normalization is performed on the first training set to obtain a first normalized training set after z-score normalization, wherein the z-score normalization formula is as follows:

[0118]

[0119] Among them, μ is the characteristic mean of this dimension, and σ is the characteristic standard deviation of this dimension.

[0120] updating the first training set according to the first normalized training set to obtain a second training set, wherein the second training set includes a first feature vector set and a first category set;

[0121] Based on the feature mean and the feature standard deviation, the first test set is z-score normalized to obtain a first normalized test set after z-score normalization; based on the first normalized test set, the first test set is updated to obtain a second test set, wherein the second test set includes a second feature vector set and a second category set.

[0122] As an example in this embodiment: for the training set where d i ={x i ,y i}, based on the above-set inter-group amplitude ratio parameter and zero-sequence voltage direction variation error characteristic parameter, the first inter-group amplitude ratio parameter and the first zero-sequence voltage direction variation error of the training set are extracted to obtain the first training set T = [A zj B zj ], perform z-score normalization on the first training set, record the corresponding μ, σ in the first training set, and apply it to the test set D test The first normalized test set after normalization is T′=[A zj ′ B zj ′], where the feature vector set of a single sample in the first normalized test set can be expressed as t i =[A′ zji B′ zji ], then the corresponding training set D training Become the second training set where d i ′={t i ,y i}, then the corresponding first feature vector set of the second training set The corresponding first category set is Sample usage of the test set and training set D training Similarly, μ and σ are normalized to obtain the second feature vector set X of all samples in the second test set. test and the corresponding second category set Y test .

[0123] Step 103: optimizing the first RBF model based on the Optics algorithm to obtain a second RBF model, and training the second RBF model based on the first capacitor breakdown data set to obtain an optimal RBF model.

[0124] In one embodiment, the first RBF neural network includes three layers: an input layer, a hidden layer, and an output layer. The hidden layer uses a radial basis function for nonlinear change, and the radial basis function is set to a Gaussian function. The formula is:

[0125]

[0126] In the formula, c i is the center point of the Gaussian function, σ i is the kernel width of the Gaussian function. The corresponding output layer uses the softmax function to perform nonlinear processing, and its output layer is expressed as

[0127] In one embodiment, the center of the Gaussian function is obtained based on the Optics algorithm, so that the first RBF model is optimized based on the center of the Gaussian function to obtain the second RBF model. Specifically, when the first RBF network is constructed, the center point c of the Gaussian function is i The selection of has a great influence on the accuracy of the model. In order to accurately find the center point c of the Gaussian function i , using the Optics method for clustering classification to select the center of the Gaussian function. This method gives two inputs E and minPts, where E represents a given neighborhood radius. For a point x∈X, the neighborhood of E is represented by a subset of the set X whose distance from the point x is not greater than E. That is, N E (x i )={x i ∈D|distance(x i ,i)≤E}, where the number of subsets is |N E (x i )|, minPts is expressed as the minimum number of points that need to be included in the given neighborhood radius E for point o.

[0128] The Optics algorithm includes two concepts: core distance and reachable distance. The core distance minR is: If point x is selected as a core point, the following conditions must be met: for x in a given neighborhood of E|N E (x i)|≥minPts. For x in a neighborhood with radius R, when the number of other points in the neighborhood is exactly minPts, R is the core distance minR, minR≤E. The reachable distance Dis is the maximum value Dis=max(DisE,minR) between the Euclidean distance between two points and the reachable distance of the core point, where DisE refers to the Euclidean distance between two points.

[0129] In this embodiment, the center of the Gaussian function is selected, and based on the historical capacitor breakdown data set D, the neighborhood radius E, and the minimum number of neighborhood points minPts under the E neighborhood, the sample points with reachable distance information are output and sorted.

[0130] In this embodiment, the specific steps of selecting the center of the Gaussian function are:

[0131] Step S1: Create an ordered queue and a result queue, wherein the ordered queue is used to store the core point and its directly reachable objects, and arrange them in ascending order of reachable distance, and the result queue is used to store the output order of sample points in the historical capacitor breakdown data set.

[0132] Step S2: Determine whether all sample points in the historical capacitor breakdown data set have been processed. If so, the Optics algorithm ends. If not, select an unprocessed sample point that is a core point from the historical capacitor breakdown data set, and find all its directly density-reachable sample points. If the sample point does not exist in the result queue, put it into the ordered queue and sort it by reachable distance.

[0133] Step S3: Determine whether the ordered queue is empty. If so, return to step S2. If not, take out the first sample point from the ordered queue, that is, the point with the smallest reachable distance for expansion, and save the sample point taken out to the result queue. After saving, check whether it exists in the result queue. If not, perform the following processing:

[0134] Step S31: determine whether the extension point is a core point, if not, return to step S3. If the point is a core point, find all the directly density-reachable points of the extension point.

[0135] Step S32, determine whether the directly density-reachable sample point already exists in the result queue, if yes, do not process it, otherwise proceed to step S33.

[0136] Step S33: If the directly density reachable point already exists in the ordered queue, and if the new reachable distance is smaller than the old reachable distance, the new reachable distance replaces the old reachable distance, and the ordered queue is reordered.

[0137] Step S4: Iterate steps S2 and S3 until the Optics algorithm ends and outputs the ordered sample points in the result queue.

[0138] Step S5: Obtain the ordered sample points in the output result queue, and draw an ordered list decision diagram based on the ordered sample points in the output result queue, wherein the horizontal axis in the ordered list decision diagram represents the processing order of the sample points in the result queue, and the vertical axis represents the reachable distance of the point.

[0139] According to the decision diagram, we can observe different numbers of clusters and the corresponding cluster center results under different numbers of clusters. Given an appropriate number of clusters, the corresponding cluster center is used as the center point c of the Gaussian function. i , optimize the first RBF model and obtain the second RBF model.

[0140] In one embodiment, data assignment is performed on multiple hyperparameters of the second RBF model and the Optics algorithm.

[0141] In one embodiment, the second training set is obtained, and the second training set is input into the second RBF model, so that the second RBF model is trained based on the second training set to obtain a third RBF model. The second test set is obtained, and the second test set is input into the third RBF model, so that the third RBF model outputs the prediction classification result of each sample in the second test set, and generates a prediction classification result set Y predict ; Set the predicted classification result set Y predict With the second category set Y test Compare and obtain the accuracy of the third RBF model; determine whether the accuracy of the third RBF model reaches the preset accuracy, if so, retain the third RBF model and its corresponding accuracy, if not, do not retain the third RBF model and its corresponding accuracy. Determine whether the number of the third RBF models currently obtained reaches the preset threshold number, if not, return to the step of "assigning data to the second RBF model and multiple hyperparameters of the Optics algorithm", continue to train the second RBF model by adjusting multiple hyperparameters, and obtain the newly trained third RBF model, if yes, compare the accuracy of the multiple third RBF models retained, obtain the third RBF model corresponding to the highest accuracy, and set the third RBF model with the highest accuracy as the optimal RBF model.

[0142] Step 104: acquiring capacitance breakdown data of the capacitive voltage transformer to be tested in real time, and inputting the capacitance breakdown data into an optimal RBF model, so that the optimal RBF model outputs an online diagnosis result of capacitance breakdown of the capacitive voltage transformer to be tested.

[0143] In one embodiment, based on an online monitoring system for a capacitive voltage transformer, the secondary voltage amplitude and phase of the capacitive voltage transformer to be tested at the current moment are collected in real time, the secondary voltage amplitude and phase data detected in real time are input into a trained optimal RBF model, and the online diagnosis result of the capacitor breakdown of the capacitive voltage transformer to be tested at the current moment is output, without the need for human intervention, thereby reducing errors caused by human factors.

[0144] Example 2

[0145] See also Figure 2 , Figure 2 1 is a schematic structural diagram of an embodiment of an online diagnostic device for capacitor breakdown of a capacitive voltage transformer provided by the present invention, such as Figure 2 As shown, the device includes a data set construction module 201, a first capacitor breakdown data set construction module 202, an RBF model optimization module 203 and a capacitor breakdown online diagnosis module 204, which are specifically as follows:

[0146] The data set construction module 201 is used to construct a historical capacitance breakdown data set based on historical capacitance breakdown data of capacitive voltage transformers in various capacitance breakdown states.

[0147] The first capacitor breakdown data set construction module 202 is used to extract the inter-group amplitude ratio parameter and the zero-sequence voltage direction change error characteristic parameter from the historical capacitor breakdown data set, and construct the first capacitor breakdown data set based on the inter-group amplitude ratio parameter and the zero-sequence voltage direction change error characteristic parameter.

[0148] The RBF model optimization module 203 is used to optimize the first RBF model based on the Optics algorithm and the first capacitor breakdown data set to obtain an optimal RBF model.

[0149] The capacitor breakdown online diagnosis module 204 is used to obtain the capacitor breakdown data of the capacitor voltage transformer to be tested in real time, and input the capacitor breakdown data into the optimal RBF model so that the optimal RBF model outputs the capacitor breakdown online diagnosis result of the capacitor voltage transformer to be tested.

[0150] In one embodiment, the data set constructs historical capacitor breakdown data of a capacitive voltage transformer in each capacitor breakdown state in the module, wherein each capacitor breakdown state includes a capacitor non-breakdown state, a high-voltage capacitor breakdown state, two high-voltage capacitor breakdown states, a medium-voltage capacitor breakdown state, and two medium-voltage capacitor breakdown states, and the historical capacitor breakdown data includes historical secondary voltage amplitudes and historical phases at each moment.

[0151] In one embodiment, the first capacitance breakdown data set construction module is used to extract the inter-group amplitude ratio parameter from the historical capacitance breakdown data set, specifically including: dividing the historical capacitance breakdown data set into a training set and a test set according to a preset ratio; constructing a feature vector based on the inter-group amplitude ratio parameter, and extracting the inter-group amplitude ratio parameter of the historical capacitance breakdown data set, respectively obtaining the first inter-group amplitude ratio parameter of the training set and the second inter-group amplitude ratio parameter of the test set, wherein the constructed feature vector based on the inter-group amplitude ratio parameter is as follows:

[0152]

[0153] Among them, A zj represents the amplitude ratio parameter between groups, n represents the number of capacitive voltage transformers, V Ai ,V Bi ,V Ci Represents the three-phase secondary voltage amplitude of the i-th capacitor voltage transformer.

[0154] In one embodiment, the first capacitor breakdown data set construction module is used to extract the inter-group amplitude ratio parameter of the zero-sequence voltage direction change error characteristic parameter from the historical capacitor breakdown data set, specifically including: respectively obtaining the zero-sequence voltage direction change corresponding to each capacitor voltage transformer in the training set and the test set at the time when the voltage breakdown change occurs, and obtaining the standard zero-sequence voltage direction change; calculating the error between the zero-sequence voltage direction change corresponding to each capacitor voltage transformer and the standard zero-sequence voltage direction change, and obtaining the first zero-sequence voltage direction change error characteristic parameter of the training set and the second zero-sequence voltage direction change error characteristic parameter of the test set, wherein the zero-sequence voltage direction change error characteristic parameter is as follows:

[0155] B zj =(ε1…ε n );

[0156] Among them, B zj It represents the error characteristic parameter of the zero-sequence voltage direction change, n represents the number of capacitive voltage transformers, ε i is the zero-sequence voltage direction change error of the i-th capacitive voltage transformer.

[0157] In one embodiment, the first capacitor breakdown data set construction module is used to construct a first capacitor breakdown data set based on the inter-group amplitude ratio parameter and the zero-sequence voltage direction change error characteristic parameter, specifically including: obtaining and obtaining a first training set based on the first inter-group amplitude ratio parameter and the first zero-sequence voltage direction change error characteristic parameter; obtaining and obtaining a first test set based on the second inter-group amplitude ratio parameter and the second zero-sequence voltage direction change error characteristic parameter; constructing the first capacitor breakdown data set based on the first training set and the first test set.

[0158] In one embodiment, the present invention provides an online diagnostic device for capacitor breakdown of a capacitive voltage transformer, further comprising: a first capacitor breakdown data set processing module; specifically, the first capacitor breakdown data set processing module is used to perform z-score normalization processing on the training set to obtain a first normalized training set after z-score normalization processing, wherein the z-score normalization formula is as follows:

[0159]

[0160] Among them, μ is the characteristic mean of this dimension, and σ is the characteristic standard deviation of this dimension.

[0161] The first capacitor breakdown data set processing module is further used to update the first training set according to the first normalized training set to obtain a second training set, wherein the second training set includes a first feature vector set and a first category set;

[0162] The first capacitor breakdown data set processing module is also used to perform z-score normalization processing on the first test set based on the feature mean and the feature standard deviation to obtain a first normalized test set after z-score normalization processing; according to the first normalized test set, the first test set is updated to obtain a second test set, wherein the second test set includes a second feature vector set and a second category set.

[0163] In one embodiment, the RBF model optimization module is used to optimize the first RBF model based on the Optics algorithm to obtain the second RBF model, and to perform model training on the second RBF model based on the first capacitor breakdown data set to obtain the optimal RBF model, specifically including: performing nonlinear change processing on the hidden layer of the first RBF model based on the radial basis function, wherein the radial basis function is a Gaussian function; obtaining the center of the Gaussian function based on the Optics algorithm, so as to optimize the first RBF model based on the center of the Gaussian function to obtain the second RBF model; assigning data to multiple hyperparameters of the second RBF model and the Optics algorithm; obtaining the second training set, and inputting the second training set into the second RBF model to obtain a third RBF model; obtaining the second test set, and inputting the second test set into the The third RBF model is selected so that the third RBF model outputs the prediction classification result of the second test set to generate a prediction classification result set; the prediction classification result set is compared with the second category set to obtain the accuracy of the third RBF model, and it is determined whether the accuracy of the third RBF model reaches the preset accuracy. If so, the third RBF model and its corresponding accuracy are retained; if not, the third RBF model and its corresponding accuracy are not retained; it is determined whether the number of the third RBF models currently obtained reaches the preset threshold number. If not, the step of "assigning data to the second RBF model and multiple hyperparameters of the Optics algorithm" is returned. If so, the next step is performed; the accuracy of the retained multiple third RBF models is compared to obtain the third RBF model corresponding to the highest accuracy, and the third RBF model with the highest accuracy is set as the optimal RBF model.

[0164] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0165] It should be noted that the above-mentioned embodiment of the online diagnostic device for capacitor breakdown of a capacitive voltage transformer is merely illustrative, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment.

[0166] Based on the above-mentioned embodiment of the online diagnosis method for capacitor breakdown of a capacitive voltage transformer, another embodiment of the present invention provides an online diagnosis terminal device for capacitor breakdown of a capacitive voltage transformer. The online diagnosis terminal device for capacitor breakdown of a capacitive voltage transformer comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the online diagnosis method for capacitor breakdown of a capacitive voltage transformer of any embodiment of the present invention is implemented.

[0167] Exemplarily, in this embodiment, the computer program can be divided into one or more modules, and the one or more modules are stored in the memory and executed by the processor to complete the present invention. The one or more modules can be a series of computer program instruction segments that can complete specific functions, and the instruction segments are used to describe the execution process of the computer program in the online diagnosis terminal device of the capacitor voltage transformer capacitance breakdown.

[0168] The online diagnosis terminal device for capacitor voltage transformer capacitance breakdown can be a computing device such as a desktop computer, a notebook, a palm computer, a cloud server, etc. The online diagnosis terminal device for capacitor voltage transformer capacitance breakdown can include, but is not limited to, a processor and a memory.

[0169] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the online diagnosis terminal device for capacitor breakdown of the capacitive voltage transformer, and uses various interfaces and lines to connect various parts of the online diagnosis terminal device for capacitor breakdown of the capacitive voltage transformer.

[0170] The memory can be used to store the computer program and / or module, and the processor realizes various functions of the online diagnosis terminal device of the capacitor voltage transformer capacitor breakdown by running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function, etc.; the data storage area can store data created according to the use of the mobile phone, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0171] Based on the above-mentioned embodiment of the online diagnosis method for capacitor breakdown of a capacitive voltage transformer, another embodiment of the present invention provides a storage medium, wherein the storage medium includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the online diagnosis method for capacitor breakdown of a capacitive voltage transformer of any embodiment of the present invention.

[0172] In this embodiment, the storage medium is a computer-readable storage medium, and the computer program includes computer program code, which may be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium may be appropriately increased or decreased 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 electric carrier signals and telecommunication signals.

[0173] In summary, the present invention provides an online diagnosis method and device for capacitor breakdown of a capacitor voltage transformer, by acquiring historical capacitor breakdown data of capacitor voltage transformers in various capacitor breakdown states, constructing a historical capacitor breakdown data set; extracting the inter-group amplitude ratio parameter and the zero-sequence voltage direction variation error characteristic parameter from the historical capacitor breakdown data set, and constructing a first capacitor breakdown data set based on the inter-group amplitude ratio parameter and the zero-sequence voltage direction variation error characteristic parameter; optimizing the first RBF model based on the Optics algorithm to obtain a second RBF model, and training the second RBF model based on the first capacitor breakdown data set to obtain an optimal RBF model; acquiring the capacitor breakdown data of the capacitor voltage transformer to be tested in real time, and inputting the capacitor breakdown data into the optimal RBF model, so that the optimal RBF model outputs the capacitor breakdown online diagnosis result of the capacitor voltage transformer to be tested. Compared with the prior art, the technical solution provided by the present invention can realize real-time and accurate monitoring of the capacitor breakdown fault of the capacitor voltage transformer, and improve the accuracy of the capacitor breakdown diagnosis result.

[0174] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and substitutions can be made without departing from the technical principles of the present invention. These improvements and substitutions should also be regarded as the scope of protection of the present invention.

Claims

1. An online diagnosis method for capacitor breakdown of a capacitive voltage transformer, characterized in that: include: Obtain historical capacitance breakdown data of capacitor voltage transformers in various capacitance breakdown states, and construct a historical capacitance breakdown data set; Extracting an inter-group amplitude ratio parameter and a zero-sequence voltage direction change error characteristic parameter from the historical capacitor breakdown data set, and constructing a first capacitor breakdown data set based on the inter-group amplitude ratio parameter and the zero-sequence voltage direction change error characteristic parameter, wherein the zero-sequence voltage direction change error characteristic parameter is an error between the zero-sequence voltage direction change corresponding to each capacitor voltage transformer at the time when the voltage breakdown change occurs and the standard zero-sequence voltage direction change; Optimizing the first RBF model based on the Optics algorithm to obtain a second RBF model, and training the second RBF model based on the first capacitor breakdown data set to obtain an optimal RBF model; Acquire capacitance breakdown data of the capacitive voltage transformer to be tested in real time, and input the capacitance breakdown data into an optimal RBF model, so that the optimal RBF model outputs an online diagnosis result of capacitance breakdown of the capacitive voltage transformer to be tested; The extracting of the inter-group amplitude ratio parameter from the historical capacitance breakdown data set specifically includes: Dividing the historical capacitor breakdown data set into a training set and a test set according to a preset ratio; A feature vector based on the inter-group amplitude ratio parameter is constructed, and the inter-group amplitude ratio parameter is extracted from the historical capacitor breakdown data set to obtain a first inter-group amplitude ratio parameter of the training set and a second inter-group amplitude ratio parameter of the test set, respectively. The feature vector based on the inter-group amplitude ratio parameter is constructed as follows: Among them, A zj represents the amplitude ratio parameter between groups, n represents the number of capacitive voltage transformers, V Ai ,V Bi ,V Ci Represents the three-phase secondary voltage amplitude of the i-th capacitor voltage transformer.

2. The online diagnosis method for capacitor voltage transformer capacitance breakdown according to claim 1, characterized in that: Historical capacitor breakdown data of a capacitive voltage transformer in each capacitor breakdown state, wherein each capacitor breakdown state includes a capacitor non-breakdown state, a high-voltage capacitor breakdown state, two high-voltage capacitor breakdown states, a medium-voltage capacitor breakdown state, and two medium-voltage capacitor breakdown states, and the historical capacitor breakdown data includes historical secondary voltage amplitudes and historical phases at each moment.

3. The online diagnosis method for capacitor voltage transformer capacitance breakdown according to claim 1, characterized in that: Extracting the zero-sequence voltage direction variation error characteristic parameter from the historical capacitor breakdown data set specifically includes: Respectively obtain the zero-sequence voltage direction change corresponding to the voltage breakdown change moment of each capacitive voltage transformer in the training set and the test set, and simultaneously obtain the standard zero-sequence voltage direction change; The error between the zero-sequence voltage direction variation corresponding to each capacitive voltage transformer and the standard zero-sequence voltage direction variation is calculated to obtain a first zero-sequence voltage direction variation error characteristic parameter of the training set and a second zero-sequence voltage direction variation error characteristic parameter of the test set, wherein the zero-sequence voltage direction variation error characteristic parameter is as follows: B zj =(ε1…ε n ); Among them, B zj represents the error characteristic parameter of the zero-sequence voltage direction change, n represents the number of capacitive voltage transformers, ε i is the zero-sequence voltage direction change error of the i-th capacitive voltage transformer.

4. The online diagnosis method for capacitor voltage transformer capacitance breakdown according to claim 3, characterized in that: Based on the inter-group amplitude ratio parameter and the zero-sequence voltage direction variation error characteristic parameter, a first capacitor breakdown data set is constructed, specifically including: Acquire and obtain a first training set based on the first inter-group amplitude ratio parameter and the first zero-sequence voltage direction change amount error characteristic parameter; Acquire and obtain a first test set based on the second inter-group amplitude ratio parameter and the second zero-sequence voltage direction change amount error characteristic parameter; The first capacitance breakdown data set is constructed based on the first training set and the first test set.

5. The online diagnosis method for capacitor voltage transformer capacitance breakdown according to claim 4, characterized in that: Constructing a first capacitance breakdown data set also includes: Perform z-score normalization processing on the first training set to obtain a first normalized training set after the z-score normalization processing, wherein the z-score normalization formula is as follows: Among them, μ is the characteristic mean, σ is the characteristic standard deviation; updating the first training set according to the first normalized training set to obtain a second training set, wherein the second training set includes a first feature vector set and a first category set; Based on the feature mean and the feature standard deviation, performing z-score normalization processing on the first test set to obtain a first normalized test set after the z-score normalization processing; According to the first normalized test set, the first test set is updated to obtain a second test set, wherein the second test set includes a second feature vector set and a second category set.

6. The online diagnosis method for capacitor voltage transformer capacitance breakdown according to claim 5, characterized in that: Optimizing the first RBF model based on the Optics algorithm to obtain a second RBF model, and training the second RBF model based on the first capacitor breakdown data set to obtain an optimal RBF model, specifically including: Performing nonlinear change processing on the hidden layer of the first RBF model based on a radial basis function, wherein the radial basis function is a Gaussian function; Acquire the center of the Gaussian function based on the Optics algorithm, so as to optimize the first RBF model based on the center of the Gaussian function to obtain a second RBF model; Assigning data to multiple hyperparameters of the second RBF model and the Optics algorithm; Acquire the second training set, and input the second training set into the second RBF model to obtain a third RBF model; Acquire the second test set, and input the second test set into the third RBF model, so that the third RBF model outputs a prediction classification result of the second test set, and generates a prediction classification result set; Comparing the prediction classification result set with the second category set to obtain the accuracy of the third RBF model, and determining whether the accuracy of the third RBF model reaches a preset accuracy, if so, retaining the third RBF model and its corresponding accuracy, and if not, not retaining the third RBF model and its corresponding accuracy; Determine whether the number of the third RBF models currently obtained reaches a preset threshold number. If not, return to the step of "assigning data to multiple hyperparameters of the second RBF model and the Optics algorithm". If yes, proceed to the next step; The multiple third RBF models retained are compared in terms of accuracy to obtain the third RBF model corresponding to the highest accuracy, and the third RBF model with the highest accuracy is set as the optimal RBF model.

7. An online diagnostic device for capacitor breakdown of a capacitive voltage transformer, characterized in that: include: A data set construction module, a first capacitor breakdown data set construction module, an RBF model optimization module and a capacitor breakdown online diagnosis module; The data set construction module is used to construct a historical capacitance breakdown data set based on the historical capacitance breakdown data of the capacitive voltage transformer in each capacitance breakdown state; The first capacitor breakdown data set construction module is used to extract the inter-group amplitude ratio parameter and the zero-sequence voltage direction change error characteristic parameter from the historical capacitor breakdown data set, and construct a first capacitor breakdown data set based on the inter-group amplitude ratio parameter and the zero-sequence voltage direction change error characteristic parameter, wherein the zero-sequence voltage direction change error characteristic parameter is the error between the zero-sequence voltage direction change corresponding to each capacitor voltage transformer at the time when the voltage breakdown change occurs and the standard zero-sequence voltage direction change; The RBF model optimization module is used to optimize the first RBF model based on the Optics algorithm to obtain a second RBF model, and perform model training on the second RBF model based on the first capacitor breakdown data set to obtain an optimal RBF model; The capacitor breakdown online diagnosis module is used to obtain the capacitor breakdown data of the capacitor voltage transformer to be tested in real time, and input the capacitor breakdown data into the optimal RBF model so that the optimal RBF model outputs the capacitor breakdown online diagnosis result of the capacitor voltage transformer to be tested; The first capacitance breakdown data set construction module is used to extract the inter-group amplitude ratio parameter from the historical capacitance breakdown data set, specifically including: the first capacitance breakdown data set construction module divides the historical capacitance breakdown data set into a training set and a test set according to a preset ratio; constructs a feature vector based on the inter-group amplitude ratio parameter, and extracts the inter-group amplitude ratio parameter from the historical capacitance breakdown data set to obtain a first inter-group amplitude ratio parameter of the training set and a second inter-group amplitude ratio parameter of the test set, respectively, wherein the constructed feature vector based on the inter-group amplitude ratio parameter is as follows: Among them, A zj represents the amplitude ratio parameter between groups, n represents the number of capacitive voltage transformers, V Ai ,V Bi ,V Ci Represents the three-phase secondary voltage amplitude of the i-th capacitor voltage transformer.

8. A terminal device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the online diagnosis method for capacitance breakdown of a capacitive voltage transformer as claimed in any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the online diagnosis method for capacitance breakdown of a capacitive voltage transformer according to any one of claims 1 to 6.

Images