Capacitive voltage transformer internal insulation parameter quantification method and system

Abstract

The invention relates to the technical field of electric power, and provides a capacitor voltage transformer internal insulation parameter quantification method and system, and the method comprises the steps: collecting the historical secondary output data of a three-phase capacitor voltage transformer after operation, and building an internal insulation state feature set composed of a ratio difference and an angular difference; constructing a typical internal insulation state classification standard of the capacitor voltage transformer; constructing a classification model of the internal insulation state; based on the preliminarily obtained internal insulation state classification, further establishing an evaluation model for quantitative analysis of the internal insulation parameters of the capacitor voltage transformer under the state classification; and outputting the type classification of the internal insulation state and the size of the internal insulation parameter by using the trained model. According to the method, the internal insulation parameters of the capacitor voltage transformer are accurately quantified, abnormity can be found in time in the early stage of insulation degradation, and differential maintenance can be made according to parameter changes.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of electric power, in particular to a capacitor voltage transformer internal insulation parameter quantification method and system. BACKGROUND

[0002] Under the background of continuous improvement of voltage level of power transmission and transformation network, a capacitor voltage transformer (CVT) is a widely used measurement and protection device in a high-voltage power grid. As a core conversion unit of power grid parameter collection, the measurement accuracy of the CVT is directly related to the measurement fairness of power market transaction and the safety and stability of power grid operation. However, in the actual operation process, due to the complex insulation structure of the CVT, the insulation performance of the CVT is easily affected by electrical aging and thermal aging, which causes the insulation strength of the CVT to decrease, reduces the reliability of the CVT measurement and protection, and seriously threatens the safe and stable operation of the power system. Therefore, it is particularly important to timely and accurately evaluate the internal insulation state of the capacitor voltage transformer.

[0003] The common insulation faults of CVT are breakdown of capacitors and change of dielectric loss factor. Therefore, measuring the capacitance and dielectric loss factor of the capacitor divider can timely find the insulation defects of CVT. The traditional verification method usually builds a bridge with standard capacitors under the condition of power-off, compares the current and then calculates the capacitance and dielectric loss factor of the CVT under test. However, this method needs to be carried out regularly, which is difficult to evaluate the internal insulation state of CVT in time and is easy to cause the problem of insufficient or excessive maintenance. Therefore, it is necessary to realize the online evaluation of the internal insulation state of CVT under the condition of electrification. The main methods for online evaluation of the internal insulation performance of CVT are infrared detection method, signal analysis method and data-driven method. The infrared detection method reflects the internal insulation state of CVT by using the change of temperature field caused by heat generated by insulation deterioration. However, due to the loss of heat conduction and the low sensitivity of abnormal diagnosis, the infrared detection method is not easy to find early insulation deterioration. Moreover, the infrared detection method needs to install additional infrared equipment, which makes the detection process more complicated. The signal analysis method collects the measured voltage signal and current signal and other state quantities, and realizes online detection of the capacitance and dielectric loss factor of the capacitor divider based on the correlation function method, sine wave parameter method and other signal processing methods. However, the signal analysis method needs to install high-precision signal acquisition devices and sensing equipment, which affects the overall sealing of CVT and is not conducive to the long-term safe operation of CVT. Moreover, the accuracy of signal acquisition is easily affected by complex electromagnetic environment, which affects the results of signal processing. The data-driven method extracts features according to the secondary output information of CVT, and realizes real-time evaluation of the dielectric loss factor of CVT by real-time detection of the change of feature quantity. For example, the Chinese patent application document with the publication number CN114358092A, "A Capacitor Voltage Transformer Internal Insulation Performance Online Diagnosis Method and System", in the existing scheme, the related parameters of the internal insulation fault CVT are extracted as common features, and the fault diagnosis model is obtained by estimating the common feature distribution model using Dirichlet process Gaussian mixture model; this method can realize the diagnosis of internal insulation fault. However, this method is essentially a binary classification model, which can only distinguish between normal and abnormal insulation states, and cannot quantitatively evaluate the change of insulation parameters of CVT. The Chinese patent application document with the publication number CN112710930A, "A Capacitor Voltage Transformer Internal Insulation State Online Evaluation Method", compares the output data of all capacitor voltage transformers of the same voltage level in the substation to obtain a typical state parameter group, and then performs fuzzy clustering analysis to divide the several typical states of CVT. This method can identify the number of capacitor breakdowns of capacitor voltage transformers, but it ignores the capacitance change and can only evaluate the number of capacitor breakdowns of capacitor voltage transformers and cannot identify early insulation deterioration of capacitor voltage transformers. On the other hand, the evaluation of dielectric loss only includes two typical states of 0.2% and 0.5%, and cannot be further quantitatively analyzed.However, the CVT is affected by complex environments such as temperature, humidity, and contamination during actual operation, and its insulation parameters will gradually change, and its internal insulation deterioration is a continuous process, so further fine evaluation of its internal insulation parameters is needed. SUMMARY

[0004] The technical problem to be solved by the present application is how to accurately quantify the internal insulation parameters of a capacitive voltage transformer.

[0005] The present application solves the above technical problems by the following technical means: The present application provides a capacitive voltage transformer internal insulation parameter quantification method, comprising the following steps: S1, collecting historical secondary output data of a three-phase capacitive voltage transformer after being put into operation, and establishing an internal insulation state feature set composed of ratio difference and angle difference S2, taking the ratio difference and angle difference of the capacitive voltage transformer as characteristic parameters representing the internal insulation state of the capacitive voltage transformer, and constructing a typical internal insulation state classification standard for the capacitive voltage transformer; S3, taking the historical data meeting the typical internal insulation state classification standard of the capacitive voltage transformer as a training feature set, and constructing a classification model of the internal insulation state; S4, based on the internal insulation state classification obtained in S3, further establishing an evaluation model for quantitatively analyzing the internal insulation parameters of the capacitive voltage transformer under the state classification; S5, collecting real-time data of a three-phase capacitive voltage transformer to be evaluated online, and using the trained model to output the internal insulation state type classification and the size of the internal insulation parameters.

[0006] Further, the step S1 comprises: S11, collecting voltage data of the three-phase capacitive voltage transformer and performing standardization processing to obtain a standardized voltage matrix , including a voltage amplitude matrix and a voltage phase matrix ; S12, separating the capacitive voltage transformer measurement error from the primary voltage fluctuation by using a correlation analysis method to obtain the amplitude error and the phase error in the measurement error; S13, taking the set of amplitude error and phase error in the measurement error as a characteristic parameter F for internal insulation state evaluation, as follows:

[0007] wherein, ​、 、 is the ratio difference of three-phase voltage, 、 、 is the angle difference of three-phase voltage.

[0008] Further, the amplitude error separation process in the measurement error is specifically as follows:

[0009] wherein, is the voltage amplitude projection space of time-invariant symmetric constraint relationship mapping, is the score matrix on the space, is the load matrix on the space; is the projection space of voltage amplitude comprehensive error information, is the score matrix on the space, is the load matrix on the space; the specific form of the residual space representing the amplitude information of the measurement error of the capacitive voltage transformer is:

[0010] wherein, , , are the measurement errors of the A, B and C phase voltage amplitudes respectively.

[0011] Further, the phase error separation process in the measurement error is specifically as follows:

[0012] wherein, is the projection space of voltage phase information; is the projection space of error information; the specific form of the residual space representing the phase information of the measurement error of the capacitive voltage transformer is:

[0013] wherein, 、 、 are the measurement errors of the A, B and C phase voltage phases respectively.

[0014] Further, the S2 comprises the following steps: S21, establishing an equivalent circuit model of the capacitive voltage transformer considering the internal insulation performance degradation, mapping the changes of the capacitive voltage transformer capacitance and dielectric loss factor into the changes of the ratio difference and the angle difference.

[0015] S22, based on the mapping relationship between the ratio difference and the angle difference and the internal insulation parameter, according to the measurement accuracy level requirement of the capacitive voltage transformer, the capacitance and the dielectric loss factor of the capacitive voltage transformer under 0.2 level and 0.5 level error are selected as the typical internal insulation state, and the capacitive voltage transformer typical internal insulation state classification standard is constructed as follows:

[0016] The typical internal insulation state in the table is taken as the basis for constructing the training set.

[0017] Further, the change of the capacitance and the dielectric loss factor of the capacitive voltage transformer is mapped to the change of the ratio difference, specifically:

[0018] Wherein, , are the actual capacitance values of the high-voltage and low-voltage capacitor arms respectively; , are the actual dielectric loss values of the high-voltage and low-voltage capacitor arms respectively; , are the initial capacitance values of the high-voltage and low-voltage capacitor arms respectively; , are the initial dielectric loss values of the high-voltage and low-voltage capacitor arms respectively.

[0019] Further, the change of the capacitance and the dielectric loss factor of the capacitive voltage transformer is mapped to the change of the angle difference, specifically:

[0020] Wherein, , are the actual capacitance values of the high-voltage and low-voltage capacitor arms respectively; , are the actual dielectric loss values of the high-voltage and low-voltage capacitor arms respectively; , are the initial capacitance values of the high-voltage and low-voltage capacitor arms respectively; , are the initial dielectric loss values of the high-voltage and low-voltage capacitor arms respectively.

[0021] Further, the step S3 comprises the following steps: S31, based on the typical internal insulation state classification standard, taking the historical data conforming to the typical internal insulation state as the training feature set ; S32, taking the training feature set as input data, using convolutional neural network to perform convolution operation on the input data, extracting the features of the input data, represented as feature map, and the specific calculation process is:

[0022] In the formula, It is a convolutional layer The Each feature map It is a convolutional layer The Each convolutional region, It is a convolutional layer The Each weight matrix It is a convolutional layer l The weight matrix of the r-th convolutional region and the i-th weight matrix. These are bias parameters; S33. Select features from the input signal to reduce the number of features, thereby reducing the number of parameters and generating a lower-dimensional feature map. For a region, select the maximum activity value of all neurons in that region as the representation of that region, i.e.:

[0023] In the formula, It is the output of the max pooling layer. The first in the region One value, The range of values ​​is defined as follows , W This is the width parameter of the pooling region; S34. A fully connected layer is used to integrate and classify the previously extracted features, and the Softmax function is used to convert the output of the last neuron in the fully connected layer into probabilities, establishing the model as follows:

[0024] In the formula, Indicates that the data belongs to the first The probability values ​​of each category ( ), This represents the total number of categories defined in the model. It is the first neuron in the output of the last neuron of the fully connected layer. A specific numerical value.

[0025] Further, step S4 includes the following steps: S41. Based on the preliminary identification of the internal insulation state type of the capacitive voltage transformer using the classification model, the membership degree of the data to be evaluated to the typical internal insulation state A under this internal insulation state type is calculated using the distance metric method. Membership degree with typical internal insulation state B As shown in the following formula;

[0026] In the formula, It is the distance from the data to be evaluated to typical internal insulation state A. It is the distance from the data to be evaluated to typical internal insulation state B; S42. Correspond the changes in internal insulation parameters to membership degrees, and use the expected value of membership degrees. The approximate size of is inferred, and its calculation method is as follows:

[0027] In the formula, E It is a mathematical expectation function. These are the internal insulation parameters under typical internal insulation condition A. Internal insulation parameters and The positive correlation function between them; S43. Calculate the expected membership values ​​under multiple sets of capacitance and dielectric loss variations, plot the quantitative evaluation curve of internal insulation parameters, and establish an evaluation model for the quantitative analysis of internal insulation parameters of capacitive voltage transformers.

[0028] This invention also provides a system for quantifying the insulation parameters of a capacitive voltage transformer. The system operates using the method described above and includes the following modules: The data acquisition module is used to collect historical secondary output data of the three-phase capacitive voltage transformer after commissioning, and to establish a data structure based on the ratio difference. Sum of angle difference The set of internal insulation state characteristics; Internal insulation condition classification standard module, used to classify the differential voltage of capacitive voltage transformers. Sum of angle difference As a characteristic parameter characterizing the internal insulation state of a capacitive voltage transformer, a classification standard for typical internal insulation states of capacitive voltage transformers is constructed. The internal insulation state classification model is used to construct a classification model of internal insulation state by using historical data that conforms to the typical internal insulation state classification standard of capacitive voltage transformers as the training feature set. An internal insulation parameter quantification model is used to classify the internal insulation state obtained from the internal insulation state classification model, and further establishes an evaluation model for the quantitative analysis of the internal insulation parameters of the capacitive voltage transformer under this state classification. The output module is used to collect real-time data of the three-phase capacitive voltage transformer to be evaluated online, and output the internal insulation state type classification and the magnitude of the internal insulation parameters using the trained model.

[0029] The advantages of this invention are: (1) Firstly, the inner insulation type features are extracted by using the convolutional neural network, the inner insulation state multi-classification model is obtained through training, the inner insulation state type of the capacitive voltage transformer is preliminarily determined, and the online diagnosis of the inner insulation state is realized. On this basis, the membership function is constructed by using the distance measurement method, and the quantitative evaluation curve of the capacitance and the dielectric loss factor parameter change is drawn, so that the data-knowledge double-layer driven capacitive voltage transformer inner insulation evaluation model is established, and the fine evaluation of the capacitive voltage transformer inner insulation parameter quantitative analysis is realized. Not only can the inner insulation fault type be accurately identified, but also the change of the capacitance and the dielectric loss factor parameter can be further quantitatively evaluated. It can help the staff to understand the inner insulation state of the capacitive voltage transformer, so as to take appropriate measures for differential operation and maintenance work.

[0030] (2) The capacitive voltage transformer measurement error is taken as a characteristic parameter representing the inner insulation state of the capacitive voltage transformer, and the output data of the capacitive voltage transformer to be evaluated are only analyzed to realize the online evaluation of the inner insulation state of the capacitive voltage transformer under the condition of no standard device, so that the inner insulation fault of the capacitive voltage transformer can be found and warned in time, and the maintenance personnel are reminded to repair the transformer before the insulation state is further deteriorated.

[0031] (3) The information-physical correlation analysis method is adopted to remove the influence of time-varying components and noise interference, so that the ratio difference and angle difference characteristic parameters which can fully represent the insulation characteristics of the capacitive voltage transformer are obtained, so that the non-stationary operation process of the power system can be adapted, and the change of the inner insulation parameter of the capacitive voltage transformer can be more accurately evaluated. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 It is a flowchart of a capacitive voltage transformer inner insulation parameter quantization method of an embodiment of the application; Figure 2 It is a schematic diagram of the quantitative evaluation curve of the capacitance change under the high-voltage capacitance increase state of the embodiment of the application; Figure 3 It is a quantitative evaluation curve schematic diagram of the dielectric loss factor under the high-voltage dielectric loss state of the embodiment of the application. DETAILED DESCRIPTION

[0033] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the embodiments of the present application. Obviously, the described embodiments are only some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts should fall into the scope of the present application.

[0034] Embodiment 1 The embodiment provides an internal insulation parameter quantification method of a capacitive voltage transformer, and a flowchart is as shown in Figure 1 The method comprises the following steps: S1, collecting historical secondary output data of a three-phase capacitive voltage transformer after being put into operation, and establishing an internal insulation state feature set composed of ratio difference and angle difference A specific implementation includes: S11, collecting voltage data of the three-phase capacitive voltage transformer, and performing standardization processing to obtain a standardized voltage matrix , including a voltage amplitude matrix and a voltage phase matrix ; S12, separating capacitive voltage transformer measurement error and primary voltage fluctuation by using a correlation analysis method to obtain amplitude error and phase error in the measurement error; the amplitude error separation process in the measurement error is specifically as follows:

[0035] Wherein, is a voltage amplitude projection space of a time-invariant symmetric constraint relationship mapping, is a score matrix on the space, is a load matrix on the space; is a projection space of voltage amplitude comprehensive error information, is a score matrix on the space, is a load matrix on the space; the specific form of a residual error space representing amplitude error information of the capacitive voltage transformer measurement error is:

[0036] Wherein, 、 、 The measurement error of the voltage amplitude of phase A, B and C respectively.

[0037] The phase error separation process in the measurement error is specifically as follows:

[0038] in, This is the projection space for voltage phase information; The projection space for error information; the specific form of the residual space characterizing the phase information of the measurement error of the capacitive voltage transformer is:

[0039] in, 、 、 These represent the measurement errors of the voltage phases of phases A, B, and C, respectively.

[0040] S13. The set of amplitude error and phase error in the measurement error is used as a characteristic parameter for internal insulation condition assessment. F As shown in the following formula:

[0041] in, 、 、 The ratio difference of the three-phase voltages. 、 、 This represents the angle difference between the three-phase voltages.

[0042] S2, the ratio difference of the capacitive voltage transformer Sum of angle difference As a characteristic parameter characterizing the internal insulation state of a capacitive voltage transformer, a classification standard for typical internal insulation states of capacitive voltage transformers is constructed; specific implementation methods include: S21. Establish an equivalent circuit model for a capacitive voltage transformer that considers the deterioration of internal insulation performance, and map the changes in capacitance and dielectric loss factor of the capacitive voltage transformer to the changes in ratio difference and phase difference.

[0043] In this embodiment, the parameters of the constructed equivalent circuit model of the capacitive voltage transformer are shown in the table below:

[0044] The changes in the capacitance and dielectric loss factor of the capacitive voltage transformer are mapped to changes in the ratio difference, specifically:

[0045] in, , These are the actual capacitance values ​​of the high-voltage and low-voltage capacitor arms, respectively. , actual dielectric loss value of the high-voltage capacitor arm and the low-voltage capacitor arm, respectively; initial capacitance value of the high-voltage capacitor arm and the low-voltage capacitor arm, respectively; initial dielectric loss value of the high-voltage capacitor arm and the low-voltage capacitor arm, respectively.

[0046] The change of the capacitance and the dielectric loss factor of the capacitor voltage transformer is mapped to the change of the phase difference, specifically:

[0047] actual capacitance value of the high-voltage capacitor arm and the low-voltage capacitor arm, respectively; actual dielectric loss value of the high-voltage capacitor arm and the low-voltage capacitor arm, respectively; initial capacitance value of the high-voltage capacitor arm and the low-voltage capacitor arm, respectively; initial dielectric loss value of the high-voltage capacitor arm and the low-voltage capacitor arm, respectively.

[0048] Based on the above mapping relationship, the corresponding internal insulation parameter change amount of the capacitor voltage transformer under the typical internal insulation state can be obtained based on the ratio difference and the phase difference.

[0049] S22, based on the mapping relationship between the ratio difference and the phase difference and the internal insulation parameter, according to the measurement accuracy level requirement of the capacitor voltage transformer, the capacitance and the dielectric loss factor of the capacitor voltage transformer under 0.2 level and 0.5 level error are selected as the typical internal insulation state, and the classification standard of the typical internal insulation state of the capacitor voltage transformer is constructed as follows:

[0050] The typical internal insulation state in the table is taken as the basis for constructing the training set. In this embodiment, nine typical insulation states of "normal", "high and low voltage capacitance increase 0.29%", "high and low voltage capacitance increase 0.71%", "high and low voltage dielectric loss 0.47%", and "high and low voltage dielectric loss 0.88%" are calculated, which respectively represent the normal, 0.2 level and 0.5 level measurement error of the internal insulation state.

[0051] S3, taking the historical data meeting the classification standard of the typical internal insulation state of the capacitor voltage transformer as the training feature set, a classification model of the internal insulation state is constructed; the specific implementation includes the following steps: S31, based on the typical internal insulation state classification standard, taking the historical data meeting the typical internal insulation state as the training feature set ； ​​​​​​​S32, taking the training feature set as input data, performing convolution operation on the input data by using a convolutional neural network, extracting features of the input data, and representing the features as a feature map, and the specific calculation process is as follows:

[0052] In the formula, is the i-th feature map of the convolutional layer , is the i-th feature map of the convolutional layer , is the i-th convolutional region of the convolutional layer , is the i-th weight matrix of the convolutional layer , is the i-th weight matrix of the r-th convolutional region of the convolutional layer , l is the i-th weight matrix of the r-th convolutional region of the convolutional layer is a bias parameter; S33, selecting the features of the input signal to reduce the number of features and thus reduce the number of parameters, generating a lower-dimensional feature map, for a region, selecting the maximum activity value of all neurons in the region as the representation of the region, that is:

[0053] In the formula, is the output of the max-pooling layer, represents the i-th value in the region, the value range of i is defined as , is a width parameter of the pooling region; W S34, using a fully connected layer to integrate and classify the previously extracted features, and using a Softmax function to convert the output of the last neuron of the fully connected layer into a probability, and establishing a model as follows: In the formula,

[0054] represents the probability value of data belonging to the i-th class, represents the total number of classes set by the model, is the i-th specific value in the output result of the last neuron of the fully connected layer. S4, based on the internal insulation state classification obtained in S3, further establishing an evaluation model for quantitative analysis of the internal insulation parameters of the capacitive voltage transformer under the state classification; the specific implementation includes the following steps:

[0055] S4, based on the internal insulation state classification obtained in S3, further establishing an evaluation model for quantitative analysis of the internal insulation parameters of the capacitive voltage transformer under the state classification; the specific implementation includes the following steps: ​​​S41, based on the preliminary clear classification model of the capacitive voltage transformer, the distance measurement method is used to calculate the membership degree of the to-be-evaluated data on the typical internal insulation state A and the typical internal insulation state B and the typical internal insulation state B , as follows:

[0056] In the formula, is the distance of the to-be-evaluated data to the typical internal insulation state A, is the distance of the to-be-evaluated data to the typical internal insulation state B; When the convolutional neural network classification model judges that the internal insulation state is the change of high-voltage capacitance or low-voltage capacitance or the change of high-voltage dielectric loss factor or low-voltage dielectric loss factor, the internal insulation state is limited in two states, which correspond to 0.2 level and 0.5 level measurement error respectively. In order to facilitate calculation, the above two internal insulation states are defined as A and B respectively, and the membership degrees of the typical internal insulation state A and the typical internal insulation state B are calculated and the typical internal insulation state B If is close to 1, it indicates that the internal insulation parameter of the capacitive voltage transformer is close to the state A set; if is close to 1, it indicates that the internal insulation parameter of the capacitive voltage transformer is close to the state B set; if , are both between 0 and 1, it indicates that the internal insulation parameter of the capacitive voltage transformer is between the state A set and the state B set.

[0057] S42, corresponding the change of the internal insulation parameter to the membership degree, and approximately inferring the size of the membership degree expectation value , and the calculation method is as follows:

[0058] In the formula, E is a mathematical expectation function, is the internal insulation parameter under the typical internal insulation state A, is the positive correlation function between the internal insulation parameter and . S43, calculate the membership degree expectation value under multiple groups of capacitance and dielectric loss change, draw the internal insulation parameter quantitative evaluation curve, and establish the evaluation model of the capacitive voltage transformer internal insulation parameter quantitative analysis.

[0059] In this embodiment, the internal insulation state classification is preliminarily determined as the increase of high-voltage capacitance. When the high-voltage capacitance is between 0.29% and 0.71%, the membership degree expectation value under multiple groups of different capacitances is calculated, such as Figure 2As shown, the high-voltage capacitance quantitative evaluation curve can be drawn. Similarly, the preliminary determination of the internal insulation state classification is the high-voltage dielectric loss, when the high-voltage dielectric loss is in the middle value between 0.47% and 0.88%, the membership expectation value under different dielectric losses is calculated, and the membership expectation value under different dielectric losses is calculated Figure 3 As shown, the high-voltage dielectric loss quantitative evaluation curve can be drawn.

[0060] Similarly, by drawing the quantitative evaluation curves of the low-voltage capacitance increase and the low-voltage dielectric loss, a complete and refined evaluation model of the capacitance voltage transformer is established, and the size of the internal insulation parameter is further quantitatively analyzed.

[0061] S5, collecting real-time data of a three-phase capacitance voltage transformer to be evaluated online, and outputting the internal insulation state type classification and the size of the internal insulation parameter by using the trained model.

[0062] Embodiment 2 It needs to be further explained that based on the same inventive concept, the embodiment provides a capacitance voltage transformer internal insulation parameter quantification system, which adopts the method of embodiment 1 when the system runs, and includes the following modules: A data acquisition module is configured to acquire historical secondary output data of a three-phase capacitance voltage transformer after being put into operation, and establish an internal insulation state feature set composed of ratio difference and angle difference An internal insulation state classification standard module is configured to take the ratio difference and angle difference of the capacitance voltage transformer as a characteristic parameter representing the internal insulation state of the capacitance voltage transformer, and construct a typical internal insulation state classification standard of the capacitance voltage transformer; An internal insulation state classification model is configured to take the historical data meeting the typical internal insulation state classification standard of the capacitance voltage transformer as a training feature set, and construct a classification model of the internal insulation state; An internal insulation parameter quantification model is configured to further establish an evaluation model of the capacitance voltage transformer internal insulation parameter under the state classification obtained by the internal insulation state classification model; An output module is configured to collect real-time data of a three-phase capacitance voltage transformer to be evaluated online, and output the internal insulation state type classification and the size of the internal insulation parameter by using the trained model.

[0063] ​The above examples are only used to illustrate the technical solutions of the present application, and are not intended to limit the present application; although the present application has been described in detail with reference to the foregoing examples, those skilled in the art should understand that the technical solutions recorded in the foregoing examples can be modified, or some technical features can be replaced by equivalent features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for quantifying the insulation parameters of a capacitive voltage transformer, characterized in that, Includes the following steps: S1. Collect historical secondary output data of the three-phase capacitive voltage transformer after commissioning, and establish a system based on the ratio difference. Sum of angle difference The set of internal insulation state characteristics; S2, the ratio difference of the capacitive voltage transformer Sum of angle difference As a characteristic parameter characterizing the internal insulation state of a capacitive voltage transformer, a classification standard for typical internal insulation states of capacitive voltage transformers is constructed. S3. Using historical data that conforms to the classification criteria for typical internal insulation states of capacitive voltage transformers as the training feature set, construct a classification model for internal insulation states. S4. Based on the internal insulation state classification obtained in S3, further establish an evaluation model for the quantitative analysis of the internal insulation parameters of the capacitive voltage transformer under this state classification. S5. Collect real-time data of the three-phase capacitive voltage transformer to be evaluated online, and use the trained model to output the internal insulation state type classification and the magnitude of the internal insulation parameters.

2. The method for quantifying the internal insulation parameters of a capacitive voltage transformer according to claim 1, characterized in that, Step S1 includes: S11. Collect voltage data from the three-phase capacitive voltage transformer and standardize it to obtain a standardized voltage matrix. Including voltage magnitude matrix With voltage phase matrix ; S12. The correlation analysis method is used to separate the measurement error of the capacitive voltage transformer from the primary voltage fluctuation, and the amplitude error and phase error in the measurement error are obtained. S13. The set of amplitude error and phase error in the measurement error is used as a characteristic parameter for internal insulation condition assessment. F As shown in the following formula: in, 、 、 The ratio difference of the three-phase voltages. 、 、 This represents the angle difference of the three-phase voltages.

3. The method for quantifying the internal insulation parameters of a capacitive voltage transformer according to claim 2, characterized in that, The process of separating the amplitude error in the measurement error is as follows: in, The voltage magnitude projection space is mapped by the time-invariant symmetric constraint relation. Let be the score matrix in this space. This is the load matrix in this space; This is the projection space for the comprehensive error information of voltage amplitude. Let be the score matrix in this space. The load matrix in this space; the specific form of the residual space characterizing the measurement error amplitude information of the capacitive voltage transformer is: in, 、 、 These represent the measurement errors of the voltage amplitudes of phases A, B, and C, respectively.

4. The method for quantifying the internal insulation parameters of a capacitive voltage transformer according to claim 2, characterized in that, The phase error separation process in the measurement error is as follows: in, This is the projection space for voltage phase information; The projection space for error information; the specific form of the residual space characterizing the phase information of the measurement error of the capacitive voltage transformer is: in, 、 、 These represent the measurement errors of the voltage phases of phases A, B, and C, respectively.

5. The method for quantifying the internal insulation parameters of a capacitive voltage transformer according to claim 1, characterized in that, S2 includes the following steps: S21. Establish an equivalent circuit model for a capacitive voltage transformer that considers the deterioration of internal insulation performance, and map the changes in capacitance and dielectric loss factor of the capacitive voltage transformer to the changes in ratio difference and angle difference. S22. Based on the mapping relationship between the ratio difference and angle difference and the internal insulation parameters, and according to the measurement accuracy requirements of the capacitive voltage transformer, the capacitance and dielectric loss factor of the capacitive voltage transformer under the error levels of 0.2 and 0.5 are selected as typical internal insulation states, and the classification standard of typical internal insulation states of capacitive voltage transformers is constructed as shown in the table below. The typical internal insulation states in the table are used as the basis for constructing the training set.

6. The method for quantifying the internal insulation parameters of a capacitive voltage transformer according to claim 5, characterized in that, The changes in the capacitance and dielectric loss factor of the capacitive voltage transformer are mapped to changes in the ratio difference, specifically: in, , These are the actual capacitance values ​​of the high-voltage and low-voltage capacitor arms, respectively. , These are the actual dielectric loss values ​​for the high-voltage and low-voltage capacitor arms, respectively. , These are the initial capacitance values ​​for the high-voltage and low-voltage capacitor arms, respectively. , These are the initial dielectric loss values ​​for the high-voltage and low-voltage capacitor arms, respectively.

7. The method for quantifying the internal insulation parameters of a capacitive voltage transformer according to claim 5, characterized in that, The changes in the capacitance and dielectric loss factor of the capacitive voltage transformer are mapped to changes in the phase angle difference, specifically: in, , These are the actual capacitance values ​​of the high-voltage and low-voltage capacitor arms, respectively. , These are the actual dielectric loss values ​​for the high-voltage and low-voltage capacitor arms, respectively. , These are the initial capacitance values ​​for the high-voltage and low-voltage capacitor arms, respectively. , These are the initial dielectric loss values ​​for the high-voltage and low-voltage capacitor arms, respectively.

8. The method for quantifying the internal insulation parameters of a capacitive voltage transformer according to claim 1, characterized in that, Step S3 includes the following steps: S31. Based on the typical internal insulation state classification standard, use historical data that conforms to the typical internal insulation state as the training feature set. ; S32. Using the training feature set as input data, a convolutional neural network is used to perform convolution operations on the input data to extract the features of the input data, which are represented as a feature map. The specific calculation process is as follows: In the formula, It is a convolutional layer The Each feature map It is a convolutional layer The Each convolutional region, It is a convolutional layer The Each weight matrix It is a convolutional layer l The weight matrix of the r-th convolutional region and the i-th weight matrix. These are bias parameters; S33. Select features from the input signal to reduce the number of features, thereby reducing the number of parameters and generating a lower-dimensional feature map. For a region, select the maximum activity value of all neurons in that region as the representation of that region, i.e.: In the formula, It is the output of the max pooling layer. The first in the region One value, The range of values ​​is defined as follows , W This is the width parameter of the pooling region; S34. A fully connected layer is used to integrate and classify the previously extracted features, and the Softmax function is used to convert the output of the last neuron in the fully connected layer into probabilities, establishing the model as follows: In the formula, Indicates that the data belongs to the first The probability values ​​of each category ( ), This represents the total number of categories defined in the model. It is the first neuron in the output of the last neuron of the fully connected layer. A specific numerical value.

9. The method for quantifying the internal insulation parameters of a capacitive voltage transformer according to claim 1, characterized in that, Step S4 includes the following steps: S41. Based on the preliminary identification of the internal insulation state type of the capacitive voltage transformer using the classification model, the membership degree of the data to be evaluated to the typical internal insulation state A under this internal insulation state type is calculated using the distance metric method. Membership degree with typical internal insulation state B As shown in the following formula; In the formula, It is the distance from the data to be evaluated to typical internal insulation state A. It is the distance from the data to be evaluated to typical internal insulation state B; S42. Correspond the changes in internal insulation parameters to membership degrees, and use the expected value of membership degrees. The approximate size of is inferred, and its calculation method is as follows: In the formula, E It is a mathematical expectation function. These are the internal insulation parameters under typical internal insulation condition A. Internal insulation parameters and The positive correlation function between them; S43. Calculate the expected membership values ​​under multiple sets of capacitance and dielectric loss variations, plot the quantitative evaluation curve of internal insulation parameters, and establish an evaluation model for the quantitative analysis of internal insulation parameters of capacitive voltage transformers.

10. A system for quantifying insulation parameters within a capacitive voltage transformer, characterized in that, Includes the following modules: The data acquisition module is used to collect historical secondary output data of the three-phase capacitive voltage transformer after commissioning, and to establish a data structure based on the ratio difference. Sum of angle difference The set of internal insulation state characteristics; Internal insulation condition classification standard module, used to classify the differential voltage of capacitive voltage transformers. Sum of angle difference As a characteristic parameter characterizing the internal insulation state of a capacitive voltage transformer, a classification standard for typical internal insulation states of capacitive voltage transformers is constructed. The internal insulation state classification model is used to construct a classification model of internal insulation state by using historical data that conforms to the typical internal insulation state classification standard of capacitive voltage transformers as the training feature set. An internal insulation parameter quantification model is used to classify the internal insulation state obtained from the internal insulation state classification model, and further establishes an evaluation model for the quantitative analysis of the internal insulation parameters of the capacitive voltage transformer under this state classification. The output module is used to collect real-time data of the three-phase capacitive voltage transformer to be evaluated online, and output the internal insulation state type classification and the magnitude of the internal insulation parameters using the trained model.

Images