A novel method for detecting measurement errors of wide-area capacitor voltage transformers in power systems

By constructing the topological structure and deep convolutional neural network model of the wide-area substation cluster of the power system, combining the European-style distance and symmetric point mode, the measurement error detection of the wide-area capacitive voltage transformer of the power system is realized, and the existing technology is solved The lack of domain error detection improves error recognition accuracy and model stability.

CN118688704BActive Publication Date: 2024-11-01CHINA ELECTRIC POWER RESEARCH INSTITUTE CO LTD
View PDF 10 Cites 0 Cited by

Patent Information

Application Number
CN202411170563.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-26
Publication Date
2024-11-01
Estimated Expiration
2044-08-26

AI Technical Summary

Technical Problem

The existing metering error detection methods are limited to capacitive voltage transformers in substations, and fail to form large-scale wide-area error detection, making it difficult to effectively detect metering errors between wide-area substations in power systems.

Method used

The topological structure of a new power system wide-area substation cluster was constructed, three-phase timing voltage data was collected, one-dimensional timing measurement error was calculated using the European distance method, and converted into a two-dimensional time-frequency map through symmetric point mode, combined with the deep convolutional neural network model , generate an online error identification model to realize the prediction and identification of wide-area measurement errors.

Benefits of technology

Through the ResNet50-L1M model, the online recognition accuracy of error states is improved, the stability and reliability of the model are enhanced, key features can be captured more accurately, noise interference can be reduced, and error state of interconnected CVT between wide-area substations of power system is solved. Identify the problem.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118688704B_ABST
    Figure CN118688704B_ABST
Patent Text Reader

Abstract

The present invention discloses a novel method for detecting metering errors of wide-area capacitor voltage transformers in a power system, comprising: constructing a topological structure of a wide-area substation cluster of a novel power system, and collecting A, B, and C three-phase time-series voltage data of capacitor voltage transformers at both ends of all transmission lines in the topological structure within a preset time period; calculating a one-dimensional time-series metering error of each transmission line according to the A, B, and C three-phase time-series voltage data; converting the one-dimensional time-series metering error of each transmission line into a two-dimensional time-frequency diagram by adopting a symmetrical point pattern, and generating a symmetrical point pattern spectrum of different errors; training a pre-constructed deep convolutional neural network model according to the symmetrical point pattern spectrum of different errors, and generating an error online recognition model; inputting the collected real-time A, B, and C three-phase time-series voltage data of the wide-area substation cluster of the novel power system into the error online recognition model, and outputting a metering error prediction result of the wide-area substation cluster of the novel power system.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of transformer metering error detection, and more specifically, to a novel method for detecting metering errors of wide-area capacitor voltage transformers in a power system. Background Art

[0002] The topological correlation of high-voltage electric energy metering devices in new power systems is stronger, and they are easily affected by the operating status of the opposite station with topological correlation. The existing metering error detection method is limited to the error detection of capacitor voltage transformer (capacitor voltage transformer) in the substation, and has not yet formed a large-scale wide-area error detection. Therefore, it is urgent to carry out the research on the domain-station-point capacitor voltage transformer metering error detection method based on the wide-area substation network architecture, get rid of the conditions in a single substation, greatly improve the data dimension and data volume, and realize the domain-station-point capacitor voltage transformer metering error detection. Summary of the invention

[0003] In view of the deficiencies in the prior art, the present invention provides a novel method for detecting metering errors of wide-area capacitor voltage transformers in power systems.

[0004] According to one aspect of the present invention, a novel method for detecting measurement errors of wide-area capacitor voltage transformers in a power system is provided, comprising:

[0005] Construct the topological structure of the new power system wide-area substation cluster, and collect the A, B, and C three-phase time-series voltage data of the capacitor voltage transformers at both ends of all transmission lines in the topological structure within a preset time period, where the A, B, and C three-phase time-series voltage data includes voltage amplitude time-series data and voltage phase time-series data;

[0006] The Euclidean distance method is used to calculate the one-dimensional time series measurement error of each transmission line based on the time series voltage data of the three phases A, B and C, where the one-dimensional time series measurement error includes: amplitude time series measurement error and phase time series measurement error;

[0007] The symmetrical point pattern is used to convert the one-dimensional time series measurement error of each transmission line into a two-dimensional time-frequency diagram, and the symmetrical point pattern spectrum of different errors is generated, wherein the symmetrical point pattern spectrum includes: amplitude distance characteristic spectrum and phase distance characteristic spectrum;

[0008] Training a pre-built deep convolutional neural network model according to symmetrical point pattern graphs of different errors to generate an error online recognition model, wherein the deep convolutional neural network model is a Resnet50-L1M network structure established by applying L1 attention to a Resnet50 network structure;

[0009] The collected real-time A, B, and C three-phase time-series voltage data of the new power system wide-area substation cluster are input into the error online identification model, and the metering error prediction results of the new power system wide-area substation cluster are output.

[0010] Optionally, a one-dimensional time series measurement error of each transmission line is calculated based on the time series voltage data of the three phases A, B and C using the Euclidean distance method, including:

[0011] The collected A, B, and C three-phase time-series voltage data of the capacitor voltage transformer at both ends of each transmission line are marked with coordinates in three-dimensional space according to the collection time sequence;

[0012] Perform distance calculation on each group of A, B, C three-phase time-series voltage data of capacitor voltage transformers at both ends of the transmission line in three-dimensional space to obtain distance points of capacitor voltage transformers at both ends of the transmission line, wherein the distance points include: amplitude distance points and phase distance points;

[0013] The one-dimensional time series measurement error is determined according to the distance point of each time node within the preset time period.

[0014] Optionally, the calculation formula of the distance point is expressed as:

[0015]

[0016] Where D f and D θ are the amplitude distance points and phase distance points of the capacitive voltage transformers at both ends of the transmission line in three-dimensional space; is the three-dimensional spatial coordinate point of the voltage amplitude of the three phases A, B, and C of any substation 1 in the new power system wide-area substation cluster, It is the three-dimensional spatial coordinate point of the A, B, and C three-phase voltage phase of any substation 2 in the new power system wide-area substation cluster.

[0017] Optionally, the one-dimensional time series measurement error in the symmetric point pattern spectrum x No. i Sampling points x i The polar coordinate space is S ;in,

[0018]

[0019] In the formula, and Represents the one-dimensional time series measurement error x The maximum and minimum values ​​of ; τ is the sampling time interval factor; θ is the rotation angle of the mirror symmetry plane, nis the number of mirror-symmetric faces; represents the gain of the drawing angle; is the one-dimensional time series measurement error x No. i+t sampling points.

[0020] Optionally, the process of applying L1 attention to the Resnet50 network structure to build the Resnet50-L1M network structure is as follows:

[0021] Take the pixels at the same position corresponding to different channels in the input convolution feature as a vector, calculate the L1 norm of the vector group, and get the feature map;

[0022] The feature map is convolved to obtain significant spatial information;

[0023] The sigmoid function is used to map the significant spatial information to obtain the weight matrix in space. The calculation method is:

[0024]

[0025]

[0026] in, Represents vector groups V n Middle i Vector V i The Manhattan distance of Indicates a convolution operation with a convolution kernel of 7×7. is the calculated spatial attention weight, n Represents the total number of pixels in a single-channel feature map, F M Represents a vector group Vn The feature map obtained by L1 norm calculation is: is the activation function.

[0027] Optionally, a pre-built deep convolutional neural network model is trained according to the symmetric point pattern graphs of different errors to generate an error online recognition model, including:

[0028] Mark each feature in the symmetric point pattern map and determine the error label of each feature;

[0029] The symmetric point pattern map and its corresponding error labels are input into the input layer of the Resnet50-L1M network structure, and the symmetric point pattern map is trained through the middle convolution layer of the Resnet50-L1M network structure to extract feature information. The extracted feature information is compressed through the pooling layer, and the compressed feature information is integrated in the fully connected layer, and the error value is output through the output layer.

[0030] Compare the error value with the corresponding error label to determine the training error of the Resnet50-L1M network structure;

[0031] When the training error reaches a preset value, an error online recognition model is generated.

[0032] According to another aspect of the present invention, a novel device for detecting measurement errors of wide-area capacitor voltage transformers in a power system is provided, comprising:

[0033] The acquisition module is used to construct the topological structure of the new power system wide-area substation cluster, and collect the A, B, and C three-phase time-series voltage data of the capacitor voltage transformers at both ends of all transmission lines in the topological structure within a preset time period, wherein the A, B, and C three-phase time-series voltage data includes voltage amplitude time-series data and voltage phase time-series data;

[0034] A calculation module is used to calculate the one-dimensional time series measurement error of each transmission line according to the time series voltage data of the three phases A, B and C by using the Euclidean distance method, wherein the one-dimensional time series measurement error includes: amplitude time series measurement error and phase time series measurement error;

[0035] The first generation module is used to convert the one-dimensional time series measurement error of each transmission line into a two-dimensional time-frequency diagram by using a symmetrical point mode, and generate a symmetrical point mode spectrum of different errors, wherein the symmetrical point mode spectrum includes: an amplitude distance characteristic spectrum and a phase distance characteristic spectrum;

[0036] The second generation module is used to train a pre-built deep convolutional neural network model according to symmetric point pattern graphs of different errors to generate an error online recognition model, wherein the deep convolutional neural network model is a Resnet50-L1M network structure established by applying L1 attention to the Resnet50 network structure;

[0037] The output module is used to input the collected real-time A, B, and C three-phase time-series voltage data of the new power system wide-area substation cluster into the error online identification model, and output the metering error prediction result of the new power system wide-area substation cluster.

[0038] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the storage medium stores a computer program, and the computer program is used to execute the method described in any one of the above aspects of the present invention.

[0039] According to another aspect of the present invention, an electronic device is provided, comprising: a processor; a memory for storing instructions executable by the processor; the processor is configured to read the executable instructions from the memory and execute the instructions to implement the method described in any one of the above aspects of the present invention.

[0040] Therefore, the present invention combines the ResNet50 model with the L1 paradigm attention mechanism and proposes the ResNet50-L1M algorithm model. The L1 paradigm attention mechanism improves the feature selection capability by calculating the absolute difference between features, enabling the model to capture key features more accurately and improve classification accuracy. The L1 paradigm is more robust to noise and outliers, reducing the interference of noise on the final output of the model, and improving the stability and reliability of the model; it improves the gradient flow and makes the optimization process smoother, especially at the depth position close to the output layer. Its performance in high-dimensional space is more stable, and it can more accurately measure the similarity between features, improve the quality of feature representation, and thus improve the classification accuracy of the model. This solves the problem of online identification of the error state of interconnected CVTs between wide-area substations in the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] A more complete understanding of exemplary embodiments of the present invention may be obtained by referring to the following drawings:

[0042] Figure 1 It is a flow chart of a novel method for detecting measurement errors of a wide-area capacitor voltage transformer in a power system provided by an exemplary embodiment of the present invention;

[0043] Figure 2 is a schematic diagram of a wide area power plant / substation connection provided by an exemplary embodiment of the present invention;

[0044] Figure 3 is a schematic diagram of an SDP method provided by an exemplary embodiment of the present invention;

[0045] Figure 4 is a residual connection schematic diagram provided by an exemplary embodiment of the present invention;

[0046] Figure 5 is a Resnet50 structure diagram provided by an exemplary embodiment of the present invention;

[0047] Figure 6 is a schematic diagram of a spatial attention module provided by an exemplary embodiment of the present invention;

[0048] Figure 7 It is a structural schematic diagram of a novel power system wide-area capacitor voltage transformer metering error detection device provided by an exemplary embodiment of the present invention;

[0049] Figure 8 This is a structure of an electronic device provided by an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0050] Below, the exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments of the present invention, and it should be understood that the present invention is not limited to the exemplary embodiments described here.

[0051] It should be noted that the relative arrangement of components and steps, the numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present invention unless specifically stated otherwise.

[0052] Those skilled in the art can understand that the terms "first" and "second" in the embodiments of the present invention are only used to distinguish different steps, devices or modules, etc., and neither represent any specific technical meaning nor indicate the necessary logical order between them.

[0053] It should also be understood that, in the embodiments of the present invention, “plurality” may refer to two or more than two, and “at least one” may refer to one, two or more than two.

[0054] It should also be understood that any component, data or structure mentioned in the embodiments of the present invention can generally be understood as one or more, unless explicitly limited or otherwise indicated in the context.

[0055] In addition, the term "and / or" in the present invention is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in the present invention generally indicates that the associated objects before and after are in an "or" relationship.

[0056] It should also be understood that the description of the various embodiments of the present invention focuses on the differences between the various embodiments, and the same or similar aspects thereof can be referenced to each other, and for the sake of brevity, they will not be described one by one.

[0057] At the same time, it should be understood that for the convenience of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship.

[0058] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses.

[0059] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.

[0060] It should be noted that like reference numerals and letters refer to similar items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0061] Embodiments of the present invention can be applied to electronic devices such as terminal devices, computer systems, servers, etc., which can operate with many other general or special computing system environments or configurations. Examples of well-known terminal devices, computing systems, environments and / or configurations suitable for use with electronic devices such as terminal devices, computer systems, servers, etc. include, but are not limited to: personal computer systems, server computer systems, thin clients, thick clients, handheld or laptop devices, microprocessor-based systems, set-top boxes, programmable consumer electronics, network personal computers, small computer systems, large computer systems, and distributed cloud computing technology environments including any of the above systems, etc.

[0062] Electronic devices such as terminal devices, computer systems, servers, etc. can be described in the general context of computer system executable instructions (such as program modules) executed by computer systems. Generally, program modules can include routines, programs, object programs, components, logic, data structures, etc., which perform specific tasks or implement specific abstract data types. Computer systems / servers can be implemented in a distributed cloud computing environment, where tasks are performed by remote processing devices linked through a communication network. In a distributed cloud computing environment, program modules can be located on local or remote computing system storage media including storage devices.

[0063] Exemplary Methods

[0064] Figure 1 1 is a flow chart of a novel method for detecting measurement errors of a wide-area capacitor voltage transformer in a power system provided by an exemplary embodiment of the present invention. This embodiment can be applied to electronic devices, such as Figure 1 As shown, the novel power system wide-area capacitor voltage transformer measurement error detection method 100 includes the following steps:

[0065] Step 101, constructing a topological structure of a new power system wide-area substation cluster, and collecting A, B, and C three-phase time-series voltage data of capacitor voltage transformers at both ends of all transmission lines in the topological structure within a preset time period, wherein the A, B, and C three-phase time-series voltage data includes voltage amplitude time-series data and voltage phase time-series data;

[0066] Step 102, using the Euclidean distance method to calculate the one-dimensional time series measurement error of each transmission line according to the time series voltage data of the three phases A, B, and C, wherein the one-dimensional time series measurement error includes: amplitude time series measurement error and phase time series measurement error;

[0067] Step 103, using a symmetrical point pattern to convert the one-dimensional time series measurement error of each transmission line into a two-dimensional time-frequency diagram, and generate symmetrical point pattern spectra of different errors, wherein the symmetrical point pattern spectra include: an amplitude distance characteristic spectrum and a phase distance characteristic spectrum;

[0068] Step 104, training a pre-built deep convolutional neural network model according to symmetrical point pattern graphs of different errors to generate an error online recognition model, wherein the deep convolutional neural network model is a Resnet50-L1M network structure established by applying L1 attention to a Resnet50 network structure;

[0069] Step 105, input the collected real-time A, B, C three-phase time-series voltage data of the new power system wide-area substation cluster into the error online identification model, and output the metering error prediction result of the new power system wide-area substation cluster.

[0070] Specifically, in order to solve the problem of rapid online error identification of CVTs interconnected between wide-area substations in the power system, the present invention proposes an online identification method for CVT errors based on PMU measurement data, SDP spectrum and Resnet50-L1M. First, the three-phase voltage data of the CVTs interconnected between wide-area substations is measured by PMU. Subsequently, the three-phase voltage amplitudes of the CVTs interconnected between substations are converted into Euclidean distances, and the daily Euclidean distances are converted into SDP spectrums, where different measurement errors can correspond to different types of SDP spectrums. The L1 norm is applied to the attention mechanism, and the designed L1 attention mechanism is applied to Resnet50, thereby proposing the Resnet50-L1M model. Resnet50-L1M is used to train CVT SDP spectrum images with different errors to establish an online recognition model for error states, and finally the trained model is used for online error identification of CVTs interconnected between wide-area substations. The specific contents are as follows:

[0071] When different substations are connected by transmission lines, the changes in CVT voltage and current at both ends of the line show a strong correlation. Figure 2 As shown in the figure, starting from power plant G1, due to line losses between substations, there are relative errors in the CVT voltage and current signals in two adjacent stations. If the error can be accurately calculated, the specific wide-area CVT metering error status can be determined by evaluating the combination of error results within the evaluation area.

[0072] First, the three-phase voltage data of the interconnected CVT of the wide-area substation is obtained by using the PMU device, and then the difference of the three-phase voltage of the interconnected CVT of the substations on both sides of the line is calculated by using the Euclidean distance. Then, the Euclidean distance is converted into an SDP graph. The Resnet50-L1M model is used to train the CVT SDP images with different errors to establish an error online recognition model, so as to identify the relative error state of the interconnected CVT.

[0073] 1. Feature extraction:

[0074] Euclidean distance, also known as Euclidean distance, is the true distance between two points in m-dimensional space. First, the distance d between two points in two-dimensional space is defined as:

[0075] (1)

[0076] in are the coordinates of the first point, The coordinates of the second point.

[0077] Similarly, the formula for the distance D between two points in three-dimensional space is:

[0078] (2)

[0079] Where are the coordinates of the first point and are the coordinates of the second point.

[0080] SDP analysis is a time series signal transformation method in polar coordinate system, which converts one-dimensional time domain signal into polar SDP diagram. Different SDP diagrams reflect the difference in frequency and amplitude of the collected signal. The graphic principle of SDP analysis is as follows Figure 3 shown.

[0081] Assumptions x is the measured time signal sequence, and x i is the i-th signal. According to the SDP principle, x i Transform to polar coordinate space In the polar coordinates, the radius Can be mapped to:

[0082] (3)

[0083] in and Respectively represent signals x Assume that τ is the sampling time interval. Then and The reverse rotation angle is given by The mapping is:

[0084] (4)

[0085] (5)

[0086] In the formula θ is the rotation angle of the mirror symmetry plane, n is the number of mirror symmetric faces; Indicates the gain of the drawing angle.

[0087] The point of the SDP transformation is to locate the point on the polar coordinate plane S Differences in the frequency or amplitude of the original signal can be manifested as differences in the position and curvature of the polar plane points.

[0088] 2. ResNet50-LIM model:

[0089] CNN is mainly composed of input layer, middle layer and output layer. The middle layer is connected by multiple convolutional layers and pooling layers alternately, and finally connected to the fully connected layer. CNN inputs the image information into the model through the input layer, extracts and compresses the image information through the convolutional layer, integrates the extracted features in the fully connected layer, and finally outputs the prediction results through the output layer. CNN adopts the weight sharing method, that is, the weight data contained in the corresponding convolution kernel and pooling window in each convolutional layer and pooling layer is a fixed value, which reduces the parameter information in the network to a certain extent and simplifies the model structure.

[0090] ResNet50-LIM: When building a CNN model, as the model becomes more complex and deep, gradient vanishing or gradient exploding may occur, affecting the final accuracy of the model. To address this phenomenon, a residual architecture is proposed. The core idea is to add additional jump connections to directly pass the input information to the subsequent processing layer to form a residual block. This structure enables the model to learn high-level features efficiently, alleviate the problem of gradient vanishing and gradient exploding, and thus enable deeper networks to be successfully trained and improve performance. Figure 4 shown.

[0091] Among them, the Resnet50 network structure adopts the residual connection method. Its network structure first includes the input layer and the residual blocks of four stages, such as Figure 5As shown in the figure. The input layer includes a 7x7 convolution layer with 64 convolution kernels and a 3x3 maximum pooling layer. The stride of the convolution layer and the pooling layer in this part are both 2. The residual module mainly includes BTNK1 and BTNK2. The BTNK1 block in each residual block is composed of 4 convolution layers, and each BTNK2 block is composed of 3 3 convolution layers. The last part of the Resnet50 network structure is the global average pooling layer and the fully connected layer, which are combined with the Softmax function to calculate the probability distribution.

[0092] In ResNet50-LIM, the present invention uses the calculation of norm to characterize the spatial position of the target. First, the input convolution feature ( H × W × C ) The pixels corresponding to the same position in different channels are taken as a vector, and the L1 norm of the vector group is calculated to obtain the feature map. Then, the convolution operation is performed to obtain the significant spatial information. Finally, the sigmoid function is used to map the spatial weight matrix. Figure 6 As shown in the figure Vn represents a vector group, n Represents the total number of pixels in a single-channel feature map, F M Represents the feature map of the vector group calculated by L1 norm.

[0093] By vector group Vn The norm calculation of , compares the position differences of different vectors in the vector space. The calculation method is:

[0094] (6)

[0095] (7)

[0096] in, Respectively represent i Vector V i The Manhattan distance of Indicates a convolution operation with a convolution kernel of 7×7. is the calculated spatial attention weight. Norm-based spatial attention uses norm calculation to measure the spatial position of features, and effectively utilizes spatial context information through convolution operations to construct a spatial mapping that is closer to the real image.

[0097] Therefore, the present invention combines the ResNet50 model with the L1 paradigm attention mechanism and proposes the ResNet50-L1M algorithm model. The L1 paradigm attention mechanism improves the feature selection capability by calculating the absolute difference between features, enabling the model to capture key features more accurately and improve classification accuracy. The L1 paradigm is more robust to noise and outliers, reducing the interference of noise on the final output of the model, and improving the stability and reliability of the model; it improves the gradient flow and makes the optimization process smoother, especially at the depth position close to the output layer. Its performance in high-dimensional space is more stable, and it can more accurately measure the similarity between features, improve the quality of feature representation, and thus improve the classification accuracy of the model. This solves the problem of online identification of the error state of interconnected CVTs between wide-area substations in the power system.

[0098] Exemplary Devices

[0099] Figure 7 1 is a schematic diagram of the structure of a novel power system wide-area capacitor voltage transformer measurement error detection device provided by an exemplary embodiment of the present invention. Figure 7 As shown, the apparatus 700 includes:

[0100] The acquisition module 710 is used to construct the topological structure of the new power system wide-area substation cluster, and collect the A, B, and C three-phase time-series voltage data of the capacitor voltage transformers at both ends of all transmission lines in the topological structure within a preset time period, wherein the A, B, and C three-phase time-series voltage data includes voltage amplitude time-series data and voltage phase time-series data;

[0101] A calculation module 720 is used to calculate the one-dimensional time series measurement error of each transmission line according to the time series voltage data of the three phases A, B and C by using the Euclidean distance method, wherein the one-dimensional time series measurement error includes: amplitude time series measurement error and phase time series measurement error;

[0102] The first generating module 730 is used to convert the one-dimensional time series measurement error of each transmission line into a two-dimensional time-frequency diagram by using a symmetrical point pattern, and generate a symmetrical point pattern spectrum of different errors, wherein the symmetrical point pattern spectrum includes: an amplitude distance characteristic spectrum and a phase distance characteristic spectrum;

[0103] A second generation module 740 is used to train a pre-built deep convolutional neural network model according to symmetric point pattern graphs of different errors to generate an error online recognition model;

[0104] The output module 750 is used to input the collected real-time A, B, and C three-phase time-series voltage data of the new power system wide-area substation cluster into the error online identification model, and output the metering error prediction result of the new power system wide-area substation cluster.

[0105] Optionally, the calculation module 720 includes:

[0106] The first marking submodule is used to mark the coordinates of the collected A, B, and C three-phase time-series voltage data of the capacitor voltage transformer at both ends of each transmission line in three-dimensional space according to the collection time sequence;

[0107] The calculation submodule is used to calculate the distance of each group of A, B, C three-phase time-series voltage data of the capacitor voltage transformer at both ends of the transmission line in three-dimensional space to obtain the distance points of the capacitor voltage transformer at both ends of the transmission line, wherein the distance points include: amplitude distance points and phase distance points;

[0108] The determination submodule is used to determine the one-dimensional time series measurement error according to the distance point of each time node within a preset time period.

[0109] Optionally, the calculation formula of the distance point is expressed as:

[0110]

[0111] In the formula, and are the amplitude distance points and phase distance points of the capacitive voltage transformers at both ends of the transmission line in three-dimensional space; is the three-dimensional spatial coordinate point of the voltage amplitude of the three phases A, B, and C of any substation 1 in the new power system wide-area substation cluster, It is the three-dimensional spatial coordinate point of the A, B, and C three-phase voltage phase of any substation 2 in the new power system wide-area substation cluster.

[0112] Optionally, the one-dimensional time series measurement error in the symmetric point pattern spectrum x No. i Sampling points x i The polar coordinate space is ;in,

[0113]

[0114]

[0115]

[0116] In the formula, and Represents the one-dimensional time series measurement error x The maximum and minimum values ​​of ; τ is the sampling time interval factor; θ is the rotation angle of the mirror symmetry plane, n is the number of mirror-symmetric faces; represents the gain of the drawing angle; is the one-dimensional time series measurement error x No. i+t sampling points.

[0117] Optionally, the deep convolutional neural network model is a Resnet50-L1M network structure established by applying L1 attention to a Resnet50 network structure, wherein the process of applying L1 attention to the Resnet50 network structure to establish the Resnet50-L1M network structure is as follows:

[0118] Take the pixels at the same position corresponding to different channels in the input convolution feature as a vector, calculate the L1 norm of the vector group, and get the feature map;

[0119] The feature map is convolved to obtain significant spatial information;

[0120] The sigmoid function is used to map the significant spatial information to obtain the weight matrix in space. The calculation method is:

[0121]

[0122] in, Represents vector groups V n Middle i Vector V i The Manhattan distance of Indicates a convolution operation with a convolution kernel of 7×7. is the calculated spatial attention weight, n Represents the total number of pixels in a single-channel feature map, F M Represents a vector group Vn The feature map obtained by L1 norm calculation is: is the activation function.

[0123] Optionally, the second generating module 740 includes:

[0124] The second marking submodule is used to mark each feature in the symmetric point pattern map and determine the error label of each feature;

[0125] The training submodule is used to input the symmetric point pattern map and its corresponding error labels into the input layer of the Resnet50-L1M network structure, train the symmetric point pattern map through the middle convolution layer of the Resnet50-L1M network structure, extract feature information, compress the extracted feature information through the pooling layer, integrate the compressed feature information in the fully connected layer, and output the error value through the output layer;

[0126] The comparison submodule is used to compare the error value with the corresponding error label to determine the training error of the Resnet50-L1M network structure;

[0127] The generation submodule is used to generate an error online recognition model when the training error reaches a preset value.

[0128] Exemplary Electronic Devices

[0129] Figure 8 This is a structure of an electronic device provided by an exemplary embodiment of the present invention. Figure 8 As shown, the electronic device 80 includes one or more processors 81 and a memory 82 .

[0130] The processor 81 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions.

[0131] The memory 82 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may include, for example, a read-only memory (ROM), a hard disk, a flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 81 may run the program instructions to implement the methods of the software programs of the various embodiments of the present invention described above and / or other desired functions. In one example, the electronic device may also include: an input device 83 and an output device 84, which are interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0132] In addition, the input device 83 may also include, for example, a keyboard, a mouse, etc.

[0133] The output device 84 can output various information to the outside, and can include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto.

[0134] Of course, to simplify, Figure 8 Only some of the components related to the present invention in the electronic device are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, the electronic device may further include any other appropriate components according to specific application conditions.

[0135] Exemplary computer program products and computer-readable storage media

[0136] In addition to the above-mentioned methods and devices, an embodiment of the present invention may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the method according to various embodiments of the present invention described in the above-mentioned "Exemplary Method" section of this specification.

[0137] The computer program product may be written in any combination of one or more programming languages ​​to write program code for performing the operations of the embodiments of the present invention, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0138] In addition, an embodiment of the present invention may also be a computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, enable the processor to execute the steps of the method according to various embodiments of the present invention described in the above “Exemplary Method” section of this specification.

[0139] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, system or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0140] The basic principle of the present invention is described above in conjunction with specific embodiments. However, it should be pointed out that the advantages, strengths, effects, etc. mentioned in the present invention are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. must be possessed by each embodiment of the present invention. In addition, the specific details disclosed above are only for the purpose of illustration and facilitation of understanding, rather than limitation, and the above details do not limit the present invention to being implemented by adopting the above specific details.

[0141] Each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system embodiment, since it basically corresponds to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0142] The block diagrams of the devices, systems, equipment, and systems involved in the present invention are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, systems, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and / or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The word "such as" used here refers to the phrase "such as but not limited to", and can be used interchangeably with it.

[0143] The method and system of the present invention may be implemented in many ways. For example, the method and system of the present invention may be implemented by software, hardware, firmware or any combination of software, hardware, firmware. The above order of steps for the method is only for illustration, and the steps of the method of the present invention are not limited to the order specifically described above, unless otherwise specifically stated. In addition, in some embodiments, the present invention may also be implemented as a program recorded in a recording medium, which includes machine-readable instructions for implementing the method according to the present invention. Thus, the present invention also covers a recording medium storing a program for executing the method according to the present invention.

[0144] It should also be noted that in the system, device and method of the present invention, each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. The above description of the disclosed aspects is provided to enable any technician in the field to make or use the present invention. Various modifications to these aspects are very obvious to those skilled in the art, and the general principles defined here can be applied to other aspects without departing from the scope of the present invention. Therefore, the present invention is not intended to be limited to the aspects shown here, but in accordance with the widest range consistent with the principles and novel features disclosed here.

[0145] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present invention to the forms disclosed herein. Although a number of example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A novel method for detecting measurement errors of wide-area capacitor voltage transformers in power systems, characterized in that: include: Constructing a topological structure of a new power system wide-area substation cluster, and collecting A, B, and C three-phase time-series voltage data of capacitor voltage transformers at both ends of all transmission lines in the topological structure within a preset time period, wherein the A, B, and C three-phase time-series voltage data includes voltage amplitude time-series data and voltage phase time-series data; Using the Euclidean distance method to calculate the one-dimensional time series measurement error of each of the transmission lines according to the three-phase time series voltage data of A, B, and C, wherein the one-dimensional time series measurement error includes: amplitude time series measurement error and phase time series measurement error; The one-dimensional time series measurement error of each transmission line is converted into a two-dimensional time-frequency diagram by using a symmetrical point pattern, and a symmetrical point pattern spectrum of different errors is generated, wherein the symmetrical point pattern spectrum includes: an amplitude distance characteristic spectrum and a phase distance characteristic spectrum; Training a pre-built deep convolutional neural network model according to the symmetrical point pattern graphs of different errors to generate an error online recognition model, wherein the deep convolutional neural network model is a Resnet50-L1M network structure established by applying L1 attention to a Resnet50 network structure; The collected real-time A, B, and C three-phase time-series voltage data of the new power system wide-area substation cluster are input into the error online identification model, and the metering error prediction result of the new power system wide-area substation cluster is output.

2. The method according to claim 1, characterized in that The one-dimensional time series measurement error of each of the transmission lines is calculated according to the time series voltage data of the three phases A, B and C using the Euclidean distance method, including: Mark the coordinates of the A, B, and C three-phase time-series voltage data of the capacitor voltage transformers at both ends of each transmission line in three-dimensional space according to the order of collection time; Performing distance calculation on each group of A, B, C three-phase time-series voltage data of the capacitor voltage transformer at both ends of the transmission line in three-dimensional space to obtain distance points of the capacitor voltage transformer at both ends of the transmission line, wherein the distance points include: amplitude distance points and phase distance points; The one-dimensional time series measurement error is determined according to the distance point of each time node within the preset time period.

3. The method according to claim 2, characterized in that The calculation formula of the distance point is expressed as: In the formula, and are the amplitude distance points and phase distance points of the capacitive voltage transformers at both ends of the transmission line in three-dimensional space; is the three-dimensional spatial coordinate point of the voltage amplitude of the three phases A, B, and C of any substation 1 in the new power system wide-area substation cluster, It is the three-dimensional spatial coordinate point of the A, B, and C three-phase voltage phase of any substation 2 in the new power system wide-area substation cluster.

4. The method according to claim 1, characterized in that: The one-dimensional time series measurement error in the symmetrical point pattern spectrum x No. i Sampling points x i The polar coordinate space is ;in, In the formula, and Represents the one-dimensional time series measurement error x The maximum and minimum values ​​of ; t is the sampling time interval factor; is the rotation angle of the mirror symmetry plane; represents the gain of the drawing angle; is the one-dimensional time series measurement error x No. i+t sampling points.

5. The method according to claim 1, characterized in that The process of applying L1 attention to the Resnet50 network structure to build the Resnet50-L1M network structure is as follows: Take the pixels at the same position corresponding to different channels in the input convolution feature as a vector, calculate the L1 norm of the vector group, and get the feature map; The feature map is subjected to a convolution operation to obtain significant spatial information; The salient spatial information is mapped using the Sigmoid function to obtain a spatial weight matrix, and the calculation method is: in, Represents vector groups V n Middle i Vector V i The Manhattan distance of Indicates a convolution operation with a convolution kernel of 7×7. is the calculated spatial attention weight, n Represents the total number of pixels in a single-channel feature map, F M Represents a vector group V n The feature map obtained by L1 norm calculation is: is the activation function.

6. The method according to claim 5, characterized in that The pre-built deep convolutional neural network model is trained according to the symmetrical point pattern graphs of different errors to generate an error online recognition model, including: Marking each feature in the symmetric point pattern map and determining an error label for each feature; Input the symmetrical point pattern map and its corresponding error label into the input layer of the Resnet50-L1M network structure, train the symmetrical point pattern map through the middle convolution layer of the Resnet50-L1M network structure, extract feature information, compress the extracted feature information through the pooling layer, integrate the compressed feature information in the fully connected layer, and output the error value through the output layer; Compare the error value with the corresponding error label to determine the training error of the Resnet50-L1M network structure; When the training error reaches a preset value, the error online recognition model is generated.

7. A novel power system wide-area capacitor voltage transformer measurement error detection device, characterized in that: include: The acquisition module is used to construct a topological structure of a new power system wide-area substation cluster, and to collect the A, B, and C three-phase time-series voltage data of the capacitive voltage transformers at both ends of all transmission lines in the topological structure within a preset time period, wherein the A, B, and C three-phase time-series voltage data includes voltage amplitude time-series data and voltage phase time-series data; A calculation module, used for calculating the one-dimensional time series measurement error of each of the transmission lines according to the time series voltage data of the three phases A, B and C by using the Euclidean distance method, wherein the one-dimensional time series measurement error includes: amplitude time series measurement error and phase time series measurement error; A first generating module is used to convert the one-dimensional time series measurement error of each of the transmission lines into a two-dimensional time-frequency diagram by using a symmetrical point pattern, and generate a symmetrical point pattern spectrum of different errors, wherein the symmetrical point pattern spectrum includes: an amplitude distance characteristic spectrum and a phase distance characteristic spectrum; A second generation module is used to train a pre-built deep convolutional neural network model according to the symmetrical point pattern graphs of different errors to generate an error online recognition model, wherein the deep convolutional neural network model is a Resnet50-L1M network structure established by applying L1 attention to the Resnet50 network structure; The output module is used to input the collected real-time A, B, and C three-phase time-series voltage data of the new power system wide-area substation cluster into the error online identification model, and output the metering error prediction result of the new power system wide-area substation cluster.

8. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and the computer program is used to execute the method according to any one of claims 1 to 6.

9. An electronic device, characterized in that: The electronic device comprises: processor; a memory for storing instructions executable by the processor; The processor is used to read the executable instructions from the memory and execute the instructions to implement the method described in any one of claims 1 to 6.

Citation Information

Patent Citations

  • Numerically-controlled machine tool fault diagnosis method based on symmetric point mode and spatial density clustering

    CN108490877A

  • Capacitive voltage transformer metering error prediction method and system, terminal and medium

    CN115878963A

  • Capacitive voltage transformer insulation fault diagnosis method, system, device and medium

    CN116008756A

  • Cross-site mutual inductor metering error online monitoring method and device and electronic equipment

    CN116520234A

  • Capacitive voltage transformer error prediction method, device, equipment and medium

    CN117475158A