Error evaluation method and device of current transformer, equipment and medium

Abstract

The invention provides an error evaluation method, device and equipment of a current transformer and a medium, and belongs to the technical field of transformer evaluation. The error assessment method comprises the following steps: acquiring secondary current data of a current transformer to be assessed; decomposing the secondary current data into an intrinsic mode component and a residual component by using an ensemble empirical mode decomposition method; reconstructing the intrinsic mode component to obtain a high-frequency component and a low-frequency component; and inputting the secondary current data, the high-frequency component, the low-frequency component and the residual component into an error evaluation model, and evaluating the error state of the current transformer. In order to solve the problem of large fluctuation of the current data, the secondary current data of the transformer is decomposed and reconstructed, the reconstructed data is processed in a layered iteration fusion mode, and finally the evaluation result of the current transformer is obtained, so that the evaluation precision is high.

Description

Technical Field

[0001] The present invention belongs to the technical field of transformer evaluation, and in particular relates to an error evaluation method, device, equipment and medium for a current transformer. Background Art

[0002] Current transformers (CT) are important measuring devices in power systems. The primary winding of the current transformer is connected in series in the main circuit of power transmission and transformation, and the secondary winding is connected to measuring instruments, relay protection or automatic devices according to different requirements. It is used to change the large current of the primary circuit into a small current on the secondary side for safe collection by measurement, control, protection and metering equipment. The accuracy and reliability of current transformers are of great significance to the safe operation, control and protection, electric energy measurement and trade settlement of power systems.

[0003] At present, the error assessment of current transformers usually adopts the method of offline calibration or online calibration, and obtains the ratio error and angle error of electronic current transformers by direct comparison. However, these methods have a long calibration cycle, complex on-site wiring, and low work efficiency. In addition, due to the large fluctuation of current data, frequent transient processes, large amplitude fluctuations, and no fixed rules, if the model is simply based on the current amplitude and phase characteristics, it will have an adverse effect on the evaluation accuracy. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides an error evaluation method, device, equipment and medium for a current transformer.

[0005] In order to solve one, part or all of the above technical problems, the technical solution adopted by the present invention is: A method for evaluating an error of a current transformer comprises: obtaining secondary current data of a current transformer to be evaluated; decomposing the secondary current data into an intrinsic mode component and a residual component by using a set empirical mode decomposition method; reconstructing the intrinsic mode component to obtain a high-frequency component and a low-frequency component; and inputting the secondary current data, the high-frequency component, the low-frequency component and the residual component into an error evaluation model to evaluate the error state of the current transformer.

[0006] Furthermore, the error assessment model includes a processing module, a 3×3 convolutional layer and an MLP neural network; the processing module includes a raw data processing module for processing raw secondary current data, a high-frequency data processing module for processing high-frequency components, a low-frequency data processing module for processing low-frequency components, and a residual data processing module for processing residual components; the processing modules are all provided with a convolutional layer.

[0007] Furthermore, the 3×3 convolutional layer takes the secondary current data, and the output results of the original data processing module, the high-frequency data processing module, the low-frequency data processing module and the residual data processing module as input; the MLP neural network takes the output result of the 3×3 convolutional layer as input; and the output result of the MLP neural network is the error state of the current transformer.

[0008] Furthermore, the original data processing module takes the secondary current data as input, and is provided with two convolution layers; the high-frequency data processing module takes the high-frequency component as input, and is provided with four convolution layers; the low-frequency data processing module takes the low-frequency component as input, and is provided with five convolution layers; the residual data processing module takes the residual component as input, and is provided with three convolution layers; and the processing modules all take the result of the last convolution layer as output.

[0009] Furthermore, in the convolutional layers of the original data processing module, the low-frequency data processing module and the residual data processing module, the output results of the previous layer are used as inputs to the next layer; in the convolutional layers of the high-frequency data processing module, the output results of the first layer are used as inputs to the second layer, and the output results of the third layer are used as inputs to the fourth layer, and the results of the second layer are not output to the third layer.

[0010] Furthermore, in the convolution layer of the processing module, the second layer of the original data processing module also takes the output result of the fourth layer of the high-frequency data processing module as input, the first layer of the high-frequency data processing module also takes the output result of the first layer of the original data processing module as input, the third layer of the high-frequency data processing module takes the output result of the fourth layer of the low-frequency data processing module as input, the first layer of the low-frequency data processing module also takes the output result of the second layer of the high-frequency data processing module as input, the first layer of the residual data processing module also takes the output result of the second layer of the low-frequency data processing module as input, and the third layer of the low-frequency data processing module also takes the output result of the second layer of the residual data processing module as input.

[0011] Furthermore, the specifications of the convolutional layers set by the processing modules are all 1×1.

[0012] An error evaluation device for a current transformer comprises an acquisition unit, a decomposition unit, a reconstruction unit and an evaluation unit; the acquisition unit is used to acquire secondary current data of a current transformer to be evaluated; the decomposition unit is used to decompose the secondary current data into an intrinsic mode component and a residual component by using a set empirical mode decomposition method; the reconstruction unit is used to reconstruct the intrinsic mode component to obtain a high-frequency component and a low-frequency component; the evaluation unit is used to input the secondary current data, the high-frequency component, the low-frequency component and the residual component into an error evaluation model to evaluate the error state of the current transformer.

[0013] An electronic device comprises a processor and a memory; the memory is used to store executable instructions, and the processor is used to execute the instructions to implement the error evaluation method.

[0014] A computer-readable storage medium stores instructions, and when the instructions are executed, the error evaluation method is implemented.

[0015] Compared with the prior art, the present invention has the following beneficial effects: Among the high-frequency components, low-frequency components and residual components obtained after the decomposition and reconstruction of the secondary current data, the high-frequency components fluctuate violently and have a short period, reflecting the violent fluctuations of the secondary current data, which may be caused by line faults, high-proportion penetration of distributed power sources, etc.; the low-frequency components fluctuate gently and have a long period, reflecting the secondary current fluctuations during the normal operation of the power grid; the residual components reflect the long-term change trend of the current data. The low-frequency components contain more features that can reflect the operating status of the current transformer, so five convolution layers are set in the low-frequency data processing module of the method for feature extraction; but the mutual influence of the high-frequency components and the residual components on the low-frequency components cannot be ignored. The high-frequency components and the residual components also contain features that can reflect the operating status of the current transformer, so four convolution layers are set in the high-frequency data processing module, and three convolution layers are set in the residual data processing module. At the same time, the convolution layers between different processing modules also perform data fusion, and feature extraction is performed while taking into account the association between different types of data.

[0016] The high-frequency data processing module of the present invention does not continuously use convolution layers for feature extraction, but integrates the features extracted by two convolution layers into the low-frequency data processing module for processing, and then returns the processed features to the high-frequency data processing module to continue feature extraction. In this way, the features extracted from the high-frequency components will integrate more low-frequency component features, and more features reflecting the operating status of the current transformer can be extracted.

[0017] In view of the problem of large fluctuations in current data, the present invention decomposes and reconstructs the secondary current data of the transformer, and processes the reconstructed data in a layered iterative fusion manner, and finally obtains the evaluation result of the current transformer, so the evaluation accuracy is high. The present invention realizes accurate and timely identification of out-of-tolerance transformers, thereby ensuring the stability of power grid operation. The present invention can timely discover the problem of out-of-tolerance errors of current transformers, reduce the error-over-limit operation time of current transformers, guide the detection work of transformers, and thus ensure the fairness of electric energy metering. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The present invention is further described in detail below in conjunction with the accompanying drawings.

[0019] Figure 1 : Flow chart of embodiment 1 of the present invention; Figure 2 : A schematic diagram of the structure of the error assessment model of the present invention; Figure 3 : Schematic diagram of artificially adding difference to the online monitoring data set in Example 1 of the present invention; Figure 4 : A schematic diagram of Embodiment 2 of the present invention; Figure 5 : Schematic diagram of embodiment 3 of the present invention. DETAILED DESCRIPTION

[0020] In order to better understand the present invention, the content of the present invention is further clearly described below in conjunction with the embodiments and the accompanying drawings, but the protection content of the present invention is not limited to the following embodiments. In the following description, a large number of specific details are provided to provide a more thorough understanding of the present invention. However, it is obvious to those skilled in the art that the present invention can be implemented without one or more of these details.

[0021] Example 1: See Figure 1-Figure 3 The purpose of this embodiment is to provide a current transformer error evaluation method based on decomposition and reconstruction. The error evaluation method includes: S1. Obtain secondary current data of the current transformer to be evaluated.

[0022] S2. Decompose the secondary current data of the current transformer into inherent modal components and residual components using the method of ensemble empirical mode decomposition.

[0023] Ensemble Empirical Mode Decomposition (EEMD) is an adaptive, data-driven signal analysis method used to decompose nonlinear and non-stationary data sequences.

[0024] In this step, EEMD is used to decompose the secondary current data set, and several intrinsic mode components (IMF components) and a residual component are obtained after decomposition. Each intrinsic mode component decomposed by EEMD contains local characteristic signals of the original signal at different time scales.

[0025] The EEMD algorithm is common knowledge and will not be described in detail.

[0026] S3. Reconstruct the decomposed intrinsic modal components to obtain high-frequency components and low-frequency components.

[0027] This step uses the FTC (Fine to coarse) method to reconstruct the decomposed intrinsic modal components to obtain high-frequency components and low-frequency components. Specifically, starting from the first intrinsic modal component, other intrinsic modal components are accumulated in sequence, and after each accumulation, it is determined whether the average value of the accumulated result significantly deviates from the zero point sequence; when the average value of the accumulated result begins to significantly deviate from the zero point, all the intrinsic modal components accumulated in the previous time are superimposed and reconstructed into high-frequency components, and the remaining intrinsic modal components are superimposed and reconstructed into low-frequency components.

[0028] The FTC algorithm is common knowledge and will not be elaborated on in detail.

[0029] S4. Input the secondary current data, and the residual component, high-frequency component and low-frequency component obtained by decomposing and reconstructing the secondary current data into an error evaluation model to evaluate the error state of the current transformer.

[0030] The structure of the error estimation model is as follows: Figure 2 As shown in the figure, a layered iterative fusion strategy is adopted to process the input secondary current data (original current data), high-frequency components, low-frequency components and residual components using several convolutional layers, and the convolutional layers between different input data are also fused with each other. In the model, 1×1 conv represents a 1×1 convolutional layer, and 3×3 conv represents a 3×3 convolutional layer. The convolutional layer is used to extract features from the input data and output the extracted features; MLP (Multilayer Perceptron) represents a multilayer perceptron, which is an artificial neural network with a forward structure. The output of MLP (i.e., the output of the error evaluation model) is the error state of the current transformer, such as normal, abnormal or alarm.

[0031] The error assessment model includes a processing module, a 3×3 convolutional layer and an MLP neural network. The processing module includes an original data processing module for processing original secondary current data, a high-frequency data processing module for processing high-frequency components, a low-frequency data processing module for processing low-frequency components, and a residual data processing module for processing residual components.

[0032] The error assessment model adopts a hierarchical iterative structure. Specifically, the raw data processing module takes the raw current data as input, and is equipped with two convolution layers; the high-frequency data processing module takes the high-frequency component as input, and is equipped with four convolution layers; the low-frequency data processing module takes the low-frequency component as input, and is equipped with five convolution layers; the residual data processing module takes the residual component as input, and is equipped with three convolution layers. Each processing module uses the output result of the last convolution layer as the output of the module. Except for the high-frequency data processing module, the output result of the previous convolution layer in each processing module is used as the input of the next convolution layer. In the high-frequency data processing module, the output result of the first layer is used as the input of the second layer, and the output result of the third layer is used as the input of the fourth layer, but the result of the second layer is not output to the third layer. The specifications of the convolution layers set for each processing module are all 1×1.

[0033] Data fusion is also performed between different processing modules. Specifically, the second layer of the original data processing module also uses the output result of the fourth layer of the high-frequency data processing module as input, the first layer of the high-frequency data processing module also uses the output result of the first layer of the original data processing module as input, the third layer of the high-frequency data processing module uses the output result of the fourth layer of the low-frequency data processing module as input, the first layer of the low-frequency data processing module also uses the output result of the second layer of the high-frequency data processing module as input, the first layer of the residual data processing module also uses the output result of the second layer of the low-frequency data processing module as input, and the third layer of the low-frequency data processing module also uses the output result of the second layer of the residual data processing module as input.

[0034] The 3×3 convolutional layer takes the original current data and the output results of the original data processing module, the high-frequency data processing module, the low-frequency data processing module and the residual data processing module as input, and the output result of the 3×3 convolutional layer is used as the input of the MLP.

[0035] Fine-to-coarse is used to reconstruct the high and low frequencies of each IMF subsequence to obtain a high-frequency sequence and a low-frequency sequence. The high-frequency sequence fluctuates violently and has a short period, which can be regarded as price fluctuations caused by normal market activities in a short period of time; the low-frequency sequence fluctuates gently and has a long period, which can be regarded as price fluctuations caused by uncertain sudden factors such as major policies and market environment; the trend item reflects the long-term trend of carbon market prices. In this step, after the secondary current data, as well as the high-frequency component, low-frequency component and residual component obtained by decomposing and reconstructing the secondary current data are input into the error evaluation model, the error state of the current transformer to be evaluated can be obtained.

[0036] Among the high-frequency components, low-frequency components and residual components obtained after decomposition and reconstruction of the secondary current data by this method, the high-frequency components fluctuate violently and have a short period, reflecting the violent fluctuations of the secondary current data, which may be caused by line faults, high-proportion penetration of distributed power sources, etc.; the low-frequency components fluctuate gently and have a long period, reflecting the secondary current fluctuations during normal operation of the power grid; the residual components reflect the long-term trend of current data. The low-frequency components contain more features that can reflect the operating status of the current transformer, so five convolution layers are set in the low-frequency data processing module of this method for feature extraction; but the mutual influence of high-frequency components and residual components on low-frequency components cannot be ignored. The high-frequency components and residual components also contain features that can reflect the operating status of the current transformer, so four convolution layers are set in the high-frequency data processing module, and three convolution layers are set in the residual data processing module. At the same time, the convolution layers between different processing modules also perform data fusion, and feature extraction is performed while taking into account the association between different types of data.

[0037] In particular, the high-frequency data processing module does not continuously use the convolution layer for feature extraction, but integrates the features extracted by the two convolution layers into the low-frequency data processing module for processing, and then returns the processed features to the high-frequency data processing module to continue feature extraction. In this way, the features extracted from the high-frequency components will integrate more low-frequency component features, and more features reflecting the operating status of the current transformer can be extracted.

[0038] The following numerical simulation method is used to verify the evaluation accuracy of this method. During the verification, the secondary output of qualified current transformers is first collected from the substation to construct an offline modeling and online monitoring data set; then artificial errors are added to the online monitoring data set to construct a test data set with known error states; finally, the accuracy and effectiveness of this method are verified by comparing the state evaluation results of this method with the preset error state.

[0039] Specifically, based on a 220KV substation, the 0.5-level current transformer for measurement is obtained based on the synchronous phasor measurement unit (PMU). It has a 1-input and 3-output topology. According to its offline verification work records, the actual error state of the 0.5-level CT A phase is normal. Figure 3 Verify using the artificial difference method shown.

[0040] The verification results show that the evaluation accuracy of the method is 0.951, and the evaluation recall is 0.924. Therefore, the present invention can achieve accurate identification of current transformers.

[0041] Example 2: See Figure 4The purpose of this embodiment is to provide a current transformer error evaluation device based on decomposition and reconstruction, which is used to perform the error evaluation method of embodiment 1. The error evaluation device includes an acquisition unit, a decomposition unit, a reconstruction unit and an evaluation unit.

[0042] The acquisition unit is used to acquire the secondary current data of the current transformer to be evaluated. In combination with Embodiment 1, the acquisition unit is used to perform step S1.

[0043] The decomposition unit is used to decompose the secondary current data into intrinsic mode components and residual components by using a method of ensemble empirical mode decomposition. In combination with embodiment 1, the decomposition unit is used to perform step S2.

[0044] The reconstruction unit is used to reconstruct the intrinsic modal components to obtain high-frequency components and low-frequency components. In combination with embodiment 1, the reconstruction unit is used to perform step S3.

[0045] The evaluation unit is used to input the secondary current data, the high frequency component, the low frequency component and the residual component into an error evaluation model to evaluate the error state of the current transformer. In combination with embodiment 1, the reconstruction unit is used to perform step S4.

[0046] The structure of the error assessment model is as follows Figure 2 As shown, the specific structure refers to Example 1.

[0047] Regarding the device in this embodiment, the specific manner in which each unit or module performs operations has been described in detail in Example 1 and will not be elaborated here.

[0048] Those skilled in the art should understand that for the convenience and simplicity of description, the present application only uses the division of the above-mentioned functional modules or units as examples in the device embodiments. In actual applications, the above-mentioned functions can be assigned to different functional modules or units as needed, that is, the internal structure of the device can be divided into different functional modules or units to complete all or part of the functions described above.

[0049] Example 3: See Figure 5 The purpose of this embodiment is to provide an electronic device, including at least one processor and one or more memories for storing processor executable instructions. The processor is used to execute the instructions in the memory to implement the error evaluation method described in Embodiment 1.

[0050] It should be noted that electronic devices can be desktop computers, portable computers, network servers, mobile phones, tablet computers, wireless terminals, embedded devices, chip systems or Figure 5 In addition, Figure 5The components shown in the figure do not constitute a limitation on the terminal device, except Figure 5 In addition to the components shown, the electronic device may include more or fewer components than shown, or combine certain components, or arrange the components differently.

[0051] Embodiment 4: The purpose of this embodiment is to provide a computer-readable storage medium.

[0052] All or part of the processes in the above method embodiments may be completed by computer instructions to instruct relevant hardware. The program may be stored in the computer-readable storage medium. When the program is executed, the error assessment method described in Example 1 may be implemented.

[0053] The computer-readable storage medium may be an internal storage unit of the electronic device of Example 3, such as a hard disk or a memory. The computer-readable storage medium may also be an external storage device of the electronic device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. Further, the computer-readable storage medium may also include both an internal storage unit and an external storage device of the above-mentioned electronic device. The computer-readable storage medium is used to store computer programs and other programs and data required by the electronic device. The computer-readable storage medium may also be used to temporarily store data that has been output or is to be output.

[0054] In conjunction with the several embodiments provided in this application, it should be understood that the provided devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the assessment of the modules or units is only a logical function assessment. There may be other assessment methods in actual implementation, such as multiple modules or units or components can be combined or integrated into another device, or some features can be ignored or not executed.

[0055] In addition, in the embodiment of the device of the present application, each functional module or unit can be integrated into one unit, or each module or unit can exist physically separately, or two or more modules or units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional modules.

[0056] If the above-mentioned integrated unit is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a device (which can be a single-chip microcomputer, chip, etc.) or a processor (processor) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard drives, ROM, RAM, magnetic disks or optical disks.

[0057] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Other modifications or equivalent substitutions made to the technical solution of the present invention by ordinary technicians in the field should be included in the scope of the claims of the present invention as long as they do not depart from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for evaluating an error of a current transformer, characterized in that: include: Obtaining secondary current data of the current transformer to be evaluated; Decomposing the secondary current data into intrinsic modal components and residual components by using a method of ensemble empirical mode decomposition; Reconstructing the inherent modal components to obtain high-frequency components and low-frequency components; The secondary current data, the high frequency component, the low frequency component and the residual component are input into an error evaluation model to evaluate the error state of the current transformer.

2. The error evaluation method of current transformer according to claim 1, characterized in that: The error assessment model includes a processing module, a 3×3 convolutional layer and an MLP neural network; The processing module includes an original data processing module for processing original secondary current data, a high-frequency data processing module for processing high-frequency components, a low-frequency data processing module for processing low-frequency components, and a residual data processing module for processing residual components; The processing modules are all provided with a convolution layer.

3. The error evaluation method of a current transformer according to claim 2, characterized in that: The 3×3 convolutional layer takes the secondary current data, and the output results of the original data processing module, the high-frequency data processing module, the low-frequency data processing module and the residual data processing module as input; the MLP neural network takes the output result of the 3×3 convolutional layer as input; and the output result of the MLP neural network is the error state of the current transformer.

4. The error evaluation method of a current transformer according to claim 2, characterized in that: The raw data processing module takes the secondary current data as input and is provided with two convolution layers; The high-frequency data processing module takes the high-frequency component as input and is provided with four convolutional layers; The low-frequency data processing module takes the low-frequency component as input and is provided with five convolutional layers; The residual data processing module takes the residual component as input and is provided with three convolutional layers; The processing modules all use the result of the last convolutional layer as output.

5. The error evaluation method for current transformer according to claim 4, characterized in that: In the convolutional layers of the original data processing module, the low-frequency data processing module and the residual data processing module, the output results of the previous layer are used as the input of the next layer; In the convolutional layer of the high-frequency data processing module, the output result of the first layer is used as the input of the second layer, the output result of the third layer is used as the input of the fourth layer, and the result of the second layer is not output to the third layer.

6. The error evaluation method for current transformer according to claim 5, characterized in that: In the convolutional layer of the processing module, the second layer of the original data processing module also takes the output result of the fourth layer of the high-frequency data processing module as input, the first layer of the high-frequency data processing module also takes the output result of the first layer of the original data processing module as input, the third layer of the high-frequency data processing module takes the output result of the fourth layer of the low-frequency data processing module as input, the first layer of the low-frequency data processing module also takes the output result of the second layer of the high-frequency data processing module as input, the first layer of the residual data processing module also takes the output result of the second layer of the low-frequency data processing module as input, and the third layer of the low-frequency data processing module also takes the output result of the second layer of the residual data processing module as input.

7. The error evaluation method for current transformer according to claim 2, characterized in that: The specifications of the convolutional layers set by the processing modules are all 1×1.

8. A current transformer error evaluation device, characterized in that: It includes acquisition unit, decomposition unit, reconstruction unit and evaluation unit; The acquisition unit is used to acquire secondary current data of the current transformer to be evaluated; The decomposition unit is used to decompose the secondary current data into an intrinsic mode component and a residual component by using a method of collective empirical mode decomposition; The reconstruction unit is used to reconstruct the intrinsic modal components to obtain high-frequency components and low-frequency components; The evaluation unit is used to input the secondary current data, the high-frequency component, the low-frequency component and the residual component into an error evaluation model to evaluate the error state of the current transformer.

9. An electronic device, characterized in that: It comprises a processor and a memory; the memory is used to store executable instructions, and the processor is used to execute the instructions to implement the error assessment method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The readable storage medium stores instructions, and when the instructions are executed, the error evaluation method according to any one of claims 1 to 7 is implemented.

Images