Current transformer error fitting calculation method and device

Abstract

The invention discloses a current transformer error fitting calculation method, which relates to the technical field of power system measurement, and is used for solving the problem that the impedance error is difficult to accurately analyze in the prior art. The method comprises the following steps: generating a pilot frequency signal and injecting the pilot frequency signal into a secondary loop of a current transformer, and obtaining the impedance of each frequency point after the pilot frequency signal is injected; performing data cleaning on the impedance data, and performing data fusion on the cleaned impedance data through weighted fusion to obtain total impedance; performing impedance inversion on the total impedance to obtain excitation impedance and load impedance, and extracting equivalent parameters; constructing an objective function through a least square method, and performing parameter fitting in combination with gradient descent; and inputting the fitted parameters into an analysis model for model optimization, and obtaining optimized parameters. The invention further discloses the electronic equipment. According to the invention, through pilot frequency signal injection, full-band impedance measurement is realized, and comprehensive basic data is provided for error analysis of impedance parameters.

Description

Technical Field

[0001] The present invention relates to the technical field of power system measurement, and in particular to a current transformer error fitting calculation method and device based on heterofrequency signal injection and impedance inversion technology. Background Art

[0002] The main function of a current transformer is to convert the primary current into the secondary current in proportion, providing input signals for relay protection and energy metering equipment. Its metering accuracy directly affects the operational stability of the power system and the fairness of energy trading. However, errors in current transformers are unavoidable, mainly due to the following: the nonlinear characteristics of the excitation impedance: the saturation and hysteresis effects of the current transformer core cause the excitation impedance to vary with current and frequency, thereby introducing errors; the complexity of the load impedance: the load impedance of the secondary circuit includes resistance, inductance, and capacitance, and its dynamic changes have a significant impact on the metering accuracy of the current transformer; and the diversity of operating frequencies: the impedance characteristics and error performance of the current transformer vary significantly at different operating frequencies.

[0003] Therefore, error analysis is necessary. Existing error analysis methods typically rely on laboratory environments and single-frequency measurements, which have limitations. For example, single-frequency data is insufficient to reveal full-frequency response characteristics, impedance characteristics are difficult to separate, and field applicability is limited. Therefore, a more accurate current transformer error analysis method is needed. Summary of the Invention

[0004] In order to overcome the shortcomings of the prior art, one of the objectives of the present invention is to provide a current transformer error fitting calculation method, which obtains the full-band impedance by injecting heterodyne signals, and combines impedance inversion and parameter optimization to obtain the optimized prediction results of the impedance parameters.

[0005] One of the purposes of the present invention is achieved by the following technical solution: A current transformer error fitting calculation method includes the following steps: Generate an inter-frequency signal and inject it into the secondary circuit of the current transformer, and obtain the impedance at each frequency point after the inter-frequency signal is injected; Cleaning the impedance data, fusing the cleaned impedance data by weighted fusion to obtain a total impedance; Performing impedance inversion on the total impedance to obtain excitation impedance and load impedance, and extracting equivalent parameters; The objective function is constructed by the least squares method and the parameters are fitted in combination with gradient descent; The fitted parameters are input into the analysis model for model optimization, and the optimized parameters are obtained.

[0006] Furthermore, generating a different frequency signal and injecting it into the secondary circuit of the current transformer includes: Generate a multi-frequency sinusoidal current signal in the range of 200 Hz to 1500 Hz by a signal generator; The signal is injected into the secondary circuit of the current transformer through a contactless coupler in a manner of gradually increasing frequency.

[0007] Furthermore, the impedance at each frequency point after the injection of the heterodyne signal is obtained, including: Get the voltage at each frequency point detected by the high-sensitivity miniature current transformer and current , Calculating impedance : .

[0008] Furthermore, the impedance data is cleaned, and the cleaned impedance data is fused by weighted fusion to obtain the total impedance, including: The outliers in the impedance data are detected and eliminated using the Isolation Forest algorithm; Perform interpolation processing on the frequency points where outliers are removed; According to the weight, the cleaned impedance data is fused to meet the following requirements: ,in, is the total impedance, Frequency point The weight of Frequency point The impedance, is the number of frequency points.

[0009] Furthermore, the frequency point The weight is the preset weight of the frequency point, and the weight is updated according to the signal-to-noise ratio: when the signal-to-noise ratio of the measurement signal at the frequency point is higher than the preset value, the weight is increased, otherwise the weight is reduced.

[0010] Furthermore, impedance inversion is performed on the total impedance to obtain the excitation impedance and the load impedance, including: Construct an impedance separation model to meet the following requirements: ,in, is the excitation impedance, is the load impedance; The excitation impedance satisfies: ,in, is the resistance component of the excitation impedance, is the inductive component of the excitation impedance, represents the imaginary part, is the angular frequency of the signal, is the capacitive component of the excitation impedance; The load impedance satisfies: ,in, The resistive component of the load impedance, is the inductive component of the load impedance, Represents the capacitive component of the load impedance; The equivalent parameters include 、 、 、 、 、 .

[0011] Furthermore, the objective function is constructed by the least square method and the parameters are fitted in combination with gradient descent. satisfy: ,in, Indicates frequency point The measured impedance, represents the theoretical impedance obtained by analyzing the model; The gradient descent update satisfies: ,in, Indicates the current parameter value. represents the gradient of the objective function with respect to the parameters, Represents the learning rate.

[0012] Furthermore, the analysis model is a piecewise function model, and the dynamic evaluation of the measurement error is performed through the piecewise function model.

[0013] Furthermore, the fitted parameters are input into the analysis model for training. After obtaining the optimized parameters, the following steps are also included: Obtain the actual measured value of impedance; The error calculation is performed based on the optimized parameters, satisfying the following: Error = (Optimized Parameter - Actual Measured Value) / Actual Measured Value. A second object of the present invention is to provide an electronic device for implementing one of the objects of the invention, comprising a processor, a storage medium, and a computer program. The computer program is stored in the storage medium and, when executed by the processor, implements the above-described current transformer error fitting calculation method.

[0014] Compared with the prior art, the present invention has the following beneficial effects: The present invention realizes full-band impedance measurement through heterodyne signal injection, providing comprehensive basic data for impedance parameter error analysis; reduces single-point measurement error by weighted fusion of impedance data, and improves the robustness of data fusion; accurately separates characteristic parameters through impedance separation, improves fitting accuracy, and combines machine learning to further improve the prediction accuracy of impedance parameters, providing an accurate analysis basis for error analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is a flow chart of the current transformer error fitting calculation method of embodiment 1; Figure 2 is a schematic diagram of an error curve of Example 2; Figure 3 It is a structural block diagram of an electronic device of embodiment 3. DETAILED DESCRIPTION

[0016] The present invention will be described in more detail below with reference to the accompanying drawings. It should be noted that the following description of the present invention with reference to the accompanying drawings is merely illustrative and non-limiting. Various embodiments may be combined with each other to form other embodiments not shown in the following description.

[0017] Example 1 The first embodiment provides a current transformer error fitting calculation method, which aims to obtain high-precision impedance parameter prediction results by performing impedance inversion and parameter fitting on the impedance data of each frequency point, thereby providing a basis for error analysis.

[0018] Existing error analysis methods typically rely on laboratory environments and single-frequency measurements. These methods have the following limitations: (1) Single-frequency data is not sufficient to reveal the full-frequency response characteristics: The traditional method uses power frequency signals for error testing, ignoring the impedance characteristics of the current transformer at different frequencies, making it difficult to achieve a comprehensive analysis of errors across the entire frequency band.

[0019] (2) Difficulty in separating impedance characteristics: The characteristics of the excitation impedance and the load impedance are coupled together, and the existing methods lack effective mathematical tools to analyze them independently, resulting in low accuracy of the error analysis results.

[0020] (3) Insufficient on-site applicability: Laboratory testing methods usually require disconnecting the secondary circuit and reconfiguring the test equipment, which cannot meet the needs of complex on-site operating environments.

[0021] With the increasing demand for metering accuracy in modern power systems, multi-frequency signal injection technology has begun to attract attention in impedance measurement. By injecting multi-frequency sinusoidal current signals into the secondary circuit, the dynamic response of the current transformer at different frequencies can be obtained. When combined with impedance inversion technology, it is possible to accurately separate the excitation impedance and load impedance, laying the foundation for error fitting calculations.

[0022] In summary, please refer to Figure 1As shown, based on the existing technology, the present invention uses non-contact signal injection and impedance inversion technology, combined with the least squares method and gradient descent method, to fit and calculate the impedance parameters of the current transformer and its secondary circuit, separate the excitation impedance and load impedance, and achieve accurate analysis and trend prediction of measurement errors. The present invention provides a current transformer error fitting calculation method, including the following steps: S1. Generate a frequency-varying signal and inject it into the secondary circuit of the current transformer to obtain the impedance at each frequency point after the frequency-varying signal is injected. The frequency difference signal is generated in S1 and injected into the secondary circuit of the current transformer, including: Generate a multi-frequency sinusoidal current signal in the range of 200 Hz to 1500 Hz by a signal generator; The signal is injected into the secondary circuit of the current transformer through a contactless coupler in a manner of gradually increasing frequency.

[0023] By generating multi-frequency sinusoidal current signals in the 200Hz to 1500Hz range and combining them with a contactless coupler, the frequency response characteristics of the circuit can be stimulated, stimulating a full-band impedance response. This method covers a wide frequency range, captures the dynamic characteristics of the current transformer, and provides a data foundation for full-band error analysis, overcoming the limitation of insufficient single-frequency data.

[0024] Specifically, the contactless coupler can be fixed on the secondary loop conductor of the current transformer, and the frequency range of the signal generator can be set, that is, 200Hz to 1500Hz as mentioned above, and the frequency can be gradually increased, such as in steps of 100Hz, and a steady-state sinusoidal current signal can be injected at each frequency point.

[0025] S1 obtains the impedance at each frequency point after the injection of the heterodyne signal, including: Get the voltage at each frequency point detected by the high-sensitivity miniature current transformer and current , Calculating impedance : .

[0026] The above impedance It refers to the equivalent impedance of the current transformer secondary circuit as a whole to the injected signal at a specific frequency. It includes the combined effect of the current transformer's own excitation impedance and the secondary load impedance.

[0027] Excitation impedance: reflects the impedance generated by the nonlinear magnetic properties of the current transformer core (such as hysteresis, saturation, etc.), mainly including inductance and nonlinear factors.

[0028] Load impedance: The impedance composed of the electric energy meter, wiring cables, etc. connected in the secondary circuit, usually manifested as series resistance, inductance and possible capacitance components.

[0029] S2. Cleaning the impedance data, and fusing the cleaned impedance data by weighted fusion to obtain a total impedance; In S2, the impedance data is cleaned and the cleaned impedance data is fused by weighted fusion to obtain the total impedance, including: The Isolation Forest algorithm is used to detect outliers in the impedance data and remove them. Outliers usually include mutation points, noise interference, etc., which can be selected and removed based on the actual data, and bad data can be removed based on abnormal distribution.

[0030] Interpolation is performed on the frequency points after outliers are removed to ensure data continuity. In addition, missing data can also be handled by weighted averaging.

[0031] According to the weight, the cleaned impedance data is fused to meet the following requirements: ,in, is the total impedance, Frequency point The weight of Frequency point The impedance, is the number of frequency points.

[0032] In this embodiment, the weights can be preset or updated based on the signal-to-noise ratio (SNR) of the frequency point signal, so that the weights can be dynamically allocated based on the reliability of the frequency point. If the measurement signal at a certain frequency point has a higher signal-to-noise ratio, it means that the measurement data at this frequency point is more accurate, stable, and reliable, so it should be given a larger weight. On the contrary, if the signal noise is large (such as caused by external interference, equipment failure, etc.), the data reliability of this frequency point is low, and the weight is smaller. In this embodiment, the frequency point The weight is the preset weight for the frequency point. The weight is updated based on the signal-to-noise ratio: when the measured signal at the frequency point has a higher signal-to-noise ratio, the weight is increased, and vice versa. It should be noted that the aforementioned "higher" can be determined by setting a preset value. A signal-to-noise ratio higher than the preset value is considered to be higher. The preset value can be set based on actual conditions and is not limited in this embodiment.

[0033] S3. Perform impedance inversion on the total impedance to obtain excitation impedance and load impedance, and extract equivalent parameters; S3 specifically includes: constructing an impedance separation model to meet the following requirements: ,in, is the excitation impedance, is the load impedance; The excitation impedance satisfies: ,in, is the resistance component of the excitation impedance, is the inductive component of the excitation impedance, represents the imaginary part, that is, the imaginary part of the inductance, is the angular frequency of the signal, indicating the impedance change related to frequency, satisfying , is the frequency of the signal, is the capacitive component of the excitation impedance; The load impedance satisfies: ,in, The resistive component of the load impedance, is the inductive component of the load impedance, Represents the capacitive component of the load impedance; The equivalent parameters include 、 、 、 、 、 .

[0034] S4. Construct the objective function by the least squares method and perform parameter fitting in combination with gradient descent; S4 will perform fitting optimization on the equivalent parameters in S3. By adjusting these parameters, the theoretically calculated total impedance is made as close as possible to the measured value, thereby achieving the purpose of error fitting.

[0035] S4 constructs the objective function by the least squares method and combines it with gradient descent to perform parameter fitting. The objective function is the quantity to be minimized during the fitting process, usually expressed as the error between the actual measured data and the model predicted data. In impedance inversion and fitting, the objective function is the sum of squared errors between the total impedance and the measured data. The objective function is achieved by continuously adjusting the parameters. , minimize the error, and thus obtain the best fitting impedance model. The objective function satisfy: ,in, Indicates frequency point The measured impedance, represents the theoretical impedance obtained by analyzing the model; Gradient descent is used for optimization problems by calculating the gradient of the objective function with respect to the parameters and updating the parameters along the gradient direction to minimize the objective function. The gradient descent update satisfies: ,in, Indicates the current parameter value. represents the gradient of the objective function with respect to the parameters, Represents the learning rate, which is used to control the size of the update step and can be set according to actual needs.

[0036] S5. Input the fitted parameters into the analysis model to perform model optimization and obtain the optimized parameters.

[0037] Since the impedance of the high-frequency part is measured in this embodiment, the impedance of the low-frequency part needs to be derived, so fitting is required to fit from high frequency to power frequency. The fitted parameters have deviations, so they are optimized in combination with the analytical model.

[0038] This paper presents a transformer impedance modeling and error estimation method based on the Levenberg–Marquardt (nonlinear least squares iteration) nonlinear optimization algorithm. This method constructs an impedance model using a hyperbolic tangent function, fits multiple sets of frequency and impedance measurement data, and obtains an estimated impedance value at a given frequency. The method then dynamically evaluates the measurement error using a piecewise function model, combining the current magnitude and the CT ratio.

[0039] In this embodiment, the analytical model is a piecewise function model, a rule-based empirical model or a quasi-analytical error estimation model. Numerical optimization methods are used to fit the explicit function model to address the complex nonlinear coupling relationships between parameters. By constructing a residual objective function and iteratively optimizing it using the LM algorithm, efficient convergence and high-precision fitting are achieved while ensuring model interpretability.

[0040] Among them, the fitted parameters are input into the analysis model for model optimization, and after the optimized parameters are obtained, the following steps are also included: Obtain the actual measured value of impedance; The error calculation is performed based on the optimized parameters, satisfying the following: Error = (Optimized parameters - actual measured value) / actual measured value.

[0041] The above-mentioned error analysis includes predicting the future error change trend and life of the transformer. Based on the error calculated after fitting and optimization, a long-term trend analysis is performed to predict future error changes.

[0042] It should be noted that the meaning of error analysis is: based on the fitted impedance model parameters (such as R, L, C of the excitation impedance and load impedance), the difference between the ideal and actual output is simulated at each frequency point, thereby quantitatively evaluating the error of the transformer at that frequency.

[0043] The measurement error is calculated by inverting the actual output current based on the voltage, current and impedance model parameters, and comparing it with the ideal output current to quantitatively obtain the ratio difference and phase difference, which are the core indicators of transformer error evaluation.

[0044] Example 2 Example 2 is an illustration of the actual test results of the method of Example 1.

[0045] The frequency range of the test is 200Hz to 1500Hz. For the specific method, please refer to the description of Example 1. After collecting the impedance data of the frequency points, the data is cleaned and high-quality data of 13 frequency points is retained.

[0046] After fitting and analysis model optimization, the extracted parameters are as follows: Excitation impedance: , ,

[0047] Load impedance: , ,

[0048] According to the parameter extraction results, draw Figure 2 From the error curve shown (300 / 5-1, 300 / 5-2 and 300 / 5-3 overlap in the figure), we can see that: The maximum error occurs at 400 Hz, which is 0.85%. The error trend shows a nonlinear decrease in the high frequency band.

[0049] Optimization results show that the optimized analytical model improves fitting accuracy by 20%-30% compared to traditional methods. After removing abnormal data, the fitting error is reduced by 15%.

[0050] Example 3 Figure 3 This is a structural diagram of an electronic device provided in the third embodiment of the present invention, such as Figure 3 As shown, the electronic device includes a processor 210, a memory 220, an input device 230, and an output device 240; the number of processors 210 in the computer device can be one or more. Figure 2 In the figure, a processor 210 is used as an example; the processor 210, memory 220, input device 230 and output device 240 in the electronic device can be connected via a bus or other means. Figure 3 The bus connection is taken as an example.

[0051] Memory 220, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules. Processor 210 executes the software programs, instructions, and modules stored in memory 220 to perform various functional applications and data processing of the electronic device, thereby implementing the current transformer error fitting calculation method of the first embodiment.

[0052] The memory 220 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created based on the use of the terminal, etc. In addition, the memory 220 may include a high-speed random access memory and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 220 may further include a memory remotely located relative to the processor 410, and these remote memories may be connected to the electronic device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0053] The input device 230 may be used to receive input user identity information, impedance data, etc. The output device 240 may include a display device such as a display screen.

[0054] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented with the help of software and necessary general-purpose hardware. Of course, it can also be implemented with hardware, but in many cases the former is a more preferred embodiment. Based on this understanding, the technical solution of the present invention, or the part that contributes to the existing technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., and includes a number of instructions for enabling an electronic device (such as a mobile phone, personal computer, server, or network device) to execute the methods described in various embodiments of the present invention.

[0055] Those skilled in the art can make various other corresponding changes and deformations based on the technical solutions and concepts described above, and all of these changes and deformations should fall within the scope of protection of the claims of the present invention.

Claims

1. A current transformer error fitting calculation method, characterized in that: The following steps are involved: Generate an inter-frequency signal and inject it into the secondary circuit of the current transformer, and obtain the impedance at each frequency point after the inter-frequency signal is injected; Cleaning the impedance data, fusing the cleaned impedance data by weighted fusion to obtain a total impedance; Performing impedance inversion on the total impedance to obtain excitation impedance and load impedance, and extracting equivalent parameters; The objective function is constructed by the least squares method and the parameters are fitted in combination with gradient descent; The fitted parameters are input into the analysis model for model optimization, and the optimized parameters are obtained.

2. The current transformer error fitting calculation method according to claim 1, characterized in that: Generate an out-of-frequency signal and inject it into the secondary circuit of the current transformer, including: Generate a multi-frequency sinusoidal current signal in the range of 200 Hz to 1500 Hz by a signal generator; The signal is injected into the secondary circuit of the current transformer through a contactless coupler in a manner of gradually increasing frequency.

3. The current transformer error fitting calculation method according to claim 1 or 2, characterized in that: Obtain the impedance at each frequency point after injecting the frequency-independent signal, including: Get the voltage at each frequency point detected by the high-sensitivity miniature current transformer and current , Calculating impedance : .

4. The current transformer error fitting calculation method according to claim 1, characterized in that: The impedance data is cleaned, and the cleaned impedance data is fused by weighted fusion to obtain the total impedance, including: The outliers in the impedance data are detected and eliminated using the Isolation Forest algorithm; Perform interpolation processing on the frequency points where outliers are removed; According to the weight, the cleaned impedance data is fused to meet the following requirements: ,in, is the total impedance, Frequency point The weight of Frequency point The impedance, is the number of frequency points.

5. The current transformer error fitting calculation method according to claim 4, characterized in that: The frequency point The weight is the preset weight of the frequency point, and the weight is updated according to the signal-to-noise ratio: when the signal-to-noise ratio of the measurement signal at the frequency point is higher than the preset value, the weight is increased, otherwise the weight is reduced.

6. The current transformer error fitting calculation method according to claim 1, characterized in that: Impedance inversion is performed on the total impedance to obtain the excitation impedance and the load impedance, including: Construct an impedance separation model to meet the following requirements: ,in, is the excitation impedance, is the load impedance; The excitation impedance satisfies: ,in, is the resistance component of the excitation impedance, is the inductive component of the excitation impedance, represents the imaginary part, is the angular frequency of the signal, is the capacitive component of the excitation impedance; The load impedance satisfies: ,in, The resistive component of the load impedance, is the inductive component of the load impedance, Represents the capacitive component of the load impedance; The equivalent parameters include 、 、 、 、 、 .

7. The current transformer error fitting calculation method according to claim 1, characterized in that: The objective function is constructed by the least square method and the parameters are fitted in combination with gradient descent. satisfy: ,in, Indicates frequency point The measured impedance, represents the theoretical impedance obtained by analyzing the model; The gradient descent update satisfies: ,in, Indicates the current parameter value. represents the gradient of the objective function with respect to the parameters, Represents the learning rate.

8. The current transformer error fitting calculation method according to claim 1 or 7, characterized in that: The analysis model is a piecewise function model, and the dynamic evaluation of the measurement error is performed through the piecewise function model.

9. The current transformer error fitting calculation method according to claim 1, characterized in that: After inputting the fitted parameters into the analysis model for model optimization and obtaining the optimized parameters, it also includes: Obtain the actual measured value of impedance; The error calculation is performed based on the optimized parameters, satisfying the following: Error = (Optimized parameters - actual measured value) / actual measured value.

10. An electronic device comprising a processor, a storage medium, and a computer program, wherein the computer program is stored in the storage medium, When the computer program is executed by a processor, the current transformer error fitting calculation method according to any one of claims 1 to 9 is implemented.

Images