Current transformer measurement interference suppression method under high-voltage strong electromagnetic interference

Abstract

The invention relates to the technical field of current transformer measurement, in particular to a current transformer measurement interference suppression method under high-voltage strong electromagnetic interference, and the method comprises the steps: collecting related signals used for analyzing the influence of interference on the measurement of a current transformer; aiming at high-frequency interference generated by impulse noise and instantaneous surge, analyzing the composite degree of the high-frequency interference caused by the two factors, and analyzing the coupling characteristic of the high-frequency interference so as to quantify the measurement interference intensity characteristic of the high-frequency noise suffered by the current transformer at the historical moment; and step length parameters in an LMS adaptive filtering algorithm are improved by using measured interference intensity characteristics, and a measured current signal after interference suppression is obtained. The invention aims to accurately quantify the pulse-surge space-time correlation characteristics, so that the subsequent LMS algorithm establishes the dynamic mapping relation of the interference intensity, and the risk that the measurement of the current transformer is interfered is reduced.

Description

Technical Field

[0001] The present application relates to the technical field of current transformer measurement, and in particular to a method for suppressing interference in current transformer measurement under high voltage and strong electromagnetic interference. Background Art

[0002] In high-voltage and strong electromagnetic interference environments, LMS adaptive filtering technology is usually used to suppress interference in current transformer measurement. The core is to dynamically adjust the filter weights through the gradient descent algorithm to minimize the output error. However, the unique characteristics of high-frequency interference such as pulse noise and transient surges seriously restrict the effectiveness of traditional algorithms. Pulse noise manifests as high-amplitude spikes in the microsecond range. Its suddenness causes the gradient estimate to deviate from the true error surface, causing the weight update direction to be disordered, resulting in signal overshoot or undercompensation; transient surges have wide spectrum characteristics and time-varying energy distribution.

[0003] The most important thing is that the spatiotemporal coupling effect of the two types of interference forms a composite interference pattern. Its non-stationary characteristics make the fixed step size parameter face the contradiction between convergence speed and steady-state accuracy. A small step size is difficult to track the fast rising edge of surge interference, and although a large step size can improve the transient response speed, it will lead to weight oscillation caused by pulse noise.

[0004] Since the existing technology fails to establish a dynamic mapping relationship of interference intensity, it is impossible to accurately quantify the temporal and spatial correlation characteristics of pulses and surges, which easily leads to interference in the measurement of current transformers. Summary of the invention

[0005] In order to solve the above technical problems, the present application provides a method for suppressing interference in current transformer measurement under high voltage and strong electromagnetic interference to solve the existing problems.

[0006] The present invention discloses a method for suppressing interference in current transformer measurement under high voltage and strong electromagnetic interference, which adopts the following technical solution: An embodiment of the present application provides a method for suppressing interference in current transformer measurement under high voltage and strong electromagnetic interference, the method comprising the following steps: Collect the current signal of the current transformer, as well as the voltage signal of the circuit in which it is located, the magnetic field signal and ground potential signal around it in its working state; Reconstruct various signals at high frequencies and extract high-frequency interference signals; detect different types of abnormal signals in high-frequency interference signals; The signal amplitude accumulation value of different types of abnormal signals adjacent to each other in the time domain within the corresponding time domain range is forward fused with the number ratio of signal pairs composed of different types of abnormal signals adjacent to each other in the time domain, and the fusion result is used as the composite interference feature of various signals; Calculate the covariance between the calculated current and voltage signals and the magnetic field and ground potential signals respectively; divide all types of signals into two categories, and perform regression analysis with one category as the independent variable and the other category as the dependent variable to obtain two regression coefficient vectors; positively fuse the composite interference characteristic mean of all types of signals, the sum of all calculated covariances, and the differences between the two regression coefficient vectors, and use the fusion result as the measured interference intensity characteristic. According to the measured interference intensity characteristic calculated in real time, improve the step size parameter in the LMS adaptive filtering algorithm for the current signal to obtain the current data measured by the current transformer after interference suppression.

[0007] Preferably, the detection method for different types of abnormal signals in the high-frequency interference signal is determined as: Use a preset time-domain threshold to detect pulse noise signals in the high-frequency interference signal; Use a preset envelope threshold to detect instantaneous surge signals in the analytic signal of the high-frequency interference signal.

[0008] Preferably, the use of the preset time-domain threshold to detect pulse noise signals in the high-frequency interference signal includes: Calculate the amplitude mean μ and standard deviation σ of the high-frequency interference signal; Set the time-domain threshold to μ + kσ, where k is a constant greater than zero; Detect the signal points in the high-frequency interference signal that exceed the time-domain threshold; Mark the detected signal points as pulse noise signals.

[0009] Preferably, the use of the preset envelope threshold to detect instantaneous surge signals in the analytic signal of the high-frequency interference signal includes: Perform Hilbert transform on the high-frequency interference signal to obtain the analytic signal and calculate its envelope; Set the envelope threshold to the P% number of the envelope, where P is a positive number less than 100; Detect the signals in the envelope that exceed the envelope threshold; Mark the detected signals as instantaneous surge signals.

[0010] Preferably, the method for obtaining different types of abnormal signals adjacent in time domain is determined as: When the time length between the corresponding moments of different types of abnormal signals is less than a preset time window, these two different types of abnormal signals are used as different types of abnormal signals adjacent in time domain.

[0011] Preferably, record the composite interference characteristic of the current signal as A, , where a is the number of signal pairs in the high-frequency interference signal of the current signal, and N is the number of signal points of the collected current signal. is the total number of pairwise combinations of these signal points, represents the signals within the corresponding time domain range between the two abnormal signals in signal pair i, represents the cumulative value of the signal amplitudes within the corresponding time domain range between the two abnormal signals in signal pair i.

[0012] Preferably, the measurement interference intensity feature is further determined by the product result of the mean value of the composite interference features of all types of signals, the calculated sum value of all covariances, and the difference between the two regression coefficient vectors.

[0013] Preferably, the improved formula for the step size parameter is: , where is the improved step size, is the initial step size, is the currently calculated measurement interference intensity feature, is the maximum value of the historically calculated measurement interference intensity features.

[0014] Preferably, the method for dividing all types of signals into two categories is: divide them according to internal influence signals and external influence signals, classify current and voltage signals as internal influence signals, and classify magnetic field and ground potential signals as external influence signals.

[0015] Preferably, the independent variable or the dependent variable during regression analysis is a matrix constructed from all types of signals in the corresponding category.

[0016] This application has at least the following beneficial effects: This application proposes a method for suppressing measurement interference of current transformers under high-intensity electromagnetic interference. Aiming at the problem of high-frequency interference distortion caused by pulse noise and instantaneous surges, it reconstructs the high-frequency components of various data, analyzes the spatio-temporal coupling intensity of high-frequency interference signals, and solves the measurement deviation problem caused by insufficient quantification of transient composite interference by traditional methods; aiming at the asymmetric interference caused by the coupling of internal and external electromagnetic interference, it analyzes the non-linear correlation characteristics between internal circuit parameters and the external electromagnetic environment through covariance and regression analysis, and eliminates the limitations of single interference source analysis; based on the dynamically calculated measurement interference intensity feature, it dynamically adjusts the step size parameter of the LMS adaptive filtering algorithm, realizes the weight update mechanism adaptive to the interference intensity, enables the improved LMS algorithm to establish a dynamic mapping relationship of the interference intensity, accurately quantifies the spatio-temporal correlation characteristics of pulse-surges, and reduces the risk of measurement interference of current transformers. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] To more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0018] Figure 1 It is a flowchart of a method for suppressing measurement interference of a current transformer under high - pressure electromagnetic interference provided by the present application; Figure 2 It is a flowchart of the construction process of the measurement interference intensity characteristics of high - frequency noise received by the current transformer provided by the present application; Figure 3 It is a schematic diagram of the signal convergence process of traditional LMS and improved LMS with step size in an embodiment of the present application. Detailed implementation manners

[0019] To further elaborate on the technical means and effects adopted by the present application to achieve the intended invention purpose, the following, in combination with the drawings and preferred embodiments, details the specific implementation manners, structures, features, and effects of a method for suppressing measurement interference of a current transformer under high - pressure electromagnetic interference proposed according to the present application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present application belongs.

[0021] The following specifically describes the specific solution of a method for suppressing measurement interference of a current transformer under high - pressure electromagnetic interference provided by the present application with reference to the drawings.

[0022] A method for suppressing measurement interference of a current transformer under high - pressure electromagnetic interference provided by an embodiment of the present application.

[0023] Specifically, a method for suppressing measurement interference of a current transformer under high - pressure electromagnetic interference is provided as follows. Please refer to Figure 1 , and the method includes the following steps: Step 1: Collect relevant signals for analyzing the influence of interference on the measurement of the current transformer.

[0024] To achieve the accuracy of current transformer measurement under high - pressure electromagnetic interference, this application collects the current signal of the current transformer, as well as the voltage signal of the circuit where it is located, the magnetic field signal around it during its operation, and the ground potential signal. In this embodiment, a data collector is used to collect the current measured by the current transformer, and a voltage transformer with a data collector is installed on the circuit where the current transformer is located to obtain voltage data; a three - axis magnetic field sensor is installed around the current transformer to collect the magnetic field intensity signal; a ground potential sensor is installed near the grounding point to collect the ground potential fluctuation signal.

[0025] In this embodiment, the sampling frequencies of all sensors are uniformly set to 100 kHz, and timestamp alignment is achieved through a multi - channel synchronous signal acquisition system to ensure the strict synchronous acquisition of current, voltage, magnetic field, and ground potential signals, and the collected signals are subjected to overall normalization processing to provide support for high - resolution signals for quantifying the non - stationary characteristics of high - pressure electromagnetic interference.

[0026] Sort the normalized signals of each type in ascending order of time to obtain the current signal, voltage signal, magnetic field signal, and ground potential signal. In this embodiment, the current transformer is monitored for a long time to determine its anti - interference state. Taking the power - on of the circuit as the initial acquisition moment of the signal, this embodiment collects the current, voltage, magnetic field, and ground potential signals within 1 hour, and assumes that there are N elements in each type of signal collected.

[0027] Step 2: For the high - frequency interference generated by pulse noise and instantaneous surges, analyze the composite degree of the high - frequency interference caused by the two factors and analyze its coupling characteristics to quantify the measurement interference intensity characteristics of the high - frequency noise suffered by the current transformer at historical moments.

[0028] In this application, the flowchart of the construction process of the measurement interference intensity characteristics of the high - frequency noise suffered by the current transformer is as shown in the appendix Figure 2 as follows: S1, perform high - frequency reconstruction on various signals respectively to extract high - frequency interference signals.

[0029] In a high - pressure electromagnetic interference environment, high - frequency noise in the current signal measured by the current transformer, such as pulse noise and instantaneous surges, is the main source of interference leading to measurement errors and misoperation of protection devices. The high - frequency noise will be superimposed on the current signal, causing the measured value to deviate from the true value, and will mask the low - frequency components of the useful signal, resulting in signal distortion.

[0030] To extract the high - frequency interference components in the relevant signals, perform high - frequency reconstruction on various signals respectively to extract high - frequency interference signals.

[0031] In this embodiment, four signals, namely current signal, voltage signal, magnetic field signal, and ground potential signal, are used as inputs. Taking the current signal as an example, a Coiflet wavelet basis function suitable for extracting high-frequency components is selected, the decomposition level is set to 4, and the discrete wavelet transform (DWT) is used to perform multi-scale decomposition on the current signal respectively. Four high-frequency detail coefficients and one low-frequency approximation coefficient of the current signal are output. The detail coefficients capture the high-frequency components of the signal at different scales, including high-frequency interference such as pulse noise and instantaneous surges. By reconstructing the detail coefficients, the high-frequency components in the signal can be restored. Taking the four high-frequency detail coefficients decomposed by DWT as inputs, the inverse discrete wavelet transform (IDWT) is used to reconstruct the input detail coefficients to obtain the high-frequency interference signal of the current signal.

[0032] Among them, methods suitable for extracting high-frequency components also include discrete Fourier transform, etc., which are specifically set by the implementer. In addition, the discrete wavelet transform of the signal and the inverse discrete wavelet transform for reconstructing the decomposed signal are both well-known technologies and will not be elaborated here. The setting of the decomposition level can be set by the implementer himself.

[0033] S2. Detect different abnormal signals in the high-frequency interference signal.

[0034] As a preferred implementation manner, the detection method for different abnormal signals in the high-frequency interference signal is determined as follows: using a preset time-domain threshold to detect pulse noise signals in the high-frequency interference signal; using a preset envelope threshold to detect instantaneous surge signals in the analytic signal of the high-frequency interference signal.

[0035] In other implementation manners, the detection of different abnormal signals in the high-frequency interference signal can be performed through frequency-domain analysis, 3σ principle, etc. Or directly identify different abnormal signals manually.

[0036] In the high-frequency interference signal, pulse noise usually appears as a high-amplitude spike that suddenly appears in the signal and has an extremely short duration. Therefore, in this embodiment, using a preset time-domain threshold to detect pulse noise signals in the high-frequency interference signal includes: calculating the amplitude mean of the high-frequency interference signal and the standard deviation , setting the time-domain threshold to , where k is a constant greater than zero; detecting the signal points in the high-frequency interference signal that exceed the time-domain threshold, marking each signal point that exceeds the time-domain threshold, and recording all the marked points as pulse noise signals respectively. At the same time, record the occurrence times of all pulse noise signals.

[0037] An instantaneous surge usually manifests as a short-term high-energy fluctuation in a signal, which has a longer duration compared to impulse noise. Accordingly, in this embodiment, using a preset envelope threshold, the instantaneous surge signal in the analytic signal of the high-frequency interference signal is detected, including: performing a Hilbert transform on the high-frequency interference signal, outputting the analytic signal, calculating the envelope of the analytic signal, setting the envelope threshold. In this embodiment, the envelope threshold is set to the P% of the envelope, where P is a positive number less than 100, and in this embodiment, the value of P is 95 to more flexibly control the sensitivity of instantaneous surge detection, detecting the signal in the envelope that exceeds the envelope threshold, marking it as an instantaneous surge signal, and recording the occurrence time of all instantaneous surge signals. Among them, the Hilbert transform is a well-known technology and will not be elaborated further.

[0038] S3, positively fuse the cumulative value of the signal amplitudes within the corresponding time domain range between different types of abnormal signals that are adjacent in the time domain and the proportion of the number of signal pairs composed of different types of abnormal signals that are adjacent in the time domain, and use the fusion result as the composite interference feature of various signals.

[0039] As a preferred implementation manner, the method for obtaining different types of abnormal signals that are adjacent in the time domain is determined as: when the time length between the corresponding times of different types of abnormal signals is less than a preset time window, these two different types of abnormal signals are regarded as different types of abnormal signals that are adjacent in the time domain.

[0040] Specifically, in this embodiment, the time window is used to determine whether the impulse noise signal and the instantaneous surge signal occur within the same time period, that is, whether they are adjacent in the neighborhood. In this embodiment, the size of the time window is 10 ms. For each element in the impulse noise signal, calculate the time interval between its occurrence time and the occurrence times of all elements in the instantaneous surge signal. If the time interval is less than the size of the time window, it is considered that the two are related in time, that is, the two types of abnormal signals are adjacent in the time domain, and these two signals are recorded as a signal pair. Traverse all the occurrence times of the impulse noise to obtain the signal pairs that meet the above time-domain adjacency condition. Suppose a total of a signal pairs are obtained.

[0041] For each signal pair, taking signal pair i as an example, assume that the signal within the corresponding time domain range between the two abnormal signals in signal pair i is . For example, the occurrence time of the impulse noise in signal pair i is , and the occurrence time of the instantaneous surge is . If , then is the high-frequency interference signal of the current signal between , and vice versa.

[0042] Based on the above analysis, in this application, the sum of signal amplitudes within the corresponding time domain range between different types of abnormal signals adjacent in the time domain is positively fused with the proportion of the number of signal pairs composed of different types of abnormal signals adjacent in the time domain, and the fusion result is used as the composite interference feature of various signals.

[0043] It can be understood that positive fusion refers to fusion methods such as addition and multiplication between data. The specific positive fusion method is determined by the implementer according to the actual situation to select a suitable fusion method, and this application does not impose special restrictions.

[0044] Specifically, in this embodiment, the composite interference feature of the current signal is calculated , and the specific calculation formula is: , where a is the number of signal pairs in the high-frequency interference signal of the current signal, N is the number of signal points of the collected current signal, is the total number of pairwise combinations of these signal points, represents the signal within the corresponding time domain range between the two abnormal signals in signal pair i, represents the sum of signal amplitudes within the corresponding time domain range between the two abnormal signals in signal pair i.

[0045] It should be understood that represents the proportion of the number of signal pairs to all possible combinations, reflecting the correlation intensity of impulse noise and instantaneous surge signals in time. When it is larger, it indicates that the proportion of signal pairs is relatively high. The larger the value, the stronger the composite effect of interference; is used to reflect the interference intensity of a single pulse-surge coupling event. is the secondary summation of the sum of signal amplitudes of all a signal pairs, representing the total amplitude intensity of composite interference. The larger its value, the higher the intensity of composite interference. The larger the value.

[0046] is used to quantify the composite effect of impulse noise and instantaneous surge signals in the high-voltage electromagnetic interference environment of the current signal, describing the intensity of simultaneous occurrence of impulse noise and instantaneous surge signals in the high-voltage electromagnetic interference environment and their impact on the system. The larger the A value, the stronger the composite effect of interference, which may lead to an increase in measurement error or an increase in the risk of misoperation of the protection device.

[0047] Subsequently, using the same method, the composite interference features of voltage, magnetic field, and ground potential signals are calculated, and the mean value of the composite interference features of the four types of data is calculated .

[0048] S4. Calculate the covariance between the current and voltage signals and the magnetic field and ground potential signals respectively; divide all types of signals into two categories, and perform regression analysis with one category as the independent variable and the other as the dependent variable to obtain two regression coefficient vectors; positively fuse the mean of the composite interference characteristics of all types of signals, the sum of all calculated covariances, and the differences between the two regression coefficient vectors, and use the fusion result as the measured interference intensity characteristic.

[0049] During the measurement process of current transformers, the sources of their high-frequency noise are mainly divided into two types: external electromagnetic interference and internal circuit interference. The types of external electromagnetic interference include but are not limited to high-voltage equipment operation, lightning surges, interference from electromagnetic components, and poor grounding; the types of internal circuit interference include but are not limited to parasitic parameters and power supply fluctuations.

[0050] Therefore, the four types of collected data are divided into two categories. Among them, the current and voltage signals are classified as internal influence signals, and the ground potential and magnetic field signals are classified as external influence signals; current and voltage are direct electrical quantities in the power system, reflecting the electrical state of the system, and are mainly affected by internal circuit interference. The ground potential and magnetic field are indirect physical quantities in the power system, reflecting the influence of the external environment on the system, and are mainly affected by external electromagnetic interference.

[0051] At different times, the interference caused by external electromagnetic interference and internal circuit interference to current transformers is different, that is, external electromagnetic interference and internal circuit interference are not constant. In order to analyze the respective effects of the two on current transformers, in this application, the covariance between the current and voltage signals and the magnetic field and ground potential signals is calculated respectively.

[0052] During the calculation process of covariance, there are 4 variables: current, voltage, magnetic field, and ground potential. Among them, the covariances calculated for the four groups of current and magnetic field, current and ground potential, voltage and magnetic field, and voltage and ground potential are all the covariances between internal and external interferences, reflecting the coupling relationship between internal and external interferences. Thus, the coupling strength between internal and external interferences is obtained, and its value is the sum of all calculated covariances.

[0053] Furthermore, in this application, all types of signals are divided into two categories, and regression analysis is performed with one category as the independent variable and the other as the dependent variable to obtain two regression coefficient vectors. Among them, the independent variable or dependent variable during the regression analysis is a matrix constructed by all types of signals in the corresponding category.

[0054] In this embodiment, let the current signal be I, the voltage signal be V, the magnetic field signal be G, and the ground potential signal be M. The current, voltage, ground potential, and magnetic field signals are respectively represented to form matrices X and Y, where , Subsequently, using the matrix X as the independent variable and the matrix Y as the dependent variable, the partial least squares regression algorithm is used to output the regression coefficient vector , similarly, using the matrix X as the dependent variable and the matrix Y as the independent variable, the partial least squares regression algorithm is used to output the regression coefficient vector . Among them, the partial least squares regression algorithm is a well-known technology and will not be elaborated here. The methods of regression analysis can also include multiple linear regression, principal component regression, etc.

[0055] Based on the above analysis, the present application positively fuses the mean value of the composite interference characteristics of all types of signals, the calculated sum value of all covariances, and the difference between the two regression coefficient vectors, and uses the fusion result as the measured interference intensity characteristic.

[0056] As a preferred embodiment, the measured interference intensity characteristic is further determined by the product result of the mean value of the composite interference characteristics of all types of signals, the calculated sum value of all covariances, and the difference between the two regression coefficient vectors.

[0057] In this embodiment, the specific calculation formula of the measured interference intensity characteristic B is: , where is the mean value of the composite interference characteristics of all types of signals, which reflects the composite interference intensity of impulse noise and instantaneous surge signals in the system. The larger its value, the stronger the composite interference in the system, resulting in an increase in measurement error or an increase in the risk of misoperation of the protection device, and B increases accordingly; is the calculated sum value of all covariances, which describes the interaction intensity between internal interference and external interference. The larger its value, the stronger the coupling effect between internal and external interference, the higher the complexity of the interference, and the increase in the value of B; is the difference between the two regression coefficient vectors, which is the Euclidean distance between the two regression coefficient vectors in this embodiment. It describes the difference in the influence of internal interference and external interference on the measurement result of the current transformer. The larger its value, the greater the difference in the influence of internal and external interference on the measurement result, the stronger the asymmetry of the interference, and the more complex the influence on the measurement of the current transformer, and the value of B increases accordingly.

[0058] B reflects the overall intensity of high-frequency interference in the system, the coupling effect between internal and external interference, and the asymmetry of interference. Specifically, the larger the value of B, the stronger the composite interference received by the system, the more significant the interference coupling effect, and the higher the asymmetry of the interference, which will lead to an increase in the measurement error of the current transformer and an increase in the risk of misoperation of the protection device.

[0059] Step three: Use the measured interference intensity characteristic to improve the step size parameter in the LMS adaptive filtering algorithm to obtain the measured current signal after interference suppression.

[0060] Traditional methods use the LMS adaptive filtering algorithm to suppress the measurement interference of current transformers. By dynamically adjusting the filter weights in real time, it can accurately track and cancel non-stationary electromagnetic noise signals, effectively separate the fundamental wave from high-frequency noise, improve the signal-to-noise ratio by more than 20 dB, significantly improve the measurement accuracy, and have environmental self-adaptability compared with traditional fixed-parameter filtering. The convergence speed can still be controlled within 5 ms under sudden interference.

[0061] However, the weight update mechanism of LMS based on gradient descent has a lag in response to non-stationary and high-dynamic-range transient interference. Its fixed step-size parameter is difficult to balance the convergence speed and stability. A too-small step size leads to a tracking delay for impulse noise and cannot cancel high-amplitude spikes in time; a too-large step size causes weight oscillation or even divergence due to the sparsity and strong amplitude characteristics of impulse noise, making it difficult to meet the requirements of high-precision measurement.

[0062] From the above analysis, it can be seen that the fixed step-size parameter of the traditional LMS algorithm When facing non-stationary high-frequency interference, it is prone to the contradiction between the convergence speed and the steady-state error: a too-large step size will cause oscillation, while a too-small step size cannot quickly track the change of interference. Therefore, in this application, according to the measured interference intensity characteristics calculated in real time, the step-size parameter in the LMS adaptive filtering algorithm for the current signal is improved to obtain the current data measured by the current transformer after interference suppression.

[0063] In this embodiment, the step size is dynamically adjusted using the measured interference intensity characteristic B calculated in real time to make the step size match the current interference intensity and asymmetry. The specific improvement formula is: where, is the improved step size, is the initial step size, which takes a value of 0.01 in this embodiment, is the measured interference intensity characteristic calculated currently, is the maximum value of the measured interference intensity characteristics calculated historically.

[0064] In this embodiment, the schematic diagrams of the signal convergence processes of traditional LMS and LMS with improved step size are as Figure 3 shown, Figure 3 The first figure in shows the signal amplitude filtering effect of the current signal measured by the current transformer with traditional LMS under a fixed step size The second figure shows the signal amplitude filtering effect of the current signal measured by the current transformer with LMS having an improved step size under an adaptive step size

[0065] By dynamically adjusting the step size parameter in the LMS through a step size adaptive mechanism based on the measured interference intensity characteristics, the non-stationary characteristics of the interference can be accurately sensed. When impulse noise or instantaneous surges occur, the B value rises in real time to trigger the step size to be scaled proportionally, significantly improving the tracking speed of the algorithm for sudden high-frequency interference and matching the weight update rate with the dynamic characteristics of the interference; through the historical calibrated non-linear constraint mechanism, the algorithm can still maintain the weight stability during the strong interference coupling stage (such as the internal and external interference resonance caused by switch operation), avoiding the divergence risk caused by the fixed step size of traditional algorithms due to parameter curing. The improved LMS algorithm establishes a dynamic mapping relationship of the interference intensity, accurately quantifies the spatio-temporal correlation characteristics of impulse-surges, and reduces the risk of interference in the measurement of current transformers.

[0066] The above technical features constitute the best embodiment of this application, which has strong adaptability and the best implementation effect. Non-essential technical features can be added or subtracted according to actual needs to meet the requirements of different situations.

Claims

1. A method for suppressing interference in current transformer measurement under high voltage and strong electromagnetic interference, characterized in that: The method comprises the following steps: Collect the current signal of the current transformer, as well as the voltage signal of the circuit in which it is located, the magnetic field signal around it in its working state and the ground potential signal; Reconstruct various signals at high frequencies and extract high-frequency interference signals; detect different types of abnormal signals in high-frequency interference signals; The signal amplitude accumulation value of different types of abnormal signals adjacent to each other in the time domain within the corresponding time domain range is forward fused with the number ratio of signal pairs composed of different types of abnormal signals adjacent to each other in the time domain, and the fusion result is used as the composite interference feature of various signals; Calculate the covariance between the current and voltage signals and the magnetic field and ground potential signals respectively; divide all kinds of signals into two categories, use one category as the independent variable and the other category as the dependent variable for regression analysis, and obtain two regression coefficient vectors; forward fuse the composite interference feature mean of all kinds of signals, all calculated covariance sum values, and the difference between the two regression coefficient vectors, and use the fusion result as the measurement interference intensity feature; According to the measurement interference intensity characteristics calculated in real time, the step size parameter in the LMS adaptive filtering algorithm is used to improve the current signal, and the current data measured by the current transformer after interference suppression is obtained.

2. The method for suppressing interference in current transformer measurement under high voltage and strong electromagnetic interference as claimed in claim 1, characterized in that: The detection methods for different types of abnormal signals in high-frequency interference signals are determined as follows: Detecting impulse noise signals in high-frequency interference signals using a preset time domain threshold; The preset envelope threshold is used to detect the instantaneous surge signal in the analytical signal of the high-frequency interference signal.

3. The method for suppressing interference in current transformer measurement under high voltage and strong electromagnetic interference as claimed in claim 2, characterized in that: The method of detecting the impulse noise signal in the high-frequency interference signal by using a preset time domain threshold comprises: Calculate the amplitude mean μ and standard deviation σ of the high-frequency interference signal; Set the time domain threshold to μ+kσ, where k is a constant greater than zero; Detecting signal points exceeding the time domain threshold in the high frequency interference signal; The detected signal points are marked as impulse noise signals.

4. The method for suppressing interference in current transformer measurement under high voltage and strong electromagnetic interference as claimed in claim 2, characterized in that: The method of using a preset envelope threshold to detect an instantaneous surge signal in an analytical signal of a high-frequency interference signal includes: Perform Hilbert transform on the high-frequency interference signal to obtain the analytical signal and calculate its envelope; Set the envelope threshold to P% of the envelope, where P is a positive number less than 100; detecting a signal in the envelope that exceeds the envelope threshold; The detected signal is marked as a transient surge signal.

5. The method for suppressing interference in current transformer measurement under high voltage and strong electromagnetic interference as claimed in claim 1, characterized in that: The method for obtaining different types of abnormal signals adjacent to each other in the time domain is determined as follows: When the time length between corresponding moments of different types of abnormal signals is less than a preset time window, the two different types of abnormal signals are regarded as different types of abnormal signals adjacent in time domain.

6. The method for suppressing interference in current transformer measurement under high voltage and strong electromagnetic interference as claimed in claim 1, characterized in that: The composite interference characteristic of the current signal is recorded as A. , where a is the number of signal pairs in the high-frequency interference signal of the current signal, N is the number of signal points of the collected current signal, is the total number of pairwise combinations of these signal points, represents the signal in the corresponding time domain between the two abnormal signals in signal pair i, It represents the accumulated value of the signal amplitude in the corresponding time domain range between the two abnormal signals in signal pair i.

7. The method for suppressing interference in current transformer measurement under high voltage and strong electromagnetic interference as claimed in claim 1, characterized in that: The measured interference intensity feature is further determined by the product of the composite interference feature mean of all types of signals, all calculated covariance sum values, and the difference between two regression coefficient vectors.

8. The method for suppressing interference in current transformer measurement under high voltage and strong electromagnetic interference as claimed in claim 1, characterized in that: The improved formula of the step size parameter is: ,in, is the improved step size, is the initial step length, is the currently calculated measurement interference intensity characteristic, The maximum value of the measured interference intensity characteristic calculated historically.

9. The method for suppressing interference in current transformer measurement under high voltage and strong electromagnetic interference as claimed in claim 1, characterized in that: The method of dividing all kinds of signals into two categories is: dividing according to internal influence signals and external influence signals, dividing current and voltage signals into internal influence signals, and dividing magnetic field and ground potential signals into external influence signals.

10. The method for suppressing interference in current transformer measurement under high voltage and strong electromagnetic interference as claimed in claim 9, characterized in that: The independent variables or dependent variables in the regression analysis are matrices constructed from all the signals in the corresponding class.

Images