CN-122017317-A - Wireless high-voltage current measurement method with filtering anti-interference function

CN122017317ACN 122017317 ACN122017317 ACN 122017317ACN-122017317-A

Abstract

The invention discloses a wireless high-voltage current measurement method with a filtering anti-interference function, and belongs to the technical field of robot coordination control; the method comprises the steps of obtaining an original discrete voltage signal sequence, calculating transient strength indexes of the discrete voltage signal sequence, performing variation modal decomposition on the original discrete voltage signal sequence by using modal numbers and punishment parameters to obtain a group of eigen modal functions, respectively synthesizing the classified modes into corresponding interference reference signals, respectively inputting the orthogonal interference reference signals into a filter channel to estimate various interference components and perform dynamic adjustment, and restoring pure high-voltage wire current time domain signals. The invention adopts VMD real-time sensing signal characteristics to dynamically adjust subsequent key parameters and improve denoising performance, and adopts a multi-feature mode intelligent classification method to map mathematical modes with unknown physical meaning to four physical sets of useful signals, power frequency, high frequency and transient state so as to realize decoupling of signals and noise on a time-frequency domain.

Inventors

LV LIXIANG
XU WEILUN
LIU JIAN
Rao huanyu
WANG YAN
WANG ZHE
Miao Cairan

Assignees

国网江苏省电力有限公司南京供电分公司

Dates

Publication Date: 20260512
Application Date: 20260213

Claims (10)

1. A wireless high-voltage current measurement method with a filtering anti-interference function is characterized by comprising the following steps: step S1, acquiring a voltage signal and discretizing to obtain an original discrete voltage signal sequence; s2, calculating transient strength indexes of the discrete voltage signal sequence, and determining the mode number and penalty parameters of a variation mode decomposition algorithm to realize dynamic matching of decomposition parameters and signal characteristics; S3, performing variation modal decomposition on the original discrete voltage signal sequence by using the modal number and the penalty parameter to obtain a group of eigen modal functions, and performing modal intelligent classification; s4, respectively synthesizing the classified modes into corresponding interference reference signals, and carrying out orthogonalization treatment; S5, respectively inputting the orthogonal interference reference signals into a filter channel to estimate various interference components and dynamically adjusting the interference components; And S6, subtracting the estimated various interference components from the primary purification signal, and carrying out numerical integration on the obtained signal to restore a pure high-voltage wire current time domain signal.
2. The method for measuring wireless high-voltage current with the filtering anti-interference function of claim 1, wherein the step S1 is characterized in that voltage signals are obtained and discretized to obtain an original discrete voltage signal sequence, and the method specifically comprises the following steps: step S1-1, selecting a current sensor to acquire an original current signal: based on the advantages of electrical isolation safety, no magnetic saturation risk and broadband response, an open type rogowski coil is selected as a current sensor; For the Rogowski coil to achieve the best performance, a mutual inductance is introduced The rogowski coil was optimized: ; Wherein, the Is the number of turns of the rogowski coil, For the average radius of the rogowski coil, Is the relative permeability of the magnetic core material, Is the cross-sectional area of the circle, Is vacuum permeability by optimizing 、、、 Mutual inductance coefficient The optimized value suitable for the high-voltage scene is reached, enough sensitivity can be ensured, broadband response is ensured, and the saturation of a post-stage circuit is prevented; Aligning and encircling an opening of the open type rogowski coil with a high-voltage wire, closing the opening and mechanically locking after encircling the high-voltage wire, so as to ensure that the rogowski coil is stably and tightly encircling the high-voltage wire, and forming a single-turn measuring loop; step S1-2, converting a current signal acquired by a current sensor into a voltage signal, wherein the step is specifically as follows: current signal Through the high-voltage wire, a variable magnetic field is generated around the high-voltage wire, the magnetic field is coupled with the Rogowski coil, and electromotive force is induced at two ends of the Rogowski coil The expression is: ; Wherein, the Representing the measured current signal The first derivative of time t, the rate of change of current; a termination resistor is connected in parallel with the output port of the Rogowski coil Forming a current-voltage conversion circuit, wherein the self impedance of the Rogowski coil is smaller than that of the current sensor in the normal working frequency range Output voltage signal : ; Step S1-3, converting the continuous voltage signal into a discrete voltage signal sequence through analog-to-digital conversion: ; Wherein, the The time interval of the sampling is indicated, Representing a discrete time index; Representing a sequence of discrete voltage signals.
3. The method for measuring wireless high-voltage current with filtering anti-interference function according to claim 1, wherein the step S2 is characterized in that transient strength indexes of discrete voltage signal sequences are calculated, the mode number and penalty parameters of a variation mode decomposition algorithm are determined, and dynamic matching of the decomposition parameters and signal characteristics is realized, specifically: Step S2-1, a double buffer mechanism combining long and short windows is introduced to preprocess a discrete voltage signal sequence: Creating two independent data buffers, namely a long window buffer and a short window buffer, wherein the capacity of the long window buffer is The capacity of the short window buffer is that at each sampling point The position of the short window buffer zone is completely overlapped with the tail part of the long window buffer zone; for each newly arrived discrete voltage signal sequence The following operations are performed: Updating long window by storing discrete voltage signal sequence in long window buffer zone, and maintaining the latest one in first-out strategy Sampling points; Short window updating, in which discrete voltage signal sequences are stored in a short window buffer area at the same time, and the short window buffer area maintains the latest by using a first-in first-out strategy Sampling points; Since the position of the window buffer completely overlaps the tail of the long window buffer, the window buffer is in the short window buffer A point, which is the latest in the long window buffer area Sampling points of the number; The long window data stored in the long window buffer zone adopts a Hanning window function And (3) windowing calculation: ; Defining a data block stored in a long window buffer as a long window signal vector Defining the data block stored in the short window buffer as a short window signal vector ; And S2-2, acquiring a transient strength index of the discrete voltage signal sequence, and converting the transient strength index into a modal number and a punishment parameter of the VMD parameter.
4. The method for measuring wireless high-voltage current with filtering anti-interference function according to claim 3, wherein the step S2-2 is characterized in that the transient strength index of the discrete voltage signal sequence is obtained, specifically: Calculating long window signal vector Time domain feature index of (a): the time domain characteristic indexes comprise kurtosis characteristic indexes, waveform factor characteristic indexes and Teager energy characteristic indexes, and the three characteristic indexes respectively form complementary characteristic perceptions from the angles of statistical distribution, time domain shape and instantaneous energy; Calculating kurtosis characteristic indexes: ; Wherein, the Representing the variance; representing the mean; calculating characteristic indexes of the waveform factors: ; Wherein, the Represents a root mean square value; Representing the peak maximum; Calculating Teager energy characteristic indexes: ; Wherein, the , ; The transient strength index is generated by weighting and fusing kurtosis characteristic index, waveform factor characteristic index and Teager energy characteristic index, and the calculation formula of the transient strength index is as follows: ; Wherein, the Representing normalized transient intensity indicators; the function of the maximum value is represented, A minimum function; the transient strength index is converted into the mode number and punishment parameters of VMD parameters, specifically: The number of modes is: ; Wherein, the Representing the number of modes of the VMD; representing an upward rounding function; Penalty parameters are: ; Wherein, the Penalty parameters representing the VMD; the base number is 2, and the index is Is an exponential operation of (a).
5. The method for measuring wireless high-voltage current with filtering anti-interference function as set forth in claim 1, wherein in the step S3, a set of eigen mode functions is obtained by performing variation mode decomposition on an original discrete voltage signal sequence by using a mode number and a penalty parameter, and the method specifically comprises the steps of: s3-1, performing variational modal decomposition on the short window signal vector by utilizing VMD parameters to obtain an eigenmode function: the variational modal decomposition adaptively decomposes a complex noise-containing high-voltage current signal into an intrinsic modal function with definite physical meaning by establishing an optimization problem taking frequency domain compactness as an optimal criterion and pursuing that energy of each mode is concentrated on a frequency domain to surround respective center frequency, and is expressed as follows: ; constraint conditions: ; Wherein, the Represent the first A discrete sequence of the individual modes, Representing all Sets of discrete sequences, i.e. ; Represents the first The center angular frequency of the individual modal components, Representing all The collection of central angular frequencies, i.e ; Representing an input short window signal discrete sequence; Representing a frequency domain differentiation operator; , wherein, By means of a discrete fourier transform DTF, ; Representing a sampling time interval; Representing imaginary units; Representing a convolution operation symbol; representing euclidean norms; s3-2, reconstructing an optimization problem by using an augmented Lagrangian method, and adopting ADMM iteration to solve the optimization problem: The reconstruction optimization problem is as follows: ; Wherein, the Ensuring that each mode is compact in the frequency domain for a bandwidth constraint term; Reconstructing an error penalty term, and ensuring that the sum of all modes can reconstruct an original signal; For Lagrange constraint term, the Lagrange multiplier is passed Strictly reinforcing the reconstruction constraint condition to ensure that the optimal solution is accurately satisfied ; And S3-3, extracting key characteristic parameters from the obtained discretization eigenmode function sequence, and carrying out intelligent classification on modes.
6. The method for measuring wireless high-voltage current with filtering anti-interference function according to claim 5, wherein the step S3-3 is characterized in that key characteristic parameters are extracted from the obtained discretized eigenmode function sequence, and intelligent classification of modes is performed, specifically: step S3-3-1, obtaining a modal characteristic triplet, wherein the modal characteristic triplet specifically comprises the following steps: the key characteristic parameters of each mode comprise a center frequency, a correlation coefficient and an energy ratio, and the center frequency, the correlation coefficient and the energy ratio form a mode characteristic triplet ; Center frequency acquisition for each modality Center frequency of (2) Obtained by calculating the centroid of the discrete power spectrum, the expression is: ; Wherein, the Representing sequences A kind of electronic device Point discrete Fourier transform (DCT) of the first point A coefficient; Represent the first Actual frequencies corresponding to the frequency points; Representing the number of DFT points, ; Representing a sampling time interval; For each modality Pearson correlation coefficient with original signal The calculation is as follows: ; Wherein, the Represent the first A mean value of the individual modality sequences; representing the mean value of the original short window signal sequence; the energy duty cycle is calculated by calculating the energy of each mode And total energy of all modes The expression is: ; ; ; Wherein, the Represent the first The relative energy ratio of each mode is satisfied And is also provided with ; Represent the first Energy of individual modes; Representing all The sum of the energies of the individual modes; The number of modes representing the decomposition of the VMD, ; The representation takes absolute value; Step S3-3-2, classifying the modal feature triples, wherein the method specifically comprises the following operations: the modality set is defined by initializing four empty sets, which are respectively: a set of useful signal modalities; a working interference mode set; a set of high frequency interference modes; a transient interference modality set; Each modality is classified in order of priority from known interference to unknown transient: The power frequency interference mode is judged, namely the most stable and easily identified power frequency and harmonic interference thereof are preferentially identified and separated, and the rule is as follows: ; and judging a high-frequency interference mode, namely identifying deterministic high-frequency noise with frequency far higher than a useful signal frequency band, wherein the rule is as follows: ; Useful signal modality determination: ; transient interference mode determination: ; Wherein, the Representing a modality Belongs to the power frequency interference set ; Indicating the presence of one , Is a harmonic index; representing the logical and operator of the device, Representing a set union operator.
7. The method for measuring wireless high-voltage current with filtering anti-interference function as set forth in claim 1, wherein in the step S4, the classified modes are respectively synthesized into corresponding interference reference signals, and orthogonalization processing is performed, specifically: Synthesizing modalities classified as "useful signals" into a primarily purified signal sequence The expression is: ; Wherein, the Representing the signal sequence after the preliminary purification, ; Belonging to the useful signal set Is the first of (2) A sequence of modalities; Represent the first Weight coefficients of the useful signal modalities; Represent the first Correlation coefficients of the individual modes and the original signal; Represent the first The energy duty cycle of the individual modes; representing all useful signal modalities And (3) summing; using correlation-based coefficients And energy duty cycle The weighting synthesis strategy of the method comprises the steps of synthesizing the initially purified signal sequence into three types of interference reference signals, namely a power frequency interference reference signal, a high frequency noise reference signal and a transient interference reference signal; Synthesizing a power frequency interference reference signal: ; synthesizing a high-frequency noise reference signal: ; Synthesizing transient interference reference signals: ; Wherein, the The reference signal sequences respectively represent power frequency interference, high-frequency noise and transient interference; Representing interference sets to power frequency Summing all the modal sequences in the model; Using orthogonalization to eliminate correlation between three kinds of interference reference signals, the correlated interference signals are converted into mutually orthogonal independent components: determining an orthogonalization sequence according to the remarkable degree of interference characteristics by adopting a Gram-Schmidt orthogonalization method, and processing according to the sequence of the priority of the power frequency interference reference signals, the secondary of the high frequency noise reference signals and the last of the transient interference reference signals; Orthogonalization power frequency interference reference signal: ; Orthogonalizing high frequency noise reference signals: ; Orthogonalization of transient interference reference signals: 。
8. the method for measuring wireless high-voltage current with filtering anti-interference function as set forth in claim 1, wherein in the step S5, the orthogonal post-interference reference signals are respectively input into the filter channels to estimate various interference components and perform dynamic adjustment, specifically: Step S5-1, initializing the adaptive filter according to three interference channels of three types of interference reference signals: For each interference channel Sequentially executing, initializing weight vectors ; Initializing an inverse correlation matrix Initializing forgetting factors ; Step S5-2, updating forgetting factors in real time: for each sampling instant Each channel The method comprises the following steps of: Forgetting factor update: ; Wherein, the Representation channel At the moment of time Is used for the input signal energy of the (a), ; Representing a channel sensitivity parameter; Channel At the moment of time Forgetting factor of (2); S5-3, constructing an input vector to calculate an optimal weight vector updating direction, and performing Kalman gain calculation: Each channel using its own dedicated input signal Constructing an input vector: ; Kalman gain computation relies on the present channel Time inverse correlation matrix And Time forgetting factor : ; Wherein, the Representation channel At the moment of time Is a vector of inputs of (a); Representation channel At the moment of time Is a Kalman gain vector of (1); s5-4, carrying out priori error estimation and dynamic adjustment of weight coefficients based on the filtering errors; Step S5-5, updating an inverse correlation matrix and maintaining numerical stability: and S5-6, obtaining and outputting an interference signal.
9. The method for measuring wireless high-voltage current with the filtering anti-interference function according to claim 8, wherein the step S5-4 is characterized in that the prior error estimation and the dynamic adjustment of the weight coefficient are performed based on the filtering error, and specifically comprises the following steps: Each channel is independently learned based on its own error signal, for each channel The method comprises the following steps of: prior error calculation ; Weight vector update ; Wherein, the Representation channel At the moment of time Is a priori error of (2); Representation channel At the moment of time The updated weight vector; Step S5-5, updating an inverse correlation matrix and maintaining numerical stability: For each channel Sequentially performing, and updating an inverse correlation matrix: ; Wherein, the Representation channel At the moment of time The updated inverse correlation matrix; step S5-6, obtaining and outputting an interference signal: at each sampling instant The channel is used for updating the optimal weight vector at the current moment Current input vector And carrying out weighted summation to calculate an estimated value of the interference component represented by the channel at the moment, wherein the expression is as follows: ; Wherein, the Representing the original signal In, by the first The portion of the signal contributed by the interference-like signal; The calculation results of all channels at all sampling moments are arranged to form a final output data set: ; Wherein, the A power frequency interference estimation signal representing the whole process; a high frequency noise estimation signal representing the whole course; representing the transient interference estimation signal throughout.
10. The method for measuring wireless high-voltage current with filtering anti-interference function as set forth in claim 1, wherein in the step S6, the estimated various interference components are subtracted from the preliminary purified signal, and the obtained signal is subjected to numerical integration, specifically: The virtual common mode inhibition method is adopted to inhibit voltage synthesis, and the expression is as follows: ; Wherein, the Representing a denoising voltage signal; a high frequency noise estimation signal representing the whole course; A power frequency interference estimation signal representing the whole process; a transient interference estimation signal representing the whole course; a signal sequence representing the preliminary purification; recovery current by numerical integration: ; Wherein, the Representing the mutual inductance coefficient of the rogowski coil; Representing a sampling time interval; representing a trapezoidal integration method; Representing the recovered high voltage current dispersion signal; converting the high-voltage current discrete signal into a continuous time domain waveform to be output, and generating a high-voltage current time domain signal, wherein the expression is as follows: ; Wherein, the Representing a raised cosine window function, Representation of Interpolation of functions; Indicating the overlap length.

Description

Wireless high-voltage current measurement method with filtering anti-interference function Technical Field The invention belongs to the technical field of high-voltage current measurement, and particularly relates to a wireless high-voltage current measurement method with a filtering anti-interference function. Background With the rapid development of the power industry, the scale of the power system is continuously enlarged, and the current power industry is continuously improved, under the environment, a high-voltage wire is used as a 'large artery' of power transmission, and the running state of the high-voltage wire is directly related to the safety and the stability of the whole power system. In the process of testing and maintaining high-voltage wires, the conventional contact ammeter has various inconveniences and limitations in actual measurement, such as high cable height, interference of a strong electromagnetic field on signals, environment of a test site, influence on ground current, operation of operation and maintenance personnel and the like. The accurate measurement of the high-voltage wire current is not only the basis for realizing line protection, fault diagnosis and state evaluation, but also the key link for realizing advanced applications such as intelligent power grid, distribution automation, on-line monitoring and the like. In traditional engineering practice, high-voltage current measurement is mostly dependent on devices such as a closed Current Transformer (CT) or a Hall current sensor, and the devices often need complex insulation structures and installation processes under a high-voltage scene, have large volume, high weight and difficult maintenance, easily have the problems of iron core saturation, limited frequency band and the like under the working conditions of large current, wide frequency band and high transient state, and are difficult to meet the requirements of a modern power system on high-precision, wide frequency band and online real-time measurement. In recent years, rogowski coils are widely applied to the field of high-voltage current measurement because of the advantages of safe electrical isolation, no risk of iron core magnetic saturation, wide linear range, wide frequency band response and the like. Especially, the open type rogowski coil can be directly sleeved outside a high-voltage wire under the condition of no power failure, has simple structure and flexible installation, and is suitable for the transformation and the on-line monitoring of the existing power grid. However, the rogowski coil essentially outputs a voltage signal proportional to the rate of change of the measured current, and the current waveform is recovered by numerical integration. In an actual high-voltage operation environment, strong power frequency electromagnetic field interference, high-frequency switching noise, harmonic waves introduced by a power electronic device, transient interference caused by faults and operation and the like exist around a lead, and multi-source heterogeneous interference signals can be overlapped on the output voltage of a rogowski coil, so that integral drift, waveform distortion and measurement errors are obviously increased. The traditional filtering method, such as a band-pass/low-pass filter with fixed parameters, wavelet denoising or Empirical Mode Decomposition (EMD), has obvious limitations in processing complex and changeable interference in high-voltage current measurement, namely, on one hand, the power frequency and harmonic waves thereof, broadband high-frequency noise and fault transient characteristics are overlapped in frequency spectrum and are difficult to realize effective separation through simple fixed frequency band segmentation, on the other hand, the on-site working condition presents obvious non-stationary characteristics, the frequency band, amplitude and time sequence characteristics of the interference can dynamically change along with factors such as loads, switching operation, fault types, weather conditions and the like, the fixed parameter filtering strategy is difficult to be adaptive to match with signal characteristics, and insufficient filtering or excessive filtering easily occurs, so that useful signal information loss is caused. In addition, most of the existing methods focus on single-type noise suppression, lack of a unified and cooperative suppression mechanism for power frequency interference, high-frequency noise and transient interference, and are difficult to meet the requirements of long-term stable operation and high-precision measurement of a wireless high-voltage current measuring device in a complex field environment. Therefore, how to solve the problems of multi-type signal interference suppression and useful signal reconstruction in the high-voltage wire current signal and improve the anti-interference capability and measurement accuracy of a wireless high-voltage current measurement system in a compl