Search

CN-121978446-A - Self-adaptive power quality monitoring method for power supply cabinet

CN121978446ACN 121978446 ACN121978446 ACN 121978446ACN-121978446-A

Abstract

The invention relates to the technical field of measuring electric variables, in particular to a self-adaptive power quality monitoring method for a power cabinet, which comprises the following steps of: and starting a low-frequency acquisition unit arranged at a monitoring node of the power cabinet, and acquiring a low-frequency rough scanning voltage signal of the power cabinet in the current monitoring period at a preset low-frequency sampling rate. According to the invention, the compressed observation vector is transmitted, the sparse basis inversion calculation is executed by utilizing the orthogonal matching pursuit algorithm, the high-precision electric energy quality time domain waveform data is reconstructed, abundant high-frequency transient details are restored under the low data volume transmission condition, distortion characteristic analysis is carried out based on the reconstructed waveform, disturbance types are identified, a monitoring analysis report is generated and uploaded, real-time dynamic adjustment of a sampling scale and a matrix structure along with the signal fluctuation state is realized, the optimal balance of the data compression ratio and the signal reconstruction quality is achieved, the contradiction between massive high-frequency data storage and the limited transmission bandwidth in the complex power grid environment is solved, and the transient abnormal acuity is improved.

Inventors

  • LIU KUN

Assignees

  • 浙江欧日力电气有限公司

Dates

Publication Date
20260505
Application Date
20260407

Claims (7)

  1. 1. The self-adaptive power quality monitoring method for the power cabinet is characterized by comprising the following steps of: starting a low-frequency acquisition unit arranged at a monitoring node of a power cabinet, acquiring a low-frequency rough scanning voltage signal of the power cabinet in a current monitoring period at a preset low-frequency sampling rate, and extracting the time domain fluctuation characteristic of the low-frequency rough scanning voltage signal; Calculating an adaptive observation dimension for a current monitoring period according to the time domain fluctuation feature, and constructing a dynamic Gaussian random observation matrix based on the adaptive observation dimension, wherein the number of lines of the dynamic Gaussian random observation matrix is equal to the adaptive observation dimension, and the number of columns of the dynamic Gaussian random observation matrix is equal to a preset high-frequency reconstruction length; the dynamic Gaussian random observation matrix is utilized to configure a compression sampling circuit at the analog front end, the real-time analog voltage signal of the power cabinet is subjected to compression perception analog information conversion, and a compression observation vector is obtained, wherein the dimension of the compression observation vector is consistent with the dimension of the self-adaptive observation; Transmitting the compressed observation vector to a digital signal processing unit, performing sparse basis inversion calculation on the compressed observation vector by using an orthogonal matching pursuit algorithm, and reconstructing to obtain high-precision power quality time domain waveform data; Performing waveform distortion characteristic analysis on the high-precision power quality time domain waveform data, identifying the power quality disturbance type of the power cabinet, generating a power quality monitoring analysis report and uploading the power quality monitoring analysis report to a monitoring terminal; Starting a low-frequency acquisition unit arranged at a monitoring node of a power cabinet, acquiring a low-frequency rough scanning voltage signal of the power cabinet in a current monitoring period at a preset low-frequency sampling rate, and extracting time domain fluctuation characteristics of the low-frequency rough scanning voltage signal specifically comprises the following steps: Initializing an analog-to-digital converter in the low-frequency acquisition unit, and setting the sampling frequency of the analog-to-digital converter to be a preset multiple of power frequency; Polling and sampling a three-phase voltage port of the power cabinet through the analog-to-digital converter, obtaining a discrete voltage sampling sequence within a preset time window length, and marking the discrete voltage sampling sequence as the low-frequency rough scanning voltage signal; Calculating the voltage amplitude change rate of the low-frequency rough scanning voltage signal in the preset time window length and the total energy value of the voltage waveform; combining the voltage amplitude change rate and the total energy value into the time domain fluctuation feature.
  2. 2. The adaptive power quality monitoring method for a power cabinet according to claim 1, wherein the step of calculating an adaptive observation dimension for a current monitoring period according to the time domain fluctuation feature and constructing a dynamic gaussian random observation matrix based on the adaptive observation dimension is specifically: Reading the voltage amplitude change rate and the total energy value of the voltage waveform contained in the time domain fluctuation characteristic, and acquiring a rated voltage reference value, a reference energy value and a reference voltage amplitude change rate of the power cabinet which are stored in advance; Processing the voltage amplitude change rate and the total energy value by using an adaptive dimension calculation model, and calculating to obtain the adaptive observation dimension, wherein a calculation formula in the adaptive dimension calculation model is as follows: ; Wherein, the Representing the dimensions of the adaptive observation, Representing the preset basic observation dimensions of the system, Representing a down-rounding operation, The energy weighting coefficients are represented by a set of coefficients, Representing the ratio of the total energy value to the reference energy value, Which represents the value of the total energy value, Representing the value of the reference energy in question, Representing the sensitivity coefficient of the rate of change, A logarithmic function with a base of 10 is shown, Representing the absolute value of the rate of change of the voltage amplitude, Representing the reference voltage amplitude change rate; initializing a zero matrix with the number of lines as the self-adaptive observation dimension and the number of columns as the preset high-frequency reconstruction length; And generating a Gaussian random number sequence by using a Gaussian distribution random number generator, and filling the Gaussian random number sequence into the zero matrix to obtain the dynamic Gaussian random observation matrix.
  3. 3. The adaptive power quality monitoring method for a power supply cabinet according to claim 1, wherein the step of configuring a compressed sampling circuit at an analog front end by using the dynamic gaussian random observation matrix to perform compressed sensing analog information conversion on a real-time analog voltage signal of the power supply cabinet, and obtaining a compressed observation vector specifically comprises the steps of: loading the dynamic Gaussian random observation matrix into a random demodulator controller of the analog front end; controlling a mixer switch state in the compressed sampling circuit according to the sequential logic of the dynamic Gaussian random observation matrix by using the random demodulator controller; Performing analog multiplication operation on the real-time analog voltage signal of the power cabinet and a pseudo-random sequence through the mixer to obtain a mixed analog modulation signal, wherein the pseudo-random sequence is obtained by converting the dynamic Gaussian random observation matrix; controlling an integrator to perform integral operation of a preset time period on the analog modulation signal to obtain an integral analog voltage value; And at the end time of each integration period, triggering a low-speed analog-to-digital converter to quantize the integrated analog voltage value to obtain a digitized observed value, and arranging the observed values according to a time sequence to form the compressed observed vector.
  4. 4. The adaptive power quality monitoring method for a power cabinet according to claim 1, wherein the step of transmitting the compressed observation vector to a digital signal processing unit, performing sparse basis inversion calculation on the compressed observation vector by using an orthogonal matching pursuit algorithm, and reconstructing to obtain high-precision power quality time domain waveform data is specifically as follows: Inputting the compressed observation vector into an operation memory of the digital signal processing unit, and loading a preset discrete cosine transform sparse basis matrix; Initializing a residual vector as the compressed observation vector, initializing an index set as an empty set, and initializing an iteration counter; Calculating the product of the dynamic Gaussian random observation matrix and the discrete cosine transform sparse base matrix to obtain a sensing matrix; In each iteration process, calculating the inner product correlation coefficient of each column of the sensing matrix and the current residual vector, and finding out the column index corresponding to the column with the largest absolute value of the inner product correlation coefficient; Adding the column index into the index set, and extracting a corresponding sub-matrix from the sensing matrix according to the index set; solving an approximate solution of the submatrix and the compressed observation vector by using a least square method to obtain a current sparse coefficient estimated value; Updating the residual vector, judging whether the norm of the residual vector is smaller than a preset reconstruction stopping threshold, stopping iteration if the norm is smaller than the preset reconstruction stopping threshold, and otherwise, continuing the next iteration; And performing inverse transformation by using the finally obtained sparse coefficient estimation value and the discrete cosine transform sparse base matrix, and synthesizing to obtain the high-precision power quality time domain waveform data.
  5. 5. The method for adaptively monitoring power quality for a power cabinet according to claim 4, wherein in each iteration process, calculating an inner product correlation coefficient between each column of the sensing matrix and the current residual vector, and finding a column index corresponding to a column with the largest absolute value of the inner product correlation coefficient comprises the following steps: Normalizing each column vector of the sensing matrix to obtain a normalized sensing column vector set; respectively calculating dot products of each vector in the normalized sensing column vector set and the residual vector to obtain a correlation coefficient vector; traversing the related coefficient vector, and positioning the maximum value position in the related coefficient vector through comparison operation; and determining a matrix column sequence number corresponding to the maximum value position as the column index.
  6. 6. The adaptive power quality monitoring method for a power cabinet according to claim 1, wherein the step of performing waveform distortion feature analysis on the high-precision power quality time domain waveform data, identifying a power quality disturbance type of the power cabinet, generating a power quality monitoring analysis report, and uploading the power quality monitoring analysis report to a monitoring terminal comprises the following steps: Performing Fourier transform on the high-precision power quality time domain waveform data to obtain power quality frequency spectrum data; Calculating fundamental component amplitude and harmonic component amplitude in the electric energy quality frequency spectrum data, and calculating total harmonic distortion rate according to the fundamental component amplitude and the harmonic component amplitude; Detecting voltage sag features, voltage sag features and voltage interruption features in the high-precision power quality time domain waveform data; if the total harmonic distortion rate exceeds a preset harmonic threshold, judging that the electric energy quality disturbance type is harmonic pollution; if the voltage sag characteristic is detected, judging that the electric energy quality disturbance type is voltage sag; If the voltage transient rise characteristic is detected, judging that the electric energy quality disturbance type is voltage transient rise; If the voltage interruption characteristic is detected, judging that the electric energy quality disturbance type is voltage interruption; packaging the determined power quality disturbance type, the total harmonic distortion rate and the disturbance time stamp into a data packet in a standard format, and taking the data packet as the power quality monitoring analysis report; and sending the power quality monitoring analysis report to the remote monitoring terminal through an industrial Ethernet interface.
  7. 7. The adaptive power quality monitoring method for a power cabinet according to claim 2, wherein the step of generating a gaussian random number sequence by using a gaussian distributed random number generator, and filling the gaussian random number sequence into the zero matrix to obtain the dynamic gaussian random observation matrix comprises the following steps: Calling a hardware random number generator to generate an original random seed, and injecting the original random seed into the Gaussian distribution random number generator; the uniformly distributed random numbers are converted into standard normal distributed random numbers through a Box-Muller conversion algorithm; scaling the standard normal distribution random number to ensure that the variance meets the numerical requirement of dividing 1 by the adaptive observation dimension; Filling each element position of the zero matrix in sequence until all positions are filled; And carrying out orthogonalization check on the filled matrix, and regenerating if the check is not passed, until the dynamic Gaussian random observation matrix meeting the limited equidistant property is obtained.

Description

Self-adaptive power quality monitoring method for power supply cabinet Technical Field The invention relates to the technical field of measuring electric variables, in particular to a self-adaptive power quality monitoring method for a power cabinet. Background The existing power quality monitoring mainly relies on a fixed high-frequency sampling mechanism meeting the Nyquist sampling theorem to capture transient disturbance, constant high-speed data acquisition is required to be maintained in actual operation to prevent missing high-frequency harmonic waves or transient pulses no matter whether a power grid load is in a severe fluctuation or stable operation state, the indiscriminate full-period high-frequency sampling generates redundant data in a long-period stable operation process, load pressure is brought to a front-end analog-digital conversion device and a storage unit, high-frequency sampling has high requirements on hardware specifications, high equipment manufacturing cost and high operation power consumption are caused, and when a multi-node distributed monitoring scene is involved, communication network congestion is easily caused by concurrent transmission of original high-sampling-rate data, data packet loss or transmission delay is caused, and key fault information is prevented from being uploaded in real time. Therefore, improvements are needed. Disclosure of Invention The invention aims to solve the defects in the prior art, and provides a self-adaptive power quality monitoring method for a power cabinet. In order to achieve the above purpose, the invention adopts the following technical scheme that the self-adaptive power quality monitoring method for the power cabinet comprises the following steps: starting a low-frequency acquisition unit arranged at a monitoring node of a power cabinet, acquiring a low-frequency rough scanning voltage signal of the power cabinet in a current monitoring period at a preset low-frequency sampling rate, and extracting the time domain fluctuation characteristic of the low-frequency rough scanning voltage signal; Calculating an adaptive observation dimension for a current monitoring period according to the time domain fluctuation feature, and constructing a dynamic Gaussian random observation matrix based on the adaptive observation dimension, wherein the number of lines of the dynamic Gaussian random observation matrix is equal to the adaptive observation dimension, and the number of columns of the dynamic Gaussian random observation matrix is equal to a preset high-frequency reconstruction length; the dynamic Gaussian random observation matrix is utilized to configure a compression sampling circuit at the analog front end, the real-time analog voltage signal of the power cabinet is subjected to compression perception analog information conversion, and a compression observation vector is obtained, wherein the dimension of the compression observation vector is consistent with the dimension of the self-adaptive observation; Transmitting the compressed observation vector to a digital signal processing unit, performing sparse basis inversion calculation on the compressed observation vector by using an orthogonal matching pursuit algorithm, and reconstructing to obtain high-precision power quality time domain waveform data; And carrying out waveform distortion characteristic analysis on the high-precision power quality time domain waveform data, identifying the power quality disturbance type of the power cabinet, generating a power quality monitoring analysis report and uploading the power quality monitoring analysis report to a monitoring terminal. Preferably, the step of starting a low-frequency acquisition unit arranged at a monitoring node of the power cabinet, acquiring a low-frequency rough scanning voltage signal of the power cabinet in a current monitoring period at a preset low-frequency sampling rate, and extracting a time domain fluctuation characteristic of the low-frequency rough scanning voltage signal comprises the following steps: Initializing an analog-to-digital converter in the low-frequency acquisition unit, and setting the sampling frequency of the analog-to-digital converter to be a preset multiple of power frequency; Polling and sampling a three-phase voltage port of the power cabinet through the analog-to-digital converter, obtaining a discrete voltage sampling sequence within a preset time window length, and marking the discrete voltage sampling sequence as the low-frequency rough scanning voltage signal; Calculating the voltage amplitude change rate of the low-frequency rough scanning voltage signal in the preset time window length and the total energy value of the voltage waveform; combining the voltage amplitude change rate and the total energy value into the time domain fluctuation feature. Preferably, the step of calculating the adaptive observation dimension for the current monitoring period according to the time domain fluctuation