A method for evaluating the partial discharge state of substation switchgear based on high-dimensional indicators

Abstract

The present invention discloses a method for evaluating the partial discharge state of a substation switchgear based on high-dimensional indicators, comprising the following steps: collecting phase decomposition partial discharge PRPD pattern data of the substation switchgear; performing spectrogram construction and statistical feature calculation on the obtained partial discharge PRPD pattern data; preprocessing the statistical features calculated from the spectrogram data to construct a spatio-temporal data set matrix; based on the random matrix theory, using the moving sliding window method to sequentially select matrices on the spatio-temporal data set matrix, constructing a statistic using the Chebyshev polynomial T<subgt;3< / subgt> to determine whether there are outliers, calculating the statistic of the matrix selected by the sliding window, and evaluating the state of the substation switchgear. The present invention can monitor the operating state of the substation switchgear and improve the efficiency and accuracy of substation switchgear monitoring.

Description

Technical Field

[0001] The present invention relates to the technical field of substation equipment, and in particular to a method for evaluating the partial discharge state of a substation switchgear cabinet with high-dimensional indicators, an electronic device, and a readable storage medium. Background Art

[0002] Partial discharge (PD) is a long-term and slow damage process of insulating materials. The switchgear cabinet is a circuit used to control and protect electrical appliances in substation equipment. During operation, it is prone to insulation deterioration due to the effects of chemicals, heat, and other abnormal factors, and even explosion and insulation breakdown may occur. Further, it may lead to faults in the power system, which is not conducive to the stable operation of the power system.

[0003] Therefore, those skilled in the art are committed to developing a method for evaluating the partial discharge state of a substation switchgear cabinet with high-dimensional indicators to improve the efficiency and accuracy of partial discharge monitoring of the switchgear cabinet. Summary of the Invention

[0004] In view of the above-mentioned defects of the prior art, the technical problem to be solved by the present invention is how to provide a method, system, and readable storage medium for evaluating the partial discharge state of a substation switchgear cabinet, conduct feature analysis on a high-dimensional spatio-temporal data set, deeply mine the statistical information of the operating characteristics of the substation switchgear cabinet, and monitor and evaluate the operating state of the switchgear cabinet through operations such as establishing a spatio-temporal data set and analyzing a random matrix model.

[0005] To achieve the above object, the present invention provides a method for evaluating the partial discharge state of a substation switchgear cabinet based on high-dimensional indicators, including the following steps:

[0006] Step 1: Collect the partial discharge PRPD pattern data of the substation switchgear cabinet;

[0007] Step 2: Conduct spectrogram construction and statistical feature calculation on the obtained partial discharge PRPD pattern data;

[0008] Step 3: Preprocess the statistical features calculated from the spectrogram data and construct a spatio-temporal data set matrix;

[0009] Step 4: Based on the random matrix theory, use the moving sliding window method to sequentially select matrices on the spatio-temporal data set matrix, use the Chebyshev polynomial T3 to construct a statistic to determine whether there are outliers, calculate the statistic of the matrix selected by the sliding window, and evaluate the state of the substation switchgear cabinet.

[0010] Further, the sampling frequency for collecting the partial discharge PRPD pattern data of the substation switchgear cabinet in Step 1 is 10 - 20 min / time.

[0011] Further, the statistical features calculated in step 2 include discharge amplitude, discharge pulse value, skewness and kurtosis, cross-correlation coefficient, shape parameter and scale parameter of Weibull distribution.

[0012] Further, the way of preprocessing the statistical features calculated from the atlas data in step 3 includes one or more of missing value filling and outlier cleaning.

[0013] Further, constructing the spatio-temporal dataset matrix in step 3 includes establishing a vector of statistical feature quantities of atlas data, and forming the spatio-temporal dataset matrix with the vector of statistical feature quantities.

[0014] Further, evaluating the state of the substation switchgear in step 4 includes constructing a linear eigenvalue index system for partial discharge of the switchgear by using the statistics of the selected matrix, and monitoring the operating state of partial discharge of each switchgear according to the preset index system.

[0015] Further, monitoring the operating state of partial discharge of each switchgear according to the preset index system includes selecting the substation switchgear state index from the pre-constructed linear eigenvalue index system, and judging whether the index exceeds the threshold when an abnormal state occurs.

[0016] Further, step 4 further includes: converting the matrix selected by the sliding window into a standardized matrix with a mean of 0 and a variance of 1, obtaining a singular value equivalent matrix for the standardized matrix, and performing unitary processing.

[0017] The present invention also provides an electronic device, including a processor and a memory, where a computer program is stored on the memory, and when the computer program is executed by the processor, the above-mentioned method is implemented.

[0018] The present invention also provides a readable storage medium, in which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned method is implemented.

[0019] Compared with the current monitoring method for the partial discharge state of switchgear, the present invention has the following advantages:

[0020] 1) Extract partial discharge atlas features, construct a spatio-temporal dataset, and adopt the M-P law and circular law of random matrix theory, making full use of the data stream information of the switchgear, and realizing efficient and accurate monitoring of the operating state of the switchgear based on a data-driven method;

[0021] 2) By constructing an eigenvalue index system, a reasonable eigenvalue function can be better adjusted and constructed according to the performance evaluation threshold of the switchgear operating state.

[0022] The concept, specific structure and technical effects of the present invention will be further described below in conjunction with the accompanying drawings to fully understand the purpose, features and effects of the present invention. Description of the Drawings

[0023] Figure 1 is a flowchart of a method for evaluating the partial discharge state of a substation switchgear based on high-dimensional indicators in a preferred embodiment of the present invention. Detailed Embodiment

[0024] The preferred embodiments of the present invention will be introduced below with reference to the accompanying drawings of the specification to make its technical content clearer and easier to understand. The present invention can be embodied in many different forms of embodiments, and the protection scope of the present invention is not limited to the embodiments mentioned in the text.

[0025] The present invention conducts feature monitoring, analysis and evaluation on spatio-temporal data sets based on high-dimensional statistical characteristic analysis and evaluation technology. Through in-depth mining of the statistical information of the operating characteristics of switchgears, it can provide effective monitoring and evaluation means for the metrological characteristics of intelligent switchgears. The present invention is beneficial to improving the efficiency and accuracy of switchgear metrological monitoring, timely and accurately identifying changes in the partial discharge state of switchgears, and is of great significance for carrying out work such as partial discharge monitoring of switchgears and equipment insulation aging protection.

[0026] First, the theoretical basis, data model and algorithm involved in the present invention will be introduced below:

[0027] 1. Collect the phase-resolved partial discharge (PRPD) patterns of the switchgear at T moments, and calculate the statistical characteristic quantities of the patterns. The obtained statistical characteristic quantity vector includes: discharge amplitude q 1 ×T , discharge pulse value f 1×T , skewness λ sk 1×T and kurtosis λ ku 1×T , cross-correlation coefficient λ c 1×T , scale parameter α of the Weibull distribution 1×T and shape parameter β 1×T , to form a spatio-temporal data set matrix X N×T , where N = 7. Further, by introducing tools such as random matrix modeling, eigenvalue (spectrum) analysis, and random matrix splicing, the statistical information of the statistical characteristic quantity spatio-temporal data set of the pattern is deeply mined, and then it is associated with the engineering problems of concern, and the corresponding index system is designed to assist the function design.

[0028] 2. Random matrix theory model

[0029] Analyze the inherent uncertainty and periodicity in data using a random matrix model, and refine relevant metrics for subsequent signal detection, correlation analysis, etc.

[0030] For engineering big data, its spatial dimension N and time dimension T are often not the same. Study the law of large numbers, central limit theorem, etc. of the Wishart matrix and its linear eigenvalue statistics (LES) that conform to such characteristics. Based on the above research, obtain the limit value of LES a priori and evaluate the convergence rate of the empirical spectral density of the data matrix. They can be used as a comparison reference and analysis basis for experimental data, and serve as the starting point for realizing data-driven object cognition.

[0031] Analyze the collected spatio-temporal data using the M-P law (Marchenko–Pastur law) and the circular law, monitor the occurrence of abnormal states, and introduce a linear eigenvalue evaluation system to evaluate the abnormal states.

[0032] M-P law: For an N×T linear unbiased estimation (LUE) matrix, when N, T → ∞ and Its spectral density function satisfies:

[0033]

[0034] where

[0035] Circular law: Consider the product of the singular value equivalent matrices of L independent random matrices where Furthermore, Z can be normalized to Then when N, T → ∞ and N / T = c ∈ (0, 1], The empirical spectral density function of

[0036]

[0037] Develop an analysis algorithm for the data model, extract the statistics of the data model and calculate its theoretical limit, and then extract the statistical information contained in the data by means of hypothesis testing, etc.

[0038] 3. Linear eigenvalue statistic LES

[0039] Give the definition of the linear eigenvalue statistic LES and study its statistical properties - the law of large numbers and central limit theorem of LES.

[0040] Definition:

[0041] Among them, is a continuous testing function, and λ i is the matrix eigenvalue.

[0042] 4. Design of the testing function The design of

[0043] The testing function is the core of optimizing the LES performance, and it is sufficient to satisfy sufficient continuity. Common testing functions are as follows:

[0044] Chebyshev Polynomials T3(x): 4x 3 - 3x.

[0045] Similar to a filter, different can obtain different perception effects to adapt to specific applications, and then construct an effective criterion.

[0046] Next, the method for evaluating the partial discharge state of a substation switchgear based on high-dimensional indicators provided by the present invention will be introduced. As Figure 1 shown, it includes the following steps:

[0047] S1. Collect the partial discharge PRPD pattern data of the substation switchgear at T moments.

[0048] Preferably, in order to improve the granularity of the collected data, the sampling frequency can be 15 min / time.

[0049] S2. Perform spectrogram construction and statistical feature calculation on the obtained partial discharge PRPD pattern data of the switchgear, including the following statistical features of the spectrogram: discharge amplitude, discharge pulse value, skewness and kurtosis, cross-correlation coefficient, shape parameter and scale parameter of the Weibull distribution.

[0050] S3. Preprocess the features calculated from the spectrogram data to construct a spatio-temporal dataset matrix;

[0051] Preferably, the preprocessing methods include missing value filling and outlier cleaning. Through preprocessing, the standardization of the data can be improved.

[0052] S4. Based on the random matrix theory, use the moving sliding window method to sequentially select matrices on the spatio-temporal dataset, and use the Chebyshev polynomial (T3(x): 4x 3 - 3x) to construct a statistic to determine whether there are outliers, calculate the statistic of the matrix framed by the sliding window, and evaluate the state of the substation switchgear.

[0053] By introducing tools such as random matrix models and eigenvalue (spectrum) analysis, the statistical information of the spatio-temporal data set is deeply mined, and then it is associated with the partial discharge state of the switchgear, and a corresponding index system is designed to assist in the function design.

[0054] Specifically, analyze the empirical spectral distribution of the matrix through a random matrix model, and judge whether there are outliers through the M-P law and the circular law.

[0055] Select a switchgear operating state index from the pre-constructed linear eigenvalue index system, and judge whether the index exceeds the threshold when an abnormal state occurs.

[0056] The technical solution of the present invention will be described below with a specific embodiment.

[0057] Step 1: Collect the partial discharge PRPD spectrograms of the switchgear at T moments, perform spectrogram construction and statistical feature calculation on the obtained partial discharge PRPD spectrogram data of the switchgear. The statistical feature quantity vector includes: discharge amplitude q 1×T , discharge pulse value f 1×T , skewness λ sk 1×T and kurtosis λ ku 1×T , cross-correlation coefficient λ c 1×T , scale parameter α of Weibull distribution 1×T and shape parameter β 1×T ;

[0058] Step 2: Stack the statistical feature quantity vectors calculated in Step A spatio-temporal data set matrix X is formed, N×T where N = 7;

[0059] Step 3: Adopt a moving sliding window with a width of t, and frame out X on the spatio-temporal set matrix X in Step N×T and determine the moving window trend; N×t Step 4: Standardize the matrix X obtained in Step It can be transformed into a standardized matrix with a mean of 0 and a variance of 1 through the basic transformation formula

[0060] where, N×t Step 5: Calculate the singular value equivalent matrix of the standardized matrix according to the formula where U is a unitary matrix; Step 6: For the singular value equivalent matrix X

[0061] Step 5: For the standardized matrix According to the formula Find its singular value equivalent matrix, where U is a unitary matrix;

[0062] Step 6: For the singular value equivalent matrix X uPerform unitization processing to obtain a unitization matrix The calculation formula is:

[0063] Step 7: Perform empirical spectral distribution analysis on the unitization matrix through the M-P law and the circular law of the random matrix model to detect whether there are outliers.

[0064] Step 8: Calculate the linear eigenvalue index system of the matrix and monitor and evaluate the partial discharge state of the switchgear in combination with the partial discharge operation state threshold of the switchgear.

[0065] The preferred specific embodiments of the present invention have been described in detail above. It should be understood that those of ordinary skill in the art can make many modifications and variations according to the concept of the present invention without creative labor. Therefore, all technical solutions that can be obtained by those skilled in the art in the technical field based on the concept of the present invention through logical analysis, reasoning, or limited experiments on the basis of the prior art should be within the protection scope determined by the claims.​

Claims

1. A method for evaluating the partial discharge state of a substation switchgear based on high-dimensional indicators, characterized in that, It includes the following steps: Step 1: Collect the partial discharge PRPD pattern data of the substation switchgear cabinet; Step 2: Conduct spectrogram construction and statistical feature calculation on the obtained partial discharge PRPD pattern data, where the statistical features are discharge amplitude, discharge pulse value, skewness, kurtosis, cross-correlation coefficient, shape parameter and scale parameter of Weibull distribution; Step 3: Preprocess the statistical features calculated from the spectrogram data and construct a spatio-temporal dataset matrix; Step 4: Based on the random matrix theory, use the moving sliding window method to sequentially select matrices on the spatio-temporal dataset matrix, construct a statistic using the Chebyshev polynomial T3 to determine whether there are outliers, calculate the statistic of the matrix selected by the sliding window, and evaluate the state of the substation switchgear cabinet.

2. The method for evaluating the partial discharge state of a substation switchgear based on high-dimensional indicators according to claim 1, wherein In Step 1, the sampling frequency for collecting the partial discharge PRPD pattern data of the substation switchgear cabinet is 10 - 20 min / time.

3. The method for evaluating the partial discharge state of a substation switchgear based on high-dimensional indicators according to claim 1, characterized in that In Step 3, the preprocessing method for the statistical features calculated from the spectrogram data includes one or more of missing value filling and outlier cleaning.

4. The method for evaluating the partial discharge state of a substation switchgear based on high-dimensional indicators according to claim 1, characterized in that, In Step 3, constructing the spatio-temporal dataset matrix includes establishing a vector of statistical feature quantities of the spectrogram data and forming the spatio-temporal dataset matrix with the vector of statistical feature quantities.

5. The method for evaluating the partial discharge state of a substation switchgear based on high-dimensional indicators according to claim 1, wherein In Step 4, evaluating the state of the substation switchgear cabinet includes constructing a linear eigenvalue index system for the partial discharge of the switchgear cabinet using the statistic of the selected matrix, and monitoring the operating state of the partial discharge of each switchgear cabinet according to the preset index system.

6. The method for evaluating the partial discharge state of a substation switchgear based on high-dimensional indicators according to claim 5, wherein Monitoring the operating state of the partial discharge of each switchgear cabinet according to the preset index system includes selecting the substation switchgear cabinet state index from the pre-constructed linear eigenvalue index system and determining whether the index exceeds the threshold when an abnormal state occurs.

7. The method for evaluating the partial discharge state of a substation switchgear based on high-dimensional indicators according to claim 1, wherein Step 4 further includes: converting the matrix selected by the sliding window into a standardized matrix with a mean of 0 and a variance of 1, obtaining a singular value equivalent matrix for the standardized matrix, and performing unitary processing.

8. An electronic device, characterized in that, It includes a processor and a memory, and a computer program is stored on the memory. When the computer program is executed by the processor, the method described in any one of claims 1 to 7 is implemented.

9. A readable storage medium, characterized in that, A computer program is stored in the readable storage medium. When the computer program is executed by the processor, the method described in any one of claims 1 to 7 is implemented.

Images