DC power system fault detection method, system and device, and storage medium
By collecting node data from a DC power system to generate column vectors, and then performing standardization processing and eigenvalue analysis, the speed and reliability issues of fault detection in DC power systems are resolved, achieving efficient and accurate fault diagnosis.
Patent Information
- Application Number
- CN202510942666.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-09
- Publication Date
- 2025-10-17
AI Technical Summary
Existing fault detection methods for DC power systems suffer from insufficient speed, susceptibility to transition resistance, and difficulty in balancing economy and reliability in multi-terminal power supply systems. Furthermore, machine learning models are highly complex and require a large amount of training data.
By collecting node data from the DC power system to generate column vectors, standardizing the data, calculating the standard non-Hermitian matrix product and eigenvalues, using the one-loop theorem to determine the eigenvalue distribution, and constructing random variables to determine whether the system is faulty or fault-free.
It enables early detection of subtle changes in fault trends, reduces early warning response time by more than 30%, achieves fault diagnosis accuracy of over 95%, has a false detection rate of less than 2%, and is adaptable to multi-dimensional data environments.
Smart Images
Figure CN120801791A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of fault detection, in particular to a DC power system fault detection method, system, device and storage medium. BACKGROUND
[0002] The fault detection technology of DC power system has attracted widespread attention in recent years, but due to the characteristics of fast rising speed of DC fault current, lack of natural zero point, etc., the traditional method based on the amplitude of electrical quantity (such as overcurrent, overvoltage) has the problems of insufficient speed and being easily affected by transition resistance. The existing technology mainly includes two types: one is the method based on signal processing (such as Fourier transform, wavelet transform), which can capture fault singular points, but depends on manual feature extraction; the other is the fault detection based on communication, which needs synchronous communication equipment at both ends, with high cost and synchronization challenge. In addition, the traditional single-end fault location has inherent defects in multi-terminal power supply system due to insufficient equations, and it is difficult to balance economy and reliability. In recent years, machine learning (such as autoencoder, support vector machine) and deep learning (such as CNN-LSTM) have been introduced into fault detection, but they require a large amount of training data and have high model complexity. SUMMARY
[0003] The purpose of the present application is to provide a DC power system fault detection method, system, device and storage medium, which aims to solve or improve at least one of the above technical problems.
[0004] To achieve the above purpose, the present application provides the following solutions:
[0005] A DC power system fault detection method, comprising:
[0006] Collecting node data in the DC power system and generating a column vector according to time sequence; the node data includes voltage variation and current variation of each node;
[0007] Standardizing the column vector to obtain a standard non-Hermitian matrix;
[0008] Calculating the standard matrix product of the standard non-Hermitian matrix and the corresponding eigenvalue, and determining the eigenvalue distribution of the standard matrix product according to the single ring theorem; the eigenvalue distribution is limited by the inner ring radius and the outer ring radius;
[0009] Based on the eigenvalue distribution and the average spectral radius, a random variable is constructed;
[0010] When the random variable is less than the inner ring radius, it is determined that the system has a fault; when the random variable is greater than or equal to the inner ring radius, it is determined that the system has no fault.
[0011] Optionally, the column vector is normalized to obtain a standard non-Hermitian matrix, and the normalization process specifically includes:
[0012] Based on the column vector, a data matrix is obtained from a data source by using an N w ×T w separation window, and the elements in the data matrix are normalized to obtain a standard non-Hermitian matrix;
[0013] The formula of the normalization process is:
[0014]
[0015] wherein, represents a standard non-Hermitian matrix with a specification of N w ×T w , and respectively represent the mean value and the standard deviation of , and respectively represent the mean value and the standard deviation of , N w represents the number of all measurement nodes, and T w represents the length of a measurement period.
[0016] Optionally, the standard matrix product of the standard non-Hermitian matrix and the corresponding eigenvalue are calculated, and the calculation specifically includes:
[0017] First, the singular value equivalent matrix of the standard non-Hermitian matrix is calculated, then the matrix product corresponding to multiple arbitrary standard non-Hermitian matrices is calculated, and each row of the matrix product is processed to obtain a standard matrix product;
[0018] According to the standard matrix product, eigenvalue calculation is performed to obtain the eigenvalue corresponding to the standard matrix product.
[0019] Optionally, the eigenvalue distribution of the standard matrix product is determined according to the single ring theorem, and the determination specifically includes:
[0020] On the basis of setting the sample data N and the time T, the ratio of the two is calculated as c=N / T, and according to the single ring theorem, it is determined that the eigenvalue of the standard matrix product is distributed between two rings on the complex plane, wherein the inner ring radius is L represents the number of independent variables, and the outer ring radius is 1.
[0021] Optionally, the expression of the random variable is:
[0022]
[0023] wherein, N represents a set sample data; represents an eigenvalue of the i-th standard matrix product.
[0024] The application further provides a direct-current power system fault detection system, comprising:
[0025] a data acquisition module, configured to acquire node data in the direct-current power system and generate a column vector according to a time sequence; the node data comprises voltage variation and current variation of each node;
[0026] a matrix standardization processing module, configured to perform standardization processing on the column vector to obtain a standard non-Hermitian matrix;
[0027] a characteristic distribution calculation module, configured to calculate a standard matrix product of the standard non-Hermitian matrix and a corresponding eigenvalue, and determine an eigenvalue distribution of the standard matrix product according to a single ring theorem; the eigenvalue distribution is limited by an inner ring radius and an outer ring radius;
[0028] a fault diagnosis module, configured to construct a random variable based on the eigenvalue distribution and an average spectral radius; when the random variable is smaller than the inner ring radius, it is determined that a fault occurs in the system; when the random variable is greater than or equal to the inner ring radius, it is determined that no fault occurs in the system.
[0029] The application further provides an electronic device, comprising a memory and a processor, wherein the memory is configured to store a computer program, and the processor is configured to run the computer program to enable the electronic device to execute the direct-current power system fault detection method according to the above.
[0030] The application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the direct-current power system fault detection method as described above.
[0031] According to the embodiments of the application, the following technical effects are achieved:
[0032] The application discloses a DC power system fault detection method, system, device and storage medium, the method comprises collecting node data in a DC power system, and generating a column vector according to a time sequence; the column vector is standardized to obtain a standard non-Hermitian matrix; the standard matrix product of the standard non-Hermitian matrix and the corresponding eigenvalue are calculated, and the eigenvalue distribution of the standard matrix product is determined according to the single ring theorem; the eigenvalue distribution is limited by the inner ring radius and the outer ring radius; a random variable is constructed based on the eigenvalue distribution and the average spectral radius; when the random variable is less than the inner ring radius, it is determined that the system has a fault; when the random variable is greater than or equal to the inner ring radius, it is determined that the system has no fault. The application can quantify the linear change trend of the fault correlation characteristics, and solve the deficiencies of traditional methods in speed, anti-interference and multi-terminal system adaptability. BRIEF DESCRIPTION OF DRAWINGS
[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings described below are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.
[0034] Figure 1 The flowchart of the standard matrix product in the present embodiment is shown. DETAILED DESCRIPTION
[0035] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of the present application.
[0036] The purpose of the present application is to provide a DC power system fault detection method, system, device and storage medium, aiming to solve or improve at least one of the above technical problems.
[0037] In order to make the above purposes, features and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.
[0038] As shown in Figure 1 The present application provides a DC power system fault detection method, comprising:
[0039] Step 100: collecting node data in a DC power system, and generating a column vector according to a time sequence; the node data includes the voltage change and current change of each node.
[0040] Step 200: normalizing the column vector to obtain a standard non-Hermitian matrix.
[0041] Step 300: calculating the standard matrix product of the standard non-Hermitian matrix and the corresponding eigenvalue, and determining the eigenvalue distribution of the standard matrix product according to the single ring theorem; the eigenvalue distribution is defined by an inner ring radius and an outer ring radius.
[0042] Step 400: constructing a random variable based on the eigenvalue distribution and the average spectral radius; when the random variable is less than the inner ring radius, it is determined that the system has a fault; when the random variable is greater than or equal to the inner ring radius, it is determined that the system has no fault.
[0043] As a specific embodiment, the following specific processing procedures of each step are provided.
[0044] First, the voltage variation and the current variation of each node in the direct current power system are selected as the spatial samples of a complex system, and the sample number is n. At the sampling time t i , the measurement data of n variables can form a column vector.
[0045]
[0046] The measurement data vectors of each sampling time are arranged in time sequence to form a matrix.
[0047]
[0048] The matrix is the data source for fault diagnosis of the direct current power system. In the data source Ω, an N w ×T w separation window can be used to obtain the data matrix The elements in are normalized according to the following formula to obtain a standard non-Hermitian matrix
[0049]
[0050] wherein are the mean and standard deviation of respectively; are the mean and standard deviation of respectively, and The singular value equivalent matrix of
[0051]
[0052] wherein
[0053] In order to realize real-time data analysis, a specific separation window can be used to obtain the raw data matrix, which is called real-time separation window. The real-time separation window can obtain the measurement data of the current moment and its historical sampling moments from the data source. At the sampling time t i , the raw data matrix obtained using the real-time separation window is:
[0054]
[0055] in, is the sampling time t j The measurement data vector.
[0056] The linear eigenvalue statistic (LES) can reflect the eigenvalue distribution of a DC power system. For a random matrix X, its linear eigenvalue statistic is defined as:
[0057]
[0058] Among them, λ i for The eigenvalues of For the test function.
[0059] Define the ratio of sample number N and time T as c = N / T. According to the single ring theorem, in the complex plane, The eigenvalues of are roughly distributed between the two rings, and the radius of the inner ring is The outer ring radius is 1.
[0060] The average spectral radius is a linear statistic, which is defined as the average value of the distribution radius of all eigenvalues of the matrix on the complex plane, expressed as:
[0061]
[0062] Among them, λ i (i=1,2,…,N) are all eigenvalues of the matrix, |λ i | is λ i The distribution radius on the complex plane. This embodiment uses the average spectrum radius to describe the standard matrix product. The eigenvalue distribution of :
[0063]
[0064] because Each eigenvalue of is a random matrix The complex function of κ is a highly correlated random variable. MSR is a random variable.
[0065] When κ MSR is less than , then the system fails; otherwise, the system is fault-free.
[0066] Therefore, the present application significantly improves the efficiency and reliability of fault detection through linear feature statistics. The technical effects are embodied in: 1) large-scale data analysis can capture the early trend of small changes in faults, and the response time is shortened by more than 30% compared with the traditional threshold method; 2) through linear feature quantity and combined with circuit topology modeling, the fault diagnosis accuracy can reach more than 95%, especially when the transition resistance is as high as 500Ω, it still maintains stable performance; 3) the system is compatible with multi-dimensional data of power parameters and environmental parameters, and the fault type identification covers typical scenarios such as single ground and double short circuit, with a false detection rate of less than 2%.
[0067] Each embodiment in the specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between each embodiment can be referred to each other.
[0068] The principles and implementation modes of the present application are described by applying specific examples in this paper, and the above description of the embodiments is only used to help understand the core idea of the present application; at the same time, for those skilled in the art, according to the idea of the present application, the specific implementation mode and application range will be changed. In view of the above, the content of the specification should not be understood as a limitation of the present application.
Claims
1. A DC power system fault detection method, characterized in that: include: Collecting node data in a DC power system and generating a column vector according to a time series; the node data includes a voltage change and a current change of each node; Normalizing the column vector to obtain a standard non-Hermitian matrix; Calculating the standard matrix product and the corresponding eigenvalue of the standard non-Hermitian matrix, and determining the eigenvalue distribution of the standard matrix product according to the single ring theorem; the eigenvalue distribution is limited by the inner ring radius and the outer ring radius; constructing a random variable based on the eigenvalue distribution and the average spectral radius; When the random variable is smaller than the inner ring radius, it is determined that the system has failed; when the random variable is greater than or equal to the inner ring radius, it is determined that the system has no failure.
2. The DC power system fault detection method according to claim 1, characterized in that: The step of normalizing the column vector to obtain a standard non-Hermitian matrix specifically includes: Based on the column vector, N w ×T w A separation window is used to obtain a raw data matrix from a data source, and elements in the raw data matrix are standardized to obtain a standard non-Hermitian matrix; The formula for standardization is expressed as: in, Indicates specification N w ×T w The standard non-Hermitian matrix of and Respectively The mean and standard deviation of and Respectively The mean and standard deviation, N w Represents the number of all measurement nodes, T w Indicates the length of the measurement period.
3. The DC power system fault detection method according to claim 1, characterized in that: The calculating of the standard matrix product and the corresponding eigenvalue of the standard non-Hermitian matrix specifically includes: First, a singular value equivalent matrix of the standard non-Hermitian matrix is calculated, and then corresponding matrix products are calculated for a plurality of arbitrary standard non-Hermitian matrices, and then each row of the matrix product is processed to obtain a standard matrix product; Eigenvalues corresponding to the standard matrix product are obtained by performing eigenvalue calculation according to the standard matrix product.
4. The DC power system fault detection method according to claim 1, characterized in that: Determining the eigenvalue distribution of the standard matrix product according to the single ring theorem specifically includes: Based on the set sample data N and time T, the ratio of the two is calculated as c = N / T, and according to the single ring theorem, it is determined that on the complex plane, the eigenvalues of the standard matrix product are distributed between two rings, where the radius of the inner ring is L represents the number of independent variables, and the outer ring radius is 1.
5. The DC power system fault detection method according to claim 1, characterized in that: The expression of the random variable is: Wherein, N represents the set sample data; represents the eigenvalue of the i-th standard matrix product.
6. A DC power system fault detection system, characterized in that: include: A data acquisition module is used to collect node data in the DC power system and generate column vectors according to the time series; the node data includes the voltage change and current change of each node; A matrix normalization processing module, used for normalizing the column vector to obtain a standard non-Hermitian matrix; a characteristic distribution calculation module, configured to calculate the standard matrix product and the corresponding eigenvalue of the standard non-Hermitian matrix, and determine the eigenvalue distribution of the standard matrix product according to the single ring theorem; the eigenvalue distribution is limited by the inner ring radius and the outer ring radius; A fault diagnosis module is used to construct a random variable based on the eigenvalue distribution and the average spectrum radius; when the random variable is smaller than the inner ring radius, it is determined that the system has a fault; when the random variable is greater than or equal to the inner ring radius, it is determined that the system has no fault.
7. An electronic device, characterized in that: The electronic device comprises a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to perform the DC power system fault detection method according to any one of claims 1 to 5.
8. A computer-readable storage medium, characterized in that A computer program is stored therein, and when the computer program is executed by a processor, the DC power system fault detection method according to any one of claims 1 to 5 is implemented.