WO2023216553A1 - 一种配电网多重故障诊断方法及系统 - Google Patents

一种配电网多重故障诊断方法及系统 Download PDF

Info

Publication number
WO2023216553A1
WO2023216553A1 PCT/CN2022/133889 CN2022133889W WO2023216553A1 WO 2023216553 A1 WO2023216553 A1 WO 2023216553A1 CN 2022133889 W CN2022133889 W CN 2022133889W WO 2023216553 A1 WO2023216553 A1 WO 2023216553A1
Authority
WO
WIPO (PCT)
Prior art keywords
fault
distribution network
line
sampling
branch
Prior art date
Application number
PCT/CN2022/133889
Other languages
English (en)
French (fr)
Inventor
罗金满
邹钟璐
戴喜良
翟柱新
刘飘
余凌
姚子汭
高承芳
邓雄荣
刘卓贤
刘丽媛
叶思淇
周彦吉
Original Assignee
广东电网有限责任公司东莞供电局
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 广东电网有限责任公司东莞供电局 filed Critical 广东电网有限责任公司东莞供电局
Publication of WO2023216553A1 publication Critical patent/WO2023216553A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/08—Locating faults in cables, transmission lines, or networks
    • G01R31/081—Locating faults in cables, transmission lines, or networks according to type of conductors
    • G01R31/086—Locating faults in cables, transmission lines, or networks according to type of conductors in power transmission or distribution networks, i.e. with interconnected conductors
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/08—Locating faults in cables, transmission lines, or networks
    • G01R31/088—Aspects of digital computing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
    • Y04S10/52—Outage or fault management, e.g. fault detection or location

Definitions

  • This application relates to the field of distribution network fault diagnosis technology, and specifically relates to a distribution network multiple fault diagnosis method and system.
  • Electric power has become an indispensable and important resource in human life. Once the power grid fails during operation, it will bring huge losses to social production and people's lives. Therefore, when the power grid fails, necessary measures should be taken to quickly and accurately Locate the area where the fault occurs, find out the specific fault line, and improve the speed and detection accuracy of restoring power supply after a power grid failure.
  • the distribution network fault detection system also has the following defects:
  • the signal acquisition devices and other protection devices in the actual distribution network are easily affected by man-made or mutual influence between various electromagnetic equipment, resulting in many uncertain factors in the distribution network.
  • the fault detection method is based on the dynamic changes of the line. When the fault occurs, the fault cannot be located and detected in time. When the fault information contains misoperation information, the fault tolerance is poor.
  • the purpose of this application is to provide a multiple fault diagnosis method and system for a distribution network to solve the problem of inability to detect faults in a timely manner when dynamic changes in lines occur in related technologies.
  • the fault information contains misoperation information, the fault tolerance is relatively low. Poor, tedious technical issues in topology planning.
  • a multiple fault diagnosis method for distribution network including the following steps:
  • Step S1 Use the MATLAB platform to conduct short-circuit fault analysis on the distribution network line, and obtain a fault information decision table based on electrical quantities;
  • Step S2 Use the Simulink platform to model and simulate the fault information decision table based on electrical quantities, and obtain the output training data of the distribution network line neural network;
  • Step S3 Use the wavelet transform method to denoise and collect the output training data of the neural network to form a relevant fault information decision table as a training sample for the neural network;
  • Step S4 Use the improved Artificial Tree intelligent optimization algorithm to optimize the weights and thresholds of the neural network, achieve the relevant error accuracy requirements through iteration, select part of the data as fault data and use the trained neural network to detect faults.
  • step S1 the MATLAB platform collects the zero-sequence current, zero-sequence power and zero-sequence admittance signals generated after a small current grounding short-circuit fault occurs on the distribution network line. Based on these three The information forms a fault information decision table based on electrical quantities.
  • fault analysis is performed based on the zero-sequence current, zero-sequence power and zero-sequence admittance signals using a sampling mode that mixes discrete sampling time and continuous sampling time.
  • the specific steps are as follows:
  • Step S201 Sample and store the positive and negative zero-sequence currents at one of the two terminals of each line;
  • Step S202 Select the sampling current within a fixed-length time window as the line reference current, compare the dynamic sampling current with the line reference current, and use the improved Pearson method to solve the correlation coefficient between the two in real time;
  • Step S203 Set the correlation coefficient threshold, determine the size of the correlation coefficient and the set threshold, and output the line where the fault is located and the fault type.
  • the improved Pearson method uses the accumulated value of the line correlation coefficient to perform correlation coefficient analysis for fault detection.
  • the specific steps of the improved Pearson method are as follows:
  • Step S2021 Define movable window values T det and T loc within a time window with a fixed length.
  • the fixed end times of the T det and T loc windows are t 1 and t 2 respectively.
  • the sampling current in T det is Compare it with the steady-state reference current to detect faults in the DC microgrid.
  • the relationship between T det and T loc is:
  • f s is the sampling frequency of the line current
  • n s is the number of sampling points
  • T 0 is the time from the time when the fault occurs t 0 to the time when the fault is detected t 2 ;
  • Step S2022 Use the accumulated value of the line correlation coefficient to perform correlation coefficient analysis for fault detection, and express the i-th comparison value of the steady-state reference current curve and the sampling current curve as P 1_i and P 2_i respectively:
  • I ref_j is the i-th transient line current
  • I sam_j is the i-th sampling line current
  • Step S2023 Introduce the adjustment factor p to optimize the Pearson correlation coefficient.
  • the expression of the adjustment factor p is:
  • e is a mathematical constant, expressed as the base of the natural logarithm, and n is the total number of contrast points between the two curves;
  • Step S2023 Optimize the correlation coefficient r according to the adjustment factor p.
  • the expression of the correlation coefficient r is:
  • the fault location and fault resistance are obtained based on the fault analysis and a model is constructed through the Simulink platform to output training data, and the wavelet transform method is used to process the training data.
  • Step S301 Perform wavelet transformation on the noisy training data signal f(t) to obtain a set of wavelet decomposition coefficients
  • Step S302 The sampling soft threshold function performs threshold processing on the obtained wavelet decomposition coefficients to obtain a set of estimated wavelet coefficients.
  • the expression of the soft threshold function ⁇ ( ⁇ ) is:
  • sign is the sign function
  • ⁇ is the wavelet coefficient
  • T is the threshold
  • ⁇ is the noise intensity
  • m is the signal length
  • d is the amplitude factor
  • a and b are adjustment factors
  • Step S303 Extract the denoised signal by adjusting d to make it continuous at the threshold point T and reduce the oscillation error caused to the original signal.
  • step S4 the mean square error of the oscillation error is optimized using the Artificial Tree algorithm to obtain the line neural network.
  • the Artificial Tree algorithm optimization steps are as follows:
  • Step S401 Using the mean square error MSE of the oscillation error as the objective function, organize the collected data and prepare training samples and test samples;
  • Step S402 Set the relevant parameters of the Artificial Tree algorithm, determine the number of branches SN, the spatial dimension D, and the number of iterations t max , and perform initialization operations on the branches to generate initial branches;
  • Step S402 Calculate the function value corresponding to each branch, and select the best branch x best and the function value f(x best ) corresponding to the best branch according to the function value;
  • Step S403 Determine whether the error accuracy requirements are met or whether the maximum number of iterations is reached. If neither meets the requirements, prepare to enter the iterative optimization process of the IAT algorithm;
  • Step S404 If the number of searches reaches the upper limit of the number of searches, a random operation will be performed to generate a new branch, and the newly generated branch will be compared with the old branch. If the new branch is better than the old branch, the new branch will replace the old branch. branch, otherwise the old branch will still be the optimal branch;
  • Step S405 Determine again whether the optimal branch meets the error accuracy or whether the current number of iterations has reached the set maximum number of iterations t max .
  • the relevant parameters of the Artificial Tree algorithm form corresponding neural network training samples based on the action information of the distribution network line protection device and the corresponding fault area settings.
  • the action information of the distribution network line protection device mainly collects power grid signals based on switching values, and uses 0 and 1 to number the power grid switching actions.
  • a system for multiple fault diagnosis methods of distribution network including: circuit sampling module, line protection device, fault analysis module, simulation module and detection output module
  • the circuit sampling module is used to collect and preprocess various data imported to the MATLAB platform and Simulink platform;
  • the line protection device is used to perform corresponding protection actions such as tripping or alarming when a fault occurs in the distribution network. Based on the different protection actions of the protection device, the approximate interval in which the fault occurs is determined, and the intervals in the power grid are numbered;
  • the fault analysis module is used to analyze the location of the fault and the resistance value of the fault resistor, and determine the approximate area of the fault;
  • the simulation module uses the Simulink platform to obtain fault detection data through the Artificial Tree intelligent optimization algorithm and displays the simulation results.
  • the detection output module selects a part of the data to detect the line neural network based on the simulation results to analyze the adaptability report of the accuracy of power grid fault detection.
  • This application uses the MATLAB platform to use the improved wavelet threshold denoising method to remove noise sources in the signal to establish a power grid fault model, improve the accuracy of signal collection, run simulation software for simulation, and establish a power grid fault model based on switching quantities to obtain power grid fault information.
  • Decision tables train neural networks, detect fault occurrence intervals, and adopt a divide-and-conquer strategy for complex power grids to make the line topology flexible and changeable. This not only improves the accuracy and reliability of power grid fault diagnosis, but also improves the accuracy and reliability of power grid fault diagnosis when fault information exists. The fault is detected in the case of misoperation information, thereby achieving the effect of locating the power grid fault area and effectively enhancing the power grid fault detection capability.
  • Figure 1 is a flow chart of a multiple fault diagnosis method for a distribution network provided by an embodiment of the present application
  • Figure 2 is a schematic diagram of the dynamic fault detection timeline provided by the embodiment of the present application.
  • Figure 3 is a structural block diagram of a fault detection system provided by an embodiment of the present application.
  • 1-Circuit sampling module 2-Line protection device; 3-Fault analysis module; 4-Simulation module; 5-Detection output module.
  • this application provides a multiple fault diagnosis method for distribution network, including the following steps:
  • Step S1 Use the MATLAB platform to conduct short-circuit fault analysis on the distribution network line, and obtain a fault information decision table based on electrical quantities;
  • Step S2 Use the Simulink platform to model and simulate the fault information decision table based on electrical quantities, and obtain the output training data of the distribution network line neural network;
  • Step S3 Use the wavelet transform method to denoise and collect the output training data of the neural network to form a relevant fault information decision table as a training sample for the neural network;
  • Step S4 Use the improved Artificial Tree intelligent optimization algorithm to optimize the weights and thresholds of the neural network, achieve the relevant error accuracy requirements through iteration, select part of the data as fault data and use the trained neural network to detect faults.
  • a power grid fault model is established based on the distribution network line data, and the required neural network input training data is collected. While collecting information, wavelet transform theory is used to denoise the collected data to remove interference to the accuracy of the data. noise source, and then collect and process the denoised data to form a relevant fault information decision table as a training sample for the neural network, and then apply the Artificial Tree intelligent optimization algorithm to optimize the weights and thresholds of the neural network, and achieve the relevant error through iteration For accuracy requirements, select part of the data as fault data and use the trained neural network to detect faults.
  • step S1 the MATLAB platform collects the zero-sequence current, zero-sequence power and zero-sequence admittance signals generated after a small current grounding short-circuit fault occurs on the distribution network line, and forms fault information based on electrical quantities based on these three types of information. Decision table.
  • the action information of the protection device is used as the input value of the neural network, and the corresponding fault area is used as the output value to form the corresponding neural network training sample, which is used to Build a detection model.
  • fault analysis is performed using a sampling mode that mixes discrete sampling time and continuous sampling time. The specific steps are as follows:
  • Step S201 Sample and store the positive and negative zero-sequence currents at one of the two terminals of each line;
  • Step S202 Select the sampling current within a fixed-length time window as the line reference current, compare the dynamic sampling current with the line reference current, and use the improved Pearson method to solve the correlation coefficient between the two in real time;
  • Step S203 Set the correlation coefficient threshold, determine the size of the correlation coefficient and the set threshold, and output the line where the fault is located and the fault type.
  • the line is considered to have a polar short circuit fault; if the positive and negative poles of the line If the current correlation degree is lower than the threshold and the current sum of the two poles in the line is approximately zero, then the line is considered to have an inter-pole short circuit fault; otherwise, the line is considered to have no fault.
  • power grid fault protection consists of three parts: fault detection, fault isolation and fault location, as shown in Figure 2: Assume that a short-circuit fault occurs at time t 1 , and the current of the fault line increases rapidly. The fault detection equipment will detect the fault at a later time, which is defined as t 2 . Starting from time t 2 , fault isolation and fault location are performed simultaneously: on the one hand, the fault isolation equipment starts to act and succeeds at the following time t 3 Break the faulty line; on the other hand, the fault location equipment begins to predict the location of the fault, and gives the predicted location of the fault at time t 4. By analyzing the time sequence of each part of the fault protection, it can be found:
  • Fault location is started simultaneously with fault isolation after the fault is detected, and the two processes of fault location and fault isolation are decoupled from each other, so the system does not have high requirements for the rapidity of fault location;
  • the improved Pearson method uses the accumulated value of the line correlation coefficient to perform correlation coefficient analysis for fault detection.
  • the specific steps of the improved Pearson method are as follows:
  • Step S2021 Define movable window values T det and T loc within a time window with a fixed length.
  • the fixed end times of the T det and T loc windows are t 1 and t 2 respectively.
  • the sampling current in T det is Compare it with the steady-state reference current to detect faults in the DC microgrid.
  • the relationship between T det and T loc is:
  • f s is the sampling frequency of the line current
  • n s is the number of sampling points
  • T 0 is the time from the time when the fault occurs t 0 to the time when the fault is detected t 2 ;
  • Step S2022 Use the accumulated value of the line correlation coefficient to perform correlation coefficient analysis for fault detection, and express the i-th comparison value of the steady-state reference current curve and the sampling current curve as P 1_i and P 2_i respectively:
  • I ref_j is the i-th transient line current
  • I sam_j is the i-th sampling line current
  • Step S2023 Introduce the adjustment factor p to optimize the Pearson correlation coefficient.
  • the expression of the adjustment factor p is:
  • e is a mathematical constant, expressed as the base of natural logarithms, and n is the total number of contrast points between the two curves;
  • Step S2023 Optimize the correlation coefficient r according to the adjustment factor p.
  • the expression of the correlation coefficient r is:
  • step S2022 if the sampling current and the steady-state calculated current are directly compared, the correlation coefficient will change with the change of the current ripple, and a large number of false detections may occur.
  • the The cumulative value of the line is used for correlation coefficient analysis of fault detection to eliminate the interference caused by this phenomenon, which leads to low fault location accuracy.
  • the sampling current and the steady-state reference current are compared during fault detection, and the improved Pearson method is used to solve the correlation coefficient between the two in real time.
  • the correlation coefficient drops below the set threshold, a short-circuit fault is considered to have occurred, and Output the line where the fault is located and the fault type;
  • an iterative method based on a genetic algorithm is used in fault location to generate and update the predicted fault location and fault impedance.
  • the fault current curve corresponding to the predicted fault location and fault impedance is calculated, and the improved Pearson method is used to solve the problem.
  • the magnitude of the correlation coefficient between the sampling curve and the calculated curve is used to solve the problem.
  • step S2 the fault location and fault resistance are obtained based on the fault analysis, a model is constructed through the Simulink platform, and training data is output, and the wavelet transform method is used to process the training data.
  • Step S301 Perform wavelet transformation on the noisy training data signal f(t) to obtain a set of wavelet decomposition coefficients
  • Step S302 The sampling soft threshold function performs threshold processing on the obtained wavelet decomposition coefficients to obtain a set of estimated wavelet coefficients.
  • the expression of the soft threshold function ⁇ ( ⁇ ) is:
  • sign is the sign function
  • ⁇ is the wavelet coefficient
  • T is the threshold
  • ⁇ is the noise intensity
  • m is the signal length
  • d is the amplitude factor
  • a and b are adjustment factors
  • Step S303 Extract the denoised signal by adjusting d to make it continuous at the threshold point T and reduce the oscillation error caused to the original signal.
  • the value of the wavelet coefficient caused by the noise source will be less than the preset critical threshold, and the wavelet coefficient caused by the original signal will be greater than the preset critical threshold.
  • the wavelet coefficient is smaller than the critical threshold.
  • the original signal whose wavelet coefficient is greater than the critical threshold is retained, and then the inverse wavelet transform is performed to reconstruct the signal to achieve the purpose of denoising.
  • the flexibility of using the threshold function is enhanced by adjusting the values of a and b.
  • the parameter d determines the degree of approximation of the threshold. By adjusting d, it is continuous at the threshold point T and reduces the oscillation error caused to the original signal, so that The extracted original signal after denoising is more accurate and more credible, thereby improving the accuracy of fault signal denoising in power grid fault detection, thus improving the accuracy of power grid fault detection.
  • step S4 the mean square error of the oscillation error is optimized using the Artificial Tree algorithm to obtain the line neural network.
  • the Artificial Tree algorithm optimization steps are as follows:
  • Step S401 Using the mean square error MSE of the oscillation error as the objective function, organize the collected data and prepare training samples and test samples;
  • Step S402 Set the relevant parameters of the Artificial Tree algorithm, determine the number of branches SN, the spatial dimension D, and the number of iterations t max , and perform initialization operations on the branches to generate initial branches;
  • Step S402 Calculate the function value corresponding to each branch, and select the best branch x best and the function value f(x best ) corresponding to the best branch according to the function value;
  • Step S403 Determine whether the error accuracy requirements are met or whether the maximum number of iterations is reached. If neither meets the requirements, prepare to enter the iterative optimization process of the IAT algorithm;
  • Step S404 If the number of searches reaches the upper limit of the number of searches, a random operation will be performed to generate a new branch, and the newly generated branch will be compared with the old branch. If the new branch is better than the old branch, the new branch will replace the old branch. branch, otherwise the old branch will still be the optimal branch;
  • Step S405 Determine again whether the optimal branch meets the error accuracy or whether the current number of iterations has reached the set maximum number of iterations t max .
  • the relevant parameters of the Artificial Tree algorithm form corresponding neural network training samples based on the action information of the distribution network line protection device and the corresponding fault area settings.
  • the action information of the distribution network line protection device mainly collects power grid signals based on switching values, and uses 0 and 1 to number the power grid switching actions.
  • a system for multiple fault diagnosis methods of distribution network including: circuit sampling module 1, line protection device 2, fault analysis module 3, simulation module 4 and detection output module 5
  • the circuit sampling module 1 is used to collect and preprocess various data imported to the MATLAB platform and Simulink platform;
  • the line protection device 2 is used to perform corresponding protection actions such as tripping or alarming when a fault occurs in the distribution network. Based on the different protection actions of the protection device, it determines the approximate interval in which the fault occurs, and numbers the intervals in the power grid;
  • the fault analysis module 3 is used to analyze the location of the fault and the resistance value of the fault resistor, and determine the approximate area of the fault;
  • the simulation module 4 uses the Simulink platform to obtain fault detection data through the Artificial Tree intelligent optimization algorithm and displays the simulation results.
  • the detection output module 5 selects a partial data detection line neural network based on the simulation results to report on the accuracy of the power grid fault detection adaptability analysis.
  • This application uses the MATLAB platform to use the improved wavelet threshold denoising method to remove noise sources in the signal to establish a power grid fault model, improve the accuracy of signal collection, run simulation software for simulation, and establish a power grid fault model based on switching quantities to obtain power grid fault information.
  • Decision tables train neural networks, detect fault occurrence intervals, and adopt a divide-and-conquer strategy for complex power grids to make the line topology flexible and changeable. This not only improves the accuracy and reliability of power grid fault diagnosis, but also improves the accuracy and reliability of power grid fault diagnosis when fault information exists. The fault is detected in the case of misoperation information, thereby achieving the effect of locating the power grid fault area and effectively enhancing the power grid fault detection capability.

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Computer Hardware Design (AREA)
  • Evolutionary Computation (AREA)
  • Geometry (AREA)
  • General Engineering & Computer Science (AREA)
  • Locating Faults (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

一种配电网多重故障诊断方法及系统，方法包括：S1、利用MATLAB平台对配电网线路进行短路故障分析，得出基于电气量的故障信息决策表；S2、利用Simulink平台对故障信息决策表进行建模仿真；S3、采用小波变换法对神经网络的输出训练数据进行去噪、采集处理，形成相关故障信息决策表作为神经网络的训练样本；S4、采用改进的Artificial Tree智能优化算法优化神经网络的权值和阈值，选择部分数据作为故障数据将训练好的神经网络用来检测故障，不仅能提高电网故障诊断准确性与可靠性，还可在故障信息存在误动时对故障进行检测，从而达到定位电网故障区域的效果，有效增强了电网故障检测的能力。

Description

一种配电网多重故障诊断方法及系统
本申请要求在2022年05月11日提交中国专利局、申请号为202210506365.6的中国专利申请的优先权，该申请的全部内容通过引用结合在本申请中。
技术领域
本申请涉及配电网故障诊断技术领域，具体涉及一种配电网多重故障诊断方法及系统。
背景技术
电力能源已成为人类生活中不可缺少的重要资源，在电网运行时一旦出现故障将会给社会生产和人民生活带来巨大的损失，因此当电网发生故障时，应采取必要的措施，快速准确地定位故障发生的区域，找出具体的故障线路，提高电网故障后恢复供电的速度和检测精度。
配电网故障检测系统还存在以下缺陷：
(1)实际配电网中的采集信号装置和其他保护装置容易受到人为或者各种电磁设备之间产生的相互影响，使得配电网中存在着很多不确定因素，故障检测方法针对线路动态变化时，无法及时对故障定位检测，在故障信息存在误动信息的情况下，容错能力较差。
(2)针对故障检测的拓扑结构规划完成后，当遇到实际电网的结构和相关保护器件发生改变时，将很难对网络的拓扑结构进行改变，而是需要重新进行设计，会使得实际操作非常繁琐。
发明内容
本申请的目的在于提供一种配电网多重故障诊断方法及系统，以解决相关技术中针对线路动态变化时，无法及时对故障定位检测，在故障信息存在误动信息的情况下，容错能力较差，拓扑结构规划繁琐的技术问题。
为解决上述技术问题，本申请具体提供下述技术方案：
一种配电网多重故障诊断方法，包括以下步骤：
步骤S1、利用MATLAB平台对所述配电网线路进行短路故障分析，得出基于电气量的故障信息决策表；
步骤S2、利用Simulink平台对所述基于电气量的故障信息决策表进行建模 仿真，获取配电网线路神经网络的输出训练数据；
步骤S3、采用小波变换法对所述神经网络的输出训练数据进行去噪、采集处理，形成相关故障信息决策表作为神经网络的训练样本；
步骤S4、采用改进的Artificial Tree智能优化算法优化神经网络的权值和阈值，通过迭代达到相关误差精度的要求，选择部分数据作为故障数据并利用训练好的神经网络来检测故障。
作为本申请的一种方案，所述步骤S1中，所述MATLAB平台采集配电网线路发生小电流接地短路故障后产生的零序电流、零序功率和零序导纳信号，基于这三种信息形成基于电气量的故障信息决策表。
作为本申请的一种方案，依据所述零序电流、零序功率和零序导纳信号用离散采样时间和连续采样时间相混合的采样模式进行故障分析，具体步骤如下：
步骤S201、在每条线路两个终端中的一个终端上采样并存储正负极零序电流；
步骤S202、选择固定长度的时间窗内的采样电流作为线路参考电流，对动态采样电流与线路参考电流进行比较，使用改进Pearson方法实时求解二者的相关系数大小；
步骤S203、设定相关系数阈值，判断相关系数与所述设定阈值的大小，并输出故障所在线路及故障类型。
作为本申请的一种方案，所述步骤S1中，所述改进Pearson方法采用线路相关系数的累加值进行故障检测的相关系数分析，所述改进Pearson方法具体步骤如下：
步骤S2021、定义具有固定长度的时间窗内可移动窗口值T det、T loc，所述T det、T loc窗口的末端固定时刻分别为t 1，t 2，将所述T det中的采样电流与稳态参考电流进行比较，检测直流微电网的故障，所述T det和T loc的关系式为：
n s/f s≤T loc≤T 0≤T det
其中f s为线路电流的采样频率，n s为采样点的数量，T 0为从故障发生时刻t 0到故障被检测到时刻t 2之间的时长；
步骤S2022、采用线路相关系数的累加值进行故障检测的相关系数分析，将稳态参考电流曲线和采样电流曲线的第i个对比值分别表示为P 1_i、P 2_i：
Figure PCTCN2022133889-appb-000001
其中I ref_j为第i个暂态线路电流，I sam_j为第i个采样线路电流；
步骤S2023、引入调整因子p优化Pearson相关系数调整因子p表达式为：
Figure PCTCN2022133889-appb-000002
其中，e为数学常数，表示为自然对数的底数，n为两条曲线所对比点的总数；
步骤S2023、依据所述调整因子p优化相关系数r，所述相关系数r的表达式为：
r＝p×r 0
Figure PCTCN2022133889-appb-000003
作为本申请的一种方案，所述步骤S2中，依据所述故障分析获取故障位置和故障电阻通过Simulink平台构建模型输出训练数据，采用小波变换法处理训练数据。
作为本申请的一种方案，所述小波变换法处理训练数据具体步骤为：
步骤S301、对含噪声的训练数据信号f(t)作小波变换，获取一组小波分解系数；
步骤S302、采样软阈值函数对得到的小波分解系数进行阈值处理，得到一组估计小波系数，所述软阈值函数υ(ω)表达式为：
Figure PCTCN2022133889-appb-000004
Figure PCTCN2022133889-appb-000005
其中，sign为符号函数，ω为小波系数，T为阈值，σ为噪声强度，m为信号长度，d为幅值因子，a、b为调节因子；
步骤S303、通过调节d使得在阈值点T处连续且降低对原信号产生的震荡误差，提取去噪后的信号。
作为本申请的一种方案，在步骤S4中，将所述震荡误差的均方误差采用Artificial Tree算法优化获取线路神经网络，所述Artificial Tree算法优化步骤如下：
步骤S401、以所述震荡误差的均方误差MSE作为目标函数，将收集到的数据进行整理，准备好训练样本和测试样本；
步骤S402、设置Artificial Tree算法的相关参数，确定分支个数SN、空间维数D、迭代次数t max，并对分支进行初始化操作，生成初始分支；
步骤S402、计算每个分支所对应的函数值，根据函数值选出最好分支x best和最好分支所对应的函数值f(x best)；
步骤S403、判断是否满足误差精度的要求或者是否达到了最大迭代次数，如果都不满足要求则准备进入IAT算法的迭代寻优过程；
步骤S404、如果搜索次数达到了搜索次数的上限，将进行随机操作生成新的分支，将新产生的分支与旧的分支进行比较，如果新的分支优于旧的分支则新的分支将取代旧分支，否则仍以旧的分支为最优分支；
步骤S405、再次判断最优分支是否满足误差精度或者当前迭代次数是否达到了设定的最大迭代次数t max。
作为本申请的一种方案，所述Artificial Tree算法的相关参数依据配电网线路保护装置的动作信息及对应发生故障的区域设定，形成相应的神经网络训练样本。
作为本申请的一种方案，所述配电网线路保护装置的动作信息主要采集基于开关量的电网信号，用0、1对电网开关动作进行编号。
一种配电网多重故障诊断方法的系统，包括：电路采样模块、线路保护装置、故障分析模块、仿真模块以及检测输出模块
所述电路采样模块，用于采集和预处理导入到所述MATLAB平台和Simulink平台的各项数据；
所述线路保护装置，用于配电网发生故障时会进行跳闸或者报警等相应的保护动作，根据保护装置不同的保护动作，判断故障发生的大致区间，并且将电网的区间进行编号；
所述故障分析模块，用于分析故障发生位置及故障电阻阻值，判断故障大致区域；
所述仿真模块，利用Simulink平台通过Artificial Tree智能优化算法获取故障检测数据，显示仿真结果。
所述检测输出模块，依据仿真结果选用部分数据检测线路神经网络对电网故障检测的准确率分析适应性报告。
本申请与相关技术相比较具有如下有益效果：
本申请利用MATLAB平台用改进后的小波阈值去噪方法去除信号中的噪声源建立电网故障模型，提高信号采集准确度，运行仿真软件进行仿真，并建立基于开关量的电网故障模型得到电网故障信息决策表，训练神经网络，检测故障发生的区间，对复杂电网采取分而治之的策略，使得线路拓扑结构能够灵活多变，这样不仅能提高电网故障诊断的准确性与可靠性，还可以在故障信息存在误动信息的情况下对故障进行检测，从而达到电网故障区域定位的效果，有效的增强了电网故障检测的能力。
附图说明
图1为本申请实施例提供的配电网多重故障诊断方法流程图；
图2为本申请实施例提供的动态故障检测时间轴示意图；
图3为本申请实施例提供的故障检测系统结构框图。
图中的标号分别表示如下：
1-电路采样模块；2-线路保护装置；3-故障分析模块；4-仿真模块；5-检测输出模块。
具体实施方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，所描述的实施例仅仅是本申请一部分实施例，而不是全部的实施例。
如图1-3所示，本申请提供了一种配电网多重故障诊断方法，包括以下步骤：
步骤S1、利用MATLAB平台对所述配电网线路进行短路故障分析，得出基于电气量的故障信息决策表；
步骤S2、利用Simulink平台对所述基于电气量的故障信息决策表进行建模仿真，获取配电网线路神经网络的输出训练数据；
步骤S3、采用小波变换法对所述神经网络的输出训练数据进行去噪、采集处理，形成相关故障信息决策表作为神经网络的训练样本；
步骤S4、采用改进的Artificial Tree智能优化算法优化神经网络的权值和阈值，通过迭代达到相关误差精度的要求，选择部分数据作为故障数据并利用训练好的神经网络来检测故障。
本实施例中，根据配电网线路数据建立电网故障模型，收集所需要的神经网络输入训练的数据，在收集信息的同时采用小波变换理论对收集的数据进行去噪处理，去除干扰数据准确性的噪声源，然后将去噪后的数据进行采集处理，形成相关故障信息决策表作为神经网络的训练样本，然后应用Artificial Tree智能优化算法来优化神经网络的权值和阈值，通过迭代达到相关误差精度的要求，选择部分数据作为故障数据将训练好的神经网络用来检测故障。
所述步骤S1中，所述MATLAB平台采集配电网线路发生小电流接地短路故障后产生的零序电流、零序功率和零序导纳信号，基于这三种信息形成基于电气量的故障信息决策表。
本实施例中，在应用神经网络解决电网故障检测问题时，将保护装置的动作信息作为神经网络的输入值，将对应发生故障的区域作为输出值，形成相应的神经网络训练样本，从而用来建立检测模型。
依据所述零序电流、零序功率和零序导纳信号用离散采样时间和连续采样时间相混合的采样模式进行故障分析，具体步骤如下：
步骤S201、在每条线路两个终端中的一个终端上采样并存储正负极零序电流；
步骤S202、选择固定长度的时间窗内的采样电流作为线路参考电流，对动态采样电流与线路参考电流进行比较，使用改进Pearson方法实时求解二者的相关系数大小；
步骤S203、设定相关系数阈值，判断相关系数与所述设定阈值的大小，并输出故障所在线路及故障类型。
本实施例中，如果正负两极中的一个极的电流相关程度低于阈值，且该线路中两个极点的电流和不为零，则认为该线路发生极地短路故障；如果线路正负极的电流相关程度均低于阈值，且该线路中两个极点的电流和近似为零，则认为该线路发生极间短路故障；否则，认为该线路没有发生故障。
本实施例中，电网故障保护由故障检测、故障隔离和故障定位三部分组成，如图2所示：假定在t 1时刻发生短路故障，故障线路的电流迅速增大。故障检测设备会在之后的某时刻检测到故障，定义该时刻为t 2，从t 2时刻开始，故障隔离和故障定位同步进行：一方面故障隔离设备开始动作，并于之后的t 3时刻成功开断故障线路；另一方面故障定位设备开始预测故障的发生位置，并于之后的t 4时 刻给出预测的故障发生位置，分析故障保护中各部分的时间顺序，可以发现：
一、故障隔离需要故障被检测到后才会动作，由于直流微电网故障隔离需要在几个毫秒内完成，因此系统对故障检测的快速性要求很高；
二、故障定位是在故障被检测到后，与故障隔离同步开始进行，且故障定位与故障隔离两个过程相互解耦，因此系统对故障定位的快速性要求不高；
三、由于故障隔离设备动作后线路电流变化与故障隔离设备的动作情况相关且难以预测，因此只有故障被检测到前的系统状态量是可供故障检测与故障定位的优质数据。
所述步骤S202中，所述改进Pearson方法采用线路相关系数的累加值进行故障检测的相关系数分析，所述改进Pearson方法具体步骤如下：
步骤S2021、定义具有固定长度的时间窗内可移动窗口值T det、T loc，所述T det、T loc窗口的末端固定时刻分别为t 1，t 2，将所述T det中的采样电流与稳态参考电流进行比较，检测直流微电网的故障，所述T det和T loc的关系式为：
n s/f s≤T loc≤T 0≤T det
其中f s为线路电流的采样频率，n s为采样点的数量，T 0为从故障发生时刻t 0到故障被检测到时刻t 2之间的时长；
步骤S2022、采用线路相关系数的累加值进行故障检测的相关系数分析，将稳态参考电流曲线和采样电流曲线的第i个对比值分别表示为P 1_i、P 2_i：
Figure PCTCN2022133889-appb-000006
其中I ref_j为第i个暂态线路电流，I sam_j为第i个采样线路电流；
步骤S2023、引入调整因子p优化Pearson相关系数调整因子p表达式为：
Figure PCTCN2022133889-appb-000007
其中，e为数学常数，表示为自然对数的底数，n为两条曲线所对比点的总 数；
步骤S2023、依据所述调整因子p优化相关系数r，所述相关系数r的表达式为：
r＝p×r 0
Figure PCTCN2022133889-appb-000008
本实施例中，在步骤S2022中，如果直接对采样电流和稳态计算电流进行对比，相关系数将随电流纹波的变化而变化，可能出现大量的误检测，为防止误检测的发生，采用线路的累加值进行故障检测的相关系数分析，消除由这种现象引起的干扰，导致的故障定位准确性不高的现象。
本实施例中，在故障检测中对采样电流与稳态参考电流进行比较，使用改进Pearson方法实时求解二者的相关系数大小，当相关系数降低到设定阈值以下时，认为发生短路故障，并输出故障所在线路及故障类型；
本实施例中，在故障定位中采用基于遗传算法的迭代方法来实现所预测故障位置和故障阻抗的生成与更新，计算所预测故障位置和故障阻抗对应的故障电流曲线，并采用改进Pearson方法求解采样曲线与所计算曲线的相关系数大小，当相关系数升高到设定阈值以上时，认为所预测故障位置和故障阻抗足够准确，并输出预测的故障位置和故障阻抗。
所述步骤S2中，依据所述故障分析获取故障位置和故障电阻通过Simulink平台构建模型输出训练数据，采用小波变换法处理训练数据。
所述小波变换法处理训练数据具体步骤为：
步骤S301、对含噪声的训练数据信号f(t)作小波变换，获取一组小波分解系数；
步骤S302、采样软阈值函数对得到的小波分解系数进行阈值处理，得到一组估计小波系数，所述软阈值函数υ(ω)表达式为：
Figure PCTCN2022133889-appb-000009
Figure PCTCN2022133889-appb-000010
其中，sign为符号函数，ω为小波系数，T为阈值，σ为噪声强度，m为信号长度，d为幅值因子，a、b为调节因子；
步骤S303、通过调节d使得在阈值点T处连续且降低对原信号产生的震荡误 差，提取去噪后的信号。
本实施例中，噪声源引起的小波系数的值将会小于预先设置的临界阈值，而原始信号所引起的小波系数将会大于预先设置的临界阈值，利用这个特性，将小波系数小于临界阈值的噪声信号去除，将小波系数大于临界阈值的原始信号保留，然后进行小波逆变换对信号进行重构，达到去噪的目的。
本实施例中，通过调节a、b的值来增强阈值函数使用的灵活性，参数d决定阈值的逼近程度，通过调节d使得在阈值点T处连续且降低对原信号产生的震荡误差，使得提取到的去噪后的原始信号更为准确，可信度更高，由此提高了电网故障检测中对于故障信号去噪的精度，从而提高了电网故障检测的准确性。
在步骤S4中，将所述震荡误差的均方误差采用Artificial Tree算法优化获取线路神经网络，所述Artificial Tree算法优化步骤如下：
步骤S401、以所述震荡误差的均方误差MSE作为目标函数，将收集到的数据进行整理，准备好训练样本和测试样本；
步骤S402、设置Artificial Tree算法的相关参数，确定分支个数SN、空间维数D、迭代次数t max，并对分支进行初始化操作，生成初始分支；
步骤S402、计算每个分支所对应的函数值，根据函数值选出最好分支x best和最好分支所对应的函数值f(x best)；
步骤S403、判断是否满足误差精度的要求或者是否达到了最大迭代次数，如果都不满足要求则准备进入IAT算法的迭代寻优过程；
步骤S404、如果搜索次数达到了搜索次数的上限，将进行随机操作生成新的分支，将新产生的分支与旧的分支进行比较，如果新的分支优于旧的分支则新的分支将取代旧分支，否则仍以旧的分支为最优分支；
步骤S405、再次判断最优分支是否满足误差精度或者当前迭代次数是否达到了设定的最大迭代次数t max。
所述Artificial Tree算法的相关参数依据配电网线路保护装置的动作信息及对应发生故障的区域设定，形成相应的神经网络训练样本。
所述配电网线路保护装置的动作信息主要采集基于开关量的电网信号，用0、1对电网开关动作进行编号。
一种配电网多重故障诊断方法的系统，包括：电路采样模块1、线路保护装置2、故障分析模块3、仿真模块4以及检测输出模块5
所述电路采样模块1，用于采集和预处理导入到所述MATLAB平台和Simulink平台的各项数据；
所述线路保护装置2，用于配电网发生故障时会进行跳闸或者报警等相应的保护动作，根据保护装置不同的保护动作，判断故障发生的大致区间，并且将电网的区间进行编号；
所述故障分析模块3，用于分析故障发生位置及故障电阻阻值，判断故障大致区域；
所述仿真模块4，利用Simulink平台通过Artificial Tree智能优化算法获取故障检测数据，显示仿真结果。
所述检测输出模块5，依据仿真结果选用部分数据检测线路神经网络对电网故障检测的准确率分析适应性报告。
本申请利用MATLAB平台用改进后的小波阈值去噪方法去除信号中的噪声源建立电网故障模型，提高信号采集准确度，运行仿真软件进行仿真，并建立基于开关量的电网故障模型得到电网故障信息决策表，训练神经网络，检测故障发生的区间，对复杂电网采取分而治之的策略，使得线路拓扑结构能够灵活多变，这样不仅能提高电网故障诊断的准确性与可靠性，还可以在故障信息存在误动信息的情况下对故障进行检测，从而达到电网故障区域定位的效果，有效的增强了电网故障检测的能力。

Claims (10)

  1. 一种配电网多重故障诊断方法，包括：
    步骤S1、利用MATLAB平台对配电网线路进行短路故障分析，得出基于电气量的故障信息决策表；
    步骤S2、利用Simulink平台对所述基于电气量的故障信息决策表进行建模仿真，获取配电网线路神经网络的输出训练数据；
    步骤S3、采用小波变换法对所述神经网络的输出训练数据进行去噪、采集处理，形成相关故障信息决策表作为神经网络的训练样本；
    步骤S4、采用改进的Artificial Tree智能优化算法优化神经网络的权值和阈值，通过迭代达到相关误差精度的要求，选择部分数据作为故障数据并利用训练好的神经网络检测故障。
  2. 根据权利要求1所述的配电网多重故障诊断方法，其中，所述步骤S1中，所述MATLAB平台采集配电网线路发生小电流接地短路故障后产生的零序电流、零序功率和零序导纳信号，基于这三种信息形成基于电气量的故障信息决策表。
  3. 根据权利要求2所述的配电网多重故障诊断方法，其中，依据所述零序电流、零序功率和零序导纳信号用离散采样时间和连续采样时间相混合的采样模式进行故障分析，包括：
    步骤S201、在每条线路两个终端中的一个终端上采样并存储正负极零序电流；
    步骤S202、选择固定长度的时间窗内的采样电流作为线路参考电流，对动态采样电流与线路参考电流进行比较，使用改进Pearson方法实时求解二者的相关系数大小；
    步骤S203、设定相关系数阈值，判断相关系数与设定阈值的大小，并输出故障所在线路及故障类型。
  4. 根据权利要求3所述的配电网多重故障诊断方法，其中，所述步骤S202中，所述改进Pearson方法采用线路相关系数的累加值进行故障检测的相关系数分析，所述改进Pearson方法包括：
    步骤S2021、定义具有固定长度的时间窗内可移动窗口值T det、T loc，所述T det、T loc窗口的末端固定时刻分别为t 1，t 2，将所述T det中的采样电流与稳态参考电流进行比较，检测直流微电网的故障，所述T det和T loc的关系式为：
    n s/f s≤T loc≤T 0≤T det
    其中f s为线路电流的采样频率，n s为采样点的数量，T 0为从故障发生时刻t 0到故障被检测到时刻t 2之间的时长；
    步骤S2022、采用线路相关系数的累加值进行故障检测的相关系数分析，将稳态参考电流曲线和采样电流曲线的第i个对比值分别表示为P 1_i、P 2_i：
    Figure PCTCN2022133889-appb-100001
    其中I ref_j为第i个暂态线路电流，I sam_j为第i个采样线路电流；
    步骤S2023、引入调整因子p优化Pearson相关系数调整因子p表达式为：
    Figure PCTCN2022133889-appb-100002
    其中，e为数学常数，表示为自然对数的底数，n为两条曲线所对比点的总数；
    步骤S2023、依据所述调整因子p优化相关系数r，所述相关系数r的表达式为：
    r＝p×r 0
    Figure PCTCN2022133889-appb-100003
  5. 根据权利要求4所述的配电网多重故障诊断方法，其中，所述步骤S2中，依据所述故障分析获取故障位置和故障电阻，通过Simulink平台构建模型输出训练数据，采用小波变换法处理训练数据。
  6. 根据权利要求5所述的配电网多重故障诊断方法，其中，所述小波变换法处理训练数据的过程包括：
    步骤S301、对含噪声的训练数据信号f(t)作小波变换，获取一组小波分解系数；
    步骤S302、采样软阈值函数对得到的小波分解系数进行阈值处理，得到一组估计小波系数，所述软阈值函数υ(ω)表达式为：
    Figure PCTCN2022133889-appb-100004
    Figure PCTCN2022133889-appb-100005
    其中，sign为符号函数，ω为小波系数，T为阈值，σ为噪声强度，m为信号长度，d为幅值因子，a、b为调节因子；
    步骤S303、通过调节d使得在阈值点T处连续且降低对原信号产生的震荡误差，提取去噪后的信号。
  7. 根据权利要求6所述的配电网多重故障诊断方法，其中，在步骤S4中，将所述震荡误差的均方误差采用Artificial Tree算法优化获取线路神经网络，所述Artificial Tree算法优化的过程包括：
    步骤S401、以所述震荡误差的均方误差MSE作为目标函数，将收集到的数据进行整理，准备好训练样本和测试样本；
    步骤S402、设置Artificial Tree算法的相关参数，确定分支个数SN、空间维数D、迭代次数t max，并对分支进行初始化操作，生成初始分支；
    步骤S402、计算每个分支所对应的函数值，根据函数值选出最好分支x best和最好分支所对应的函数值f(x best)；
    步骤S403、判断是否满足误差精度的要求或者是否达到了最大迭代次数，如果都不满足要求则准备进入IAT算法的迭代寻优过程；
    步骤S404、如果搜索次数达到了搜索次数的上限，将进行随机操作生成新的分支，将新产生的分支与旧的分支进行比较，如果新的分支优于旧的分支则新的分支将取代旧分支，否则仍以旧的分支为最优分支；
    步骤S405、再次判断最优分支是否满足误差精度或者当前迭代次数是否达到了设定的最大迭代次数t max。
  8. 根据权利要求7所述的配电网多重故障诊断方法，其中，所述Artificial Tree算法的相关参数依据配电网线路保护装置的动作信息及对应发生故障的区域设定，形成相应的神经网络训练样本。
  9. 根据权利要求8所述的配电网多重故障诊断方法，其中，所述配电网线路保护装置的动作信息包括基于开关量的电网信号，用0、1对电网开关动作进行编号。
  10. 一种采用权利要求1-9任一项所述的配电网多重故障诊断方法的系统，包括：电路采样模块(1)、线路保护装置(2)、故障分析模块(3)、仿真模 块(4)以及检测输出模块(5)；
    所述电路采样模块(1)，用于采集和预处理导入到所述MATLAB平台和Simulink平台的各项数据；
    所述线路保护装置(2)，用于配电网发生故障时会进行跳闸或者报警保护动作，根据保护装置不同的保护动作，判断故障发生的区间，并且将电网的区间进行编号；
    所述故障分析模块(3)，用于分析故障发生位置及故障电阻阻值，判断故障区域；
    所述仿真模块(4)，利用Simulink平台通过Artificial Tree智能优化算法获取故障检测数据，显示仿真结果；
    所述检测输出模块(5)，依据仿真结果选用部分数据检测线路神经网络对电网故障检测的准确率分析适应性报告。
PCT/CN2022/133889 2022-05-11 2022-11-24 一种配电网多重故障诊断方法及系统 WO2023216553A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210506365.6A CN114636900B (zh) 2022-05-11 2022-05-11 一种配电网多重故障诊断方法
CN202210506365.6 2022-05-11

Publications (1)

Publication Number Publication Date
WO2023216553A1 true WO2023216553A1 (zh) 2023-11-16

Family

ID=81953165

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/133889 WO2023216553A1 (zh) 2022-05-11 2022-11-24 一种配电网多重故障诊断方法及系统

Country Status (2)

Country Link
CN (1) CN114636900B (zh)
WO (1) WO2023216553A1 (zh)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117310353A (zh) * 2023-11-30 2023-12-29 淮安苏达电气有限公司 一种变电站一、二次回路通流加压故障测试方法及系统
CN117390407A (zh) * 2023-12-13 2024-01-12 国网山东省电力公司济南供电公司 变电站设备的故障识别方法、系统、介质和设备
CN117388641A (zh) * 2023-12-11 2024-01-12 武汉邢仪新未来电力科技股份有限公司 一种基于分布式物联网终端的线路故障远程监测系统
CN117538689A (zh) * 2024-01-10 2024-02-09 高唐县恒诚建筑工程有限公司 一种基于物联网的电路故障检测方法、系统、装置及介质
CN117668622A (zh) * 2024-02-01 2024-03-08 山东能源数智云科技有限公司 设备故障诊断模型的训练方法、故障诊断方法及装置
CN117909909A (zh) * 2024-03-19 2024-04-19 青岛鼎信通讯股份有限公司 一种配电网弧光接地故障识别方法
CN118017606A (zh) * 2024-04-10 2024-05-10 长沙润伟机电科技有限责任公司 一种变流器并离网切换控制系统及方法

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114636900B (zh) * 2022-05-11 2022-09-30 广东电网有限责任公司东莞供电局 一种配电网多重故障诊断方法

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH05298105A (ja) * 1992-04-22 1993-11-12 Mitsubishi Electric Corp 故障原因推論装置
WO2000048337A1 (de) * 1999-02-12 2000-08-17 Deutsche Telekom Ag Verfahren zur überwachung der übertragungsqualität eines optischen übertragungssystems, insbesondere eines optischen wellenlängenmultiplexnetzes
CN103778479A (zh) * 2014-01-10 2014-05-07 国网上海市电力公司 自适应信息容错保护方法
CN104360227A (zh) * 2014-10-31 2015-02-18 国家电网公司 基于行波法和暂态主频法的变电站电缆出线故障监测方法
CN110457750A (zh) * 2019-07-09 2019-11-15 国家电网有限公司 基于神经网络响应面的有载分接开关弹簧储能不足故障识别方法
CN110533549A (zh) * 2019-07-12 2019-12-03 杭州电子科技大学 一种基于ga优化rs-bp神经网络电网故障定位方法
CN111339872A (zh) * 2020-02-18 2020-06-26 国网信通亿力科技有限责任公司 一种基于分类模型的电网故障分类方法
CN112749509A (zh) * 2020-12-30 2021-05-04 西华大学 一种基于lstm神经网络的智能变电站故障诊断方法
CN114636900A (zh) * 2022-05-11 2022-06-17 广东电网有限责任公司东莞供电局 一种配电网多重故障诊断方法及系统

Family Cites Families (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102508076B (zh) * 2011-11-09 2013-09-25 东北大学 一种基于多智能体系统和小波分析的故障诊断装置及方法
EP2985613B1 (en) * 2013-03-29 2022-02-09 Beijing Inhand Networks Technology Co., Ltd. Method and system for detecting and locating single-phase ground fault on low current grounded power-distribution network
CN103713237B (zh) * 2013-12-30 2016-08-17 华北电力大学 一种电力系统输电线路短路故障诊断方法
CN104502795B (zh) * 2014-11-26 2017-11-14 国家电网公司 一种适用于微电网的智能故障诊断方法
CN106597260B (zh) * 2016-12-29 2020-04-03 合肥工业大学 基于连续小波分析和elm网络的模拟电路故障诊断方法
CN111239549A (zh) * 2020-02-18 2020-06-05 国网信通亿力科技有限责任公司 一种基于离散小波变换的配电故障快速定位方法
CN111537830A (zh) * 2020-03-26 2020-08-14 南方电网科学研究院有限责任公司 一种基于云边架构和小波神经网络的配电网故障诊断方法
CN111537841B (zh) * 2020-06-30 2021-08-06 上海交通大学 一种适用于接地故障类型识别的优化方法及系统
CN111896890B (zh) * 2020-08-06 2021-10-08 山东大学 基于希尔伯特-黄变换的微电网线路故障诊断方法及系统
WO2022064350A1 (en) * 2020-09-22 2022-03-31 Victoria Link Limited Electricity distribution network fault detection system and method
CN113902050A (zh) * 2021-08-30 2022-01-07 广东电网有限责任公司 一种适用于电网故障预案信息的检测提取方法
CN113902250A (zh) * 2021-09-03 2022-01-07 广东电网有限责任公司 一种基于人工神经网络的风险评估分析方法

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH05298105A (ja) * 1992-04-22 1993-11-12 Mitsubishi Electric Corp 故障原因推論装置
WO2000048337A1 (de) * 1999-02-12 2000-08-17 Deutsche Telekom Ag Verfahren zur überwachung der übertragungsqualität eines optischen übertragungssystems, insbesondere eines optischen wellenlängenmultiplexnetzes
CN103778479A (zh) * 2014-01-10 2014-05-07 国网上海市电力公司 自适应信息容错保护方法
CN104360227A (zh) * 2014-10-31 2015-02-18 国家电网公司 基于行波法和暂态主频法的变电站电缆出线故障监测方法
CN110457750A (zh) * 2019-07-09 2019-11-15 国家电网有限公司 基于神经网络响应面的有载分接开关弹簧储能不足故障识别方法
CN110533549A (zh) * 2019-07-12 2019-12-03 杭州电子科技大学 一种基于ga优化rs-bp神经网络电网故障定位方法
CN111339872A (zh) * 2020-02-18 2020-06-26 国网信通亿力科技有限责任公司 一种基于分类模型的电网故障分类方法
CN112749509A (zh) * 2020-12-30 2021-05-04 西华大学 一种基于lstm神经网络的智能变电站故障诊断方法
CN114636900A (zh) * 2022-05-11 2022-06-17 广东电网有限责任公司东莞供电局 一种配电网多重故障诊断方法及系统

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
"Doctoral Dissertation", 10 June 2021, ZHEJIANG UNIVERSITY, China, article KONG LIANG: "Research on Key Technologies of Short-circuit Fault Protection for DC microgrid", pages: 1 - 119, XP009550432 *
LI, ZHUOTONG: "Research of Fault Detection Based on Optimized Neural Network in Power Network", MASTER’S THESES, NORTH CHINA UNIVERSITY OF SCIENCE AND TECHNOLOGY, CHINA, 26 November 2018 (2018-11-26), China, pages 1 - 69, XP009550433 *

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117310353A (zh) * 2023-11-30 2023-12-29 淮安苏达电气有限公司 一种变电站一、二次回路通流加压故障测试方法及系统
CN117310353B (zh) * 2023-11-30 2024-02-09 淮安苏达电气有限公司 一种变电站一、二次回路通流加压故障测试方法及系统
CN117388641A (zh) * 2023-12-11 2024-01-12 武汉邢仪新未来电力科技股份有限公司 一种基于分布式物联网终端的线路故障远程监测系统
CN117388641B (zh) * 2023-12-11 2024-03-08 武汉邢仪新未来电力科技股份有限公司 一种基于分布式物联网终端的线路故障远程监测系统
CN117390407A (zh) * 2023-12-13 2024-01-12 国网山东省电力公司济南供电公司 变电站设备的故障识别方法、系统、介质和设备
CN117390407B (zh) * 2023-12-13 2024-04-05 国网山东省电力公司济南供电公司 变电站设备的故障识别方法、系统、介质和设备
CN117538689A (zh) * 2024-01-10 2024-02-09 高唐县恒诚建筑工程有限公司 一种基于物联网的电路故障检测方法、系统、装置及介质
CN117668622A (zh) * 2024-02-01 2024-03-08 山东能源数智云科技有限公司 设备故障诊断模型的训练方法、故障诊断方法及装置
CN117668622B (zh) * 2024-02-01 2024-05-10 山东能源数智云科技有限公司 设备故障诊断模型的训练方法、故障诊断方法及装置
CN117909909A (zh) * 2024-03-19 2024-04-19 青岛鼎信通讯股份有限公司 一种配电网弧光接地故障识别方法
CN118017606A (zh) * 2024-04-10 2024-05-10 长沙润伟机电科技有限责任公司 一种变流器并离网切换控制系统及方法

Also Published As

Publication number Publication date
CN114636900B (zh) 2022-09-30
CN114636900A (zh) 2022-06-17

Similar Documents

Publication Publication Date Title
WO2023216553A1 (zh) 一种配电网多重故障诊断方法及系统
CN110082640B (zh) 一种基于长短时记忆网络的配网单相接地故障辨识方法
CN110829417B (zh) 基于lstm双结构模型的电力系统暂态稳定预测方法
CN110907754B (zh) 一种基于psd-bpa的故障线路严重程度评估方法
CN111308260B (zh) 一种基于小波神经网络的电能质量监测和电器故障分析系统及其工作方法
CN111404130A (zh) 基于快速开关的新型配电网故障检测方法及故障自愈系统
CN110543921A (zh) 一种基于波形学习的电缆早期故障识别方法
CN116401532B (zh) 一种电力系统受扰后频率失稳识别方法及系统
Izadi et al. A synchronized lissajous-based method to detect and classify events in synchro-waveform measurements in power distribution networks
CN106019043B (zh) 一种基于模糊积分多源信息融合理论的电网故障诊断方法
CN113486535A (zh) 一种基于环境信息的电网信息检测分析方法
CN111953657B (zh) 一种序列-数据联合驱动的配电网cps网络攻击辨识方法
CN111369179A (zh) 一种配电自动化终端设备闭环检测方法和装置
CN116127302A (zh) 基于改进AlexNet算法的电动汽车充电装置故障电弧识别方法
Ding et al. Research on power grid fault diagnosis method based on PMU data and convolutional neural network
Zheng et al. Transformer Vibration Analysis Based on Double Branch Convolutional Neural Network
CN116702629B (zh) 一种具备可迁移能力的电力系统暂态稳定评估方法
CN114021433B (zh) 一种电力系统的主导失稳模式识别模型的构建方法及应用
Izadi Data-Driven Analysis and Applications of Time-Synchronized Waveform Measurements in Power Systems
CN110618347B (zh) 一种基于小波变换和粗糙集理论的区域电网故障诊断方法
Feng et al. A method for identifying major disturbance sources in a regional grid
Shuai et al. Design of Intelligent Low Voltage Station System Based on Edge Calculation
CN117310373A (zh) 基于空洞卷积识别配网数据的方法及系统、设备、介质
Wang et al. Fault feeder detection method of distribution network based on gramian angular field and convolutional neural network
Tan et al. On-line detection method of insulator zero-value fault based on image features and YOLOv3 algorithm

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22941488

Country of ref document: EP

Kind code of ref document: A1