摘要
电网中存在大量的非线性器件,给电网母线带来不同频率的谐波干扰。同时,由于地磁暴的影响也会给母线带来较大的直流分量。这些干扰会使得电流互感器进入饱和状态,二次侧电流畸变严重,使得继电保护装置不能正确动作。该文首先在PSCAD中仿真电力系统突发单相接地短路故障时造成的直流分量,使得保护用电流互感器饱和;然后,通过小波变换识别饱和开始以及结束的时间;最后,通过RBF神经网络将已经发生畸变的二次侧电流补偿成未畸变的期望值。仿真结果表明,基于小波变换的饱和检测可以准确识别饱和起始时间以及饱和严重程度,经过RBF神经网络补偿后的电流与期望值的误差可以控制在0.5%以内。
There are a large number of non-linear devices in the power grid,which bring harmonic interference of different frequencies to the power grid bus.At the same time,due to the influence of geomagnetic storms,it will also bring a larger DC component to the bus.These interferences will cause the current transformer to enter a saturated state,and the secondary side current will be severely distorted,making the relay protection device unable to operate correctly.This article simulates the DC component caused by the sudden single-phase grounding short-circuit fault of the power system to make the protection current transformer saturate.And then uses the wavelet transform to identify the start and end time of the saturation.Finally,the distortion current has occurred through the RBF neural network where the secondary side current is compensated to an undistorted desired value.The simulation results show that the saturation detection based on wavelet transform can accurately identify the saturation start time and saturation severity.The error between the expected value and the current after compensation by the RBF neural network can be controlled within 0.5%.
作者
胡红利
崔晨辉
员鹏宇
段羽洁
张肖
HU Hongli;CUI Chenhui;YUAN Pengyu;DUAN Yujie;ZHANG Xiao(State Key Laboratory of Electrical Insulation and Power Equipment, Xi′an Jiaotong University, Xi′an 710049, China;State Grid Hefei Electric Power Supply Company, Hefei 230022, China;Xi′an Supervision & Inspection Institute of Product Quality, Xi′an 710065, China)
出处
《西北大学学报（自然科学版）》
CAS
CSCD
北大核心
2022年第1期52-59,共8页
Journal of Northwest University（Natural Science Edition）
基金
国家自然科学基金资助项目(52177009)。
关键词
电流互感器
饱和检测
电流补偿
小波变换
RBF神经网络
current transformer
saturated detection
current compensation
wavelet transform
RBF neural networks
