白噪声干扰下的超高频局部放电信号检测方法  

Detection Method of UHF Partial Discharge Signal Under White Noise Interference

作　　者：王彦军 李宗虎 郑建志 张彦琪 马全云 WANG Yan-jun;LI Zong-hu;ZHENG Jian-zhi;ZHANG Yan-qi;MA Quan-yun(National Network Gansu Electric Power Company Jiuquan Power Supply Company,Jiuquan 735000 China;State Grid Gansu Electric Power Research Institute,Lanzhou 730000 China)

机构地区：[1]国网甘肃省电力公司酒泉供电公司,甘肃酒泉735000 [2]国网甘肃省电力公司电力科学研究院,甘肃兰州730000

出　　处：《自动化技术与应用》2024年第9期69-72,共4页Techniques of Automation and Applications

摘　　要：为及时发现电力设备绝缘故障,防止电力事故发生,提出白噪声干扰下的超高频局部放电信号检测方法。选取特宽频传感器作为检测设备,设置天线的上、下限截止频率;利用复小波构造算法,通过数字滤波器确定小波基并生成函数序列,获取重构后的复高通滤波器系数,抑制白噪声;根据相关性系数、放电量因数与峰值率提取信号特征;引入激活函数建立人工神经网络模型,设置权值参数,将特征参数作为输入,经过训练完成局部放电信号检测。仿真测试表明,所提方法传感器响应灵敏,能够抵抗白噪声干扰,精准检测出不同局部放电模型特征。In order to find the insulation fault of power equipment in time and prevent power accidents,an UHF partial discharge signal detection method under the interference of white noise is proposed.It selects the ultra wideband sensor as the detection equipment,and sets the upper and lower cut-off frequencies of the antenna.Using the complex wavelet construction algorithm,the wavelet basis is determined by the digital filter and the function sequence is generated to obtain the reconstructed complex high pass filter coefficients and suppress the white noise.The signal characteristics are extracted according to the correlation coefficient,discharge factor and peak rate.The activation function is introduced to establish the artificial neural network model,set the weight parameters,take the characteristic parameters as the input,and complete the partial discharge signal detection after training.Simulation results show that the sensor response of the proposed method is sensitive,can resist the interference of white noise,and accurately detect the characteristics of different partial discharge models.

关 键 词：白噪声 超高频信号 检测 局部放电 复小波构建 神经网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TM711[自动化与计算机技术—控制科学与工程]