相对温度分布特征与人工神经网络相结合的零值绝缘子识别方法

提出利用绝缘子串相对温度分布特征和人工神经网络模型相结合的方法识别不同污秽等级、不同湿度条件下的零值绝缘子。试验获取模拟110 kV线路悬式绝缘子的红外运行图像,经图像去噪、分割等预处理后,提取绝缘子串区域相对温度分布特征参数作为识别零值绝缘子的温度信息特征量,并结合环境相对湿度、等值附盐密度作为识别模型的输入向量,将实际测定绝缘子串是否含零值的状态分类信息作为输出向量,通过训练得到优化的识别模型,并用于零值绝缘子识别。试验结果验证该方法准确性高,可为输电线路瓷绝缘设备的故障检修提供参考和方法借鉴。

Identification of Zero Resistance Insulators by Combining Relative Temperature Distribution Characteristics With Artificial Neural Network

A method is proposed to identify zero resistance insulators under various pollution levels and humidity conditions by combining relative temperature distribution characteristics of insulator string with artificial neural network(ANN) model.The infrared image of suspension insulator string being operated in 110?kV transmission line is achieved by simulation tests and after the preprocessing of image denoising and segmentation the extracted characteristic parameters of relative temperature distribution in the region of insulator string are taken as temperature information characteristics to identify zero resistance insulator,and taking environmental relative humidity and equivalent salt deposit density as input vectors of identification model and regarding the state classification information that whether the actually measured insulation string contains zero resistance insulator as the output vector the optimized identification model is obtained by training and applied to the identification of zero resistance insulator.Testing results show that the zero resistance insulator identification by the proposed method is accurate,so it is available for reference to corrective maintenance and troubleshooting of porcelain insulation equipments for transmission lines.