基于人工神经网络的零值绝缘子诊断

Diagnosis of Faulty Porcelain Insulators Based on Artificial Neural Network

研究了高压线路零值检测中绝缘子电晕脉冲信号及现场脉冲干扰的个数、最大幅度、平均幅度对相位的分布曲线 (n—φ、qmax—φ、qave—φ) ,并提取了表示曲线特性的 5项统计指标 ,利用人工神经网络对曲线进行模式识别以实现零值绝缘子的检测 ,研究表明其识别准确率～ 80 %。

In this paper, some curves of corona signal of porcelain insulators and field noises in detecting faulty insulators on power transmission line are discussed. Eight statistic factors are extracted from these curves and used as the input of a BP artificial neural network to identify signal pattern for finding faulty insulators. According to our research, the network has good accuracy up to 80%.