基于BP神经网络和模糊逻辑的绝缘子污秽等级预测

针对复杂地理环境下的绝缘子污闪问题,提出一种基于BP人工神经网络和模糊逻辑补偿的绝缘子污秽等级预测新方法.该方法首先利用BP神经网络预测在温度、湿度、降雨量、风速等基本因素下绝缘子的污秽等级,然后充分考虑与复杂地理位置相关的海拔高度、气压比、覆冰水电导率等因素对绝缘子污秽度的影响,采用模糊逻辑补偿的方法进行校正.算例结果表明方法的正确性和有效性.

Prediction of Insulator Pollution Severity Class based on BP Neural Network and Fuzzy Logic

In view of the insulator pollution flashover problem in the complex geographical environment, this article puts forward a new method to predict insulator pollution severity class based on BP artificial neural network and fuzzy logic compensation. This method first predicts the filth of the insulator caused by the basic factors such as temperature, humidity, rainfall, wind speed using BP neural network. And then the effects of such factors associated with complex geographic location as elevation, pressure ratio, and freezing water conductivity on the insulator filthy are taken into account when the method is corrected using fuzzy logic compensation method. The results of the numerical examples show the correctness and effectiveness of the method.