基于BP神经网络的绝缘子表面污秽预测方法

Prediction of pollution severity of insulators based on BP neural network

针对传统方法清扫绝缘子污秽的弊端,提出了利用BP神经网络预测绝缘子污秽的方案。通过BP神经网络的输入、输出和参数的选取,在实验室进行数据验证。结果表明:该方法可以准确有效的预测绝缘子表面污秽程度,可以指导现场绝缘子的清扫和优化绝缘子的清扫周期。

Aimed at the insulators-clearing defects of the conventional methods, the paper introduces a new method of using the BP neural network to predict the pollution severity of insulators. Selecting inputs, outputs and parameters of the BP neural network makes it possible to prove the data in experiments. The results show that the method, capable of more accurate prediction of the pollution severity of insulators, permits effectively guiding the insulators cleaning and optimizing the cleaning cycles of insulators.