基于神经网络专家系统的高压断路器在线监测研究

Study on Online Monitoring of High Voltage Circuit Breaker Based on Neural Network and Expert System

针对传统专家系统在断路器故障诊断中存在知识获取困难问题,提出了将RBF神经网络与专家系统融合的诊断方案,并通过引入可信度的概念对RBF神经网络算法进行改进,对原已形成的网络权值和阈值进行修正,使故障诊断算法具有诊断新故障类型的功能,又解决专家系统知识获取困难的瓶颈问题。结果表明,该方案合理可行,为高压断路器的在线监测提供了一种新的途径。

Aiming at the problem, the obtain knowledge difficult in the tradition expert system in the circuit breaker failure diagnosis, a diagnostic scheme combining RBF neural network with expert system is proposed. The original network weight and the threshold values are modified by using the concept of confidence level. Then the function used to diagnosis to the new breakdown failure type can be realize. And the problem to obtain the knowledge in the tradition expert system is also solved. The simulation results indicate that it is available for this method used in the breaker failure diagnosis and it also provides a new approach which is used to monitor state of the high voltage circuit breaker online.