基于神经网络及专家系统的混合诊断法

A NEW FAULT DIAGNOSIS METHOD BASED ON NEURAL NETWORKS AND EXPERT SYSTEM

本文将神经网络与人工智能专家系统结合起来，利用神经网络的联想记忆能力对发生过的故障进行直接记忆和联想。利用专家系统的反向推理能力对联想产生的侯选故障进行验证。据此原理开发了一种更接近人类思维过程的诊断系统。

In order to diagnose the faults in trubomachinery,a new type of diagnosis system based on Expert System(ES)and Neural Network(NN) is proposed.The NN diagnosis system is used to remember directly and associate faults happened before.The NN diagnosis system gives a set of candidate faults.The back reasoning mechanism of the expert system is applied to verifying these candidate faults.This diagnosis method is proved to be reliable and superior to the traditional ones.