系统故障诊断的一种神经网络方法

A Neural Network Method for System Fault Diagnosis

该文首先提出一个通用故障示例模型，然后运用一种自适应神经网络学习算法来寻找差错属性与故障类型之间的对应关系，由此对故障进行诊断。因为网络的结构事先并不确定，而是在训练的同时进行同步构造，所以确保了训练后建立网络具有较好的适应性。

In this Paper we first present an instance model of system faults, and then use an adaptive neural network learning algorithm to diagnoses the faults by discovering the mapping betWeen errors and types of malfunction. Since the network structure is not fixed a prior, it is subject to building while learning. Therefore, aftertraining,the architecture of the network is guaranteed to be well adapted for a given system.