摘要
提出了一种基于模糊神经网络的非线性系统故障诊断方法.利用模糊C 均值聚类法对测量空间进行分割,再利用模糊规则对分割后的子空间分别采用BP网络进行逼近,从而获得不同子空间故障输出与测量输入的非线性动力学特性.计算机仿真表明该网络具有良好的泛化性能,方案可行.
A new approach to fault diagnosis of nonlinear system based on fuzzy neural network is proposed. The measurement space is divided into several subspaces by using fuzzy C\|means clustering. According to the requirements of fuzzy rules, the subspaces are fitted by BP network respectively, and the characteristics between fault outputs and measured inputs in different subspaces are obtained. The computer simulation shows that the network has good generalization performance, and it is applicable for fault diagnosis of nonlinear system.
出处
《海军工程大学学报》
CAS
2003年第4期60-64,共5页
Journal of Naval University of Engineering
