摘要
介绍了一种基于人工神经网络与专家系统相结合的智能故障诊断系统,将专家系统与神经网络结合运用取长补短,发挥各自的优势.给出了系统的结构组成,并具体描述了神经网络专家系统的工作原理.采用的RBF网络有效地克服了BP网络收敛速度慢,且易陷入局部极小的缺陷.同时利用神经网络的并行处理功能,有效地解决了传统专家系统的推理复杂、时间冗余等缺点.仿真结果证明了该方法应用于卫星姿态确定系统的故障诊断是有效的.
A fault diagnosis expert system based on artificial neural network (ANN) is introduced, which can take advantage of each technique, and avoid the weakness of each individual. The architecture of the neural network expert system is given, and the working principle is described in detail. Radial basis function (RBF) network is used to effectively overcome the weakness of backprogation (BP) network, i.e. converging very slowly and frequently trapped in local minima. At the same time, the development of neural network theory which features on nonlinear parallel distribution process supplies an effective method to solve the problems of traditional expert system. Simulation results demonstrate that the proposed approach is effective for fault diagnosis in satellite attitude determination system.
出处
《东南大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2005年第A02期181-184,共4页
Journal of Southeast University:Natural Science Edition
基金
武器装备预研基金资助项目(514090802)
关键词
故障诊断
专家系统
神经网络
卫星姿态确定系统
fault diagnosis
expert system
neural network
satellite attitude determination system