摘要
神经网络是以研究人的生物神经为根据，由大量简单元件即神经元模拟电子器件相互联接而形成的一种复杂网络本文通过比较现行发动机控制系统故障诊断的方法，针对基于神经网络的发动机控制系统故障诊断进行了推导及改善数学模型；并编制了采用径向基函数神经网络完成发动机控制系统故障诊断的学习与诊断程序，实现了仿真将神经网络应用于发动机控制系统的故障诊断技术中，以发动机的各测量传感器为基础，对其控制系统进行实时故障诊断和故障分离，不仅充分体现神经网络的容错性、抗干扰性及大规模并行处理能力、自学习能力等特点。
According to the research of human biogenical neural, neural network is a complicate network which is made up of a vast amount of simple elements-neural elements. After discussing the present fault diagnosis ways of engine control system. This article sets up a mathematical model based on neural network. The learning and diagnosis programs for fault diagnosis of engine control system were edited based on RBFNN, and the system emulation was implemented. Applying neural network in fault diagnosis technology of engine control system to complete the fault diagnosis and separation by every engine-testing sensor can express neural network's fault-tolerant, noise immunity and massively parallel processing and self-study,which ability develops a new field of fault diagnosis.
出处
《江苏理工大学学报（自然科学版）》
1999年第5期36-39,共4页
Journal of Jiangsu University of Science and Technology(Natural Science)
关键词
发动机
控制系统
故障诊断
神经网络
engines
control systems
fault diagnosis
neural networks
