摘要
模拟电路的非线性特性、连续性和元器件的容差等因素给故障的建模分析造成了诸多不确定因素,因此其智能故障诊断方法的研究至关重要。柔性神经树是一种采用树形结构和一组运算符集合构成的新型神经网络,与传统神经网络相比具有更加灵活的自动优化能力。本文将人工免疫机理融入柔性神经树,提出了一种基于柔性免疫神经树的模拟电路故障诊断方法。通过对一种典型模拟电路的故障诊断仿真试验,证明了该方法的有效性和可行性。
Nonlinearity, continuity and tolerance of components cause many uncertainties to the modeling and analysis of analog circuits fault, so its intelligent fault diagnosis is very important. Flexible neural tree （FNT） is a new neural network, which is constructed by a structured tree and a set of computation symbols, and which flexible can automatically devise an artificial neural network. This paper proposes a flexible immune neural tree based analog circuits fault diagnosis approach, whose performance and effectiveness are evaluated with a typical analog circuit.
出处
《山东科学》
CAS
2009年第1期35-39,共5页
Shandong Science
关键词
模拟电路
故障诊断
人工免疫系统
神经树
analog circuits
fault diagnosis
artificial immune system
neural trees
