L＿1范数法诊断线性电阻电路故障的神经网络法

Fault Isolation in Linear Resistive Analog Circuits Using Neural-Based L_1 Norm

研究了Ｌ＿１范数法诊断线性电阻电路故障的神经网络方法，基于Ｈｏｐｆｉｅｌｄ网络原理导出了相应的神经网络电路结构。该法具有在线计算量少、故障诊断速度快等特点．计算机模拟结果表明，本文所提方法是可行的。

A neural-based optimization algorithm for fault isolation in linear analogcircuits is developed.Based on Hopfield neural network,a circuit architecture of neuralnetwork for solving related analog circuit fault isolation problem is proposed.Themethod proposed has some advantages such as minimal on-line computation and high-speed of fault diagnosis.The computer simulation results also show its feasibility.