基于网络树的模拟电路故障诊断

根据电网络理论中树的特性，提出了一种的子网络级的模拟电路故障诊断方法，给出了在大规模电网络中，网络树状态的两个判据及根据网络树状态判据的故障诊断算法。本方法避免了进行 Ｋ Ｃ Ｌ方程校验所带来的复杂计算，且大部分计算可于测试前进行，实时性强，便于在线实现。算法同时适用于线性电路和非线性电路。

A New Fault Diagnosis Method for Analog Circuit Using Network Tree

The advent of analog VLSI makes fault diagnosis almost impossible. Even methods for handling just the much simpler problem of a subnetwork of analog VLSI still meet with great difficulties. We present a new method based on network tree that is better than exsiting methods. We give two criteria for the network tree to be free of fault. We give a lengthy proof, involving eqs.(1) through (10), for the first criterion. Proof for the second criterion is omitted. In section 2, we give the six steps of the algorithm to be used for fault diagnosis. Our algorithm avoids the solving of KCL equation, which is a main source of trouble for existing methods. Fig.1 shows an analog circuit to be diagnosed. Simulation results show that our method is feasible.Our method is applicable not only to linear analog circuits but also to nonlinear ones.