基于神经网络与证据理论的模拟电路故障诊断

论述了利用多类电量测试信息、应用神经网络与D-S证据理论实现模拟电路故障诊断的基本原理,提出了一种基于可测点电压与不同测试频率下的电路增益经决策层信息融合的故障诊断新方法.分别利用此两类测试信息,各用一个独立的改进BP网络对电路进行初步诊断,再运用所提融合诊断算法实现故障定位.模拟实验结果表明:所提方法对硬故障与元件参数偏移较小的软故障均适用,故障定位准确率高.

Fault diagnosis of analog circuits based on neural network and evidence theory

Based on neural network and D-S evidence theory, this paper discusses the fundamentals of analog fault diagnosis by means of multiform circuit responses. A new fault diagnosis method is proposed based on data fusion by measuring accessible node voltages and circuit gains of output to input under different test frequencies. Preliminary diagnosis is performed separately by an independent improved BP network employing one kind of circuit responses. Fault location is accomplished by using the proposed decision fusion algorithm according to the preliminary diagnosis results. Theoretical analysis and experimental results show that the proposed fusion diagnosis method avoids limitation of single information and has the capability to diagnose catastrophic and parametric faults of analog circuits with satisfactory accuracy.