基于故障可诊性与遗传算法的模拟电路测试激励优化方法

摘要

The selection of test stimuli has an influence on circuit fault diagnosis precision.To overcome the disadvantage of prior methods, this paper proposed a new test stimuli optimal design method based on fault detectable analysis and genetic algorithm.First, executing 10 octave AC Sweep analysis on the circuit under test (CUT) so that the frequency range of sine wave could be set, and then calculated the inner and inter class distance and sensitivity factor value based on the discrete frequency-rms voltage response data which acquired under different fault type circuit simulation, finally finished the test stimuli optimal design through the genetic algorithm with sensitivity factor as its objective function.Used a double-bandpass filter circuit to validate the proposed method.The experimental results show that the optimal test stimuli can effectively reduce the ambiguity of different fault feature distribution and reach a satisfied diagnosis result.%测试激励的选择影响电路故障诊断的精度,为克服现有方法不足,提出了一种基于电路可诊性和遗传算法的模拟电路测试激励优化方法.对待测电路进行十倍频程AC Sweep分析来确定正弦测试信号的频率优化范围,基于电路幅频响应中离散频点相应的电压有效值来计算各故障特征数据类内类间离散度,统计多频测试信号敏感度因子大小,并以此为目标函数利用遗传算法实现测试信号中的频率参数优化.以双带通滤波电路为例进行了仿真验证,实验结果表明该方法优化出的测试激励信号能有效降低电路故障响应特征分布模糊性,提高了故障诊断率.