基于频域核系数的动态模拟电路故障诊断

Fault Diagnosis of Analog Dynamic Circuit Based on Frequency-domain Kernel Coefficient

因模拟电路中可及测试节点数量的限制,基于系统辨识的思想提出了一种新的故障诊断方法,频域核系数被用来进行模拟电路参数故障诊断。根据电路传递函数和元件的容差预先计算出全部频域核系数的正常范围,然后利用电路的输入输出信息提取故障电路的频域核系数,依据超出正常范围频域核系数的故障特征可确定电路故障范围或故障元件。通过仿真实例验证了该方法的有效性。

Aiming at the limitation of accessible test nodes in analog circuits, a new fault diagnosis approach is proposed based on the idea of system recognition. Frequency-domain kernel coefficient is introduced for detecting parametric faults in analog circuit. From the circuit transfer function and component tolerance specifications, the tolerance boxes of all circuit frequency-domain kernel coefficients are pre-computed. The fault circuit frequency-domain kernel coefficients are extracted using the input/output signal information of the circuit. Fault features of these measured frequency-domain kernel coefficients that are outside their tolerance boxes can be used to diagnose the fault range or fault components. Simulation experiment example in analog circuit validates that the method is effective.