基于模糊线性规划的模拟电路软故障诊断

Diagnosis of soft-fault in analog circuit based on fuzzy linear programming

研究阐述了一种新的容差条件下基于模糊线性规划理论的模拟电路软故障诊断方法。通过灵敏度分析建立电路测试节点电压增量方程,确立各个元件的模糊隶属函数,从而建立起模拟电路故障诊断的模糊线性规划方程组。通过求解方程组来确定电路中各元件在一定故障情况下的最小变化量。在容差条件下,将方程组的解与各元件容差范围相比较来定位故障元件。从而,兼顾故障元件定位和故障元件参数偏移量确定。实验结果表明,该方法可以有效地实现模拟电路元件参数在一定容差范围内随机变化条件下的软故障诊断。

In this paper, a novel soft-fauh diagnosis approach for analog circuit under parameter tolerance based on fuzzy linear programming （FLP） is proposed. The node-voltage increment equations based on the sensitivity analysis and the fuzzy function for each element in circuit are built as the constraints of a FLP equation. Through evaluating parameters deviations from the solution of the FLP equation, it can state whether the actual parameters are within their tolerance ranges or not. Then, both the locating of faulty element and the evaluating of parameter deviation of the faulty element in analog circuit are presented by the proposed approach. The simulation example shows its effectiveness in fault diagnosis with tolerance.