基于噪声估计的模拟电路故障诊断方法

Analog circuits fault diagnosis based on noise estimation

提出了一种基于噪声估计的模拟电路故障诊断新方法。首先,参考线性系统中包含的信息建立一种故障诊断模型,该模型把被测电路的输出建模为标称值叠加系统扰动（也相当于系统噪声）,提取这种噪声信息作为故障特征;接着分析了基于RS（Ramanujan sum）的噪声估计原理,利用这一原理能准确地估计给定信号的噪声值,最终通过比较无故障电路和故障电路之间的噪声差异来实现故障检测。在国际基准电路上仿真和计算的结果表明,该方法具有计算简单、检测率高、测试成本低、适用性强等优势。

This paper presents a novel analog circuit fault diagnosis method using noise estimation. First,a fault diagnosis model is established with reference to the information inside the linear system. This model regards the output of the circuit under test（ CUT） as the nominal value of the circuit by adding system perturbation（ also the system noise）. This noise information is extracted as a fault feature. Then,the principle of noise estimation based on Ramanujan sum（ RS） is analyzed and used to accurately estimate the noise level to a given signal. Finally,fault detection is achieved by comparing the divergence of noise between the fault-free circuit and the faulty circuit. The simulated and calculated results of the international standard circuits show that the method performs well with simple calculation,high detection rate,low test cost and strong applicability.