容差电路软故障检测与定位  

Soft-Fault Detection and Location for Circuits With Tolerance

作　　者：彭敏放[1] 沈美娥[2] 贺建飚[3] 谢宏[1] 何怡刚[1] 

机构地区：[1]湖南大学电气与信息工程学院,长沙410082 [2]北京信息科技大学计算机学院,北京100101 [3]中南大学信息科学与工程学院,长沙410083

出　　处：《电工技术学报》2009年第3期222-228,共7页Transactions of China Electrotechnical Society

基　　金：国家自然科学基金(60673084);湖南省自然科学基金(06JJ4075,04JJ6034);湖南省科技计划(05FJ3008)资助项目

摘　　要：提出了一种基于统计理论与神经网络数据融合、可用于容差模拟电路软故障检测与定位的诊断方法。该方法将故障诊断分两个阶段进行,即故障检测与故障定位。通过监测可测点工作电压,利用所构建的故障阈值函数与故障判据来实现容差电路的故障在线检测。再通过离线测量电路在不同测试频率下输出对输入的增益,利用可测点工作电压与电路增益,运用所提出的数据融合方法及改进的BP算法对电路进行故障定位。模拟诊断结果表明,所提方法不仅可用于硬故障诊断,而且还能实现容差电路的软故障诊断,所需测试节点少,故障检测与定位准确率高。Based on statistics theory, neural network and data fusion, a new fault diagnosis method capable of soft-fault detection and location in analog circuits with tolerance is proposed. The proposed diagnosis strategy consists of fault detection and fault location. By monitoring accessible node voltages, on-line fault detection is performed based on the proposed fault threshold function and the fault criterion. Then circuit gains are measured under selected test frequencies. Based on circuit gains and accessible node voltages, off-line fault location is performed by the proposed data fusion method and improved BP algorithm. The simulation results show that the proposed approach has the capability to detect and locate not only catastrophic faults but also parametric faults in tolerance circuits with a small quantity of accessible nodes, and the diagnosis accuracy is satisfactory.

关 键 词：模拟电路 故障诊断 数据融合 神经网络 统计理论 容差 

分 类 号：TN707[电子电信—电路与系统]