大规模模拟电路软故障快速诊断法  

A fast approach of soft fault diagnosis for large-scale analog circuit

作　　者：厉芸[1] 何怡刚[1] 徐卫林[1] 

机构地区：[1]湖南大学,湖南长沙410082

出　　处：《福建工程学院学报》2004年第2期157-160,共4页Journal of Fujian University of Technology

基　　金：国家自然科学基金项目(50277010);湖南省杰出青年基金(03JJY1010);高校博士点基金(20020532016);湖南省科技计划项目(03GKY3115);湖南大学撷英计划资助

摘　　要：研究了大规模模拟电路软故障的故障统一特征向量的提取方法,并在此基础上提出了大规模模拟电路软故障快速诊断法。该方法根据大规模模拟电路故障诊断网络交叉撕裂搜索方法,采用改进BP算法的神经网络记载电路多次撕裂信息,从而可以迅速检测出大规模电路的软故障模块。仿真实例证明该方法能快速有效地并行处理定位故障模块,具有测前工作量小、适用范围广、诊断速度快等优点。The approach of acquiring the soft fault constant eigenvector is investigated, and a fast approach of module fault diagnosis for large-scale analog circuit based on BP (Back Propagation) neural network is presented. According to circuit decomposition search technology, the decomposition information is noted by neural networks, enabling soft fault modules to be located quickly and effectively. The availability of the method is examined by simulation.

关 键 词：大规模模拟电路 软故障诊断 网络撕裂 改进BP算法 

分 类 号：TN10[电子电信—物理电子学]