论文部分内容阅读
根据大规模电路故障诊断网络撕裂法和交叉撕裂搜索方法,采用基于误差反向传播算法的多层前向神经网络(BP神经网络)记载多次撕裂信息,提出了一种新型基于BP神经网络的大规模电路模块级快速诊断方法。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,实时诊断性强等优点。
According to large-scale circuit fault diagnosis network tearing method and cross-tearing search method, multilayer forward neural network (BPNN) based on error back propagation algorithm is used to record multiple tearing information, and a new type of BP neural network based on BP Large-Scale Circuits and Module-Level Rapid Diagnostic Methods Based on Neural Networks. The method can process the fault modules in parallel quickly and effectively, and has the advantages of small workload before measurement and strong real-time diagnosis.