Analog circuit fault diagnosis method based on improvement

• Main Authors: ZHANG XUNJIE, CHAI YI, WANG YIMING, XIONG YINGZHI, ZHANG KE
• Format: Patent
• Language: English
• Published: 19.08.2015
• Subjects: MEASURING; MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES; PHYSICS; TESTING

The invention provides an analog circuit fault diagnosis method based on improvement. The conventional technology is improved from two aspects: 1. a DAGSVA method is improved: an SVM with the maximum inter-class distance acts as the uppermost layer node of the DAGSVM. If the root node classification result is i-type, the SVM with the maximum inter-class distance to the i-type is selected to act as the child node of the layer, and if the root node classification result is j-type, the SVM with the maximum inter-class distance to the j-type is selected to act as the child node of the layer; if the classification result does not belong to the i-type or the j-type, the two types are eliminated, two types of SVM with the maximum inter-class distance Dij are selected among the rest classes to act as the nodes of the layer, and the two aforementioned steps are continued unit a diagnosis result is obtained so that a situation of error of the final result caused by high-layer node diagnosis error can be effectively avoided. 2. particle swarm optimization (PSO) is performed on each child SVM for parameter optimization in order to enhance diagnosis accuracy rate of each child node so that SVM diagnosis accuracy rate of each node is enhanced, and thus diagnosis accuracy rate of the whole DAGSVM is enhanced.