Research of Analog Circuit Fault Diagnosis Based on Data Fusion Technology

For the solution of insufficient test data and single information can not represent all the fault state in the analog circuit fault diagnosis, a fault diagnosis model and a fusion algorithm based on data fusion technology is formed. The fault diagnosis information is fused with two layers: For the feature layer, the voltage and current of testing nodes are processed by different neural network in order to acquire BPA of various faults. For the decision layer, ultimate result is acquired by D-S evidence theory. The results of simulation show that: Comparing with the result from the network as single fusion layer, this method has a smaller error and higher diagnosis reliability.