Construction of Fault Diagnosis Model for Filter Capacitor Failure in Switching Power Supply
The failure of high-voltage filtering capacitors is one of the common faults in switching power supplies,which directly affects their stable operation.To achieve efficient fault diagnosis without damaging the structure of the switching power supply,this paper proposes a fault diagnosis model based on an improved deep confidence network(DBN-DF model).This model can use data acquisition devices to isolate and filter the voltage of the power grid where the switching power supply is located.Based on the voltage distortion characteristics of the power grid,the failure percentage of the filtering capacitor in the switching power supply can be determined,and the fault category can be output to complete fault diagnosis.Through experimental verification,it has been found that the DBN-DF model has higher accuracy in diagnosing faults in high-voltage filtering capacitors compared to traditional DBN and BP models.Moreover,as the number of faults in-creases,the accuracy and accuracy of the model also correspondingly improve,ultimately stabilizing.This indicates that the model has high practical application value.
Switching power supplyFault diagnosis modelCapacitor failure
NETL
NSTL
万方数据
