A Fault Diagnosis Method for Power Systems Based on Deep Belief Networks
Currently most of the fault diagnosis nodes in the power system are set in a target structure,which leads to low diagnostic efficiency and F1 of fault diagnosis approaching 0.Therefore a power system fault diagnosis method based on deep belief networks is designed and validated.Based on the current testing requirements,a multi-level approach is adopt-ed to improve the overall diagnostic efficiency.Multi-level fault data collection nodes are deployed,and based on this,fault feature extraction is carried out.A deep belief network fault diagnosis model for power systems is established,and cross validation evaluation is used to achieve fault diagnosis processing.The test results show that through three cycles of phased measurement and comparison,the F1 value of the power system fault diagnosis is 1 ,which is the best diagnostic result.This indicates that with the assistance and support of deep belief networks,the designed diagnosis method is more accurate and efficient,and has practical application value.
deep belief networkpower systemfault diagnosiscollaborative identification
万方数据
维普
