The Research Of Parallel Fault Diagnosis Method Of Condition Information Of Electric Power Equipment

Electrical equipment is the fundamental component of the power system, and thefaults of electrical equipment will affect the safe operation of the smart grid. Faultdiagnosis of electric power equipment can ensure normal operation of power system, anddata mining is the key technology for power equipment fault diagnosis. With thedevelopment of smart grid，the number of electric power equipment will increasegradually, and the electric power equipment will produce large number of conditioninformation. In addition, the intelligent fault diagnosis techniques of electric powerequipment become more and more mature. So, the intelligent technology of electricalequipment fault diagnosis is increasing gradually. Rapid parallel data mining of massiveamounts of State information for power equipment is one of the important problems inpower equipment fault diagnosis and state assessment field of smart grid.In this pape, taking Hadoop technology as the basis technology framework, a faultdiagnosis scheme of smart gird power equipment based on cloud computing has beendesigned. It can provide an efficient parallel computing ability for power equipment faultdiagnosis and state assessment. Considering the disadvantages of slowly fault diagnosisspeed and not well adapted to the massive data sets in smart grid environment instand-alone environment, this paper gives two parallel data mining method for powerequipment state information respectively based on MapReduce and Naive Bayesalgorithm and K-means algorithm. These two methods provide a parallel computingapproach for power equipment fault diagnosis.Taking transformer fault diagnosis in the smart grid as an example, we construct anexperimental platform of electric power equipment fault diagnosis. Finally, through atransformer fault diagnosis experimentation of massive DGA data as an example, theresults show that the speed of parallel fault diagnosis is faster than standaloneenvironment, and it can meet the requirement of rapid fault diagnosis for the massiveelectric power equipment status datasets in the smart grid.