Research On Capacitance Type Equipment Fault Diagnosis Algorithm

Capacitive equipments occupy a very important position in the power transmission and transformation equipments.Their failures not only affect the safe operation of the whole substation,but also endanger other equipments and personal safety.So it is very important to realize the fault diagnosis of capacitive equipments,and the dielectric loss factor is an important index for the diagnosis of capacitive equipments.In this paper,we proposed a method based on support vector machine to modify the dielectric loss,tan ? was corrected reasonably under different climatic conditions,and the equivalent value was obtained.Due to the on-line monitoring results of capacitive equipments were often affected by temperature,humidity,pollution and so on.Accurate diagnosis of equipment state have been affected.So it was very necessary.to modify the dielectric loss factor reasonably in the different climatic conditions and obtain the equivalent value.In this paper,a capacitive equipments modified model of vector machine was set up firstly.Then,we use genetic algorithm to optimize it.Next,a method of fault diagnosis based on wavelet tree transform was proposed.In order to realize the accurate diagnosis of the equipments,the tan ? which is modified by the support vector machine was used to detect the mutation and the continuous mutation.Sudden change was detected by the lower level wavelet tree,and the upper level wavelet tree was used to detect whether there was a persistent mutation.The symbol of the occurrence of persistent mutation was the wavelet tree node which was more than the threshold value of the data segment node.The integral wavelet tree algorithm was a synthesis of the two methods mentioned earlier,which could be used to make a comprehensive diagnosis of the two types of capacitive equipment.Finally,the fuzzy neural network diagnosis algorithm was proposed.Temperature,humidity,dielectric loss factor and capacitance were used as the input of the network.The diagnosis results were gotten through the network processing.The convergence of fuzzy neural network was very fast.The actual output of the simulation was very close to the theoretical output,which can be used for reference for the further research of fault diagnosis of capacitive equipments.