Fault Diagnosis Model And Methods Of Research Of Wind Power Generation System Power Converter

Global energy shortage and environmental pollution, the cost of cleaner-energysources is the hot issue. Wind Power is very popular. Reliable performance of WindPower Generation System makes a great effect for many users electricity demand.Power electronic technology reduces the cost of wind power equipment. But itincreased the percentage of malfunctions caused by power electronic in wind powersystem. When the failure of converter occurs, accurate positioning and isolating is thekey to prevent serious accidents. Converter in wind power system is mixed withswitch variable structure system. Two kinds of methods of fault diagnosis based onbond graph model were proposed in this paper. The main research content is asfollows:The current situation of wind power and the classification of fault diagnosis weresummarized at home and abroad. The methods were introduced about fault diagnosisof power electronics, and new methods were proposed for fault diagnosis based onmodel.Analyzed several kinds of basic structure in the wind power system, and severalconverter topology structure and working principle were briefly introduced. A newkind of modeling theory was put forward to solve the problem of modeling difficulty.Details the method of switch in the bond graph modeling and builds the bond graphmodel of Converter in20-sim as the basis of fault diagnosis.As the system is complicated, it is difficult for modeling and the traditional faultdiagnosis is unobvious. The two kinds of methods of fault diagnosis based on bondgraph model are proposed. According to the causality of bond graph model, one of themethods is the use of temporal causal graph. The method overcomes the difficulty toget temporal causal graph from the system. The other method is the use of analyticalredundancy for fault diagnosis. It has been overcome the shortcoming of only suitinglinear system. The causal pathways of bond graph model are analyzed and getanalytical redundancy relations of variable. According to the residual direction,fault source is positioned and isolated. The experimental results show that timecausality diagram based on bond graph and analytical redundancy of fault diagnosismethod is feasibility and convenience.