The Study Of Converter Fault Diagnosis Based On Knowledge Base

In recent years as the problem of the environment and energy hasbecome more and more severe, clear energy developing has become acenter issue of the world, in which the wind power gradually become thefastest developing and most matures technology. Since more and morewind farms are established and utilized, people’s concerning of thosewind power system’s reliability obviously increased. Currently among theelements of the large scale wind power system, one of the main objectsthat cause the system fault is the converter, thus the fault diagnosistechnology of the wind power converter has gained a lot of focus,especially the parallel-type full scale wind power converter for the largescale wind power system.However, the converter is a complicated nonlinear power electronicsystem, which is hard to establish a completed analytical model.Therefore this paper has studied the converter fault diagnosis methodbased on a knowledge base, which uses the known knowledge of thewind converter fault to extract the fault feature, and uses these features tocompare with the practical fault data and to judge the happening of thefault, which means that this method has effectively avoided thedisadvantage of unable to establish a model for the converter. In thispaper firstly different kinds of converter fault have been analyzed withthe method of simulation from which the fault phenomenon and fault feature extracting method respectively. Then the paper show the design ofthe entire converter fault diagnosis system, including the design andrealization of the monitoring system, the construction of the knowledgebase combined with the converter fault analyzing result and the designand realization of the status detection part and inference machine. At theend of this paper the whole diagnosis system are testified with thehardware-in-loop system and the result of the test has shown the validityof the diagnosis method and the system.