Based On Rough Sets And Bayesian Network Fault Diagnosis System Of Research And Application

The power is very important in people’s daily lives and production, therefore, inrecent years, research on the power grid fault diagnosis methods has been a hot spot.For now, a lot of methods for power grid fault diagnosis have been proposed byscholars, such an expert systems, artificial neural networks, genetic algorithms and soon. Although these methods can solve the problem to a certain extent, most of themhave problems like knowledge acquisition and parameter settings.Fully studying variety of power grid fault diagnosis methods, and taking theredundancy and incompleteness of the grid’s information into account, the thesisproposed a combination of rough set theory and Bayesian network grid fault diagnosismethod. Firstly, in order to avoid the decision table of a large-scale power grid is toolarge, resulting in the process of grid fault diagnostic becoming a complex issue, thisthesis divided the grid into subnets that are both independent and interconnected. Andthen used a mutual information method based on information entropy for attributesreduce, to reduce the number of attributes in the decision table, thereby reducing thecomplexity of modeling and reasoning of the Bayesian network. When the systemreceives real-time fault information, the established Bayesian network model can beused to find which kind failure have the maximum probability of occurrence under thecondition of the real-time fault information.In order to prove the effectiveness of the grid fault diagnosis method based onrough sets and Bayesian, the thesis used the Matlab for simulation, and the simulationresults showed that the method could always get ideal results whether the informationfor fault diagnostic is complete or uncertainty. Because of its correctness and tolerancein grid fault diagnostic, the method based on rough set theory and Bayesian was used inthe fault diagnosis and positioning module of the grid visualization management systemduring its developing, and it approved to have achieved the desired results.