Data Mining And Its Application In Fault Line Selection For Non-effective Grounded System

This paper discusses the theory and methods of data mining, introducesthe applied actuality of data mining in the electric power system, analyzes theproblems in detecting phase-to-ground fault in ineffectively grounded systems, andfinds the development of data mining gives new ways for the study of line selectiondata, so that the valuable information can be acquired from a large number of data. This paper uses statistical methods to recognise the parameters and theirdistributing rules and understand well the faults, applies principal componentanalysis and clustering analysis to fault feature selection and fault classify, so that itoffers evidences for modeling accurately for ground faults and witnessing thecorrectness of fault detection. Based this, proposes the idea of setting up system offault detection for line selection, and establishes management rules to perfectingmanagement. It is very significant to solve fault modeling, advance the technique ofline selection, and deal with the problems on-site to improve the right percentage ofline selection.