Application research on over-voltage smart recognition system for 110kV substation

This essay presents an efficient intelligent over-voltage hierarchical recognition system for 110kV substation based on LSSVM optimized by gridding optimal algorithm. It is very indispensable to analyze and find the causes of over-voltage for ameliorating insulation coordination in power system. According to analyzing field over-voltage data from over-voltage on-line monitoring devices installed in some 110kV substations, an overvoltage hierarchical recognition tree is established by analyzing timedomain and time-frequency features. On the basis of the demands of over-voltage hierarchical recognition, LSSVM is adopted as the classifier to establish a hierarchical recognition model, and the gridding optimal algorithm is employed to optimize the arguments of LSSVM to enhance classification efficiency. The field data testing outcomes indicate that this method is accurate and effective.