Application of Machine Learning Methods in Active Power Security Correction of Power System

A method for active power security correction based on BP neural network is presented in this paper. The active power security correction is used to give a minimum regulation of the active power output of generators in order to prevent or alleviate overload of transmission lines and tie line groups. This paper first presents the problem of the traditional method based on sensitivity analysis, and then introduces machine learning methods, and applies the method of BP neural network in the active power security correction in power system. The mathematical model of this problem is given, and the principle of BP neural network and their application procedures are explained briefly as well. The method is proved to be valid, according to the results of active power security correction calculation for the IEEE-30 node system.