Intelligent Identification Method of Sub/Super-synchronous Oscillation Parameters Based on CNN and Synchrophasor

With the increasingly prominent “double high” characteristics of power system, the occurrence probability of sub/super-synchronous oscillation is also greatly improved, which seriously affects the safe and stable operation of the system. Accurate identification of sub/super-synchronous oscillation parameters is the basis for suppressing oscillation, so this paper proposes an identification method based on convolutional neural network. The local correlation and weight sharing of convolutional neural network make it more capable of feature learning and expression. The parameter identification of sub/super-synchronous oscillation is modeled as a regression problem, and the convolutional neural network trained offline is used for online identification. Simulation results have verified that this method can not only identify the parameters of sub/super-synchronous oscillation completely, but also reduce the complexity of the identification of oscillation parameters.