NEURAL NETWORK BASED CONTROL DESIGN TO EXTENUATE SUBSYNCHRONOUS RESONANCE

Subsynchronous resonance (SSR) is a menace in the electrical power system in which disturbances are exchanged between electrical and mechanical systems. This paper aims to examine the neural network (NN) controller incorporated in a modified unit template algorithm to alleviate SSR using a static synchronous compensator (STATCOM). To understand the problem of SSR, time domain analysis and frequency scan are presented. A single machine infinite bus system is studied and simulated in MATLAB/Simulink to substantiate the efficacy and robustness of the proposed control strategy. The deviation in generator rotor speed is as high as 10.8% due to SSR oscillations, which increases mechanical stress, causing heavy damage to the mechanical shaft system. The STATCOM equipped with the proportional and integral (PI) controller in the proposed strategy reduces the deviation to 0.19% and further to 0.006% with the annexation of the NN controller.