Damping of subsynchronous oscillations using adaptive controllers tuned by artificial neural networks

Artificial neural networks (ANNs) are utilised to adapt the controller gains of two widely used control schemes i.e. static VAr compensators (SVC) and excitation controllers (EC), for the damping of subsynchronous resonance (SSR) on a power system. To have good damping characteristics of SSR modes over a wide range of operating conditions, the parameters of the adaptive controllers are adapted based on generator loading conditions. Multilayer feedforward artificial neural networks (ANNs) are developed to serve for the purpose of controller parameter adaptation. The inputs to the ANN include the real power output P and reactive power output Q which characterise generator loading conditions. The outputs from the ANN are the desired controller gains. Time domain simulations are also performed on the IEEE first benchmark model to demonstrate the effectiveness of the proposed adaptive control schemes