An Intelligent Controller to Minimize the Torsional Oscillations

Modern interconnected power systems are highly nonlinear and their parameters vary continuously with time. Therefore, the use of intelligent techniques in controlling power systems is inevitable. This paper proposes a wavelet neural PI ontroller (WNPIC) to damp the subsynchronous and low frequency cscillations. The closed-loop control system contains two wavelet neural networks (WNNs). The first one is a feedforward wavelet neural network that is used as a predictive model of the controlled plant. The second network is a wavelet neural PI-like controller, which combines abilities of a wavelet neural network and PI controller. The gradient steepest descent method is used to train the wavelet neural networks. In the whole design process of controller, an accurate model of the system is not required. The IEEE second benchmark model is used to evaluate the performance of the proposed controller in damping the oscillations. The results show that the torsional and low frequency oscillations are easily suppressed by using the proposed controller

کلیدواژه ها:

Torsional oscillation ، subsynchronous resonance (SSR) ، Low frequency oscillation ، wavelet neural network