Intelligent control of inter-area oscillations in power systems

In this paper, a neural network-based control strategy is developed for damping the inter-area electromechanical oscillations for any operating point and/or network configuration. The power system stabilizers (PSSs) have been recognized as an effective means to eliminate the inter-area oscillations. Hence, the design of PSSs becomes an important aspect in controlling oscillations. Moreover, two specific requirements for the control strategy are very important: the coordination of the control actions and the decentralized structure of the control. A suitable and easily implementable control architecture, including layered feedforward neural network (LFNN), is proposed for the PSS of the synchronous machine. Each neural network is trained to give, as output to a pre-defined set of local measurable input variables, the desired values of the feedback parameters necessary to implement a decentralized linear feedback control law. Test results on a two-area power system prove the validity and usefulness of the suggested control strategy