Performance evaluation of a PBIL-based power system damping controller

Recently Population-Based Incremental Learning (PBIL) algorithm has been applied to a range of problems in engineering with promising results. However, in most of the literature the standard PBIL with fixed learning rate has been used. It has been shown that when applied to dynamic environment as the one encountered in power systems, adaptive learning rate is more appropriate to use. In this paper, Population-Based Incremental Learning (PBIL) algorithm with adaptive learning rate is used to tune the parameters of a power system controller for damping power oscillations in a multi-machine power system. The optimization of controller´s parameters has been performed over pre-specified range of system operating conditions. Robustness Evaluation of the proposed controller based on eigenvalue analysis and time domain simulation shows that the proposed controller is more robust than the conventional controller over the range of operating conditions considered.