Optimization of Controller for Islanded AC Microgrid System

Abstract

Global optimization algorithms are becoming popular in a fast rate as they are able to solve different real life challenges. Swarm based nature inspired optimization techniques are developing in a fast pace due to its global acceptance and their capability in replicating various real life challenges and solving them efficiently with a better computation rate. Hybridization of these swarm based intelligent techniques with multiobjective based solution techniques is creating a wide door in the field of optimization as many real problems include multiple objectives which needs to be optimized. In this dissertation, two nature inspired hybrid optimizing algorithm is proposed incorporating swarm intelligence based firefly algorithm (FA) with multi objective based non-dominated sorting technique to form Non-dominated Sorting Firefly Algorithm (NSFA) and Non-dominated Sorting Whale Optimization Algorithm (NSWOA) where swarm based intelligence technique of Whale Optimization is hybridized with nondominated sorting technique. This study also demonstrates their application in optimization of controller parameter of islanded microgrid . Moreover, how the optimized parameter affects the dynamic performance of microgrid during load variation is also demonstrated in this dissertation. The purpose of incorporating FA and WOA with non-dominated sorting technique to form NSFA and NSWOA respectively is to enhance global searching capability of conventional FA and Whale Optimization Algorithm (WOA) for complex optimization problems. Based on statistical tool, SPSS software, statistical analysis was carried out where the performance of the proposed hybrid NSWOA algorithm is compared with the performance of multi objective Nondominated sorting genetic algorithm (NSGA) and Strength Pareto Evolutionary Algorithm (SPEA) in optimizing the controller parameters for the microgrid model used in this study. Similarly performance of NSFA is also compared with NSGA-II to analyze the ability of NSFA. This dissertation shows that NSWOA and NSFA is able to stabilize the system with faster computation rate with lesser number of iteration. Moreover, the system optimized using NSWOA and NSFA techniques provides better damping performance compared to NSGA-II. From this study it is obtained that NSWOA and NSFA requires 4:033 and 2:3667 iterations respectively to reach to best optimum solution which is significantly less than other existing algorithms. Moreover, the computational time required by NSWOA and NSFA is 2:9201 sec: and 4:5459 sec: respectively which proves that they reach to convergence at a much faster rate which is significantly less than existing NSGA-II and SPEA algorithms.