Optimal placement and sizing of distributed generation

Abstract

In this thesis work we solve the problem of optimal placement and sizing of distributed generation by using an original Fuzzy Adaptive Particle Swarm Optimization algorithm and a Mixed Integer Linear Programming formulation of the problem. The goal of integrating Fuzzy Logic in Particle Swarm Optimization is to be able to overcome some of the classical disadvantages of the algorithm. Particle Swarm Optimization has been traditionally criticized for the complexity to set the acceleration constants of the algorithm and the low exploration capabilities of the algorithm. In this thesis work, it is proposed a new implementation of Particle Swarm Optimization that avoids the complex setting procedure of the acceleration constants of the algorithm, while aiming at improving the exploration capabilities of the algorithm. In this thesis work it is also analyzed a novel Mixed Integer Linear Programming formulation of the problem of optimal sitting of distributed generation in distribution networks implemented in DERCAM, a tool developed at the Lawrence Berkeley National Laboratory. The results of the two models are analyzed and contrasted against each other for a real case study of an islanded microgrid located in the north of the U.S. Results obtained in the case study depict the differences between the two analyzed approaches to solve the problem. It is found that over and under estimations of voltage magnitudes in high and low loading scenarios of distribution networks have the potential to impact investment decisions in distributed generation capacity for the linear formulation of the problem. Also the models analyzed depict the synergies between renewable energy technologies and thermal generators to increase energy savings while maintaining the operation limits of the grid