Application of evolutionary programming to optimal siting and sizing and optimal scheduling of distributed generation

Abstract

This study proposes a method for sizing, siting, and scheduling distributed generators using an evolutionary programming algorithm. The distribution system is modelled using a loadflow formulation wherein the distributed generator is modelled as injected real and reactive power. Evolutionary programming is implemented by perturbing distributed generator outputs and evaluating the fitness values of the resulting systems. Optimal locations and sizes of distributed generators are determined using a fitness value based on system loss reduction, while an optimal distributed generator schedule is determined for a 24-hour period using a fitness value based on energy cost. A mathematical model is developed to represent different customer types within the distribution system and different types of energy sources are represented based on their availability and cost. Spot market prices are modelled based on a typical 24-hour price curve. The method is found to be effective and flexible based on results for a 69-bus test system.