Optimal Sizing and Placement of Distributed Generation Using Genetic Algorithm Based on Bayesian Network

This paper focuses on optimal sizing and placement of distributed generation (DG) accessed to distribution system. Considering uncertainty and correlation between wind speed, solar irradiation and load, probabilistic power flow calculation are carried out by Monte Carlo model based on Bayesian network. It saves much computational costs, compared with calculation based on time series. Thus a probability optimization model aimed at minimizing total cost can be proposed with the constraints of voltage and branch power flow. Genetic algorithm with elite retention (GAER) is used to obtain the optimal results. By this model, annual expected generation capacity and cost of DG are estimated in better precision. Branch power flow is proved as one of the main constraints.