An Integrated Clustering-MILP-Simulation Approach for Distributed Energy Resource Sizing

A method for considering the variability of solar insolation, wind speeds, and electricity demand for sizing distributed energy resources in a hybrid renewable energy microgrid system is presented in this paper. This method incorporates to the sizing process the selection of groups of days that will represent the variability of the mentioned parameters in a year. The representative days are selected through k-means clustering and the effectiveness of the chosen days are verified through simulation. From groups of representative days, designs are produced using a mixed integer linear programming (MILP)based microgrid model. The best among these designs are selected using the net present cost and reliability of electricity supply as criteria. The method is verified by finding a design using all days in the year, and comparing the results with that derived using only the representative days.