Multiple Objective Modelling by Forming Dynamic Clusters of Peak Loads and Distributed Generations for Energy Management in Grid Connected Mode

Abstract

Implementation of microgrid formation reflects proper incorporation and organized control of distributed generations in distribution system. Unlike the conventional ways of energy management by multiple microgrid in Grid connected mode, this paper proposed an idea by forming temporary clusters following the loads when they are at their peak and supplying some of the peaks by Distributed generators (DGs) installed with the loads inside the clusters. For implementing this, k-means algorithm is used to make a number of clusters consist of loads and within those clusters probable DGs are marked. Multiple objective optimization technique, Nondominated Sorting Genetic Algorithm-II (NSGA-II) has been used to determine the trade-off of feasible sharing by DGs, defined within the clusters. For decision making from the trade-off solutions a fuzzy based method have been used. The work in this paper has been done in hourly basis for a whole day by running simulations in OpenDSS from MATLAB and also system improvements has been observed.