Improved K-means clustering based distribution planning on a geographical network

Abstract

This paper presents a distribution planning on a geographical network, using improved K-means clustering algorithm and is compared with the conventional Euclidean distance based K-means clustering algorithm. The distribution planning includes optimal placement of substation, minimization of expansion cost, optimization of network parameters such as network topology, routing of single/multiple feeders, and reduction in network power losses. The improved K-means clustering is an iterative weighting factor based optimization algorithm which locates the substation optimally and improves the voltage drop at each node. For feeder routing shortest path based algorithm is proposed and the modified load flow method is used to calculate the active and reactive power losses in the network. Simulation is performed on 54 nodes based geographical network with load points and the results obtained show significant power loss minimization as compared to the conventional K-means clustering algorithm.