Abstract

In this paper, we present the greedy reconfiguration algorithm (GRA) and the fast greedy reconfiguration algorithm (FGRA) - two reconfiguration algorithms that minimize the energy not supplied (ENS) of medium-voltage (MV)-distribution network (DNs). The algorithms are derived from a graph-theoretic model for the MV-DNs and an iterative formulation for the ENS. Given a DN composed of n loads, GRA yields an initial radial topology with minimal ENS after executing n-1 iterations of a multistage decision process. Next, GRA executes a verification algorithm to evaluate all of those candidate topologies randomly discarded during the decision process. After the second stage, GRA announces the optimal topology as well as the minimum ENS of the system. The second algorithm proposed is called FGRA and is obtained when heuristic graph compression is performed throughout the execution of GRA. The aforementioned graph compression reduces the dimension of the problem, thereby producing substantial savings in computing time at the expense of possibly achieving a suboptimal solution. © 2008 IEEE.