Novel approach to reconfiguration power loss reduction problem by simulated annealing technique

Abstract

Summary The network reconfiguration is done by changing the status of the switches, mainly for 2 reasons: an active power loss reduction and load balancing attracting the attention of distribution engineers for quite a long period of time. In this article, solving method for the active power loss reduction is given. Searching for the relevant radial configurations is done by a simulated annealing technique. To aid the search, a program for checking the connectivity of the power system, with imposed radiality constraint, is enhanced by a new approximate power flow method. The applied power flow program is very efficient and fast but can be used only as an indication for the loss estimation because of the insufficient method accuracy. At the end of the algorithm, it is necessary to run an efficient power flow program to determine the real situation concerning loss reduction. A numerical example for Baran and Wu network is analyzed. The developed method converges to the global optimum given in a numerical example. The time duration of the used method is of an hour order and does not depend on the fast manipulation of incoming data files. This advantage makes the method interesting in the planning stage as well as the application in real time. Main contribution of this paper is a novel approach that exploits mostly the following mechanisms: network connectivity checking matrix, Lavorato et al criterion for imposing radiality constraint, and efficient power flow algorithm combined with precise power flow for exact loss calculation.

Cross‐entropy method for distribution power systems reconfiguration

“…Furthermore, they have the ability to avoid the convergence to local optima of the feasible region and go to a global solution, but they are very time consuming and their solutions are not stable. In recent years, the DPS reconfiguration is also achieved through metaheuristics based tools such as genetic algorithms, simulated annealing, artificial ant colony, tabu search, particle swarm optimization, hybrid differential evolution (HDE), and the artificial immune systems . The drawback of these approaches is that they are time consuming and their operation depends on careful selection of parameters.…”