Using ant colony optimization for loss minimization in distribution networks

Distribution systems have to bear different loading patterns at different times. This change in load causes distribution feeders to be overloaded at sometimes and lightly loaded at others. With this load variation, operating conditions of distribution system also vary. If not compensated well, voltage at different buses goes out of nominal range and real loss on the feeders also increases, leading to high operating cost of the system. However, with the advancement in automation of distribution networks, systems like SCADA have made possible to change the topology of the distribution network in real time for minimizing real loss in the system and for improving voltage profile at the buses. The configuration of the system is changed by changing the status of switches such that load is transferred from heavily loaded feeders to lightly loaded feeders. This paper proposes using ant colony optimization (ACO) for solving reconfiguration problem for loss minimization. Ant system, one of the ACO algorithms, has been implemented in a novel hypercube framework on a 33-bus test system and the results obtained show that ACO performs as well as any other proposed method for loss minimization. In fact, this ACO algorithm found the most optimum solution found so far by any other method proposed in the literature for the 33-bus test system considered.