Applications of hybrid EP-ACO for power distribution system loss minimization under load variations

Abstract

Distribution system reconfiguration is one of the methods that is commonly used in order to improve the system reliability and reduce the system real power loss. The distribution system reconfiguration is normally done by altering the status of switches. When the status of switches on the distribution system is changed, the topology of the system is also changed. During the normal operations, load currents change at different time due to the power usages of different customers. The operation conditions of distribution systems change when the load currents vary. As a result, the topology of distribution system needs to be altered from time to time so the system real power loss can be reduced and the reliabilities can be improved. In this paper, integration of Evolutionary Programming (EP) and Ant Colony Optimization (ACO) is proposed. The objective is to minimize the system real power loss with minimum number of switch operations by considering the load currents variations. Each day is divided into 12 time slots. The peak load current is used for each time slot. By considering the number of switch operations and system real power loss simultaneously, the proposed approach is able to identify the proper switch operations so the system real time loss within a period of time can be minimized. Instead of using constant values as the desirability of state transition during the search process of ACO, application of the grey correlation grade is proposed. Verification is performed using the loading information and the topology information obtained from the Outage Management System (OMS) of Taiwan Power Company (TPC). The results show that by applying the integrated EP and ACO with proposed desirability calculation, the distribution system real power loss can be reduced with proper switch operations for variable load currents.