Optimal placement of DGs in a multi-feed radial distribution system using actor-critic learning algorithm

Abstract

Multi-feed radial distribution systems are used to reduce the losses in the system using reconfiguration techniques. Reconfiguration can reduce the losses in the system only to a certain extent. Introduction of distributed generators has vastly improved the performance of distribution systems. Distributed generators can be used for reduction of loss, the improvement of the voltage profile, and reliability enhancement. Distribution generators play a vital role in reducing the losses in the distribution system. Placement of distributed generators in a multifeed system is a complex task to be solved using classical optimization methods. Classical optimization techniques may sometimes fail to provide a converged solution. Installation of distributed generators at suitable locations in a multi-feed system is found in this paper using the actor-critic learning algorithm. Actorcritic learning approach uses temporal difference error as a signal in making judgements regarding actions to be taken for future states in accordance with rewards that have been obtained by applying the present policy. The approach is applied to a standard 16-bus distribution system for reduction of system losses, and the results are discussed.