Phase Balancing in Distribution Systems Using Greedy Randomized Adaptive Search Procedure (GRASP)

In distribution networks, the phase unbalancing is one of the major problems for electric utilities, since this can cause heavy loading in neutral and protection functions misoperation (among other disturbances). These problems, however, can be solved through phase balancing which is sorted out efficiently through the mixed nonlinear integer programming. Thus, in this work, the GRASP meta-heuristic, aided by a local search through the VNS algorithm, is formulated to balance a large electrical network and to foster a system planning. For this purpose, strategies of the balancing problem are used to guide the search within the space of solutions and, thus, to find configurations which present a minimum unbalancing phase voltage. The tests were carried out in a three-phase 29-bus network and the results obtained were useful to prove the feasibility of the methodology, by means of the improvement of the operation indexes.