Cross decomposition for multi-area optimal reactive power planning

This paper presents a novel decomposition based algorithm to solve the reactive power planning problem in large scale multi-area power systems. Extensive studies on the application of the Benders decomposition method (BDM) have been conducted to solve this complex problem. Nevertheless, these applications are based on the difficult evaluation of the plane cuts from the solution of the master problem. In this paper, the authors introduce the possibility of solving reactive power planning based on an iterative optimization process between the BDM subproblem and the subproblem of a Lagrangean relaxation decomposition method (LRDM). The overall process is referred to as a cross decomposition algorithm (CDA). The merits of the proposed approach are shown by the evaluation of plane cuts based on the solution of easy-to-solve subproblems. Moreover, the proposed approach is extremely efficient in multi-area (zone decomposition) power systems. Test results are demonstrated for the 3-area IEEE 30-bus and the 8-area 180-bus systems