A computational comparison of two different approaches to solve the multi-area optimal power flow problem

In the increasingly competitive electric industry there's a need to carry out an optimal management of the interconnected systems. This paper deals with a computational comparison of two different approaches to solve a multi-area optimal power flow (MOPF) problem. The first approach considers all the interconnected sub-systems as a single system; the optimal operating strategy that is obtained is for the whole system and the sub-systems are operated in a centralized fashion; a one step interior-point algorithm is employed in this approach. The second approach is based in decomposition techniques that lead to an iterative algorithm; this approach is computationally more expensive than the former but enables the independent operation of each interconnected sub-system while the global optimum is also achieved. In this approach, the MOPF problem is solved by a combined Lagrangian-relaxation/interior-point (LR/IP) method. In both cases, an AC network model is used, and the problem in the framework of sequential linear programming (SLP).