An Integer Programming Approach and Implementation for an Electric Utility Capacity Planning Problem with Renewable Energy Sources

This paper presents an integer programming model and algorithm for an electric utility capacity expansion problem which considers the option of investing in nondispatchable or renewable energy sources. A branch-and-bound algorithm is proposed for this problem in which the continuous relaxation of the subproblem associated with each node in the enumeration tree is solved via an efficient two-phase procedure. This procedure solves a deterministic approximation of the problem in the first phase in order to determine a quick near-optimal solution. The resulting solution is subsequently refined in a second phase using more accurate techniques to represent the negative load due to the renewable sources, and to perform the probabilistic production costing. This technique conserves about 80% of the effort which would be required without the deterministic phase. An implementation of this approach is described for the Tijuana-Mexicali subsystem of the Mexican utility, which is not connected to the rest of the Mexican electric system, and which faces a very high summer peak load and a comparatively low winter load. The results suggest that along with a prescribed capacity expansion of conventional equipments, the utility should invest in some solar cooling systems, and, more pertinently, should involve itself intensely in conservation measures in homes of individual customers.