Day-ahead reserve scheduling approaches under wind uncertainty

Abstract

Due to the variability of wind generation, its integration presents a significant challenge to power system operators to maintain adequate reliability levels and ensure least cost operation. This paper compares deterministic and stochastic unit commitment (UC) approaches for handling wind generation uncertainty at the moment of scheduling reserves for the next day. The role of including an explicit reserve constraint (ERR) in the UC formulation is discussed for different levels of wind penetration. The different reserve scheduling approaches are tested on the Chilean Northern Interconnected System, SING. Our results for the SING suggest that when modeling generator and transmission line outages the benefits in terms of expected total cost reduction by using stochastic UC are rather modest if an adequate level of ERR has been enforced. However, stochastic UC may still offer significant benefits in terms of the robustness of the solution (less variability across samples) and in case the ERR has been underestimated.