Are optimal planning models for power systems able to deliver the flexibility they promise? The impact of high-quality models and formulations


Speaker


Abstract

Approximations made in traditional optimal energy planning (unit commitment) models can result in suboptimal or even infeasible schedules for the system. This is a natural consequence of a hidden and systematic overestimation of the actual system flexibility, resulting in higher errors than expected of actual costs and curtailment variable renewable sources. With increasing wind and solar penetration, these errors will become economically more significant. In this presentation, we show the impact of 1) the quality of the optimal energy planning models, i.e.,  inaccuracies due to (unconscious) approximations, and 2) the quality of the underlying mixed-integer program (MIP) formulation, i.e., tightness. Improving these two aspects leads to better (“more optimal”) solutions, i.e., lower costs, due to a more accurate representation of the actual system flexibility, while solving the resulting problem even faster.