Data-Driven Adjustable Robust Unit Commitment of Integrated Electric-Heat Systems

Abstract

This paper proposes a data-driven adjustable robust unit commitment (UC) model for integrated electric-heat systems (IEHSs), where the data-driven, and robustness adjustable features are endowed by the novel uncertainty modeling method. The devised uncertainty set is the intersections of a series of hyperplanes. It combines the merits of the traditional box-like uncertainty set, whose underlying computation burden is moderate, and the convex hull formed by historical data, which would lead to less conservative operation strategies. The boundaries of the bounding high-dimensional rectangle of the uncertainty set are set as decision variables to achieve the balance between the operation costs, and risk. The two-stage decision-making framework is employed in the robust UC model of IEHSs, in which the district heating networks are operated at the variable-flow-variable-temperature mode. The modified column-and-constraint generation (C, and CG) algorithm is employed to solve the proposed robust program, where an additional feasible region reduction step is embedded to decrease the computation burden of the worst case identification. Simulation results on two test systems verify the effectiveness of the proposed methods.