Data-Driven Assisted Chance-Constrained Energy and Reserve Scheduling with Wind Curtailment

Xingyu Lei,Zhifang Yang,Junbo Zhao,Juan Yu
DOI: https://doi.org/10.1016/j.apenergy.2022.119291
IF: 11.2
2022-01-01
Applied Energy
Abstract:Chance-constrained optimization (CCO) has been widely used for uncertainty management in power system operation. With the prevalence of wind energy, it becomes necessary to consider wind curtailment as a dispatch variable in CCO. However, the wind curtailment will cause an impulse for the uncertainty distribution, yielding challenges for explicit modeling of chance constraint. In this paper, a novel data-driven framework is developed to handle this issue. By modeling the wind curtailment as a cap enforced on the wind power output, the proposed framework constructs a data-driven surrogate to describe the relationship between wind curtailment and the tightening boundary of chance constraints. This allows us to reformulate the CCO with wind curtailment as a mixed-integer second-order cone programming (MI-SOCP) problem. An error correction strategy is developed by solving a convex linear programming (LP) to improve the modeling accuracy. The notable characteristic of the proposed method is that the chance constraints with wind curtailment can be explicitly and accurately handled without any distribution assumptions, which provides great convenience for the power dispatch under a high percentage of renewables. Case studies performed on the PJM 5-bus and IEEE 118-bus systems demonstrate that the proposed method is capable of accurately accounting the influence of wind curtailment dispatch in CCO. It shows that the proposed method outperforms existing methods by producing more economic and secure solutions for the chance-constrained energy and reserve scheduling problem under uncertainties.
What problem does this paper attempt to address?