A Novel Stochastic Unit Commitment Characterized by Closed-Loop Forecast-and-Decision for Wind Integrated Power Systems
Haotian Wu, Deping Ke, Lin Song, Siyang Liao, Jian Xu, Yuanzhang Sun, Ke Fang
Abstract
In general, the day-ahead wind power prediction and the probability distribution of relevant forecast error critically determines the solution of stochastic unit commitment (SUC) applied for wind-integrated power systems. However, in a conventional open-loop forecast-and-decision (OFD) framework, the accuracy-oriented forecaster independently computes the predictive information and unidirectionally pass them to the SUC model. In order to intrinsically upgrade the solution quality of SUC, this article proposes a closed-loop forecast-and-decision framework-based SUC (CFD-SUC) model where the predicted wind power scenarios are elaborately expressed as functions of adjustable parameters of wind power forecast (WPF) models. Thus, compared to the conventional OFD framework-based SUC (OFD-SUC) model that only optimizes dispatching variables, the additional searching space spanned by the WPF model parameters enables the CFD-SUC model existing superior solutions that can further remarkably reduce the risk cost-based objective function. A customized small-step iteration-based solving method (SSISM) is then proposed to find optimal solution of the CFD-SUC model. Statistical tests implemented on two systems verify the efficiency and effectiveness of SSISM. Moreover, the scheduling results derived from the CFD-SUC model are statistically predominant for risk operation of the system, compared with those from the OFD-SUC model.