A data-adaptive robust unit commitment model considering high penetration of wind power generation and its enhanced uncertainty set

Wind power generation is increasingly penetrating into the power grid, which brings great challenges to the dispatch of power systems. With the popularization of data mining technology, further exploration of the random characteristics of wind power based on the available wind power data can significantly improve the applicability of scheduling decisions. In this paper, a novel data-adaptive robust unit commitment model under high penetration of wind power is proposed, which derives a robust dispatch solution with minimal generation cost while hedging against the worst case in the uncertainty set. Firstly, copula theory is carried out to formulate a joint probabilistic distribution function and capture the correlation of power outputs among multiple wind farms. A large number of wind power scenarios are then generated and the imprecise Dirichlet model (IDM) is applied to derive the boundaries of wind power generation, which helps to construct a more practical polyhedron uncertainty set. Moreover, due to the correlation of adjacent wind farms, the auxiliary variables which determine the fluctuation of wind power have a synchronous trend. Here, the synchronous characteristic is introduced to the enhanced polyhedron uncertainty set by means of the synchronous volatility of the auxiliary variables in adjacent wind farms. Experimental studies are conducted out on a modified IEEE-118 bus system and the obtained scheduling solution is turned out to be superior under wind power uncertainties, which verifies the effectiveness of the proposed data-adaptive robust unit commitment model.