Unit Commitment With Wind Power: Model And Algorithm

Day-ahead unit commitment is a fundamental optimization problem for power system. It is able to improve economics and security level of power system based on practical experience. With large-scale wind power integration, uncertainty becomes strong even threats security of power system. Simultaneously, unit commitment is able to improve security level of power system. Therefore, this paper researches how to keep system secure and economic as absorbing wind power by reasonable unit commitment.This paper focus on the unit commitment (UC) in power system with high wind power penetration. Based on literatures, uncertainty and variability are the most obvious features of wind power generation. Uncertainty means the predicted wind power generation is inaccurate and variability means the real-time output of wind power rapidly fluctuates. Based on these two features of wind power, several reserves are proposed and considered in this paper. They are online generation capacity reserve, ramping reserve and transmission capacity reserve. An algorithm based on the interval programming theory is used to determine these reserves. It guarantees all wind power can be dispatched.To balance the relationship between reserve level and economics, a hierarchical unit commitment (HUC) model based on confidence level is proposed. HUC divides wind power output into a high possibility interval and a low possibility interval based on confidence level. And it uses different scheduling strategies for different intervals of wind power output. HUC is able to guarantee system security over the interval of wind power output while decreasing reserve level.In a practical power system, there are several forms of power generation. The feature of different forms of power generation is different. The feature affects the ability of power generation reserve that is important for wind power integration. To improve accuracy of the proposed UC model, this paper researches the hydro-thermal-wind coordination in day-ahead unit commitment. Based on the features of different power generations, a hydro-thermal-wind coordination method is proposed to dispatch wind power in safety. The hydro power is scheduled to deal with variability in wind power, and the hydro-thermal coordination is proposed to compensate power shortage or surplus caused by wind power uncertainty. Besides, the coordination in cascaded hydropower system is also considered to guarantee reservoir secure while fully taking advantage of water head benefit.Wind power brings uncertainty in not only active power distribution but also node voltage. The traditional stochastic unit commitment (SUC) only consider direct current power flow constraints but not node voltage constraints. So, the schedule result cannot guarantee node voltage safety. To improve the feasibility of the schedule result in real-time dispatch, a stochastic unit commitment with AC power flow constraints is constructed in this paper. This SUC model is a large-scale mixed-Integer nonlinear programming (MINLP) problem and very complicated. It is hard to solve this MINLP problem directly. Therefore, a decomposition algorithm, based on the augment lagrangian relaxation with variable duplication and the auxiliary problem principle, is employed to decompose the primal MINLP problem into two sub-problems with separate structures. By utilizing the separate structures of the sub-problems, an efficient solution based on parallel processing techniques has been developed to substantially decrease computational time.