计及源荷不确定性的智能小区能源系统两阶段鲁棒优化调度

Two-Stage Robust Optimal Scheduling of Energy Systems in Smart Community Considering Source-Load Uncertainty

为解决电力系统中电源侧和负荷侧的不确定性对电网调度计划的影响,针对电源侧,考虑风电与光伏出力的不确定性,分别建立风电与光伏的概率密度函数模型,通过拉丁超立方采样方法生成场景并进行缩减,从而得出风电与光伏出力区间;针对负荷侧,考虑柔性负荷对电网消峰填谷的作用,提出基于智能小区的综合需求响应两阶段鲁棒优化模型。在日前阶段,以电网系统运行成本和碳交易成本最小为优化目标,考虑源荷的不确定性,基于价格需求响应模型,从而确定日前调度方案。在日内阶段,基于日前阶段优化结果,以智能小区运行成本和碳交易成本最小为优化目标,建立两阶段鲁棒优化模型,通过列约束生成(column-and-constraint generation,C&CG)算法将目标函数进行转换,采用Karush-Kuhn-Tucker条件和Big-M约束方法将max-min形式优化问题转化为混合整数线性规划(mixed integer linear programming,MILP)模型。最终,通过算例验证了所提模型的正确性以及算法的有效性。

To address the impact of uncertainties in both the generation and demand sides of the power system on grid scheduling plans,a two-stage robust optimization model is proposed for integrated source-load uncertainty management.For the generation side,the uncertainties in wind and solar power outputs are considered.Probability density function models are established for wind and solar power,and the Latin hypercube sampling method is employed to generate scenarios and perform scenario reduction,resulting in power output intervals for wind and solar generation.For the demand side,the role of flexible loads in peak shaving and valley filling is considered,and an integrated demand response model based on smart communities is proposed.In the day-ahead stage,aiming to minimize the system operating cost and carbon trading cost while considering source-load uncertainties,a price-based demand response model is developed to determine the day-ahead scheduling plan.In the intra-day stage,based on the optimized results from the day-ahead stage,a two-stage robust optimization model is formulated with the objective of minimizing the operating cost and carbon trading cost of smart communities.The column-and-constraint generation(C&CG)algorithm is employed to convert the objective function,and the Karush-Kuhn-Tucker(KKT)conditions and Big-M constraint method are utilized to transform the max-min optimization problem into a mixed integer linear programming(MILP)model.The correctness of the proposed model and the effectiveness of the algorithm are verified through case studies.