基于动态场景集和需求响应的二阶段随机规划调度模型

Two-stage Stochastic Programming Scheduling Model Based on Dynamic Scenario Sets and Demand Response

风电及负荷场景集作为随机规划调度的输入,对决策结果影响大。为使场景集刻画出风电及负荷的波动性,采用考虑随机变量相关性的动态场景法生成风电、负荷及净负荷的场景集,并从与调度接轨的角度提出相应的指标对场景质量进行评价。建立了一种基于here-and-now和waitand-see(HN-WS)的二阶段随机规划调度模型。该模型包含了日前与实时的关联性,决策过程融入了对实时场景已实现情况下的考虑,优化目标涵盖了日前阶段的燃料成本及实时阶段的平衡矫正期望成本。此外,为减少弃风,在调度模型中引入了激励型需求响应,与火电机组的备用容量进行协同优化。最后,采用爱尔兰电网的风电数据、Elia电网的负荷数据及改进的IEEE-118节点系统验证了所提调度模型的有效性。

As the input of stochastic programming scheduling, wind power and load scenarios have great influence on the decision results. In order to make the scenario set reflect the characteristics of wind power and load, a dynamic scenario method is adopted which considers the correlation of random variables to generate wind power, load and net load scenario sets. And the corresponding index for evaluating the quality of scenarios is presented for the purpose of being in line with scheduling. A scheduling model based on the two-stage stochastic programming of here-and-now and wait-and-see （HN-WS） is established, which contains the relevance between the day-ahead and real-time scheduling, and the realization of the real-time scenario is integrated into the decision process, the optimization goal covers the fuel cost during the day-ahead stage and the expected cost of balance correction during the real-time stage. In addition, in order to reduce wind power curtailment, the incentive-based demand response is introduced into the model, which is cooperatively optimized with the reserve capacity of thermal units. Finally, the effectiveness of the proposed scheduling model is verified by adopting the wind power data of the EirGrid, the load data of the Elia power grid and the revised IEEE 118-node system.