考虑电热综合需求响应的虚拟电厂优化调度

Optimal Dispatching of Virtual Power Plants Considering Comprehensive Demand Response of Electricity and Heat Loads

针对配电网中以电、热为代表的多类型负荷的快速增长,以及可控机组、储能装置、风机 等分布式能源的协调调度问题,提出了考虑电热综合需求响应的虚拟电厂(virtual power plant,VPP)优化调度模型。首先,将风机、热电联产系统、多种储能装置、电锅炉、电热负荷集成为 虚拟电厂,在用户侧,将基于电价型和激励型需求响应措施相结合,建立电热综合需求响应模型;然后, 以最大化虚拟电厂运营利润为目标,采用机会约束模型描述风机、负荷预测的不确定性和内部功率平 衡,并考虑各机组运行约束和网络安全约束;在合理控制和协调各组件出力的基础上生成调度方案,最 后采用量子粒子群算法对模型进行求解。在算例中比较了不同需求响应方案对热电负荷曲线优化结 果、网络安全、虚拟电厂经济性的影响,比较了不同置信水平下虚拟电厂的调度结果,从而验证了模 型的可行性。

Aiming at the rapid grow th of multi-type loads represented by electricity and heat load in distribution netw ork,and the coordinated scheduling of distributed energy such as controllable units,energy storage devices and fans,an optimal scheduling model of virtual pow er plant( VPP) considering multi-type load comprehensive demand response is proposed.Firstly,w ind turbines,cogeneration system,various energy storage devices,electric boilers and electric heating load are integrated into a virtual pow er plant. On the user side,a comprehensive demand response model for electric heating load is established on the basis of the combination of electricity price type and incentive demand response measures. Then,aiming at maximizing the operating profit of the virtual pow er plant,the opportunity constraint model is used to describe the uncertainty of the w ind turbine,load forecasting and internal pow er balance,and the operational constraints and netw ork security constraints of each unit are considered. Scheduling scheme is generated on the basis of reasonable control and coordination of the output of each component. The quantum particle sw arm optimization algorithm w ith adaptive inertia w eight adjustment is used to solve the model. In the example,the effects of different demand response schemes on load curve optimization results,netw ork security and virtual pow er plant economy are compared. The dispatching results of virtual pow er plant under different confidence levels are compared. Therefore,the feasibility of the model is verified.