含光储充的配网虚拟电厂二次调频随机模型预测控制策略

Secondary Frequency Control Strategy for Photovoltaic-storage-charging Distribution-level Virtual Power Plant Based on Stochastic Model Predictive Control

含光储充的主动配电网作为一种灵活的新型调节资源,可通过聚合其内部的多类分布式电源组成配网虚拟电厂(distribution-level virtual power plant,DVPP)进而整体参与系统二次调频,具有广阔的发展前景。但实际配电网内部情况复杂,存在储能配置有限导致可调容量不足、光伏短时出力难以预测导致控制不精确、功率分配不合理导致内部电压越限的问题,传统虚拟电厂调频策略难以适用。针对上述问题,提出了一种基于随机模型预测控制(stochastic model predictive control,SMPC)的DVPP二次调频策略,通过聚合模型设计、光伏不确定功率的场景树模型搭建、SMPC策略设计的方式实现了分布式电源出力扰动情况下DVPP快速、准确响应自动发电控制(automatic generation control,AGC)指令的目标,并有效降低了功率调节对系统内部电压的影响,为DVPP参与系统二次调频提供了理论基础。

As a flexible new type of regulation resource,an active distribution network containing photovoltaics,energy storages,and charging piles can participate in the system's secondary frequency control by aggregating multiple types of distributed power sources within it to form distribution-level virtual power plant(DVPP).It has a broad development prospect.However,the actual distribution network's internal situation is complex;there are problems,including limited energy storage configuration,which leads to insufficient adjustable capacity;photovoltaic short-time output is difficult to predict,which leads to inaccurate control;power allocation is unreasonable,which leads to internal voltage overrun.Therefore,applying the traditional virtual power plant frequency control strategy is difficult.Aiming at the above problems,this paper proposes a DVPP secondary frequency control strategy based on stochastic model predictive control(SMPC).Using aggregation model design,scenario tree model construction of PV uncertain power,and SMPC strategy design,the goal of the fast and accurate response of DVPP to automatic generation control(AGC)commands under distributed power output perturbation is achieved,and the impact of power regulation on the internal voltage of the system is effectively reduced,which provides a theoretical basis for the participation of DVPP in the secondary frequency control of the system.