【摘要】 随着光伏渗透率的增加，从电网经济性和稳定性的角度考量，调度光伏的需求也愈发强烈。然而，传统电网调度都是基于可用容量固定的发电设备，对于光伏的调度则需要考虑实际可用容量不确定性变化带来的影响。目前的调度算法都是将光伏可用容量的不确定性量化为概率函数或可放缩的变量处理，或是用训练后的模型自适应光伏波动，无法实现对光伏功率偏差的准确响应。因此，提出了一种多时间尺度的光伏调度算法，在长时间尺度上，用集中式算法根据最优潮流模型求解网损最小和平均电压最小的光伏调度策略；在短时间尺度上，用分布式一致性算法实现功率偏差在各光伏电源之间的平均分担，以备用功率补足功率偏差。给出了集中式算法的潮流模型及其基于二阶锥松弛的凸优化模型，以及分布式算法的实现原理和伪代码。最后，在标准IEEE-5节点和IEEE-14节点系统上进行了功能验证。更多还原

【Abstract】 With the increase of PV permeability, the demand of dispatching PV becomes more and more urgent on the behalf of power grid economy and stability. However, traditional grid dispatching is based on fixed-capacity power generation equipment, for the PV dispatching needs to consider the impact of available capacity uncertainty. The current dispatching algorithms quantify the uncertainty of PV available capacity into probability function or scalable variable processing, or use the trained model to adapt to PV fluctuations. However, the accurate response to the PV power deviation can not be realized. Therefore, a multi-time-scale PV dispatching algorithm is proposed in this paper. On the long-time scale, the centralized algorithm is utilized to solve the PV dispatching strategy with the minimum network loss and the minimum average voltage according to the optimal power flow model. In the short time scale, the distributed consistent averaging algorithm is adopted to achieve the average sharing of power deviation among all PV generations, and the standby power is used to make up the power deviation. The power flow model of the centralized algorithm and its convex optimization model based on the second-order cone relaxation are given in this paper. Moreover, the implementation principle and pseudo-code of the distributed algorithm is illustrated. Finally, the algorithm is verified on the IEEE-5 node and IEEE-14 node grids.更多还原