Optimal Scheduling of Virtual Power Plants Considering Distributed PV Volatility

In this paper, we address the optimal scheduling problem of a virtual power plant (VPP) while accounting for the inherent volatility of distributed photovoltaic (DPV) systems. To tackle the challenges posed by the uncertainty in DPV output prediction and the difficulties in coordinating and controlling the DPV, energy storage systems (ESS), and controllable loads, we propose a Bayesian spatio-temporal model for DPV output forecasting. This model is integrated with the demand response of ESS and controllable loads to establish an optimal scheduling framework for the VPP, considering DPV fluctuations. The proposed scheduling model operates in two phases: the day-ahead scheduling phase and the real-time scheduling phase. In the day-ahead scheduling phase, optimization is performed with the objective of maximizing VPP operational revenue, based on the DPV day-ahead forecast. In the real-time scheduling phase, the actual DPV output is used to adjust the day-ahead plan, and real-time optimization is conducted with the goal of minimizing the net load fluctuations within the system. Simulation experiments validate the feasibility and effectiveness of the proposed model. The results demonstrate that the model significantly enhances VPP operational efficiency, smooth the system's net load curve, and mitigates the impact of DPV volatility on system performance.