A Data-Driven Dynamic Assessment Method for the Ancillary Service Capability of Virtual Power Plants

Abstract

To address the challenge that static capacity assessment methods fail to accurately capture the dynamic operational characteristics of Virtual Power Plants (VPPs) in ancillary service markets, this paper proposes a data-driven dynamic assessment method for VPP's regulation capabilities. The method first establishes a dynamic availability model for aggregated resources, incorporating multi-source real-time data such as meteorological forecasts, the state of charge (SOC) of energy storage, and demand response propensity. Subsequently, based on the aggregated capacity of individual units, the effective regulation capacity is calculated by considering the overall ramp rate constraints of the VPP. Finally, a capacity deviation index is introduced to enable closed-loop validation of the assessment results. The results of a case study demonstrate that, compared to conventional static methods, the proposed approach reduces the average deviation in VPP regulation capacity assessment from 14.8% to 3.6%, significantly enhancing evaluation accuracy. The findings provide crucial technical support for the capacity certification and reliable market participation of VPPs.