Data-driven approaches for optimizing EV aggregator power profile in energy and reserve market.

• A data-driven prediction method of EVs' charging behavior is developed. • Uncertainties are analyzed to estimate the performance-based payment of providing ancillary service. • In optimization model, the non-parametric distribution function is converted to a solvable multivariable polynomial function. Electric vehicle batteries can be a flexible resource for providing regulation reserves in ancillary service markets. But uncertainty exists both in EVs' charging behavior and system's real-time regulation demand, which may cause the deficiency of regulation power of EVs, thus lowering the performance in ancillary service provision. Based on historical data of EVs' charging behavior and system's regulation signal, this paper proposes data-driven approaches to optimize EVs participation in ancillary service market. Firstly, the EV charging behavior uncertainty is described using a probability prediction model, and the uncertainty of regulation signal is analyzed as well. Secondly, based on the uncertainty description, a day-ahead schedule model is proposed to maximize the income, in which the estimated performance-based payment of providing ancillary service is considered. The model is tested with three EV aggregators, and the results are compared to analyze the ability and characteristics of the EV aggregators when providing ancillary service. [ABSTRACT FROM AUTHOR]