Multi-stage day-ahead and intra-day resource scheduling and market bidding strategy for integrated PV-ESS-EV station under multiple uncertainties

Integrated PV-ESS-EV Stations (IPEES) are emerging as significant market players that integrate variable photovoltaic (PV) generation with time-coupled energy storage system (ESS) and the flexible demand of electric vehicles (EVs), participating in day-ahead and intra-day energy and ancillary services markets. However, current studies on charging station operations and bidding often employ short-term or two-stage models, which overlook the temporal correlations in intra-day bidding. In addition, the intricate interactions among various physical energy systems and the diverse energy market bidding options are not comprehensively captured. To fill the above research gaps, this paper proposes a multi-stage joint energy–ancillary market bidding and on-site scheduling model for IPEES, accounting for correlated heterogeneous uncertainties. Firstly, a non-homogeneous Markov process is proposed, combined with a Density-Based Spatial clustering discretization strategy, to uncover the fluctuation patterns of multiple stochastic series, such as PV generation, prices, and aggregated charging flexibilities. Then, a market bidding and operation scheduling optimization model is developed, integrating day-ahead and intra-day markets, as well as energy and ancillary services, while considering all dynamic security constraints of ESS and EV operations. To address the challenging multi-stage mixed-integer stochastic optimization problem, a Markov chain stochastic dual dynamic integer programming algorithm (MC-SDDiP) is employed, which constructs cuts from Monte Carlo resampled Markov trajectories to ensure scalability and applicability. Case studies on a PJM-calibrated IPEES demonstrate high accuracy (0.102% gap) and significant economic benefits from joint participation: downward reserve activation reduces procurement costs — often through opportunistic ESS charging — shifting outcomes from near break-even to positive returns, while extensive upward activation increases intra-day purchases and diminishes profits. The findings provide a practical approach for IPEES to convert PV/ESS/EV flexibility into multi-market revenue under realistic settlement frictions and correlated, time-evolving uncertainties. • A multi-stage IPEES bidding–scheduling strategy across day-ahead and intra-day. • Correlated PV/price/EV uncertainties captured with non-homogeneous Markov chains. • MC-SDDiP solves the mixed-integer model with tight cuts and scalable sampling. • Joint energy–ancillary bidding raises profits versus energy-only participation.