Multi-Timescale Low-Carbon Economic Dispatch of V2G-Integrated Energy Systems Considering Battery Physical Characteristics

Abstract

: To address dispatch inaccuracies caused by neglecting battery physical boundaries and cyclic degradation during vehicle-to-grid (V2G) interactions, alongside power imbalances from stochastic wind fluctuations, this paper proposes a multi-timescale low-carbon economic dispatch strategy for V2G-integrated energy systems (IES). In the day-ahead stage, a 300 MW-level multi-entity scheduling model encompassing wind power, an electricity-heat system, and Electric Vehicle (EV) clusters is established. By explicitly incorporating battery State of Health (SOH) and lifecycle degradation costs, a dynamic charging power attenuation constraint is formulated to prevent invalid V2G arbitrage. In the intraday stage, a real-time model predictive control (MPC) rolling framework with a quadratic deviation penalty is proposed to mitigate wind uncertainties, smoothly regulating flexible resources and ensuring day-ahead schedule fidelity. Simulations under a winter day profile validate the methodology. Ablation studies reveal that internalizing degradation costs shifts EV operations from blind arbitrage to rational response, reducing the average daily equivalent cycle number by 50.68%. The day-ahead collaborative dispatch leverages V2G capabilities to offset high-carbon grid generation, reducing total operating costs by 50.54% and net carbon emissions by 37.08% compared to uncoordinated charging. Comprehensive benchmarking against robust optimization (RO), stochastic optimization (SO), greedy response, and linear-penalty strategies demonstrates the proposed framework’s superiority. While RO incurs economic redundancy and SO heavily over-utilizes batteries with a 22.03% cycling surge to combat uncertainty, the proposed MPC synergistically overcomes these limitations. It decisively curbs abrupt power jumps, compressing the maximum power deviation to 6.98 MW, yielding a 59.71% reduction in default penalty costs compared to the linear-penalty baseline, while achieving an actual operating cost identical to SO. Sensitivity analyses further verify that quasi-static physical limits stringently dictate system economic boundaries. Ultimately, this approach establishes an optimal synergistic trade-off among robust disturbance rejection, macroscopic schedule fidelity