Intraday rolling early-warning optimal dispatch algorithm for power systems considering emerging climate change risks

Abstract

This study addresses low-carbon and secure dispatch in power systems under growing wind-power uncertainty caused by climate change. It proposes a day-ahead and intraday rolling early-warning dispatch method that integrates Coupled Model Intercomparison Project 6 (CMIP6) climate scenarios, wind-power scenario reduction, carbon-capture power plants, and demand response. Using wind-speed data from SSP126, SSP245, and SSP585, 24-hour wind-power curves are generated via the Weibull distribution, and the original scenarios are reduced to five representative cases with probabilities using Kantorovich distance. A multi-time-scale optimization model is then developed for carbon-capture units with flexible operation, price-based demand response, and Class-A/Class-B incentive-based demand response. In the day-ahead stage, unit commitment, price-based demand response, and Class-A dispatch plans are determined. In the intraday stage, decisions are rolled forward every 15 minutes to update unit outputs and Class-B demand response. The model is transformed into a mixed-integer linear program using piecewise linearization and the big-M method and solved with Gurobi. Results show that the proposed method improves adaptability to wind uncertainty, reduces total cost by 31.1% versus no demand response, and lowers total cost, carbon emissions, and load-shedding cost by 18.4%, 13.2%, and 48.6%, respectively, versus deterministic dispatch.