Multi-Time-Scale Optimization Strategy of Electric Heating Integrated Energy System Considering Wind Power Uncertainty

Background: Wind power integration challenges and the limited flexibility of combined heat and power (CHP) systems hinder the efficient operation of integrated energy systems. Objective: To enhance system adaptability to wind power uncertainties and achieve economical operation by addressing the above issues. Method: An integrated energy system coordination model centered on CHP units with electric heat storage boilers and thermal storage tanks is established. A multi-time-scale robust optimization strategy is proposed: day-ahead scheduling uses a hybrid reinforcement learning–genetic algorithm approach, with deep Q networks for discrete operation decisions and genetic algorithms for continuous power allocation; intra-day scheduling adopts a 4-hour prediction horizon and a 15- minute time step; real-time scheduling uses a 1-hour dispatch cycle and a 5-minute interval. Results: The model expands the feasible domain of electricity–heat decoupling, avoids local optimization associated with single algorithms, reduces total operating costs by approximately 10%, and significantly improves the renewable energy accommodation rate and system flexibility. Discussion: The proposed multi-time-scale scheduling strategy with electric heat storage effectively balances wind power volatility and CHP operational rigidity, reducing costs and enhancing flexibility through a hybrid algorithm. However, it does not consider multi-energy network constraints and lacks large-scale validation. Conclusion: The proposed approach achieves dynamic complementarity between electricity and heat resources, providing a feasible solution for the economical and efficient operation of integrated energy systems.