Due to limited energy storage capacity and deviations in renewable energy forecasting, park-level integrated electricity and heat systems (PIEHS) with a high proportion of renewable energy integration are confronted with severe challenges arising from operational uncertainties. To address the issues of insufficient fixed energy storage and forecasting deviations, this study proposes a two-stage, precomputing-enhanced day-ahead dispatch framework that accounts for virtual energy storage (VES), which significantly improves the operational efficiency and robustness of PIEHS. The proposed framework considers equipment start-stop characteristics, network dynamic properties, time-varying operating conditions, and fluctuating meteorological environments. Through high-precision prediction of output time periods and output power, together with VES management, a bilinear Benders decomposition algorithm is developed to reduce the dimensionality of the optimization problem and improve the computational efficiency. Experimental results show that the framework is applicable and efficient. • A multi-time scale prediction framework integrating ARIMA, LSTM, and Transformer is proposed, which achieves high-precision forecasting of renewable energy output by capturing long-term seasonal trends, meso-scale start-stop cycles, and micro-scale short-term fluctuations. • The district heating network is modeled as virtual energy storage (VES), enabling cross-temporal heat transfer without additional physical storage investment, thus significantly expanding the operational flexibility of PIEHS. • A two-stage day-ahead dispatch framework incorporating chance constraints is established, which explicitly quantifies the impact of renewable energy uncertainty and balances the economy and robustness of system operation. • A bilinear Benders decomposition algorithm with parallel computing capability is developed to solve the large-scale mixed-integer linear programming (MILP) problem efficiently, reducing the computational burden caused by multi-scenario uncertainty. • Case studies on small-scale PIEHS and the P39H44 benchmark system verify that the proposed method improves renewable energy accommodation rate by 2.71% and reduces operational costs, with strong engineering applicability.
He et al. (Wed,) studied this question.