Time-varying external characteristic model of distributed energy systems considering internal multi-energy conversion constraints

Abstract

Under the dual-carbon strategy, distributed energy systems (DES) have become important carriers for multi-energy integration. However, the strong coupling among electricity, heat, cooling, and gas, together with time-varying operating conditions, makes accurate modeling challenging. Traditional energy hub models usually assume constant equipment efficiency, resulting in deviations between optimized scheduling and actual operation. To address this issue, this paper proposes a distributed energy hub time-varying external characteristic modeling method (DEH-TE) combining physical mechanism analysis and data-driven techniques. A multi-energy flow coupling topology is established based on the energy hub framework, while the efficiencies of key devices are corrected using load-rate and ambient-temperature functions. In addition, a two-layer LSTM network is employed to capture the dynamic response of heating networks and natural gas pipelines. The obtained dynamic compensation term is integrated with the time-varying coupling matrix and embedded into a rolling optimization framework solved by CPLEX. Simulation results show that the proposed method reduces the RMSE of electrical, heating, and cooling power prediction by 63.4%, 66.5%, and 64.9%, respectively, with R2 values of at least 0.96. Relative to the constant-efficiency static EH baseline without LSTM, the system efficiency calculation error decreases from 8.7% to 1.2%, and the typical-day total operating-cost prediction error is 1.8%. The proposed DEH-TE model effectively characterizes equipment operating variations and dynamic hysteresis, providing reliable support for real-time economic dispatch of DES.