基于代理模型加速的园区综合能源系统双目标滚动运行优化

Bi-objective Rolling Operation Optimization Based on Surrogate Model Acceleratiy of Community-Level Integrated Energy Systems

高比例分布式能源的大量接入显著增加了园区综合能源系统运行的不确定性。同时,园区能源系统运行需要兼顾绿色、经济运行目标,成为典型的双目标优化问题。为此,提出了一种基于代理模型加速的园区综合能源系统双目标滚动优化调度算法。首先,以基于三角分解空间搜索的双目标优化算法为基础,利用代理模型对待搜索空间是否有解进行快速预判,有效提高了空间搜索优化效率;其次,进一步将代理模型加速的双目标优化方法应用于模型预测控制框架,在日内运行中根据不断更新的预测信息滚动优化双目标运行策略,提高了对源荷不确定性的应对能力;最后,以某实际园区综合能源系统为例,验证了所提方法的可行性和有效性。

Uncertainties in community-level integrated energy systems(CIESs)have been increasing due to the high penetration of distributed generation.Besides,the economic and environmental operation of CIESs should be considered at the same time,making its optimal operation problem a typical bi-objective one.In this paper,a bi-objective rolling scheduling method for CIES is proposed,based on the triangle splitting searching algorithm.Firstly,the surrogate model is employed to explore the feasible region.Then the surrogate model accelerated bi-objective optimization method is applied in the framework of model predictive control(MPC),in which the operation strategy is optimized according to the updated forecasting information.Finally,the proposed method is verified by a case study of a practical CIES to demonstrate its effectiveness and feasibility.