计及源荷不确定性的数据中心综合能源系统鲁棒优化配置

PDF(1612 KB)

分布式能源 ›› 0

杨东俊明1, 范俊秋2, 李庆生1, 张裕1, 杜刃刃1

作者信息 +

Robust optimization of configurations of integrated energy system for data center considering source-load uncertainties#br#

YANG Dongjunming1, FAN Junqiu2 , LI Qingsheng1, ZHANG Yu1, DU Renren1

Author information +

文章历史 +

摘要

为应对多重不确定性对数据中心综合能源系统配置方案的影响，提出一种基于信息间隙决策理论引导的双层优化的三阶段决策框架。首先，构建了基于电-热-氢多元储能的多能互补综合能源系统，建立以年化综合成本和年 CO₂排放量为双目标的双层优化模型，上层采用 NSGA‑II 算法优化设备容量，下层以日运行成本最小为目标进行混合整数线性规划调度。其次，基于信息间隙决策理论分析分别计算鲁棒性半径与机会性半径，量化各方案承受恶劣不确定性的能力与捕捉有利不确定性的潜力。最后，依据决策者风险偏好，通过直接筛选或鲁棒性感知再优化生成最终推荐配置。以某大型数据中心园区为例进行仿真，结果表明：与确定性方案相比，风险规避型策略整体上导致成本和碳排放均上升，风险逐利型策略则实现成本与排放双降。典型最优解下，风险规避型碳排放增至 3.8 倍，风险逐利型成本与排放分别降低 4.3%和 16.5%。可见，所提框架能够有效量化不确定性对系统性能的影响，为不同风险偏好的决策者提供兼顾经济性与低碳性的配置方案。

Abstract

To address the impacts of multiple uncertainties on configuration schemes of integrated energy systems for data centers, a three-stage decision framework integrating information gap decision theory and bi-level optimization is proposed. First, a bi-level optimization model is established with levelized annual cost and annual CO₂ emissions as the two objectives. In the upper level, the NSGA-II algorithm is used to optimize equipment capacities, while in the lower level, mixed-integer linear programming is employed to minimize daily operating costs. Then, based on information gap decision theory analysis, the robustness radius and opportuneness radius are calculated to quantify each scheme's ability to withstand unfavorable uncertainties and its potential to capitalize on favorable uncertainties. Finally, according to the decision-maker's risk preference, a final recommended configuration is adjusted and generated. A simulation was conducted using a typical data center as a case study, and the results indicate that the risk-averse scheme achieves robustness at the cost of low-carbon performance, with carbon emissions increasing to 3.8 times that of the deterministic scheme; while the risk-opportunistic scheme achieves a dual reduction in both cost and emissions, with decreases of 4.3% and 16.5%, respectively.

导出引用

杨东俊明1, 范俊秋2, 李庆生1, 张裕1, 杜刃刃1. 计及源荷不确定性的数据中心综合能源系统鲁棒优化配置[J]. 分布式能源, 0

YANG Dongjunming1, FAN Junqiu2 , LI Qingsheng1, ZHANG Yu1, DU Renren1. Robust optimization of configurations of integrated energy system for data center considering source-load uncertainties#br#[J]. Distributed Energy, 0