@Article{ee.2025.073720,
AUTHOR = {Qian Yang, Shenglei Du, Boyang Chen, Yalu Sun, Ding Li, Zhiheng Zhang},
TITLE = {Collaborative Scheduling Strategy for Computation and Power among Multiple Data Centers Based on New Energy State Recognition},
JOURNAL = {Energy Engineering},
VOLUME = {},
YEAR = {},
NUMBER = {},
PAGES = {{pages}},
URL = {http://www.techscience.com/energy/online/detail/24966},
ISSN = {1546-0118},
ABSTRACT = {In recent years, as the core infrastructure of the digital economy, data centers have witnessed increasingly prominent issues of energy consumption and carbon emissions. To achieve the goals of “carbon peak” and “carbon neutrality”, data centers have gradually introduced new energy power such as wind and photovoltaic power. However, the randomness and volatility of their output pose challenges to efficient absorption. Based on the spatiotemporal complementary characteristics of new energy output in multiple data centers and the spatiotemporal migration capability of computing tasks, this paper proposes a new energy-aware adaptive collaborative scheduling strategy for computation and power. The strategy first constructs a regionally differentiated load model to accurately depict the characteristic differences among the Jiangsu-Zhejiang-Shanghai mixed computing power hub, the Gansu high-efficiency computing power base, and the coastal green computing power nodes. Then, a dual-mode scheduling algorithm based on Lyapunov optimization is designed, integrating a prediction-reaction mechanism to achieve dynamic balance between system stability and new energy absorption rate. Furthermore, a V-parameter adaptive adjustment mechanism and a hierarchical fault-tolerant guarantee system are proposed to cope with new energy fluctuations and improve system robustness. Simulation results show that the proposed strategy achieves an average new energy absorption rate of 62.3% and 52.8% in normal weather and severe weather scenarios, respectively. The carbon emission per unit computing power is reduced by 20.9%, and the computing power-electricity efficiency is improved by 9.1%, which is significantly better than the static scheduling strategy. This verifies its effectiveness and practicability in improving new energy utilization, ensuring service quality, and reducing carbon emissions.},
DOI = {10.32604/ee.2025.073720}
}