Read the full-text of this article for free

Title: Optimised bidding strategy for data centres participating in the electrical energy and fast frequency regulation market under the background of carbon peak and carbon neutrality

Authors: Hongyin Chen; Wenrui Zhang; Songcen Wang; Jie Tong; Huafeng Zhang; Jianfeng Li; Xiaoqiang Jia; Yan Sun; Yujie Li

Addresses: National Key Laboratory of Power Grid Safety, Beijing, 100192, China; China Electric Power Research Institute, Beijing, 100192, China ' State Grid Gansu Electric Power Company, Lanzhou, 730010, Gansu, China ' National Key Laboratory of Power Grid Safety, Beijing, 100192, China; China Electric Power Research Institute, Beijing, 100192, China ' National Key Laboratory of Power Grid Safety, Beijing, 100192, China; China Electric Power Research Institute, Beijing, 100192, China ' State Grid Gansu Electric Power Company, Lanzhou, 730010, Gansu, China ' National Key Laboratory of Power Grid Safety, Beijing, 100192, China; China Electric Power Research Institute, Beijing, 100192, China ' National Key Laboratory of Power Grid Safety, Beijing, 100192, China; China Electric Power Research Institute, Beijing, 100192, China ' State Grid Gansu Electric Power Company, Lanzhou 730010, Gansu, China ' State Grid Gansu Electric Power Company, Lanzhou 730010, Gansu, China

Abstract: The existing bidding strategies for data centres do not fully consider electricity price fluctuations, load regulation capabilities, and carbon emission limitations, making it difficult to strike a balance between the interests of the electricity and rapid frequency regulation markets and low-carbon goals. This paper constructs a two-layer optimisation model based on Stackelberg game theory. The upper layer optimises the bidding price for the data centre, while the lower layer models load regulation, energy storage charging and discharging, and carbon emission constraints using mixed integer linear programming (MILP). Lagrange relaxation method is used for decomposition and solution, while deep Q-network (DQN) algorithm is used for dynamic simulation of market electricity price fluctuations. The experimental results show that after using MILP combined with Lagrange relaxation and DQN optimisation strategies, the market revenue of the data centre increased to 158300 yuan, which is 25.4% higher than traditional methods.

Keywords: carbon neutrality; data centre bidding; electrical energy market; DQN; deep Q-network; Stackelberg game; carbon peak.

DOI: 10.1504/IJEP.2026.153610

International Journal of Environment and Pollution, 2026 Vol.76 No.6, pp.106 - 128

Received: 30 May 2025
Accepted: 24 Oct 2025
Published online: 18 May 2026 *