Shared energy storage planning based on the adjustable potential of data center based on visual IOT platform
Abstract To address the challenges of low utilization and poor economic efficiency associated with decentralized energy storage configurations in data centers, this study proposes a shared energy storage planning method for data center groups based on the adjustable potential of data center based on...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-08-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-14205-7 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract To address the challenges of low utilization and poor economic efficiency associated with decentralized energy storage configurations in data centers, this study proposes a shared energy storage planning method for data center groups based on the adjustable potential of data center based on visual IOT platform, which leverages differences and complementarities in energy storage requirements under different scenarios to minimize investment costs while maximizing operational benefit. First, we establish a shared energy storage operation framework governed by a capacity allocation, cost-sharing mechanisms, and a Nash bargaining-based profit distribution model under different scenarios to ensure equitable benefits for alliance members. Second, a two-stage stochastic optimization model is developed to coordinate shared storage planning and alliance operations, which considers uncertainties on renewable energy output, workload and ambient temperature. Within this framework, a room-level energy management model is designed, integrating adjustable potential for batch-computing workloads and air conditioning systems to optimize time-of-use power consumption. Based the two-stage stochastic optimization model, a improved L-shaped algorithm is proposed to solve the planning model effectively, reducing computational complexity through problem decomposition. The proposed model is demonstrated on a test case involving four comparative scenarios, and the simulation results verified the effectiveness of the proposed approach. |
|---|---|
| ISSN: | 2045-2322 |