What is it about?
Data centers consume large amounts of electricity for computing and cooling, while renewable generation and weather conditions can vary unpredictably. This study develops a new scheduling method for data center microgrids that combine renewable energy with hydrogen energy storage. The proposed approach jointly considers operating cost, renewable energy curtailment, and uncertainties in both renewable generation and weather-related cooling demand. Instead of planning only for the most extreme situations, it allows a controlled level of statistical risk so that schedules can remain reliable without becoming unnecessarily conservative. The framework also evaluates both favorable and unfavorable renewable-energy conditions through interval optimization. A specially designed uncertainty-set construction method further reduces the computational burden, making the overall scheduling problem much faster to solve while preserving statistical feasibility.
Featured Image
Photo by imgix on Unsplash
Why is it important?
As data centers expand rapidly, reducing their energy use and carbon footprint is becoming increasingly important. Renewable energy and hydrogen storage offer a promising solution, but their operation is complicated by uncertain renewable output, weather conditions, workload demand, and cooling requirements. Conventional robust optimization can protect against uncertainty, but it often plans for overly pessimistic situations, leading to higher costs and wasted renewable energy. This work provides a more balanced alternative by combining statistical reliability with interval-based decision-making. In the numerical studies, the proposed method reduced mean renewable energy curtailment by 79.17%–87.19% compared with several benchmark approaches. The new uncertainty-set construction also reduced average solution time from more than 30,000 seconds to only several seconds. These improvements make uncertainty-aware scheduling much more practical for large-scale or time-sensitive energy management.
Perspectives
What I find particularly valuable about this work is that it addresses uncertainty from both a modeling and a computational perspective. Many energy scheduling methods either provide strong robustness at the cost of excessive conservatism, or achieve better economic performance by accepting weaker reliability. This study attempts to bridge that gap by allowing uncertainty to be treated statistically while still maintaining explicit feasibility guarantees. The SGET algorithm is also an important practical contribution because sophisticated uncertainty models are useful only if they can be solved efficiently enough for realistic applications. By greatly reducing the number of constraints associated with the optimization reformulation, the proposed method moves this type of robust scheduling closer to real operational use. More broadly, the framework could also be useful for other energy systems that combine renewable generation, flexible demand, storage, and thermal dynamics, where decision-makers must continuously balance efficiency, uncertainty, and reliability.
Chair, IEEE PES EICC Task Force on AI-Enabled Resilience of CPES|Clarivate HCR|AE: IEEE TSG/TSTE/TII Yang Li
Northeast Electric Power University
Read the Original
This page is a summary of: Interval optimization coupled with statistical feasibility robust optimization for hydrogen data center microgrid scheduling, European Journal of Operational Research, August 2026, Elsevier,
DOI: 10.1016/j.ejor.2026.08.043.
You can read the full text:
Contributors
The following have contributed to this page
