en.Wedoany.com Reported - Mitsubishi Heavy Industries (MHI) has successfully demonstrated cooling system optimization technology at the Fujitsu AKASHI Data Center, reducing cooling energy consumption by 2.3% while maintaining stable operation. The project validated the feasibility of system-level energy-saving pathways by holistically controlling shared cooling equipment under multi-vendor infrastructure and air handling units within server rooms.
Traditional cooling optimization typically targets single equipment, whereas this demonstration adopted a vendor-agnostic system-level holistic control strategy. The research team identified temperature distribution within server rooms as a key bottleneck affecting energy efficiency. By simulating and optimizing air conditioning unit operations to rebalance airflow, the temperature distribution improved by 2 degrees Celsius. Adjusting the operating points of shared cooling infrastructure accordingly while maintaining cooling water at appropriate temperatures reduced overall cooling system energy consumption by 2.3%. Additionally, the coefficient of performance (COP) of the centrifugal chiller improved by more than 1.2 points.
This demonstration was conducted only in a portion of the data center's server rooms. MHI estimates that extending this approach to all server rooms would achieve cooling system energy savings of 7.6%, significantly improving overall Power Usage Effectiveness (PUE). PUE is a key metric for measuring data center power efficiency, with values closer to 1.0 indicating higher efficiency.
Shoji Yamasaki, General Manager of MHI's Data Center and Energy Management Division, stated that operating data centers requires improving energy efficiency while utilizing existing equipment. This demonstration proves that system-level cooling optimization—especially in multi-vendor environments—can deliver quantifiable energy savings under real-world operating conditions.
MHI noted that global data center demand is surging, with cooling systems alone accounting for over 60% of non-IT power consumption (source: International Energy Agency, IEA). Existing facilities often prioritize operational stability, making energy efficiency improvements difficult to balance. As AI workloads increase, operating conditions become more stringent, and traditional energy-saving methods have limited effectiveness in multi-vendor environments. Going forward, MHI plans to extend this approach to decarbonized energy, resilient power systems, and digital solutions to support sustainable data center operations.
