Environmental impacts of ever-growing computational requirements have raised worldwide concerns and significant efforts have been dedicated to reducing power consumption, water usage, and eventually the carbon footprint. Cloud architecture and services have undergone substantial development to become more sustainable and reliable with implementing advanced cooling solutions. In this research, a bottom-up approach has been presented to investigate the energy optimization opportunities of an air-liquid hybrid cooling method compared to the pure air cooling for a 1.7 MW data center. A gradual transition from 100% air cooling to 25%–75% air and liquid cooling has been studied to capture the changes in IT, fan, facility, and the total data center power consumption. Various system design optimizations such as supply air temperature (SAT), facility chiller water temperature, economization and secondary fluid temperature are embedded in this work to highlight the importance of proper setpoint conditions on both primary and secondary sides. Computational fluid dynamics (CFD) and flow network modeling (FNM) are utilized to precisely assess the performance of air and liquid cooling by evaluating the required flow rate, pressure drop, and critical case temperature of computing components as well as temperature change of cooling medium. Energy consumption of the selected cooling equipment is measured based on the BIN data for CRAH and CDU’s performance models. Power usage effectiveness (PUE) measured and compared with Total Usage Effectiveness (TUE) which appears to be a more suitable metric to weigh a data center’s design efficiency by not limiting the fan power to the IT boundary. For the most optimized case, we obtained up to 27% lower consumption in the facility power and 15.5% lower usage in the whole data center site. Increasing the percentage of liquid cooling contribution significantly diminishes the power intake which addresses concerns about natural resources limit as one of the most critical requirements of a sustainable design.

This content is only available via PDF.
You do not currently have access to this content.