Thermal Profiling of a Small Operational Data Center

Abstract

Abstract This study highlights experimental challenges of setting up a computational fluid dynamics (CFD) model of a 180-m2 small-size high-performance computing (HPC) data center (DC) in a university campus and the validation procedure leading to a reasonably accurate CFD model for the investigation of the thermal environment. Experimental results based on temperature, airflow, pressure, and power measurements help to confirm and identify inefficient design and operational practices as well as problematic areas regarding the thermal management, define the boundary conditions and validate CFD models. Leaky floors, uneven placement of computing equipment and perforated tiles preventing separation of hot and cold air, low-temperature operation, excessive cooling capacity and fan power were identified sources of energy inefficiency in the DC. CFD model predictions were gradually improved by using experimental measurements in various boundary conditions and detailed geometrical representation of large leakage openings. The performance indicator of the CFD model during the validation process was the temperature prediction error at the rack inlets and exits. After all improvements, the CFD model can make estimations with an error less than 1°C (RMSE < 1.0) at the rack inlet and less than 5°C (RMSE < 5.0) at the rack outlet. The validated CFD model tests the feasibility of various energy efficiency measures. These measures are in the form of operational or design changes in line with the best practices. Impact of leakage between the raised floor and the room, reduced airflow rate, cold-aisle and hot-aisle separation, workload consolidation, and higher temperature operation were among the phenomena tested by using the validated CFD model. The estimated power usage effectiveness (PUE) value reduced from 1.95 to 1.40 with the implemented energy efficiency measures.