EVALUATION OF DATA CENTER COOLING EFFICIENCY: A CASE STUDY APPROACH

Abstract

Number of data centers has increased dramatically over the past decade to meet the ever-increasing global demand for digitality. Not only has the number of data centers has increased, the amount of energy consumption by each of the data center has also been rising. Higher demand for digitality led to higher power density required for each server in the data center. As servers with higher power density generates more heat during its processing, more energy is needed to cool them down. This inevitably leads to a demand for a more optimized cooling system with higher cooling efficiency so as to achieve energy saving. In light of this, Computational Fluid Dynamics (CFD) software has been used to see effectiveness of various optimizing strategies. A real data center, CINTECH II, has been modelled using CoolSim (one of the CFD software) to analyze its cooling performance. Different metrics were used to evaluate its performance – Power Usage Effectiveness (PUE), Return Temperature Index (RTI), Rack Cooling Index (RCI), airpath diagram and thermal diagram. Then, initial results were compared against results of first improvement (closing gaps between racks and adjusting number of CRAC units) and results of second improvement (containment strategy and increased depth of underfloor plenum). Moreover, effectiveness of each optimized strategy was evaluated to recommend energy-saving opportunity for large and complex data center like CINTECH II. There were positive findings where each strategy could improve airflow management and cooling efficiency of data center. However, more in-depth research is essential to increase reliability and accuracy of the results.

Keyword: Data center, Power Usage Effectiveness (PUE), Return Temperature Index (RTI), Rack Cooling Index (RCI), optimizing strategies, energy saving, cost saving