Energy and economic analysis of the bypass method for the computer room air conditioning units of a small data center

Abstract

Earlier optimization studies utilized reduced-order modeling approach combining flow network modeling at room level and thermodynamic modeling at the cooling infrastructure level to investigate the energy saving potential of the computer room air handling (CRAH) bypass (BP) method for data centers. Flow network modeling worked well in predicting the cooling infrastructure power use at various BP fractions. While existing studies are either on hypothetical data centers or experimental setups, there is no considerable work on the feasibility of the method in real and operational data center. Economic requirements of the method is another aspect missing in the literature. This study includes assessments on the feasibility of the CRAH BP methods (forced and induced) in an operational data center using reduced-order modeling approaches suggested in the literature. As a result of the energy and economic analysis of various scenarios, it has been concluded that the feasibility of the CRAH BP method is directly related to the number of fans required by the design and recommendations for application have been proposed.