Simulation and study on cooling plant efficiency in green data centres

• Main Author: Yeo, Keat Kee.
• Other Authors: Wong Yew Wah
• Format: Final Year Project
• Language: English
• Published: 2013
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/54072
• Institution: Nanyang Technological University

The availability of data centre is critical in keeping the basic infrastructure of many cities running. The critical nature of data centres often requires elevated design criteria to factor in the need for redundancy. As a result, data centres can typically consume as much as forty times the amount of electricity consumed in a typical office environment. With diminishing natural resources and rising utility cost, it is imperative to devise energy saving strategies to improve the energy efficiency of data centres With the energy consumption of the cooling plant just second to the amount of energy consumed by the datacom equipment in the data centre, the cooling plant is a prime target for energy efficient design. The need for redundancy often requires spare capacity components to be factored into the design of the cooling plant, resulting in low chiller efficiency at part load. This study is aimed at developing an understanding of the efficiency of the cooling plant under different redundancy tier requirements and formulating suggestions to implement different chiller configurations to strike a balance between the need for redundancy and energy efficiency. The objectives are achieved through a combination of background research, system simulation and parametric studies. By conducting background research on the leading standards in data centre design and benchmarking standards for energy efficiency, a model of a data centre and its cooling plant was designed and simulated using TRNSYS. The water cooled centrifugal chiller model used in TRNSYS was validated against the design of similar chillers used in other leading simulation software and data from the chiller manufacturer. Using the simulation model, parametric studies were conducted to model different chiller configurations under different redundancy tier requirements and installed load in the data centre. The optimum chiller configuration with the highest efficiency was then suggested for each case study. The main contribution of this study is to highlight the inefficiencies of the cooling plant under low load conditions brought about by a need for system redundancy and uncertainties in the installed capacity of the datacom equipment as opposed to the design capacity. By using multiple units of smaller sized chillers, cooling plant are able to match the installed capacity in the data centre to the optimum efficiency operating point of the chiller, reducing energy inefficiencies.