Other Research Projects
Research Title: Energy Performance of Data Centres
With the rapid growth of the information-based economy, data centres have become a prevalent phenomenon in both the public and private sectors. This study addresses the issue of energy performance of data centres by closely examining the energy use of six data centres of commercial office buildings in Singapore. The primary objectives of the study are: to examine an empirical energy use pattern of data centres under tropical climatic conditions; develop whole system based model of data centre¡¦s energy performance and; give guidance for data centres¡¦ design, operation and maintenance, and retrofits to achieve better energy performance. Methodology of energy performance evaluation in data centres is discussed in detail.
The study concludes that data centres are highly energy consuming facilities, and present highly potential energy and cost savings. Their power-demand provisions are often grossly over-provided. Disparity in energy performance among such facilities demonstrates the need for design guidelines and benchmarking for such facilities.
Output
Fig. 1Profile of data center energy use
Fig. 2 Breakdown of energy end-use in two data centres
Fig. 3 Benchmark of total data center power load
Fig.4 Benchmark of IT equipment energy use (% of total)
Fig. 5 Overdesign of data center power demand
Fig.6 Structure of Neuro-Fuzzy Network model
Benefits of the Research Works for Commercial and Industrial Use
-
The benefits to the data centre operators and designers are to provide actual and accurate energy consumption and power load information, give recommendations for improving data centre¡¦s design and operational energy performance, especially in the areas of energy conservation, power shading, contingency plan and savings of capital and running cost, and better ideas of design.
-
The preliminary benchmarking conducted among data centres investigated and in other relevant studies provides the management with a better understanding of the position of his/her facilities and the motivation to carry out energy retrofits.
-
Neuro-Fuzzy Network is characterized by high capability of modeling data centre¡¦s energy performance, in the advance of integrating performances of systems, as well as various affecting factors. By using this tool, the potential energy and cost savings can be predicted quantitatively within acceptable range of accuracy.
^Back to top |
