| Abstract | The rapid growth of the AI industry has caused a surge in demand for data centers, leading to
increasingly high-density data centers. Consequently, the proportion of energy consumption
in data centers dedicated to cooling has also increased significantly, and this has made the
development of efficient cooling systems an urgent task. However, cooling systems utilizing
outdoor air, such as cooling tower and air-cooled chiller (ACC), have limited performance
under hot and humid conditions. To address this performance limitation, a system was
developed to coolthe air in advance supplied to ACC. In this system, the outdoor air is
dehumidified by a liquid desiccant system and then cooled through evaporative cooling. This
process lowers the condensing temperature of ACC and consequently reduces compressor
power consumption. Additionally, an auxiliary condenser was installed to recover the heat
required for regenerating the liquid desiccant, which also reduces the load on the air-cooled
condenser. To validate the proposed system, a simulation model was developed, and week-
long simulations were performed under hot and humid outdoor air conditions. The results
show that the proposed system could achieve a 5.6% reduction in energy consumption for
data center cooling compared with a conventional ACC. |
|---|
