To Simulate the Air Flow and Temperature Distribution for a Data Centre

Abstract The building of data centres has risen sharply around the world because of the quick technological advances in fields including the internet, mobile internet, cloud computing, and big data applications. But the problems that this expansion has brought about in terms of the environment and energy are becoming more and more of a worry. The term “data centre” gained popularity in the 1990s as IT operations expanded, and affordable networking equipment became widely available. The study aims to optimize air distribution, improve energy efficiency, and address challenges in achieving optimal airflow for HVAC systems in data centres, especially small ones. The project is conducted numerically using ANSYS Fluent CFD solver. The results from the experimental simulations are compared to industrial simulation data from the Perunding Shanu Sdn. Bhd., both qualitatively and quantitatively. From the experimental simulation and industrial simulation data, the temperature and airflow performances are very comparable. There was a good agreement between the experimental simulation and industrial simulation results. Three different iterations were tested in this paper, 500, 1000, and lastly using the longest test duration at 2000 iterations. However, the case with the longest test duration at 2000 iterations caused instability in the simulation. With 500 and 1000 iterations producing an average temperature range of 20°C (293 K) to 23°C (296 K), which is in line with projected values, the experiment effectively achieved its goals. But when the iterations were increased to 2000, instability developed, and room temperatures rose to about 32°C (305 K). By the time the simulation was run, there was a reversed flow occurring. Despite being higher than usual, this temperature is still within a range that is suitable for human activities. Interestingly, even though the heat source temperature was only 100°C, the simulation showed a high temperature of 199°C (472 K) after 2000 iterations. This value depicts the worst-case situation in a data centre setting, when substantial heat is created; average server operation temperatures are anticipated to be lower. Results from the experimental simulation showed that having the current Precision Air Conditioner unit specification (30 kW, at 5000 CFM and 18°C) is overcooling the room. Hence, the PAC unit specification could be reduced to a more suitable and environmentally friendly one.