CFD Simulation-based Optimization Of Airflow Distribution In A Datacenter

With the rapid development of 5G technology,the number and scale of datacenters are growing sharply.However,the problem of high energy consumption in datacenters has attracted more and more attention and the management of airflow in data centers plays a key role in reducing energy consumption.In order to reduce the energy consumption of the datacenters and to improve the efficiency of the cooling system,this thesis employed CFD numerical simulation to optimize the airflow in datacenters in order to decrease the mixing of hot and cold airflow and to improve the uniformity of air supply.Firstly,this paper proposed an improved rack model.Validated by experimental data,the improved rack model can reflect the velocity and temperature change of the airflow through the rack more accurately.Therefore,the model will be used to simulate the velocity and temperature in datacenters in this study.Based on improved rack model,the velocity field and temperature field in a datacenter were studied under different forms of air supply,different forms of cold aisle containment and different positions of outlet.A preferred combination of the three design parameters of the datacenter was obtained.Based on the preferred combination of design parameters,the thesis further investigated the influence of the layout of air supply,the porosity of the tiles on the floor and the height of the raised floor on the uniformity of air supply,which is measured by the non-uniformity of velocity and non-uniformity of temperature.Results show that the two-sided air supply from the front has significant advantage in the uniformity of air supply.Reduction in the porosity can improve the uniformity of air inlet from the floor,but it has little impact on the racks.The ideal porosity is 25%-45%.Increasing the height of the raised floor can effectively improve the uniformity of air supply,while the improvement decreases with the increase in height.The recommended height of the raised floor is 0.6 m-0.8 m.Finally,orthogonal experimental design method was used to optimize the uniformity of air distribution and air supply in the datacenter.Non-uniformity of air velocity,non-uniformity of temperature and area average temperature were utilized for the evaluation.The most significant experimental factor is the layout of air supply,which is followed by the height of the raised floor,and then the porosity of the air inlet on the floor.Considering the interrelations between the experimental factors,the optimal design parameters of the datacenter regarding airflow management are two-sided air supply from the front,the height of the raised floor being 0.8m and the porosity of the air inlet on the floor being 25%.