Research On Airflow Optimization At Different Scales Of Data Center

With the rapid development of computers and information technology,the demand for data processing and data storage is increasing.Large-scale data centers are springing up,and their high-energy consumption problems have become increasingly prominent.The optimization of airflow has become the key factor of energy saving.Aiming at the problems of hot spots,mixing of hot and cold air currents,and disordered air flow in the data room,this paper studies the optimization of air distribution at the room and rack scales,respectively.At the same time,a new numerical simulation method is used to increase the simulation accuracy and convergence speed.In this paper,the research progress of data centers is reviewed from two aspects: the current research status of data center air flow and indoor environment.And then the numerical methods widely used in the data center air flow organization research is briefly introduced.Based on relevant research status and current problems,the method of combining theoretical analysis,field test investigation and airflow simulation is adopted.The main content and the technical route of this study are determined too.In the room level,two types of air distribution methods,namely the under-floor air supply with the cold aisle opening and the in-row air conditioning with the cold aisle containing,are applied to a data center in Chongqing as an example.Adopting the field tests mathod to evaluate thermal environment and energy consumption by means of relevant evaluation indicators.The experimental results show that,compared with the under-floor air supply with the cold aisle opening system,the the in-row air conditioning with the cold aisle containing system can effectively improve the utilization of cold air,greatly reduce the hot air bypass and cold air short circui.At the same time,the PUE and MLC values are significantly reduced,and both the overall energy efficiency and the energy efficiency of the air conditioning system are significantly improved.In order to further study the specific effects of the under-floor air supply system and the in-row air supply system,the opening and containing of the cold aisle on the thermal environment and energy efficiency of the data center,two sets of numerical models are established and the results are analyzed by the CFD simulations.At the rack level,a mathematical model of the power consumption of IT equipment and air temperature is established based on the quantitative relationship between the power of IT equipment and the change of air temperature,and the coupling relationship between the two is established with Fluent.Further using the existing experimental data to verify the reliability of the simulation model.By listing four typical configuration strategies of 12 servers in a rack,the optimal server configuration is found using the thermal resistance evaluation indicators.In the simulation method,since the length scales of data centers at different levels span a large scale,this paper abandons the traditional “black box” model,and uses the method of segmented simulation to achieve the similar purpose of the length scale at each level.The convergence speed is significantly increased,so users can quickly choose the equipment and layout that suits them.