Font Size: a A A

Research On Optimization Of Air Distribution In Cold Channel Of Data Center With Row-level Air Conditioner Under Raised Floor

Posted on:2024-08-18Degree:MasterType:Thesis
Country:ChinaCandidate:S X WangFull Text:PDF
GTID:2532307067476874Subject:Civil Engineering (HVAC) (Professional Degree)
Abstract/Summary:
With the increasing digitalization trend in the world,the strengthening of large and complex data centers is becoming more and more important,and improving refrigeration efficiency is a key means to reduce the overall energy consumption of data centers.In the traditional air supply mode,the air supply distance is long,the cooling capacity of IT devices is unevenly distributed,and energy is wasted and the temperature inside the equipment room is too high.Therefore,a new cooling technology must be used to improve the air distribution in the data center and the thermal environment of the cabinet.In order to solve the above problems,this paper introduces and compares the advantages and disadvantages of three different air conditioning systems,namely room level,row level and cabinet level,and puts forward a new row level air conditioning system,and describes its structure,installation and layout,functions and other characteristics.Based on the reliability model of data center air conditioning system,the air supply velocity field under different redundancy conditions was measured.The data center without redundancy backup(hot backup mode)was taken as the research object for CFD simulation,and the simulation results were compared with the measured results to verify the reliability of the simulation.The following characteristics are found in the simulation: 1.The air supply from the left and right sides of the data center generally presents a vertically upward unsteady air supply process after jet collision effect occurs in the cold pool.The right air supply and the left air supply offset to the back side and the front side of the cold passage respectively.The swing and offset of the air supply cause the two asymmetrical air flows to form a moving vortex flow field at the top of the closed cold passage.2.The maximum and minimum average inlet temperature of a cabinet are 24 ℃ and 22.4 ℃ respectively,which generally meet the standard requirements.However,some parts of the cabinet are overheating due to hot air reflux.3.The hot spots of Layer 1 and Layer 2 servers(at the bottom of the cabinet)and Layer 6 and Layer 7 servers(at the top of the cabinet)are relatively serious.The average maximum inlet temperature of servers is 29.5℃,27.6℃,25.7℃,and 28.8℃ respectively.The reasons for the formation of the characteristics are analyzed,and the optimization measures of adding the baffle + blast type refrigeration bearing base in the cold tank(in terms of structure)and appropriately increasing the total air supply(in terms of working parameters)are put forward.The results are as follows: 1.The optimization of the structure can basically eliminate the unsteady air supply process of the data center,all the wind speed points with high air supply speed disappear,and the air distribution can be effectively optimized;2.The structural optimization can effectively solve the overheating phenomenon at the bottom of the cabinet.The average maximum inlet temperature of the server on the first layer is reduced by3.5℃,and that on the second layer by 2.1℃,but the overheating at the top of the cabinet is more serious.3.An independent increase in the total air supply can solve the overheating phenomenon at the top of the cabinet.The average maximum inlet temperature of the server on the sixth floor decreases by 4℃ and that on the seventh floor decreases by 7.1℃.4.The two optimization measures can solve the local hot spot problems of overheating at the bottom and top of the cabinet at the same time.After optimization,the average maximum inlet temperature of the servers on the first,second,sixth and seventh floors is reduced to 24.5℃,24.1℃,22.5℃ and 22.5℃ respectively,and the average local overheating index is reduced from 2.5 to 1.5.Greatly improved the thermal environment level of the data center.
Keywords/Search Tags:Data center, Numerical simulation, Row air conditioning, Air distribution
Related items