- Email: [email protected]
Cooling performance of a pump-driven two phase cooling system for free cooling in data centers
Applied Thermal Engineering 95 (2016) 143–149
Contents lists available at ScienceDirect
Applied Thermal Engineering j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / a p t h e r m e n g
Research Paper
Cooling performance of a pump-driven two phase cooling system for free cooling in data centers Yuezheng Ma, Guoyuan Ma *, Shuang Zhang, Feng Zhou College of Environmental and Energy Engineering, Beijing University of Technology, Beijing 100124, China
H I G H L I G H T S
• • •
A magnetic pump is used in the two-phase cooling system. The maximum cooling capacity reaches 3.429 kW and EER is 12.9 at the indoor and outdoor temperature difference of 10 °C. The maximum cooling capacity almost increases linearly with the indoor and outdoor temperature difference.
A R T I C L E
I N F O
Article history: Received 10 May 2015 Accepted 3 November 2015 Available online 12 November 2015 Keywords: Two-phase cooling system Magnetic pump driven Free cooling Data center
A B S T R A C T
The energy consumption of data centers is increasing rapidly with the development of technology. In order to reduce the energy consumption, free cooling technology instead of the conventional refrigeration system is feasible in the cold season. In the present work, a pump-driven two phase cooling system which is an excellent free cooling technology was developed and the prototype was investigated experimentally. The experimental results show that the efficiency of the magnetic pump was 7.7% at the motor frequency of 22 Hz, the cooling capacity of the system was able to reach 3.429 kW and 9.241 kW when the temperature differences between indoor and outdoor were 10 °C and 25 °C, respectively, and the EER were 12.9 and 29.7, which are extremely higher than the conventional refrigeration system. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction With the rapid development of information network and computer technologies, the energy consumption of data center is rising sharply. Electricity used in global data centers in 2010 likely accounted for between 1.1% and 1.5% of total electricity use [1]. At the same time, electricity consumed by data centers worldwide increased by about 56% from 2005 to 2010 [1]. In order to ensure the reliable operation of the data center, the cooling system must work all year round for controlling the indoor temperature and humidity. Among the energy consumption units of data center, the energy consumption of the cooling system accounts for 30~50% of the total consumption [2,3]. During the cool and cold seasons in most areas of China, the outdoor temperature is much lower than the required indoor temperature of the data centers. For the conventional cooling system, the compressor is needed to operate uninterruptedly no matter what the outdoor temperature is. Therefore, the compressor consumed a huge amount of energy at the low outdoor temperature. When the outdoor temperature is below a certain amount compared to
* Corresponding author. Tel.: +86 10 67391613; fax: +86 10 67391983. E-mail address: [email protected] (M. Guoyuan). http://dx.doi.org/10.1016/j.applthermaleng.2015.11.002 1359-4311/© 2015 Elsevier Ltd. All rights reserved.
the indoor temperature of data center, using free cooling technology by the type of directly [4,5] or indirectly [6–10], instead of the conventional cooling system, is an economic method to decrease the power consumption of data center. Zhang et al. [11] reviewed the advancements of free cooling system in data centers from air side, water side and heat pipe. It was concluded that the heat pipe system has a great application potential due to its excellent heat transfer performance under the operation conditions of small temperature difference. Many significant studies have been focused on the cooling performance of twophase cooling system [8–10]. For example, Tong et al. investigated a thermosyphon loop used to cool a data center, and the EER is 10.96 at the temperature difference of 10 °C [12]. Normally, heat pipe is mainly driven by gravity or capillary force. Although the cooling performance of heat pipe has been improved by optimizing the physical and geometrical parameters of the capillary structure [13], there are still many limitations in the high capacity and complex structures. In order to enhance the dynamic behaviors of the working medium, the driving force has been improved continuously by using various modes, such as electrokinetic force [14,15], hydromagnetic force [16], centrifugal force [17] and pump [18–21]. Among the four driving modes, the electrokinetic force or hydromagnetic force is still limited, and it is difficult to apply them in the cooling system of data center. Although the centrifugal force can provide
144
Y. Ma et al./Applied Thermal Engineering 95 (2016) 143–149
enough power to the devices, this driving mode is only suitable for specific occasions, such as motor or generator cooling, etc. Pumpdriven two phase cooling system is an excellent thermal control cooling system and has many advantages, such as active temperature control, long heat transport distance and significant structural flexibility. The design and testing of the two-phase fluid thermal loops driven by pump for future large spacecraft have been investigated from 1982 to 1993 [18]. Liu et al. investigated the cooling performance of a mechanically pumped two-phase cooling loop for space thermal control [19,20]. These studies show that the cooling loop exhibits excellent performances and flexibility of structures. However, the structure and heat transfer mechanism are quite different from those in the ground compared with the space microgravity environment. In the past several years, active twophase cooling technology for spray cooling has been investigated for the laser diode array cooling and other high heat flux application [21,22]. However, spray cooling is very complex due to its dependence on many factors. The fluid control of two phase mixed flow is one of the factors. Different concepts of two-phase on-chip cooling cycles were proposed and compared for cooling data center servers [23–25]. The results show that, when compared with traditional air cooling systems, the energy consumption of the data center could be reduced by as much as 50% when using a liquid pumping cycle. Further investigation indicates that a maximum EER of the liquid pump cycle is about 19 and that of the vapor compressor cycle is about 4, respectively. Zhang et al. designed and investigated a pump driven loop heat pipe for data center cooling
[26]. The experimental results indicated that the EER is 3.75 at the temperature difference of 10 °C, and it increases to 9.37 at the temperature difference of 25 °C. Whereas, the existing system has the disadvantages which affect the performance as follows: (1) The heat dissipated by the canned pump is nearly absorbed by the working medium which increased the heat load of the system and affected the safe and steady operation of pump. (2) The flow resistance of the system is large which caused a higher sensible heat transfer. However, previous study of pump-driven two phase cooling system for data centers mainly focused on the on-chip cooling and the application with multiple evaporators. The objective of the present study is primarily to analyze the cooling performance in data center to obtain a higher cooling capacity and EER. For this purpose, a magnetic pump was used in the system to avoid the effect of the heat load by motor, and the designed heat exchanger and connecting pipe were considered to decrease the flow resistance of the system. Based on these, the performance characteristics of the magnetic pump and the system performances with mass flow rate and temperature difference were experimentally investigated. 2. Experiment setup Fig. 1 illustrates the schematic diagram of experimental setup and the p–h diagram of the proposed cooling system. As shown in Fig. 1(a), the cooling system is mainly composed of a magneticdriving pump, a tube–fin condenser, a tube–fin evaporator, a liquid reservoir and some connecting pipes. Fig. 1(b) shows the p–h diagram
(a) Schematic diagram of experimental setup
(b) p-h diagram of the proposed cooling system Fig. 1. Schematic diagram of experimental setup and p–h diagram of the proposed cooling system. (a) Schematic diagram of experimental setup. (b) p–h diagram of the proposed cooling system.
Y. Ma et al./Applied Thermal Engineering 95 (2016) 143–149
145
Table 1 Specification of the tube–fin heat exchanger. Geometric parameters
Value
Geometric parameters
Value
Outer tube diameter/mm Number of tube rows in air direction Transverse tube spacing/mm Fin thickness/mm
9.52 3 25 0.15
Inner tube diameter/mm Number of tube rows perpendicular to air direction Longitudinal tube spacing/mm Fin spacing/mm
8.85 36 21.65 2.15
(1)
3. Results and discussion 3.1. Performance characteristics of the magnetic pump The performance characteristics of the magnetic pump influence greatly on the performance and the control strategy of the cooling system. Therefore the operating parameters were first investigated. The pressure head and mass flow rate of the magnetic pump varying with the motor frequency are shown in Fig. 2, when the fluid temperature set as 15 °C ± 0.5 °C. It can be seen that the pressure head and the mass flow rate increase rapidly with increase of the motor frequency from 6 Hz to 22 Hz. The increment of the pressure difference between the pump inlet and outlet is greater than that of the mass flow rate. This is because the internal leakage through the gaps between the impeller and the pump
1.2
500
Δp m
1.0
400
0.8 300 0.6 200
m / kg·h-1
W S = W p + W p ,fc + W con ,f + W con ,fc + W eva ,f + W eva ,fc
with a standard pressure source with accuracy of ±0.25% over the full scale range and the temperature sensors are Pt100 thermometers. The flow rate of the fluid is measured using a Coriolis mass flow rate meter with an accuracy of ±0.2%. Two sight glasses are installed in the inlet and outlet of the Coriolis mass flow rate meter to observe the case of two-phase flowing state. For steady state testing, the sampling interval was 5 s. All the measurement data are recorded and stored on a computer by using an Agilent 34970A series data acquisition system. In data centers, the temperature and humidity play an essential role to provide a proper running environment for servers. In order to investigate the actual characteristics of the cooling system driven by magnetic pump and the effect of the outdoor temperature on its cooling capacity, the performance of the cooling system under typical data center operating condition was tested. When testing, the DB of the indoor test room set as 25 °C and the WB of the evaporator inlet is 17 °C according to a Chinese standard [29]. The DB of the outdoor test room is from 0 °C to 15 °C.
Δp / bar
of the pump-driven two-phase cooling system. The liquid from the reservoir flows to the evaporator through the liquid pipe by the pump driving (processes 1–3). The fluid at the evaporator inlet is usually in sub-cooled state as both the heat leakage and the flowing resistance in the liquid pipe are relatively small. The fluid in the evaporator absorbs heat from the hot air of the indoor room and vaporizes partly (processes 3–5). And it changes into two-phase fluid in the evaporator and the two-phase fluid from the evaporator outlet flows then into the condenser through the two-phase pipe (processes 5 and 6). The fluid in the condenser rejects heat to the cold air of the outdoor room and is cooled into liquid again, and then it returns into the reservoir (process 6–1). The next working process of the cooling cycle restarts. R22 was used as the refrigerant, and the filling ratio is 0.64 at the temperature of 15 °C. The magneticdriving pump is used in the system to avoid the leaks from the shaft seal in conventional pump. The pump is driven by the motor through the magnetic coupler, and its working chamber is completely sealed. The two sight glasses are installed in the inlet and outlet of the magnetic pump to see the fluid flowing state such as the case of cavitations. Both the liquid pipe and two-phase pipe in the cooling cycle are the standard copper tube. The specification of the liquid line is 19.05 mm (Diameter) × 4.5 m (Length), and that of the twophase line is 19.05 mm (Diameter) × 3 m (Length). The evaporator and condenser are placed in the same horizon one meter above the ground and used the same construction of tube–fin heat exchanger except for the headers. The specification of the tube–fin heat exchanger is listed in Table 1. Tests were conducted in a standard psychrometric chambers with two insulated rooms by using the air enthalpy method based on the ASHRAE Standard [27]. The evaporator is installed in the indoor test room, and the cooling capacity was estimated by using the air flow rate and the enthalpy difference of air across the evaporator. Four platinum resistance thermometers (Pt100) made by INOR with an uncertainty of ±0.05 °C are set to measure the dry-bulb temperature (DB) and wet-bulb temperatures (WB) of the air at the evaporator inlet and outlet. The air flow rate of the evaporator is obtained by a nozzle flow rate meter, across which the pressure difference is measured by a differential pressure transducer with an uncertainty of ±0.25% according to ANSI/AMCA 210 [28]. The condenser, liquid receiver and the magnetic pump assembly are installed in the outdoor test room. The DB and WB of the inlet air are measured. The input power of the magnetic pump, frequency converters and fans are measured by a power analyzer of WT230, with an uncertainty of ±0.02%. The total input power of the cooling system is given by the following equation:
0.4
where WS is the input power of the cooling system, kW. Wp is the input power of the magnetic pump, kW. Wcon,f and Weva,f are the input power of the condenser fan and evaporator fan, respectively, kW. Wp,fc, Wcon,fc and Weva,fc are the input power of the frequency converters for the magnetic pump, the condenser fan and evaporator fan, respectively, kW. P and T, as shown in Fig. 1, represent the pressure sensors and temperature sensors to measure the working fluid. The pressure in the cooling cycle is measured by MEAS pressure sensors calibrated
100
0.2
0.0 6
8
10
12
14
16
18
20
0 22
frequency / Hz Fig. 2. Pressure difference and mass flow rate of the pump varying with the frequency.
146
Y. Ma et al./Applied Thermal Engineering 95 (2016) 143–149
0.15
10
10
Wp
9
We
0.12
8
8 6
0.06
4
0.03
7
Q / kW
0.09
η/%
W / kW
η
2
0.00 6
8
10
12
14
16
18
20
ΔT=10 oC ΔT=15 oC ΔT=20 oC ΔT=25 oC
6 5 4 3
0 22
2
frequency / Hz
100
200
300
400
500
m / kg·h-1
Fig. 3. Input power, effective output power, efficiency of the pump varying with the frequency.
Fig. 4. Cooling capacity at different temperature differences varying with mass flow rate.
We =
1 m ⋅ Δp ⋅ 36 ρ
W η= e Wp
(2)
(3)
where We is the effective power of the magnetic pump, kW. Δp is the pressure difference of the working fluid at the pump inlet and outlet, bar. m is the mass flow rate of liquid R22, kg h−1. ρ is the density of the liquid R22 at the pump inlet, which is obtained from REFPROP [30], kg m−3. η is the efficiency of the magnetic pump. Wp is the input power of the magnetic pump, kW. The input power, effective power and efficiency of the magnetic pump varying with the motor frequency are shown in Fig. 3. It can be seen that the input power of the magnetic pump increased rapidly with the increase of motor frequency. The input power increased from 23 W to 143 W when the motor frequency increased from 6 Hz to 22 Hz. The effective power obtained by R22 is very small which is only 0.21 W at the motor frequency of 6 Hz and is 11 W at the motor frequency of 22 Hz. The efficiency of the magnetic pump is 0.92%–7.7% when the motor frequency is from 6 Hz to 22 Hz. 3.2. Effects of the mass flow rate The cooling capacity of the cooling system varying with the mass flow rates at different temperature differences is shown in Fig. 4. The cooling capacity becomes higher when the temperature difference is growing. The maximum cooling capacity reaches 3.429 kW and 9.241 kW when the temperature differences are 10 °C and 25 °C, respectively. The cooling capacity increases rapidly with the increase of the mass flow rate when the flow rate is relatively small, and then it decreases slowly when the mass flow rate increases continuously. Fig. 5 shows the inside coefficient of heat transfer at different temperature differences varying with mass flow rate. The inside coefficient of heat transfer of the evaporator is increased greatly with the flow velocity due to the flow velocity of the fluid increases rapidly
with increase of the mass flow rate. It is also one of the reasons that the cooling capacity increases rapidly with the increase of the mass flow rate. The inside coefficient of heat transfer is given by Eq. (4):
ki =
Q A ⋅ Δt
(4)
where ki is the inside coefficient of heat transfer, kW m−2 K−1. Q is the cooling capacity of the system, kW. A is the internal surface area of the evaporator, m2. Δt is the inside logarithmic temperature difference of the evaporator, °C. Fig. 6 shows the sub-cooling degree at different temperature differences varying with mass flow rate in the evaporator inlet. The sub-cooling degree of the evaporator inlet is higher when the indoor and outdoor temperature difference is larger. And it nearly increases linearly with the mass flow rate. It is because the flow resistance from the evaporator inlet to the pump inlet of the fluid
0.6
0.5
0.4
ki / kW·m-2·K-1
body increases with increase of the pressure difference between the pump inlet and outlet. And then it will cause the increment of the mass flow rate to slow down gradually. The effective power and efficiency of the magnetic pump are given by Eqs. (2) and (3):
0.3
ΔT=10 oC ΔT=15 oC ΔT=20 oC ΔT=25 oC
0.2
0.1
0.0 100
200
300
m / kg·h
400
500
-1
Fig. 5. Inside coefficient of heat transfer at different temperature differences varying with mass flow rate.
Y. Ma et al./Applied Thermal Engineering 95 (2016) 143–149
input power / kW
2.5
ΔT=10 C ΔT=15 oC ΔT=20 oC ΔT=25 oC
2.0
1.5
ΔT=10 oC ΔT=15 oC ΔT=20 oC ΔT=25 oC
0.35
0.30
0.25 35 30 25
1.0
EER
degree of supercooling / oC
o
147
20 15
0.5
10 100
200
300
400
500
m / kg·h-1
0.0 100
200
300
400
500 Fig. 8. Input power and EER at different temperature differences varying with mass flow rate.
m / kg·h-1 Fig. 6. Sub-cooling degree at different temperature differences varying with mass flow rate in the evaporator inlet.
is increased with the increase of the mass flow rate. Therefore, the pressure of the evaporator inlet is also increased when the flow resistance from the evaporator inlet to the pump inlet becomes bigger, which would increase the fluid sub-cooling degree at the evaporator inlet. In the evaporator, the proportion of sensible heat capacity becomes larger, and the evaporation temperature of the working fluid increased gradually with the increase of sub-cooling degree at the evaporator inlet. Accordingly, the cooling capacity decreased when the mass flow rate continues to increase. Therefore, the sub-cooling degree at the evaporator inlet becoming great can cause a decrease in the cooling capacity of the cooling system. The vapor quality of the working fluid at the evaporator outlet varying with the mass flow rates at different temperature differences is shown in Fig. 7. Under all conditions of the temperature difference, the quality at the evaporator outlet always goes up when the temperature difference between indoor and outdoor become larger, and goes down when the mass flow rate becomes bigger.
EER =
Q Ws
(5)
where EER is the energy efficiency ratio of the cooling system. The EERs corresponding to the maximum cooling capacity reach 12.9 and 29.7 when the temperature differences are 10 °C and 25 °C, respectively. However these values are not the maximum EER. When the temperature difference is 10 °C, the maximum EER is 13.4 while the mass flow rate is 185.2 kg h−1 and the cooling capacity is 3.23 kW. If the temperature difference rises to 25 °C, the maximum EER is 32.1 while the mass flow rate is 244.7 kg h−1 and the cooling capacity is 8.35 kW. It is advantageous that the cooling capacity corresponding to the maximum EER is selected as design value to meet the need of data center cooling as a better energy saving effect can be gotten for the data center.
1.0
ΔT=10 oC ΔT=15 oC ΔT=20 oC ΔT=25 oC
0.8
When the temperature differences are 10 °C and 25 °C, the quality corresponding to the maximum cooling capacity is 0.22 and 0.49, respectively. The input power and EER of the cooling system at different temperature differences varying with the mass flow rates are shown in Fig. 8. It can be seen that the input power of the cooling system has variation tendency similar to the one in Fig. 3. This is because the input power of fans and frequency converters of the condenser and the evaporator remain nearly constant, but the pump varies significantly under different operation conditions. Therefore, the input power curves of the cooling system changes with the input power of pump. The EER of the cooling system grows up with the increase of the mass flow rate when the flow rate is low, and then it decreases rapidly if the mass flow rate continues to increase. And the EER becomes higher when the indoor and outdoor temperature difference is growing. The EER is given by Eq. (5):
x
0.6
0.4
3.3. Effects of the temperature difference 0.2
0.0 100
200
300
m / kg·h
400
500
-1
Fig. 7. Vapor quality at different temperature differences varying with mass flow rate.
The maximum cooling capacity and its latent, sensible components and sub-cooling degree at evaporator inlet varying with the indoor and outdoor temperature difference are shown in Fig. 9. The maximum cooling capacity, latent component and sensible component almost display a linear growth with the temperature difference, and the sub-cooling degree at the evaporator inlet also increases linearly with the temperature difference approximately, which is mainly caused by the pressure drop of the cooling system from the evaporator inlet to the pump inlet. The sensible component
Y. Ma et al./Applied Thermal Engineering 95 (2016) 143–149
10
Q sub-cooling degree latent heat transfer sensible heat transfer
8
4
drop of the cooling system reaches 0.5 bar and 0.79 bar when the temperature differences are 10 °C and 25 °C, respectively.
3
4. Conclusions
Q / kW
6 2 4 1 2
0
sub-cooling degree / oC
148
0 10
15
20
25
ΔT / oC Fig. 9. Maximum cooling capacity and its latent, sensible components and subcooling degree at evaporator inlet varying with temperature difference.
changes with the sub-cooling degree of the fluid at the evaporator inlet. When the temperature difference is 10 °C, the sensible heat is 0.077 kW while the mass flow rate is 290.1 kg h−1 and the sensible heat ratio is 2.2%. If the temperature difference rises to 25 °C, the sensible heat is 0.225 kW while the mass flow rate is 339.4 kg h−1 and the sensible heat ratio is 2.4%. Fig. 10 shows the pressure distribution along the axis of the cooling system. Due to the influence of the outdoor temperature, the pressure in the cooling system is higher when the indoor and outdoor temperature difference becomes smaller. The pressure of the fluid increases rapidly in the pump and then it decreases continuously in the following parts such as liquid line, evaporator, twophase line and condenser, etc. The pressure drop of the fluid in the liquid line and two-phase line accounted for about 60% of the total pressure drop. Due to the effect of gravity, the pressure drop of the evaporator accounted for about 26% of the total pressure drop, and that of the condenser accounted for about 14% of the total pressure drop. The total pressure drop of the cooling system almost increases linearly with the temperature difference. The total pressure
10
ΔT=10 oC ΔT=20 oC
ΔT=15 oC ΔT=25 oC
Pressure / bar
9
8
The performance of the pump-driven two phase cooling system is experimentally investigated and the main results are summarized as follows: 1. The pump efficiency is only 7.7% when the mass flow rate is 465.9 kg h−1 and the pressure difference between the pump inlet and outlet is 1.04 bar. 2. The cooling capacity increases rapidly with the increase of the mass flow rate, and then decreases slowly when the mass flow rate continues to increase. When the temperature differences are 10 °C and 25 °C, the maximum cooling capacity reaches 3.429 kW and 9.241 kW, and the EER is 12.9 and 29.7, respectively. The subcooling degree at the evaporator inlet is not advantageous to increasing the cooling capacity of the cooling system. 3. The maximum cooling capacity and its latent and sensible components almost increase linearly with the indoor and outdoor temperature difference.
Acknowledgements This project was supported by the National Natural Science Foundation of China (Grant No. 51376010). Nomenclature A EER Q W ki m x
Area (m2) Energy efficiency ratio Cooling capacity (kW) Power dissipation (kW) Inside coefficient of heat transfer (kW m−2 K−1) Mass flow rate (kg h−1) Outlet vapor quality of evaporator
Greek symbols ΔT Indoor and outdoor temperature difference (°C) Δt Logarithmic temperature difference (°C) Δp Pressure difference (bar) η Efficiency of pump ρ Refrigerant density (kg m−3) Subscripts con Condenser e Effective power eva Evaporator f Fan fc Frequency converter p Pump s Cooling system
References 7
pump 6
liquid line evaporator two-phase condenser line
Fig. 10. Pressure distribution corresponding to the maximum cooling cycle capacity at different temperature differences.
[1] K. Jonathan, Growth in data center electricity use 2005 to 2010, A report by Analytical Press, completed at the request of The New York Times, 2011. [2] K. Ebrahimi, G.F. Jones, A.S. Fleischer, A review of data center cooling technology, operating conditions and the corresponding low-grade waste heat recovery opportunities, Renew. Sustain. Energy Rev. 31 (2014) 622–638. [3] J. Peter, T. Marker. Data centre energy efficiency product profile, Pitt & Sherry, report to equipment energy efficiency committee (E3) of The Australian Government Department of the Environment, Water, Heritage and the Arts (DEWHA), 2009.
Y. Ma et al./Applied Thermal Engineering 95 (2016) 143–149
[4] Y. Chen, Y. Zhang, Q. Meng, Study of ventilation cooling technology for telecommunication base stations in Guangzhou, Energy Build. 41 (7) (2009) 738–744. [5] J. Siriwardana, S. Jayasekara, S.K. Halgamuge, Potential of air-side economizers for data center cooling: a case study for key Australian cities, Appl. Energy 104 (2013) 207–219. [6] Z. Potts Free cooling technologies in data centers applications. Manchester, SUDLOWS White Paper, 2011. [7] L. Bao, J. Wang, L. Kang, The applied effect analysis of heat exchanger installed in a typical communication base station in Beijing of China, Energy Procedia 14 (2012) 620–625. [8] D. Zhu, D. Yan, Z. Li, Modeling and applications of annual energy-using simulation module of separated heat pipe heat exchanger, Energy Build. 57 (2013) 26–33. [9] R. Khodabandeh, Thermal performance of a closed advanced two-phase thermosyphon loop for cooling of radio base stations at different operating conditions, Appl. Therm. Eng. 24 (17–18) (2004) 2643–2655. [10] A. Samba, H. Louahlia-Gualous, S.L. Masson, D. Nörterhäuser, Two-phase thermosyphon loop for cooling outdoor telecommunication equipments, Appl. Therm. Eng. 50 (1) (2013) 1351–1360. [11] H. Zhang, S. Shao, H. Xu, H. Zou, C. Tian, Free cooling of data centers: a review, Renew. Sustain. Energy Rev. 35 (2014) 171–182. [12] Z. Tong, T. Ding, Z. Li, X.-H. Liu, An experimental investigation of an R744 two-phase thermosyphon loop used to cool a data center, Appl. Therm. Eng. 90 (2015) 362–365. [13] V.M. Kiseev, V.V. Vlassov, I. Muraoka, Experimental optimization of capillary structures for loop heat pipes and heat switches, Appl. Therm. Eng. 30 (11–12) (2010) 1312–1319. [14] S. Ebrahimi, A. Hasanzadeh-Barforoushi, A. Nejat, F. Kowsary, Numerical study of mixing and heat transfer in mixed electroosmotic/pressure driven flow through T-shaped microchannels, Int. J. Heat Mass Transf. 75 (2014) 565– 580. [15] C.-O. Ng, C. Qi, Electroosmotic flow of a power-law fluid in a non-uniform microchannel, J. Non-Newtonian Fluid Mech. 208–209 (2014) 118–125. [16] A.E. Kabeel, E.M.S. El-Said, S.A. Dafea, A review of magnetic field effects on flow and heat transfer in liquids: present status and future potential for studies and applications, Renew. Sustain. Energy Rev. 45 (2015) 830–837. [17] F. Song, D. Ewing, C.Y. Ching, Heat transfer in the evaporator section of moderate-speed rotating heat pipes, Int. J. Heat Mass Transf. 51 (7–8) (2008) 1542–1550.
149
[18] I.Y. Chen, M.-C. Chyu, Two-phase active thermal control systems for spacecrafts, in: Energy Conversion Engineering Conference, 1996, IECEC 96, Proceedings of the 31st Intersociety 2, 1996, 1488–1493. [19] J. Liu, N. Pei, K. Guo, Z. He, T. Li, M. Lv, Experimental investigation on mechanical pumped cooling loop for application in future space missions, Energy Convers. Manag. 49 (10) (2008) 2704–2710. [20] J. Liu, N. Pei, K. Guo, Z. He, T. Li, Experimental investigation on a mechanically pumped two-phase cooling loop with dual-evaporator, Int. J. Refrig. 31 (7) (2008) 1176–1182. [21] E.A. Silk, J. Kim, K. Kiger, Spray cooling of enhanced surfaces: impact of structured surface geometry and spray axis inclination, Int. J. Heat Mass Transf. 49 (2006) 4910–4920. [22] E.A. Silk, E.L. Golliher, R. Paneer Selvam, Spray cooling heat transfer: technology overview and assessment of future challenges for micro-gravity application, Energy Convers. Manag. 49 (2008) 453–468. [23] J.B. Marcinichen, J.A. Olivier, J.R. Thome, On-chip two-phase cooling of datacenters: cooling system and energy recovery evaluation, Appl. Therm. Eng. 41 (2012) 36–51. [24] J.B. Marcinichen, J.A. Olivier, V. de Oliveira, J.R. Thome, A review of on-chip micro-evaporation: experimental evaluation of liquid pumping and vapor compression driven cooling systems and control, Appl. Energy 92 (2012) 147–161. [25] J.B. Marcinichen, D. Wua, S. Paredes, J.R. Thome, B. Michel, Dynamic flow control and performance comparison of different concepts of two-phase on-chip cooling cycles, Appl. Energy 114 (2014) 179–191. [26] S. Zhang, G. Ma, F. Zhou, Experimental study on a pump driven loop-heat pipe for data center cooling, J. Energy Eng. (2014) doi: 10.1061/(ASCE)EY.19437897.0000253 (in press). [27] ASHRAE Standard 116, Methods of Testing for Seasonal Efficiency of Unitary Air-conditioners and Heat Pumps, ASHRAE, Atlanta, GA, 2007.( [28] ANSI/AMCA 210, Laboratory Methods of Testing Fans for Rating, ANSI, Arlington, VA, 2007. [29] AQSIQ (General Administration of Quality Supervision, Inspection and Quarantine of the People’s Republic of China), GB 50174-2008: National standard for code for design of electronic information system Room, 2010 (in Chinese). [30] E.W. Lemmon, M.L. Huber, M.O. McLinden, NIST Standard Reference Database 23: Reference fluid thermodynamic and transport properties – REFPROP, Version 8.0, National Institute of Standards and Technology, Standard Reference Data Program, Gaithersburg, 2007.
Contents lists available at ScienceDirect
Applied Thermal Engineering j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / a p t h e r m e n g
Research Paper
Cooling performance of a pump-driven two phase cooling system for free cooling in data centers Yuezheng Ma, Guoyuan Ma *, Shuang Zhang, Feng Zhou College of Environmental and Energy Engineering, Beijing University of Technology, Beijing 100124, China
H I G H L I G H T S
• • •
A magnetic pump is used in the two-phase cooling system. The maximum cooling capacity reaches 3.429 kW and EER is 12.9 at the indoor and outdoor temperature difference of 10 °C. The maximum cooling capacity almost increases linearly with the indoor and outdoor temperature difference.
A R T I C L E
I N F O
Article history: Received 10 May 2015 Accepted 3 November 2015 Available online 12 November 2015 Keywords: Two-phase cooling system Magnetic pump driven Free cooling Data center
A B S T R A C T
The energy consumption of data centers is increasing rapidly with the development of technology. In order to reduce the energy consumption, free cooling technology instead of the conventional refrigeration system is feasible in the cold season. In the present work, a pump-driven two phase cooling system which is an excellent free cooling technology was developed and the prototype was investigated experimentally. The experimental results show that the efficiency of the magnetic pump was 7.7% at the motor frequency of 22 Hz, the cooling capacity of the system was able to reach 3.429 kW and 9.241 kW when the temperature differences between indoor and outdoor were 10 °C and 25 °C, respectively, and the EER were 12.9 and 29.7, which are extremely higher than the conventional refrigeration system. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction With the rapid development of information network and computer technologies, the energy consumption of data center is rising sharply. Electricity used in global data centers in 2010 likely accounted for between 1.1% and 1.5% of total electricity use [1]. At the same time, electricity consumed by data centers worldwide increased by about 56% from 2005 to 2010 [1]. In order to ensure the reliable operation of the data center, the cooling system must work all year round for controlling the indoor temperature and humidity. Among the energy consumption units of data center, the energy consumption of the cooling system accounts for 30~50% of the total consumption [2,3]. During the cool and cold seasons in most areas of China, the outdoor temperature is much lower than the required indoor temperature of the data centers. For the conventional cooling system, the compressor is needed to operate uninterruptedly no matter what the outdoor temperature is. Therefore, the compressor consumed a huge amount of energy at the low outdoor temperature. When the outdoor temperature is below a certain amount compared to
* Corresponding author. Tel.: +86 10 67391613; fax: +86 10 67391983. E-mail address: [email protected] (M. Guoyuan). http://dx.doi.org/10.1016/j.applthermaleng.2015.11.002 1359-4311/© 2015 Elsevier Ltd. All rights reserved.
the indoor temperature of data center, using free cooling technology by the type of directly [4,5] or indirectly [6–10], instead of the conventional cooling system, is an economic method to decrease the power consumption of data center. Zhang et al. [11] reviewed the advancements of free cooling system in data centers from air side, water side and heat pipe. It was concluded that the heat pipe system has a great application potential due to its excellent heat transfer performance under the operation conditions of small temperature difference. Many significant studies have been focused on the cooling performance of twophase cooling system [8–10]. For example, Tong et al. investigated a thermosyphon loop used to cool a data center, and the EER is 10.96 at the temperature difference of 10 °C [12]. Normally, heat pipe is mainly driven by gravity or capillary force. Although the cooling performance of heat pipe has been improved by optimizing the physical and geometrical parameters of the capillary structure [13], there are still many limitations in the high capacity and complex structures. In order to enhance the dynamic behaviors of the working medium, the driving force has been improved continuously by using various modes, such as electrokinetic force [14,15], hydromagnetic force [16], centrifugal force [17] and pump [18–21]. Among the four driving modes, the electrokinetic force or hydromagnetic force is still limited, and it is difficult to apply them in the cooling system of data center. Although the centrifugal force can provide
144
Y. Ma et al./Applied Thermal Engineering 95 (2016) 143–149
enough power to the devices, this driving mode is only suitable for specific occasions, such as motor or generator cooling, etc. Pumpdriven two phase cooling system is an excellent thermal control cooling system and has many advantages, such as active temperature control, long heat transport distance and significant structural flexibility. The design and testing of the two-phase fluid thermal loops driven by pump for future large spacecraft have been investigated from 1982 to 1993 [18]. Liu et al. investigated the cooling performance of a mechanically pumped two-phase cooling loop for space thermal control [19,20]. These studies show that the cooling loop exhibits excellent performances and flexibility of structures. However, the structure and heat transfer mechanism are quite different from those in the ground compared with the space microgravity environment. In the past several years, active twophase cooling technology for spray cooling has been investigated for the laser diode array cooling and other high heat flux application [21,22]. However, spray cooling is very complex due to its dependence on many factors. The fluid control of two phase mixed flow is one of the factors. Different concepts of two-phase on-chip cooling cycles were proposed and compared for cooling data center servers [23–25]. The results show that, when compared with traditional air cooling systems, the energy consumption of the data center could be reduced by as much as 50% when using a liquid pumping cycle. Further investigation indicates that a maximum EER of the liquid pump cycle is about 19 and that of the vapor compressor cycle is about 4, respectively. Zhang et al. designed and investigated a pump driven loop heat pipe for data center cooling
[26]. The experimental results indicated that the EER is 3.75 at the temperature difference of 10 °C, and it increases to 9.37 at the temperature difference of 25 °C. Whereas, the existing system has the disadvantages which affect the performance as follows: (1) The heat dissipated by the canned pump is nearly absorbed by the working medium which increased the heat load of the system and affected the safe and steady operation of pump. (2) The flow resistance of the system is large which caused a higher sensible heat transfer. However, previous study of pump-driven two phase cooling system for data centers mainly focused on the on-chip cooling and the application with multiple evaporators. The objective of the present study is primarily to analyze the cooling performance in data center to obtain a higher cooling capacity and EER. For this purpose, a magnetic pump was used in the system to avoid the effect of the heat load by motor, and the designed heat exchanger and connecting pipe were considered to decrease the flow resistance of the system. Based on these, the performance characteristics of the magnetic pump and the system performances with mass flow rate and temperature difference were experimentally investigated. 2. Experiment setup Fig. 1 illustrates the schematic diagram of experimental setup and the p–h diagram of the proposed cooling system. As shown in Fig. 1(a), the cooling system is mainly composed of a magneticdriving pump, a tube–fin condenser, a tube–fin evaporator, a liquid reservoir and some connecting pipes. Fig. 1(b) shows the p–h diagram
(a) Schematic diagram of experimental setup
(b) p-h diagram of the proposed cooling system Fig. 1. Schematic diagram of experimental setup and p–h diagram of the proposed cooling system. (a) Schematic diagram of experimental setup. (b) p–h diagram of the proposed cooling system.
Y. Ma et al./Applied Thermal Engineering 95 (2016) 143–149
145
Table 1 Specification of the tube–fin heat exchanger. Geometric parameters
Value
Geometric parameters
Value
Outer tube diameter/mm Number of tube rows in air direction Transverse tube spacing/mm Fin thickness/mm
9.52 3 25 0.15
Inner tube diameter/mm Number of tube rows perpendicular to air direction Longitudinal tube spacing/mm Fin spacing/mm
8.85 36 21.65 2.15
(1)
3. Results and discussion 3.1. Performance characteristics of the magnetic pump The performance characteristics of the magnetic pump influence greatly on the performance and the control strategy of the cooling system. Therefore the operating parameters were first investigated. The pressure head and mass flow rate of the magnetic pump varying with the motor frequency are shown in Fig. 2, when the fluid temperature set as 15 °C ± 0.5 °C. It can be seen that the pressure head and the mass flow rate increase rapidly with increase of the motor frequency from 6 Hz to 22 Hz. The increment of the pressure difference between the pump inlet and outlet is greater than that of the mass flow rate. This is because the internal leakage through the gaps between the impeller and the pump
1.2
500
Δp m
1.0
400
0.8 300 0.6 200
m / kg·h-1
W S = W p + W p ,fc + W con ,f + W con ,fc + W eva ,f + W eva ,fc
with a standard pressure source with accuracy of ±0.25% over the full scale range and the temperature sensors are Pt100 thermometers. The flow rate of the fluid is measured using a Coriolis mass flow rate meter with an accuracy of ±0.2%. Two sight glasses are installed in the inlet and outlet of the Coriolis mass flow rate meter to observe the case of two-phase flowing state. For steady state testing, the sampling interval was 5 s. All the measurement data are recorded and stored on a computer by using an Agilent 34970A series data acquisition system. In data centers, the temperature and humidity play an essential role to provide a proper running environment for servers. In order to investigate the actual characteristics of the cooling system driven by magnetic pump and the effect of the outdoor temperature on its cooling capacity, the performance of the cooling system under typical data center operating condition was tested. When testing, the DB of the indoor test room set as 25 °C and the WB of the evaporator inlet is 17 °C according to a Chinese standard [29]. The DB of the outdoor test room is from 0 °C to 15 °C.
Δp / bar
of the pump-driven two-phase cooling system. The liquid from the reservoir flows to the evaporator through the liquid pipe by the pump driving (processes 1–3). The fluid at the evaporator inlet is usually in sub-cooled state as both the heat leakage and the flowing resistance in the liquid pipe are relatively small. The fluid in the evaporator absorbs heat from the hot air of the indoor room and vaporizes partly (processes 3–5). And it changes into two-phase fluid in the evaporator and the two-phase fluid from the evaporator outlet flows then into the condenser through the two-phase pipe (processes 5 and 6). The fluid in the condenser rejects heat to the cold air of the outdoor room and is cooled into liquid again, and then it returns into the reservoir (process 6–1). The next working process of the cooling cycle restarts. R22 was used as the refrigerant, and the filling ratio is 0.64 at the temperature of 15 °C. The magneticdriving pump is used in the system to avoid the leaks from the shaft seal in conventional pump. The pump is driven by the motor through the magnetic coupler, and its working chamber is completely sealed. The two sight glasses are installed in the inlet and outlet of the magnetic pump to see the fluid flowing state such as the case of cavitations. Both the liquid pipe and two-phase pipe in the cooling cycle are the standard copper tube. The specification of the liquid line is 19.05 mm (Diameter) × 4.5 m (Length), and that of the twophase line is 19.05 mm (Diameter) × 3 m (Length). The evaporator and condenser are placed in the same horizon one meter above the ground and used the same construction of tube–fin heat exchanger except for the headers. The specification of the tube–fin heat exchanger is listed in Table 1. Tests were conducted in a standard psychrometric chambers with two insulated rooms by using the air enthalpy method based on the ASHRAE Standard [27]. The evaporator is installed in the indoor test room, and the cooling capacity was estimated by using the air flow rate and the enthalpy difference of air across the evaporator. Four platinum resistance thermometers (Pt100) made by INOR with an uncertainty of ±0.05 °C are set to measure the dry-bulb temperature (DB) and wet-bulb temperatures (WB) of the air at the evaporator inlet and outlet. The air flow rate of the evaporator is obtained by a nozzle flow rate meter, across which the pressure difference is measured by a differential pressure transducer with an uncertainty of ±0.25% according to ANSI/AMCA 210 [28]. The condenser, liquid receiver and the magnetic pump assembly are installed in the outdoor test room. The DB and WB of the inlet air are measured. The input power of the magnetic pump, frequency converters and fans are measured by a power analyzer of WT230, with an uncertainty of ±0.02%. The total input power of the cooling system is given by the following equation:
0.4
where WS is the input power of the cooling system, kW. Wp is the input power of the magnetic pump, kW. Wcon,f and Weva,f are the input power of the condenser fan and evaporator fan, respectively, kW. Wp,fc, Wcon,fc and Weva,fc are the input power of the frequency converters for the magnetic pump, the condenser fan and evaporator fan, respectively, kW. P and T, as shown in Fig. 1, represent the pressure sensors and temperature sensors to measure the working fluid. The pressure in the cooling cycle is measured by MEAS pressure sensors calibrated
100
0.2
0.0 6
8
10
12
14
16
18
20
0 22
frequency / Hz Fig. 2. Pressure difference and mass flow rate of the pump varying with the frequency.
146
Y. Ma et al./Applied Thermal Engineering 95 (2016) 143–149
0.15
10
10
Wp
9
We
0.12
8
8 6
0.06
4
0.03
7
Q / kW
0.09
η/%
W / kW
η
2
0.00 6
8
10
12
14
16
18
20
ΔT=10 oC ΔT=15 oC ΔT=20 oC ΔT=25 oC
6 5 4 3
0 22
2
frequency / Hz
100
200
300
400
500
m / kg·h-1
Fig. 3. Input power, effective output power, efficiency of the pump varying with the frequency.
Fig. 4. Cooling capacity at different temperature differences varying with mass flow rate.
We =
1 m ⋅ Δp ⋅ 36 ρ
W η= e Wp
(2)
(3)
where We is the effective power of the magnetic pump, kW. Δp is the pressure difference of the working fluid at the pump inlet and outlet, bar. m is the mass flow rate of liquid R22, kg h−1. ρ is the density of the liquid R22 at the pump inlet, which is obtained from REFPROP [30], kg m−3. η is the efficiency of the magnetic pump. Wp is the input power of the magnetic pump, kW. The input power, effective power and efficiency of the magnetic pump varying with the motor frequency are shown in Fig. 3. It can be seen that the input power of the magnetic pump increased rapidly with the increase of motor frequency. The input power increased from 23 W to 143 W when the motor frequency increased from 6 Hz to 22 Hz. The effective power obtained by R22 is very small which is only 0.21 W at the motor frequency of 6 Hz and is 11 W at the motor frequency of 22 Hz. The efficiency of the magnetic pump is 0.92%–7.7% when the motor frequency is from 6 Hz to 22 Hz. 3.2. Effects of the mass flow rate The cooling capacity of the cooling system varying with the mass flow rates at different temperature differences is shown in Fig. 4. The cooling capacity becomes higher when the temperature difference is growing. The maximum cooling capacity reaches 3.429 kW and 9.241 kW when the temperature differences are 10 °C and 25 °C, respectively. The cooling capacity increases rapidly with the increase of the mass flow rate when the flow rate is relatively small, and then it decreases slowly when the mass flow rate increases continuously. Fig. 5 shows the inside coefficient of heat transfer at different temperature differences varying with mass flow rate. The inside coefficient of heat transfer of the evaporator is increased greatly with the flow velocity due to the flow velocity of the fluid increases rapidly
with increase of the mass flow rate. It is also one of the reasons that the cooling capacity increases rapidly with the increase of the mass flow rate. The inside coefficient of heat transfer is given by Eq. (4):
ki =
Q A ⋅ Δt
(4)
where ki is the inside coefficient of heat transfer, kW m−2 K−1. Q is the cooling capacity of the system, kW. A is the internal surface area of the evaporator, m2. Δt is the inside logarithmic temperature difference of the evaporator, °C. Fig. 6 shows the sub-cooling degree at different temperature differences varying with mass flow rate in the evaporator inlet. The sub-cooling degree of the evaporator inlet is higher when the indoor and outdoor temperature difference is larger. And it nearly increases linearly with the mass flow rate. It is because the flow resistance from the evaporator inlet to the pump inlet of the fluid
0.6
0.5
0.4
ki / kW·m-2·K-1
body increases with increase of the pressure difference between the pump inlet and outlet. And then it will cause the increment of the mass flow rate to slow down gradually. The effective power and efficiency of the magnetic pump are given by Eqs. (2) and (3):
0.3
ΔT=10 oC ΔT=15 oC ΔT=20 oC ΔT=25 oC
0.2
0.1
0.0 100
200
300
m / kg·h
400
500
-1
Fig. 5. Inside coefficient of heat transfer at different temperature differences varying with mass flow rate.
Y. Ma et al./Applied Thermal Engineering 95 (2016) 143–149
input power / kW
2.5
ΔT=10 C ΔT=15 oC ΔT=20 oC ΔT=25 oC
2.0
1.5
ΔT=10 oC ΔT=15 oC ΔT=20 oC ΔT=25 oC
0.35
0.30
0.25 35 30 25
1.0
EER
degree of supercooling / oC
o
147
20 15
0.5
10 100
200
300
400
500
m / kg·h-1
0.0 100
200
300
400
500 Fig. 8. Input power and EER at different temperature differences varying with mass flow rate.
m / kg·h-1 Fig. 6. Sub-cooling degree at different temperature differences varying with mass flow rate in the evaporator inlet.
is increased with the increase of the mass flow rate. Therefore, the pressure of the evaporator inlet is also increased when the flow resistance from the evaporator inlet to the pump inlet becomes bigger, which would increase the fluid sub-cooling degree at the evaporator inlet. In the evaporator, the proportion of sensible heat capacity becomes larger, and the evaporation temperature of the working fluid increased gradually with the increase of sub-cooling degree at the evaporator inlet. Accordingly, the cooling capacity decreased when the mass flow rate continues to increase. Therefore, the sub-cooling degree at the evaporator inlet becoming great can cause a decrease in the cooling capacity of the cooling system. The vapor quality of the working fluid at the evaporator outlet varying with the mass flow rates at different temperature differences is shown in Fig. 7. Under all conditions of the temperature difference, the quality at the evaporator outlet always goes up when the temperature difference between indoor and outdoor become larger, and goes down when the mass flow rate becomes bigger.
EER =
Q Ws
(5)
where EER is the energy efficiency ratio of the cooling system. The EERs corresponding to the maximum cooling capacity reach 12.9 and 29.7 when the temperature differences are 10 °C and 25 °C, respectively. However these values are not the maximum EER. When the temperature difference is 10 °C, the maximum EER is 13.4 while the mass flow rate is 185.2 kg h−1 and the cooling capacity is 3.23 kW. If the temperature difference rises to 25 °C, the maximum EER is 32.1 while the mass flow rate is 244.7 kg h−1 and the cooling capacity is 8.35 kW. It is advantageous that the cooling capacity corresponding to the maximum EER is selected as design value to meet the need of data center cooling as a better energy saving effect can be gotten for the data center.
1.0
ΔT=10 oC ΔT=15 oC ΔT=20 oC ΔT=25 oC
0.8
When the temperature differences are 10 °C and 25 °C, the quality corresponding to the maximum cooling capacity is 0.22 and 0.49, respectively. The input power and EER of the cooling system at different temperature differences varying with the mass flow rates are shown in Fig. 8. It can be seen that the input power of the cooling system has variation tendency similar to the one in Fig. 3. This is because the input power of fans and frequency converters of the condenser and the evaporator remain nearly constant, but the pump varies significantly under different operation conditions. Therefore, the input power curves of the cooling system changes with the input power of pump. The EER of the cooling system grows up with the increase of the mass flow rate when the flow rate is low, and then it decreases rapidly if the mass flow rate continues to increase. And the EER becomes higher when the indoor and outdoor temperature difference is growing. The EER is given by Eq. (5):
x
0.6
0.4
3.3. Effects of the temperature difference 0.2
0.0 100
200
300
m / kg·h
400
500
-1
Fig. 7. Vapor quality at different temperature differences varying with mass flow rate.
The maximum cooling capacity and its latent, sensible components and sub-cooling degree at evaporator inlet varying with the indoor and outdoor temperature difference are shown in Fig. 9. The maximum cooling capacity, latent component and sensible component almost display a linear growth with the temperature difference, and the sub-cooling degree at the evaporator inlet also increases linearly with the temperature difference approximately, which is mainly caused by the pressure drop of the cooling system from the evaporator inlet to the pump inlet. The sensible component
Y. Ma et al./Applied Thermal Engineering 95 (2016) 143–149
10
Q sub-cooling degree latent heat transfer sensible heat transfer
8
4
drop of the cooling system reaches 0.5 bar and 0.79 bar when the temperature differences are 10 °C and 25 °C, respectively.
3
4. Conclusions
Q / kW
6 2 4 1 2
0
sub-cooling degree / oC
148
0 10
15
20
25
ΔT / oC Fig. 9. Maximum cooling capacity and its latent, sensible components and subcooling degree at evaporator inlet varying with temperature difference.
changes with the sub-cooling degree of the fluid at the evaporator inlet. When the temperature difference is 10 °C, the sensible heat is 0.077 kW while the mass flow rate is 290.1 kg h−1 and the sensible heat ratio is 2.2%. If the temperature difference rises to 25 °C, the sensible heat is 0.225 kW while the mass flow rate is 339.4 kg h−1 and the sensible heat ratio is 2.4%. Fig. 10 shows the pressure distribution along the axis of the cooling system. Due to the influence of the outdoor temperature, the pressure in the cooling system is higher when the indoor and outdoor temperature difference becomes smaller. The pressure of the fluid increases rapidly in the pump and then it decreases continuously in the following parts such as liquid line, evaporator, twophase line and condenser, etc. The pressure drop of the fluid in the liquid line and two-phase line accounted for about 60% of the total pressure drop. Due to the effect of gravity, the pressure drop of the evaporator accounted for about 26% of the total pressure drop, and that of the condenser accounted for about 14% of the total pressure drop. The total pressure drop of the cooling system almost increases linearly with the temperature difference. The total pressure
10
ΔT=10 oC ΔT=20 oC
ΔT=15 oC ΔT=25 oC
Pressure / bar
9
8
The performance of the pump-driven two phase cooling system is experimentally investigated and the main results are summarized as follows: 1. The pump efficiency is only 7.7% when the mass flow rate is 465.9 kg h−1 and the pressure difference between the pump inlet and outlet is 1.04 bar. 2. The cooling capacity increases rapidly with the increase of the mass flow rate, and then decreases slowly when the mass flow rate continues to increase. When the temperature differences are 10 °C and 25 °C, the maximum cooling capacity reaches 3.429 kW and 9.241 kW, and the EER is 12.9 and 29.7, respectively. The subcooling degree at the evaporator inlet is not advantageous to increasing the cooling capacity of the cooling system. 3. The maximum cooling capacity and its latent and sensible components almost increase linearly with the indoor and outdoor temperature difference.
Acknowledgements This project was supported by the National Natural Science Foundation of China (Grant No. 51376010). Nomenclature A EER Q W ki m x
Area (m2) Energy efficiency ratio Cooling capacity (kW) Power dissipation (kW) Inside coefficient of heat transfer (kW m−2 K−1) Mass flow rate (kg h−1) Outlet vapor quality of evaporator
Greek symbols ΔT Indoor and outdoor temperature difference (°C) Δt Logarithmic temperature difference (°C) Δp Pressure difference (bar) η Efficiency of pump ρ Refrigerant density (kg m−3) Subscripts con Condenser e Effective power eva Evaporator f Fan fc Frequency converter p Pump s Cooling system
References 7
pump 6
liquid line evaporator two-phase condenser line
Fig. 10. Pressure distribution corresponding to the maximum cooling cycle capacity at different temperature differences.
[1] K. Jonathan, Growth in data center electricity use 2005 to 2010, A report by Analytical Press, completed at the request of The New York Times, 2011. [2] K. Ebrahimi, G.F. Jones, A.S. Fleischer, A review of data center cooling technology, operating conditions and the corresponding low-grade waste heat recovery opportunities, Renew. Sustain. Energy Rev. 31 (2014) 622–638. [3] J. Peter, T. Marker. Data centre energy efficiency product profile, Pitt & Sherry, report to equipment energy efficiency committee (E3) of The Australian Government Department of the Environment, Water, Heritage and the Arts (DEWHA), 2009.
Y. Ma et al./Applied Thermal Engineering 95 (2016) 143–149
[4] Y. Chen, Y. Zhang, Q. Meng, Study of ventilation cooling technology for telecommunication base stations in Guangzhou, Energy Build. 41 (7) (2009) 738–744. [5] J. Siriwardana, S. Jayasekara, S.K. Halgamuge, Potential of air-side economizers for data center cooling: a case study for key Australian cities, Appl. Energy 104 (2013) 207–219. [6] Z. Potts Free cooling technologies in data centers applications. Manchester, SUDLOWS White Paper, 2011. [7] L. Bao, J. Wang, L. Kang, The applied effect analysis of heat exchanger installed in a typical communication base station in Beijing of China, Energy Procedia 14 (2012) 620–625. [8] D. Zhu, D. Yan, Z. Li, Modeling and applications of annual energy-using simulation module of separated heat pipe heat exchanger, Energy Build. 57 (2013) 26–33. [9] R. Khodabandeh, Thermal performance of a closed advanced two-phase thermosyphon loop for cooling of radio base stations at different operating conditions, Appl. Therm. Eng. 24 (17–18) (2004) 2643–2655. [10] A. Samba, H. Louahlia-Gualous, S.L. Masson, D. Nörterhäuser, Two-phase thermosyphon loop for cooling outdoor telecommunication equipments, Appl. Therm. Eng. 50 (1) (2013) 1351–1360. [11] H. Zhang, S. Shao, H. Xu, H. Zou, C. Tian, Free cooling of data centers: a review, Renew. Sustain. Energy Rev. 35 (2014) 171–182. [12] Z. Tong, T. Ding, Z. Li, X.-H. Liu, An experimental investigation of an R744 two-phase thermosyphon loop used to cool a data center, Appl. Therm. Eng. 90 (2015) 362–365. [13] V.M. Kiseev, V.V. Vlassov, I. Muraoka, Experimental optimization of capillary structures for loop heat pipes and heat switches, Appl. Therm. Eng. 30 (11–12) (2010) 1312–1319. [14] S. Ebrahimi, A. Hasanzadeh-Barforoushi, A. Nejat, F. Kowsary, Numerical study of mixing and heat transfer in mixed electroosmotic/pressure driven flow through T-shaped microchannels, Int. J. Heat Mass Transf. 75 (2014) 565– 580. [15] C.-O. Ng, C. Qi, Electroosmotic flow of a power-law fluid in a non-uniform microchannel, J. Non-Newtonian Fluid Mech. 208–209 (2014) 118–125. [16] A.E. Kabeel, E.M.S. El-Said, S.A. Dafea, A review of magnetic field effects on flow and heat transfer in liquids: present status and future potential for studies and applications, Renew. Sustain. Energy Rev. 45 (2015) 830–837. [17] F. Song, D. Ewing, C.Y. Ching, Heat transfer in the evaporator section of moderate-speed rotating heat pipes, Int. J. Heat Mass Transf. 51 (7–8) (2008) 1542–1550.
149
[18] I.Y. Chen, M.-C. Chyu, Two-phase active thermal control systems for spacecrafts, in: Energy Conversion Engineering Conference, 1996, IECEC 96, Proceedings of the 31st Intersociety 2, 1996, 1488–1493. [19] J. Liu, N. Pei, K. Guo, Z. He, T. Li, M. Lv, Experimental investigation on mechanical pumped cooling loop for application in future space missions, Energy Convers. Manag. 49 (10) (2008) 2704–2710. [20] J. Liu, N. Pei, K. Guo, Z. He, T. Li, Experimental investigation on a mechanically pumped two-phase cooling loop with dual-evaporator, Int. J. Refrig. 31 (7) (2008) 1176–1182. [21] E.A. Silk, J. Kim, K. Kiger, Spray cooling of enhanced surfaces: impact of structured surface geometry and spray axis inclination, Int. J. Heat Mass Transf. 49 (2006) 4910–4920. [22] E.A. Silk, E.L. Golliher, R. Paneer Selvam, Spray cooling heat transfer: technology overview and assessment of future challenges for micro-gravity application, Energy Convers. Manag. 49 (2008) 453–468. [23] J.B. Marcinichen, J.A. Olivier, J.R. Thome, On-chip two-phase cooling of datacenters: cooling system and energy recovery evaluation, Appl. Therm. Eng. 41 (2012) 36–51. [24] J.B. Marcinichen, J.A. Olivier, V. de Oliveira, J.R. Thome, A review of on-chip micro-evaporation: experimental evaluation of liquid pumping and vapor compression driven cooling systems and control, Appl. Energy 92 (2012) 147–161. [25] J.B. Marcinichen, D. Wua, S. Paredes, J.R. Thome, B. Michel, Dynamic flow control and performance comparison of different concepts of two-phase on-chip cooling cycles, Appl. Energy 114 (2014) 179–191. [26] S. Zhang, G. Ma, F. Zhou, Experimental study on a pump driven loop-heat pipe for data center cooling, J. Energy Eng. (2014) doi: 10.1061/(ASCE)EY.19437897.0000253 (in press). [27] ASHRAE Standard 116, Methods of Testing for Seasonal Efficiency of Unitary Air-conditioners and Heat Pumps, ASHRAE, Atlanta, GA, 2007.( [28] ANSI/AMCA 210, Laboratory Methods of Testing Fans for Rating, ANSI, Arlington, VA, 2007. [29] AQSIQ (General Administration of Quality Supervision, Inspection and Quarantine of the People’s Republic of China), GB 50174-2008: National standard for code for design of electronic information system Room, 2010 (in Chinese). [30] E.W. Lemmon, M.L. Huber, M.O. McLinden, NIST Standard Reference Database 23: Reference fluid thermodynamic and transport properties – REFPROP, Version 8.0, National Institute of Standards and Technology, Standard Reference Data Program, Gaithersburg, 2007.