Experimental investigation on the operating characteristics of a dual compensation chamber loop heat pipe subjected to acceleration field

Applied Thermal Engineering 81 (2015) 297e312

Contents lists available at ScienceDirect

Applied Thermal Engineering journal homepage: www.elsevier.com/locate/apthermeng

Research paper

Experimental investigation on the operating characteristics of a dual compensation chamber loop heat pipe subjected to acceleration field Yongqi Xie a, *, Jie Zhang a, Liyao Xie a, Yin Yu a, Hongwei Wu b, Hongxing Zhang c, Hongxia Gao a a

School of Aeronautic Science and Engineering, Beihang University, Beijing 100191, China Institute of Engineering and Energy Technologies, School of Engineering and Computing, University of the West of Scotland, Paisley PA1 2BE, United Kingdom c Beijing Key Laboratory of Space Thermal Control Technology, China Academy of Space Technology, Beijing 100094, China b

h i g h l i g h t s
The DCCLHP operating performance is studied experimentally in acceleration field.
Acceleration effects have notable impacts on the DCCLHP performance in some cases.
Temperature fluctuation and reverse flow phenomenon are observed in the tests.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 6 October 2014 Accepted 9 February 2015 Available online 17 February 2015

High power and high local heat flux electronic devices employed in aircraft and spacecraft sustain the high acceleration condition in maneuvers and take-off stage. Loop heat pipe (LHP) are promising in dissipating high heat load to meet the increasing cooling needs. This article presents an experimental investigation on the operating characteristics of a dual compensation chamber loop heat pipe (DCCLHP) under elevated acceleration conditions. A centrifuge with a 2 m-long arm is used to provide the acceleration up to 7 g with four different acceleration directions. The heat load applied on the evaporator ranges from 80 W to 300 W. The typical performances in terrestrial were obtained and the influence of the different acceleration direction and magnitude on the operating characteristics was analyzed. Experimental results show that the change of the vaporeliquid distributions induced by the acceleration force results in some specific operating characteristics of the DCCLHP. The operating temperature becomes lower as the effect of the acceleration force improves the liquid returning. The operation of the DCCLHP demonstrates the sensitive behavior to the acceleration direction at small heat load and insensitive behavior at large heat load. It was also found that the acceleration magnitude can alter the operating mode. A number of unstable phenomena are observed under both terrestrial gravity and elevated acceleration conditions. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Loop heat pipe Dual compensation chamber Operating characteristics Elevated accelerated force Electronic cooling

1. Introduction With the development of packing technology and miniaturization, the electronic devices with high power level and large component density are widely used in avionics and space applications. Their thermal management, however, has become a critical issue because of the high heat load and hot spots. It is recognized that the typical cooling techniques for avionics, such as conduction

* Corresponding author. Tel.: þ86 10 82338081; fax: þ86 10 82338952. E-mail address: [email protected] (Y. Xie). http://dx.doi.org/10.1016/j.applthermaleng.2015.02.014 1359-4311/© 2015 Elsevier Ltd. All rights reserved.

and forced or natural convection, cannot meet the cooling needs. Therefore, there is an urgent demand to seek new cooling techniques or design concept to keep the components within the temperature limits [1e4]. Loop heat pipes (LHPs) could become one of the most effectively used cooling methods to prevent thermal failure. LHPs use the evaporation and condensation of a working fluid to transfer the heat and the capillary forces developed in the fine porous wicks to circulate the fluid [5,6]. The advantages of the accurate temperature control capability, long distance heat transport capability and flexibility in installation make themselves successfully apply in thermal control system of spacecraft [7e9]. As a two-phase heat

298

Y. Xie et al. / Applied Thermal Engineering 81 (2015) 297e312

Nomenclature a A c g G I L m Q r T U Dp

l q s

CC

radial acceleration, m/s2 area, m2 specific heat capacity, J/(kg K) gravitational acceleration, 9.81 m/s2 thermal conductance, W/k output current, A distance, m mass flow rate, kg/s heat load, W pore radius, m temperature, K output voltage, V pressure difference, Pa thermal conductivity, W/(m K) contact angle, arc degree surface tension, N/m2 compensation chamber

transfer unit, the LHP operation embodies a variety of dynamical responses due to the complex interactions of the numerous thermodynamics forces, interfacial forces and viscos forces. When the operating condition varies, the operating temperature will change during the transient state [10e12]. However, the aboard electronic devices always suffer from a variety of acceleration forces when the fighter aircraft maneuvers and combats. The operating performance of the conventional LHP with a single compensation chamber (CC) will be influenced by the orientation of the evaporator and compensation chamber under the terrestrial gravity and acceleration environments [13e15]. During literature surveys it was found that there were limited research investigated the acceleration effects on the conventional LHPs. Ku et al. [16,17] experimentally studied the startup and operating characteristics of a miniature aluminum-ammonia LHP to examine the effects of the varying acceleration. Several different situations were considered, including the LHP startup before the acceleration was applied and vice versa, several acceleration profiles were obtained with the value of radial acceleration ar from 0 g to 4.8 g and varying heat loads. Their results presented that the steady-state acceleration forces significantly influenced the liquidevapor distributions in the LHP and temperature oscillation. Fleming et al. [18] conducted an experimental study to investigate the behavior of a titanium-water LHP under the standard and elevated acceleration fields over the heat load ranges at the evaporator 100e600 W and at the compensation chamber 0e50 W with radial acceleration 0e10 g. It was found that the radial acceleration made the dry-out conditions occur more readily at a low heat input (Q < 400 W) while make the evaporator wall superheat turns higher when compared that with ar ¼ 0 g. The acceleration force changed the fluid distribution within the LHP and resulted in the operating temperature increases over those at ar ¼ 0.1 g in all instances, but the results obtained by Ku et al. [17] showed that the acceleration could either increase or decrease the operating temperature of LHP. Yerkes et al. [12] experimentally investigated the transient performance of a titanium-water LHP subjected to a sine wave acceleration field with the radial acceleration value ranging from 0.5 g to 10 g and with frequency from 0.01 Hz to 0.1 Hz. They found that the acceleration driven forces complimented the thermodynamic forces which improved the LHP dynamical performance and also countered the thermodynamic forces in some cases. It resulted

CCM DCCLHP LHP RTD VCM

constant conductance mode dual compensation chamber loop heat pipe loop heat pipe resistance temperature detector variable conductance mode

Subscripts cw cooling water dV Voltage drop loss e evaporator in at inlet loss heat loss mcap maximum capillary pressure out at outlet r radial sink sink total total w wick t tangential

in immediate total failure of the LHP to operate, delayed total failure, or stable operation but in degraded condition. Whether the LHP operated prior to or started up after an acceleration functions can induce various LHP dynamical performance characteristics. Based on the conventional LHP with a single compensation chamber, the dual compensation chamber loop heat pipe (DCCLHP) was developed by configuring two CCs on the two ends of the evaporator to solve the problem of the liquid supply difficulty for the primary wick under terrestrial gravity. Wolf and Bienert [19] investigated the temperature control characteristics of the LHPs with a single compensation chamber and dual compensation chambers. It was found that the LHP exhibited two operational modes, characterized by either constant conductance mode (CCM) or variable conductance mode (VCM). Gerhart et al. [20,21] verified that the DCCLHP operated normally when the evaporator and CCs were at different positions and presented different experimental results. Lin et al. [22e25] carried out some studies on the start-up behavior, operating characteristics, operating instability and visualization. They presented a detailed analysis in understanding the operating mechanism. It appears from the previous investigations that there are only limited reports on the operating performance of DCCLHP. To the best of the authors' knowledge, however, compared with the conventional LHP, both theoretical and experimental investigations on the DCCLHP have been far from complete and there is still much room to be enhanced in this area. There is also a lack of available experimental data concerning the operating characteristics of the DCCLHP under acceleration fields. As such, the present research work is aimed to address the operating characteristics of the DCCLHP subjected to various heat loads and radial acceleration forces. In the current study, both transient and steady-state performances of the DCCLHP are studied under terrestrial gravity and acceleration conditions. Four different directional acceleration configurations at different acceleration magnitudes and heat loads are applied in the study. 2. Experimental apparatus and procedure 2.1. Experimental apparatus A new experimental test apparatus was constructed to determine the operating performance of a DCCLHP under elevated

Y. Xie et al. / Applied Thermal Engineering 81 (2015) 297e312

acceleration fields. The system mainly consisted of a centrifuge in a circular pit, acceleration control system, water cooling circulation system, test section, control and data acquisition system. Fig. 1 shows the schematic diagram of the experimental apparatus and Fig. 2 shows a photo of the centrifuge and the test section. In the water cooling circulation system, the water temperature is kept at 19  C by a recirculating thermostatic water tank. Water circulation in the loop was maintained by a gear pump via a variable-frequency driver under both stationary and elevated acceleration conditions. A Coriolis force mass flow meter (DMF-1-2) is chosen to provide accurate flow measurement with an accuracy of ±0.5%. The test section is fixed on the rotating arm of the centrifuge. After the circulating water flowed out of the condenser, it entered a heat exchanger where it was cooled to a low temperature before recycling back to the thermostatic water tank. The signals measured by the temperature transducers and mass flow meter are achieved by utilizing a digital interface and recorded by a remote computer. During the experiment, all the connecting tubes in the water loop, signal wires and electrical wires used for heating go through the centrifuge axis to the test section. As the centrifuge is rotating during the experiment, the liquid collecting rings and the electric slip rings in the axis are specially designed to keep the flow and electric current working properly. The rotational speed of the centrifuge could be well controlled by the acceleration control system with an accuracy of ±5% of the indicated speed. The centrifuge is driven by an electric motor which allows the rotational acceleration at the end of the rotating arm to be varied up to 7 g with a continuous operation for no more than an hour. When the test section assembly is mounted onto the rotating arm, the nonuniform radial acceleration fields are induced because of the configuration of the DCCLHP. A percentage of the acceleration ranges from 90% to 130% compared to the center over the test

299

Fig. 2. The centrifuge and the rotating system including test section.

section can be achieved by adjusting the rotating radius setting of the centrifuge and the results meet the requirement of GB/T 2423.15. Fig. 3 shows a picture of the test DCCLHP and the internal structure of the evaporator and CCs. Its outline envelope dimension is 565 mm  469 mm  25 mm. The evaporator casing is made of stainless steel and houses a primary nickel wick with a pore radius of 1.5 mm. All the transport lines are stainless steel smooth-wall tubes with an outer diameter of 3.0 mm. Thin-film electric resistance heater capable of delivering 400 W is attached to the outer

Fig. 1. The schematic diagram of the experimental apparatus.

300

Y. Xie et al. / Applied Thermal Engineering 81 (2015) 297e312 Table 1 Major design parameters of the test DCCLH. Component

Parameter

O.d./i.d.  length of casing 20 mm/18 mm  209 mm Material Stainless steel (primary) Wick Pore radius 1.5 mm Porosity >5  1014 m2 O.d./i.d.  length 18 mm/6 mm  190 mm Material Nickel Vapor line O.d./i.d.  length 3 mm/2.6 mm  225 mm Material Stainless steel Liquid line O.d./i.d.  length 3 mm/2.6 mm  650 mm Material Stainless steel Condenser line O.d./i.d.  length 3 mm/2.6 mm  2200 mm Material Stainless steel Compensation chamber O.d./i.d.  length 27 mm/25 mm  64 mm Material Stainless steel Number 2 Working fluid Ammonia Evaporator

Fig. 3. A picture of the DCCLHP and inner structure of the evaporator and CCs. (a) current DCCLHP (b) inner structure of the evaporator and CCs.

surface of the evaporator to apply heat load. This can be adjusted by both changing the output voltage and current of the DC power in the range of 0e250 V and 0e5 A, respectively. The cold plate is made of aluminum (type 6061) and is cooled by the circulating water. The condenser line is welded to the copper heat expansion plate which is adhered to the surface of the cold plate with thermal grease. All the components are wrapped with multilayer insulation materials. The DCCLHP is installed in an enclosure filled with glass wool for the purpose of thermal insulation. The test is conducted to avoid the influence of non-condensable gas. The major design parameters of the DCCLHP are summarized in Table 1. Sixteen resistance temperature detectors (RTDs) Pt100 are attached to the tube wall to monitor the temperature profile along the loop, ambient temperature, as well as the inlet and outlet temperature of the cold plate. The RTD locations are presented in Fig. 4. RTD16 is used to monitor the ambient temperature.

test section assembly needs to be mounted horizontally at the proper position along the rotational arm upon the requirement of the acceleration direction. Four different test configurations are employed in this study, as shown in Fig. 5. Different directional configurations can be set by rotating the whole enclosure by 90 using the arm as a reference. For configurations A and B, the longitudinal axis of the evaporator and CCs is aligned with the direction of the acceleration force. For the cases of configurations C and D, the longitudinal axis is perpendicular to the direction of the acceleration force. For each configuration, the acceleration is applied just after the LHP had successfully started and operated stably. Four main magnitudes (ar ¼ 1 g, 3 g, 5 g, 7 g) and five different heat loads (Qe ¼ 80 W, 150 W, 200 W, 250 W, 300 W) on the evaporator are applied. It is noted that the gravity is always present. In all the tests, the cold   plate inlet temperature is maintained from 20.1 C to 22.2 C and the ambient temperature is controlled in the range from 24.1 to 26.4  C by air conditioning. 3. Validation and uncertainty analysis 3.1. Validation Prior to conducting the aimed test, the experimental apparatus was validated with thermal conductivity of pure copper bar test under terrestrial gravity conditions. In the validation tests, an electric resistance heater was utilized to heat it at the end of the copper bar. A simple heat exchanger was used to cool the bar at the other end. Two RTDs were attached to the surface wall to monitor

2.2. Test procedure Prior to the real test, RTD calibrations were conducted over two temperature ranges depending on the anticipated operating temperature. A constant temperature bath and standard RTD are utilized during the calibration. Both inlet and outlet temperatures of the circulating water, as well as the ambient temperature are cali  brated over the anticipated range of 12e30 C in 2 C intervals. Thirteen RTDs attached on the DCCLHP are calibrated over the full  range of 12e56  C in 2 C increments. Prior to each experiment, the

Fig. 4. The arrangement of RTDs on the DCCLHP.

Y. Xie et al. / Applied Thermal Engineering 81 (2015) 297e312

301

the temperature of the bar. The distance between both RTDs locations was 60 mm. The ambient temperature was 14.4  C. The mass flow rate of cooling water was 3.31 kg/h. The other two RTDs located at the inlet and outlet of the heat exchanger to record the temperature of cooling water. In order to decrease heat loss from the surface to the environment, the bar was wrapped with insulation materials. When the steady state reached, the Fourier's law could be used to calculate the thermal conductivity l by the following equation:

l ¼ Qcw L=ðADTÞ

(1)

where Qcw is the quantity of heat conducted to the cooling water, L is the distance between the test locations of the surface temperature of the bar, A is the section area of the bar, DT is the temperature difference of surface temperatures of the bar. The test value obtained by Eq. (1) is 357.3 W/(m$K). Compared with the known value from Refs. [26], the test error is approximately 6.3%.

3.2. Heat loss calculation As the test section was mounted on the rotational arm and the DC power supply was located in the control room outside the pit, there was a voltage drop between the control room and the test section. In other words, the electric resistance of the electric wires cannot be neglected owing to the wire length. The voltage drop loss can be accurately calculated by measuring the voltage between the thin-film electric resistance heater. Since the wall temperatures of the evaporator, CCs and vapor line are higher than the ambient temperature generally, the conduction from the wall to the environment contributed to the heat loss. In addition, the most heat load applied on the evaporator was dissipated through the cooling water. Therefore, the heat loss Qloss can be estimated approximately by Eq. (2)

Qloss ¼ UI  QdV  mcp ðTout  Tin Þ

(2)

Qe ¼ UI  QdV

(3)

where U is the output voltage, I is the output current, QdV is the power loss produced by the voltage drop, m is the mass flow rate of the cooling water. In all tests, the maximum value Qloss/Qe is not more than 14.5%.

3.3. Uncertainty analysis In the current study, the overall uncertainty of the temperature measurements is approximately ±0.5  C when considering the effects of RTD, slip rings, electric wireless, junction end and the data logger. The maximum uncertainty of the temperature is 2.6%. The maximum uncertainty of the voltage and current is 2.9% and 1.9%, respectively. As a result, the maximum uncertainty of the heat load is 3.4%. The thermal conductance G of the DCCLHP is defined as

G ¼ Qe =ðTe  Tsink Þ

(4)

where Tsink ¼ 0.5(Tout þ Tin), is the average value of the cold plate inlet temperature Tin and outlet temperature Tout. The maximum uncertainty is 8.9%.

Fig. 5. Different configurations with radial acceleration direction. (a) Configuration A (b) Configuration B (c) Configuration C (d) Configuration D.

302

Y. Xie et al. / Applied Thermal Engineering 81 (2015) 297e312

4. Results and discussion

4.2. Operating characteristics under elevated acceleration fields

4.1. Typical characteristics under terrestrial gravity

4.2.1. Effect of different acceleration directions During the experiment, the acceleration is just applied as the operation of the DCCLHP reaches a steady state in terrestrial. Fig. 8 shows the loop temperature profiles with the heat load of 150 W at 7 g under four configurations. The vaporeliquid distributions in the evaporator core and the CCs subjected to the effect of the acceleration force for configurations A, B and C are illustrated in Fig. 9. In Fig. 8, it can be found that the temperatures of the evaporator and vapor line decrease to a minimum value and then keep constant after the acceleration force is applied for all configurations. The steady-state operating temperature difference under between terrestrial gravity and acceleration fields becomes increasingly  apparent in order of configuration A, B, C and D, which is 4.8 C,    7.6 C, 9.5 C and 17.5 C, respectively. For the case of configuration A shown in Fig. 8(a), the CC1 temperature decreases slowly and the CC2 temperature decreases sharply as the acceleration is applied. It can be deduced that the liquid located in the evaporator core and the CC1 is pushed towards the CC2 due to the good hydraulic coupling between the core and the CCs under the effect of acceleration condition. Consequently, the vaporeliquid distribution changes to the state shown in Fig. 9(a) and the previous thermalehydrodynamic equilibrium is destroyed. The heat leak from the evaporator to the CCs decreases at the new vaporeliquid distribution. Simultaneously, because of the convection and cold bayonet inside the CC2, the CC2 is largely subcooled and the subcooled temperature of RTD11 away from the  evaporator is 16.9 C at steady state. It takes approximately 2700 s that the DCCLHP reaches steady state and the temperatures of the    evaporator, CC1 and CC2 are 46.9 C, 41.5 C and 30 C, respectively. The temperatures of RTD5 and RTD6 are obviously higher than those of RTD2, RTD3 and RTD4, which indicates that the vaporeliquid interface is located between RTD5 and RTD4. The condenser is not fully utilized, which shows that the DCCLHP operated at the VCM at 150 W and 7 g. Under configuration B, the acceleration direction is perpendicular to the axial direction of the evaporator which makes the liquid be pushed near the sidewall of the evaporator and the CCs, as shown in Fig. 9(b). This vaporeliquid distribution is similar to that in terrestrial condition for cylindrical evaporator and CCs and the effect of the acceleration is similar to that of the adverse elevation.

The objective of this study is to illustrate the operating characteristics of the stainless steel-ammonia DCCLHP subjected to acceleration force which functions just after its operation achieves steady state under terrestrial gravity. For the purpose of comparison, a few typical characteristics are obtained under the steadystate operating temperature in terrestrial (i.e. ar ¼ 0 g). Fig. 6 shows the variation of the steady-state temperatures of the evaporator, CC1 and CC2 at the heat load of 150 W for all test data without acceleration. It can be clearly seen that the operating temperatures show fluctuations even though the working conditions are kept the same. Chuang [27] also found such phenomenon for a single CC LHP. Combined with the visualization results of Lin et al. [24] and Okamoto et al. [28], it could be explained by the fact that the vapor bubble generation and different vapor quality in the evaporator liquid core with two CCs contributes to the variation of the steady-state operating temperature. It is worthy to note from Fig. 6 that the CC1 temperature is obviously higher than that of the CC2. It could be the reason that different vapor quality resulted in inconstant heat leak from the evaporator to the CCs. It is also believed that the cooling effect of the returning liquid leads to a lower temperature of the CC2 and further results in a lower pressure in the CC2 than that in the CC1. Fig. 7 presents the variation of the evaporator temperature (Te) with the heat load at the steady state in terrestrial. The ambient   temperature is 26.4 C and the sink temperature Tsink is 20.6 C. The DCCLHP is placed horizontally and the evaporator and condenser are at the same gravitational height. It can be seen that the temperature curve exhibits the typical ‘V’ shape at heat loads from 150 W to 300 W. It operates at the VCM when heat load is less than 250 W and translates to the CCM for heat loads more than 250 W. The difference of the operating temperature for all tests becomes smaller than that at the VCM. It should be pointed out that the fluid inventory of the DCCLHP is insufficient, namely, the CCs and the evaporator core cannot be flooded with liquid under all conditions. When the heat load is 80 W, the loop did not achieve a steady state and the evaporator temperature exceeded the maximum tolerance temperature.

Fig. 6. Steady-state temperatures of the evaporator, CC1 and CC2 at 150 W.

Fig. 7. Operating temperature vs heat load at the steady state in terrestrial.

Y. Xie et al. / Applied Thermal Engineering 81 (2015) 297e312

303

Fig. 8. Loop temperature profiles at 150 W and 7 g under configurations A, B, C and D. (a) configuration A (b) Configuration B (c) Configuration C (d) Configuration D.

During the initial period when the application of the acceleration of 7 g, the value of the tangential acceleration at is very small, but it may bring a small perturbation which changes the vaporeliquid distribution in the CCs like the case in configuration A and further influences the temperatures of the CCs. As a result, the temperatures of the evaporator and the CCs decrease. However, the external loop pressure drop increases with increasing acceleration magnitudes. It requires increasing the pressure difference across the wick, which in turn increase the saturation temperature. Therefore, the evaporator temperature needs to increase in order to provide enough pressure. However, the results are contrary to the analyses above, as shown in Fig. 8(b). The reason could be that the actual liquid evaporation area reduces due to the effect of the acceleration force at low heat load, thus the evaporation temperature decreases. It is a particular phenomenon obviously distinguished with those in the other configurations and the mechanism needs to be studied further. It takes about 2400 s to reach the thermal equilibrium state. The  temperatures of the evaporator and CC1 at steady state are 43.4 C  and 39.2 C, respectively. The tendency of the temperatures of RTD5 and RTD4 indicate that the working fluid becomes a subcooled liquid after RTD4. In addition, when the acceleration is stopped, the

temperature of the evaporator drops initially and then goes up to near the steady-state temperature under the terrestrial gravity. However, the temperature of the CC2 increases initially and then drops. For the case of configuration C, the DCCLHP is subjected to the adverse effect of the acceleration relative to configuration A. It can be seen that the temperatures of the evaporator, the CC1 and vapor line reduce sharply, whereas the CC2 temperature increases initially and then drops with the acceleration, as shown in Fig. 8(c). The reason for this phenomenon may be that the variations of the vaporeliquid distributions inside the loop result in the liquid volume level increases in the CC1 (see Fig. 9(b)) and then the heat leak from the evaporator to the CC1 decreases, thus the temperature of the CC1 drops at the effect of the colder liquid from the bayonet exit concurrently. Moreover, the temperature of RTD4 shows increasing behavior. This indicates that the effect of the acceleration changes the heat transfer performance of the condenser. The local heat transfer performance worsens and the condensation area augments, which leads to the temperature of the returning liquid (RTD12, RTD13) increase. Consequently, the CC2 temperature increases. The liquid derived from the region is pushed into the CCs. Under the combinative effect of the heat leak and the cooling of the

304

Y. Xie et al. / Applied Thermal Engineering 81 (2015) 297e312

Fig. 9. Vapor-liquid distributions in the evaporator core and CCs under configurations A, B and C. (a) Configuration A (Side view) (b) Configuration B (Top view) (c) Configuration C (Side view).

returning liquid, the temperature of the CC2 decreases gradually. Finally, the cooling effect of the returning liquid makes the temperatures of the CCs keep almost the same. At about 2350 s, thermal equilibrium in the evaporator and the CCs is achieved. The stable temperatures of the evaporator, the CC1 and CC2 are 37.2  C, 32.9  C and 32.5  C, respectively. Furthermore, it is interesting to note that the outlet temperature of the liquid line (RTD12) starts to fluctuate at approximately 800 s whereas the temperatures of RTD11and RTD13 show no fluctuations. The reason will be discussed in detail later. Under configuration D, the effect of the acceleration is contrary to that at configuration B, which is similar to the gravity-assisted effect. In Fig. 8(d), the stable temperatures of RTD10, RTD 9 and RTD11 under terrestrial gravity are 50.0  C, 45.5  C and 37.0  C, respectively. The temperatures of RTD7, RTD6, RTD5 and RTD4 are

48.1  C, 47.2  C, 43.5  C and 22.5  C, respectively, which presents the vaporeliquid interface is located between RTD5 and RTD4. As the acceleration is applied, the temperatures of RTD5, RTD6, RTD7, RTD9 and RTD11 drop quickly but the temperatures of RTD2, RTD3, RTD4 and liquid line increase. It takes approximately 300 s for the thermal equilibrium to be achieved. It is apparent that the temperature of the evaporator decreases to 32.4  C and the temperatures of all the other components locate in the range from 29.2  C to 30.7  C at the steady state. The phenomena at configuration D are significantly distinguished from those at others. The mechanism can be explained as the following. Based on the LHP theory, circulation of working fluid is possible only if the total loop pressure drop Dptotal does not exceed the maximum possible capillary pressure difference Dpmcap across the wick, which can be expressed by the inequality

Y. Xie et al. / Applied Thermal Engineering 81 (2015) 297e312

Dpmcap  Dptotal

(5)

where the maximum capillary pressure is given by the LaplaceeYoung equation

Dpmcap ¼ 2s cosðqÞ=rw

(6)

where rw is the pore radius of the wick and q is the contact angle. The variation of the capillary pressure is implemented by the change of the contact angle q which also varies with the evaporation temperature and evaporation pressure. As the acceleration is applied, the external loop pressure drop decreases due to the effect of the acceleration and the maximum pressure drop is achieved as the acceleration goes up to 7 g. This requires decreasing the capillary pressure difference to balance the total loop pressure drop. Correspondingly, the temperature difference across the wick also decreases. This reduces the heat leak, resulting in a lower operating temperature. Furthermore, the loop pressure equilibrium does not reach if the capillary pressure drop changes to 0 Pa (i.e., q ¼ 90 ), the effect of the acceleration will drive a portion of the liquid circulating in the loop. As a result, two-phase flow occurs in vapor line and both the vapor and the liquid are saturated, which can be confirmed that the temperatures of vapor line and the condenser are nearly equal. Moreover, the decrease of the actual liquid evaporation area due to the effect of the acceleration may also reduce the decrease of the evaporation temperature, as the case of configuration B. The combination of these effects above leads to significantly decreasing the operating temperature. Fig. 10 presents the loop temperature evolutions with the heat load of 300 W at 7 g under four different configurations. The inlet  temperature of the cold plate is maintained at 21.2 C. It is clearly seen that the temperatures of the liquid line fluctuate in terrestrial condition for all configurations in Fig. 10. During the initial period of the acceleration, the temperatures of liquid line initially decrease rapidly and then increase. The evaporator temperature at steady state under the acceleration condition is smaller than that under terrestrial gravity condition for each of the configurations except for configuration B. And the temperature drop of the evaporator at 300 W is less than that at 150 W due to the effect of the acceleration from Figs. 8 and 10. It indicates that the direction of the acceleration has little influence on the operating temperature at large heat load. In Fig. 10(a), the high heat load can make the vapor flow out of the condenser under terrestrial condition and the liquid derived from the condenser is pushed towards the CCs. The temperature fluctuations of the liquid lines imply that the vaporeliquid interface flows out of the condenser and returns again and again. As the acceleration is applied, the heat transfer performance of the condenser is changed due to the effect of the acceleration, where a thinner liquid layer in the vaporeliquid two phase region may be formed and the heat transfer coefficient increases. As a result, the temperatures of RTD1 and RTD2 drop rapidly which indicates that the liquid subcooling at the condenser exit increases. Consequently, the temperatures of the liquid line and the CCs drop. When the acceleration magnitude keeps 7 g, heat transfer coefficient of the condenser decreases. And the temperature of RTD2 increases to 35.8  C gradually. Thus the temperatures of the liquid line and the CC2 also increase. It can be seen that the temperature fluctuation of the liquid line under the acceleration condition becomes weak relative to that under terrestrial condition. It takes about 600 s for the evaporator to reach a steady state and the relative temperature  is 41.2 C under acceleration field. Both CCs temperatures are equal  to 35.6 C finally. When the acceleration is stopped, the loop temperatures increase. The temperatures of the liquid line start to fluctuate intensely. The stable temperature of the evaporator goes up to 43.4  C again.

305

For the case of configuration B, the effect of the acceleration in Fig. 10(b) is the same as that in Fig. 8(b), which results in the increase of the loop pressure drop and the heat leak to the core. Owing to the mass flow rate of the returning liquid being much larger at 300 W, the effect of the inertia force dominates. Thus the evaporator temperature increases slightly. Therefore, the temperatures of RTD6 and RTD7 also increase. In addition, the effect of the acceleration may change the thermal physical property of the working fluid and enhance the heat transfer performance of the condenser. The temperature of RTD2 drops sharply which shows that the liquid subcooling at the condenser exit increases. As a result, the temperatures of the liquid lines drop sharply and the temperatures of the CC2 and the evaporator decrease. It takes only about 250 s for the thermal equilibrium to be achieved. The stable temperatures of RTD10, RTD1, RTD12 and RTD13 at 7 g are 43.9  C,    26.9 C, 27.2 C and 27.0 C, respectively. It is interesting to note that the DCCLHP operates at VCM rather than CCM at 300 W under 7 g condition. As the acceleration is stopped, the situation is similar to that under configuration A. Under configuration C, it can be clearly seen that the temperatures of RTD10, RTD7 and RTD6 initially increase slightly and then drop after the acceleration is applied in Fig. 10(c). The reason may be that the effect of the acceleration induces to increase the loop pressure difference, which requires a higher operating temperature. On the other hand, the effect improves the heat transfer performance of the condenser, indicated by the sharp temperatures drop of RTD2, RTD1 and RTD12. The subcooling of the returning liquid increases. Then the temperatures of the liquid in the CCs and the core decrease, resulting in the decrease of the temperature of the evaporator. Under 7 g condition, the temperatures of the liquid line also fluctuate. At 460 s, the temperatures of both CCs are close to 36.1  C and the evaporator temperature is 43.7  C. In Fig. 10(d), the effect of the acceleration leads to the decrease of the loop pressure drop. Simultaneously, the effect of the inertia force weakens the impacts of the acceleration force at high heat load. The combinative effect of both forces in turn results in the slight drop of the evaporator temperature compared with the case in Fig. 8(d). The temperatures of the liquid line and RTD2 drop initially and then rise. The temperature of the CC2 decreases and fluctuates slightly. It takes about 70 s for the evaporator temperature to reach 41.9  C. Fig. 11 presents the steady state operating temperature and thermal conductance under four configurations at both 3 g and 7 g. It is useful to illustrate that the DCCLHP does not reach a steady state under configuration A during the maximum operating time of the centrifuge when the heat load is less than 300 W and the acceleration magnitude is less than 7 g. In Fig. 11(a), it can be found that the higher the heat load is, the smaller the temperature difference is among different configurations. The operating temperature is higher for configuration B than that for the other configurations at 150 W and 200 W under 3 g condition. At 150 W, the operating temperature is the lowest under  configuration C and the largest temperature difference is 4.8 C for different configurations. However, the maximum temperature difference is 2.1  C at 300 W. In Fig. 11(b), as the heat load is large, the thermal conductance is also large for all configurations. The largest value of the thermal conductance is 16.1 W/K at 300 W under configuration A. It can be seen that the DCCLHP operates at CCM at 250 W and 300 W for configuration B and C. However, it operates at VCM as the heat load from 150 W to 250 W and at CCM at 300 W under configuration D. This indicates that the direction of the acceleration can change the operating mode of the DCCLHP. As the acceleration magnitude is 7 g, the maximum operating  temperature difference is 14.1 C at 150 W between configurations A and D, as shown in Fig. 11(c). The smaller the heat load is, the

306

Y. Xie et al. / Applied Thermal Engineering 81 (2015) 297e312

Fig. 10. Loop temperature evolutions with heat load 300 W at 7 g for four configurations. (a) Configuration A (b) Configuration B (c) Configuration C (d) Configuration D.

larger the operating temperature difference is, which is the same to that at 3 g. From Fig. 11(d), the range of the heat load at VCM is from 150 W to 250 W under configuration A and B, but it is not more than 200 W under configuration C. Especially under configuration D, it operates at CCM for all heat loads. This indicates that the acceleration magnitude has a significant influence on the operation, which will be discussed in the following section. Based on the analyses above, it can be concluded that the operating temperature decreases as the acceleration force acts in most cases. This direction of the acceleration under configuration C and D can result in a lower operating temperature and a higher thermal conductance compared with that at configuration A and B as it operates at VCM. That is, the operation of the DCCLHP is sensitive to the direction of the acceleration at small heat load, especially under large acceleration condition. However, the insensitive behavior is observed at large heat load. 4.2.2. Effect of different acceleration magnitudes The loop temperature evolutions with the acceleration magnitudes of 1 g, 3 g, 5 g and 7 g at 200 W for four configurations are illustrated in Fig. 12. It can be clearly seen that the stable

temperature of the evaporator increases with the increase of the acceleration magnitude under configuration A, B and C, but on the contrary under configuration D. Generally, the duration of reaching a thermal equilibrium becomes shorter at high acceleration magnitude except for configuration B. According to the analyses above, the effect of the acceleration force changes the vaporeliquid distributions of the evaporator and the CCs as well as the heat transfer performance of the condenser. When it improves the subcooling liquid return, the operation of the DCCLHP is easier to reach a steady state within a short time. In each test, the temperature of the CC2 is much lower than the temperature of the CC1. In Fig. 12(a), the operation of the DCCLHP can reach a steady state only at 7 g for configuration A and the operating temperature is lower than that in terrestrial. This indicates that the high acceleration magnitude can improve its operation along the acceleration direction. The temperatures of RTD2 and RTD3 are less than 24.0  C and the temperature of RTD 5 is not less than 33.0  C. It shows that the condenser is not fully used. In Fig. 12(b), it is interesting to note that the stable operating temperature under the terrestrial condition is higher than that at 1 g and 3 g, but lower than that at 5 g and 7 g. The temperature difference of the CC2 between under

Y. Xie et al. / Applied Thermal Engineering 81 (2015) 297e312

307

Fig. 11. Operating temperature and thermal conductance at 3 g and 7 g under four configurations. (a) Operating temperature at 3 g (b) Thermal conductance at 3 g (c) Operating temperature at 7 g (d) Thermal conductance at 7 g.

terrestrial gravity and acceleration conditions augments with the increase of the magnitude of the acceleration. The effective condensation area of the condenser decreases at 7 g relative to the other cases. The condenser is also not fully used in each test. In the case of configuration C shown in Fig. 12(c), the steadystate temperature of the evaporator is 36.8  C, 37.2  C, 37.6  C and 37.7  C at 1 g, 3 g, 5 g and 7 g, respectively. This presents that the effect of the acceleration is not significant at all the configurations. The temperatures of the CC2 under the acceleration condition are very close in all tests. Under 1 g condition, the temperature of RTD2 goes up to 30  C with the condenser is fully utilized. With the increase of the acceleration magnitude, the final temperature of RTD2 decreases. From Fig. 12(d), it can be found that the higher the magnitude of the acceleration is, the lower the steady-state operating temperature is as well as the larger the effective condensation area. This indicates that the effect of the acceleration under high acceleration condition becomes more

significant at the direction of the acceleration. At 7 g, the condenser is fully utilized according to the temperature evolution of RTD2. The   stable temperature of the evaporator is 34.4 C, which is 4.8 C smaller than that at 1 g. In addition, the loop temperature evolutions at 150 W under different acceleration magnitude behave a similar vibrational trend for four configurations. At 5 g and 7 g, the condenser is also fully utilized at 150 W for configuration D. The results present that the operation mode is the function of the acceleration direction, magnitude and heat load. In some cases, the effect of the acceleration acts can be regarded as an additive heat load. However, in the other case, it acts as a cold load. Fig. 13 shows the loop temperature evolutions with the acceleration magnitude of 1 g, 3 g, 5 g and 7 g at 300 W for four configurations. In general, the steady-state operating temperature under acceleration condition decreases slightly with the increase of the acceleration magnitude for each configuration. The large

308

Y. Xie et al. / Applied Thermal Engineering 81 (2015) 297e312

Fig. 12. Loop temperature evolutions with different acceleration magnitude at 200 W for four configurations. (a) Configuration A (b) Configuration B (c) Configuration C (d) Configuration D.

acceleration magnitude will make the temperature of RTD2 and RTD11 vary significantly. The condenser is fully utilized under both terrestrial gravity and acceleration conditions. Thus the loop operates under a CCM and the operating temperature is dominated by the subcooling of the liquid exiting the condenser. The influence of the effect of the acceleration is much more significant on its heat transfer performance under configuration B and D compared with that under configuration A and C. In Fig. 13(a), the steady-state temperature of the evaporator is     42.1 C, 41.7 C, 41.6 C and 41.4 C at 1 g, 3 g, 5 g and 7 g, respectively. This indicates that the effect of the acceleration has a few impacts on the operating temperature. During the initial period of the acceleration, decreasing the temperature of RTD2 and RTD11 becomes larger with the increase of the acceleration magnitude. The temperature of the CC1 is close to that of the CC2 under acceleration condition. From Fig. 13(b), it can be clearly seen that the temperatures of the evaporator and the CCs fluctuate slightly in terrestrial gravity. As the acceleration is applied, the temperature of the evaporator increases firstly and then drops, whereas the temperature of RTD2 drops rapidly and then increases. However, the

temperature fluctuations of the evaporator and the CCs are suppressed under acceleration condition. At equilibrium, the temper    ature of the evaporator is 42.9 C, 42.8 C, 42.7 C and 43.5 C at 1 g, 3 g, 5 g and 7 g, respectively. After the centrifuge stops operate, the loop temperatures rise gradually and the temperature fluctuations occur again. For the case of configuration C, it can be found that the stable temperatures of the evaporator at different acceleration magnitude are all close to 43.9  C, as shown in Fig. 13(c). The temperature of each point on the condenser decreases much more with the acceleration magnitude increase. The temperature of the CC1 is lower than that of the CC2 at 1 g, 3 g and 5 g but higher than that at 7 g. In Fig. 13(d), it can be clearly seen that temperatures of the CC2 drops obviously as the acceleration is applied, whereas the temperature of the CC1 increases slightly. The maximum temperature  difference between the CCs is 3.8 C at 3 g. This could be the result of the increase of the heat leak to the CC1 and subcooling of the returning liquid. The temperature change of RTD2 indicates the heat transfer performance is improved by the effect of the acceleration force. The evaporator temperatures at steady state are

Y. Xie et al. / Applied Thermal Engineering 81 (2015) 297e312

309

Fig. 13. Loop temperature evolutions with different acceleration magnitude at 300 W for four configurations. (a) Configuration A (b) Configuration B (c) Configuration C (D) Configuration D.









42.8 C, 42.6 C, 42.3 C and 41.9 C at 1 g, 3 g, 5 g and 7 g for configuration D, respectively. Fig. 14 depicts the steady-state operating temperature and thermal conductance as a function of the heat load and acceleration magnitude for both configurations C and D. From Fig. 14(a) and (b), it can be found that the operating temperature under terrestrial condition is higher than that under acceleration field, whereas the thermal conductance shows a reverse behavior. As the heat load is large, the operating temperature difference under different acceleration magnitude is small, which indicates the effect of the acceleration magnitude is weak. As the heat load is small, the difference is large. When the heat load is not more than 250 W, the loop operates at VCM under all acceleration magnitudes. However, because the condenser is fully utilized at 1 g as shown in Fig. 12(c), the loop operates at CCM at 200 W under such a configuration. This presents that the effect of the acceleration force confuses the heat load range of CCM or VCM. It is hard to estimate the operating mode only by the operating temperature profiles or the value of thermal conductance.

In Fig. 14(c), the operating temperature at 150 W and 200 W increases with the increase of the acceleration magnitude for the case of configuration D. But an opposite trend appears at configuration C. When the acceleration magnitude is large, the heat load ranges of CCM become wide. Under 5 g and 7 g conditions, the DCCLHP operates at CCM with heat load from 150 W to 300 W, as shown in Fig. 14(d). However, it operates at VCM with the heat load of 150 W and 200 W under terrestrial gravity and 3 g and 5 g conditions. This indicates that the acceleration magnitude has significant impacts on the operating mode at small heat load. Since the heat load exceeds 250 W, the large acceleration magnitude results in the large thermal conductance. Furthermore, the effect of the acceleration becomes insignificant at high heat load. 4.3. Unstable phenomena A number of temperature fluctuation and reverse flow phenomena are observed in terrestrial gravity and elevated acceleration tests, which will bring adverse influence on the smooth

310

Y. Xie et al. / Applied Thermal Engineering 81 (2015) 297e312

Fig. 14. Operating temperature and thermal conductance at configuration C and D. (a) Operating temperature at configuration C (b) Thermal conductance at configuration C (c) Operating temperature at configuration D (d) Thermal conductance at configuration D.

operation of the DCCLHP. Several examples of such instabilities are analyzed in this section. In all tests, temperature fluctuations appear only at configurations A, B and C and easily happen at 300 W. Fig. 15 illustrates the transient temperature profiles at 300 W under configuration B. Before the acceleration was applied, the temperatures of the liquid line, CCs and evaporator fluctuated. The DCCLHP operated at the quasi-steady state. This phenomenon is similar to that described by Feng [25]. It might be resulted from the instability of the vaporeliquid two phase flow. After the acceleration of 5 g acts, the effect of the acceleration increases the external loop pressure loss, changes the thermal physical property of the working fluid and enhances the heat transfer performance of the condenser. The temperature of the evaporator increases to provide much higher pressure. The increased pressure will result in the condensation of the vapor. As a result, the temperatures of liquid line and RTD2 drop rapidly. The subcooling of returning liquid increases and the

evaporator temperature decreases until a new thermal equilibrium reaches. In this case, the effect of the acceleration suppressed the temperature fluctuation. Fig. 16 shows the transient temperature profiles at 250 W under configuration C. It can be clearly seen that the temperatures of the liquid line and RTD2 increase and then drop sharply as the acceleration is applied from Fig. 16. This phenomenon can be explained as the following. The vaporeliquid distribution in the loop changes and the heat leakage from evaporator core to the CC1 decreases under acceleration condition. Therefore the temperatures of the CC1 and the evaporator decrease. Considering the compressibility of the vapor and the pressure change in the loop, the original liquid inside the liquid line is pushed into the condenser by the effect of the acceleration. Synchronously, the working fluid with high temperature is derived from the core to the condenser exit along the liquid line. As a result, the temperatures of liquid line and RTD2 increase rapidly.

Y. Xie et al. / Applied Thermal Engineering 81 (2015) 297e312

Fig. 15. Transient temperature profiles at 300 W under configuration B.

On the other hand, this reverse flow results in the external loop pressure increasing. To balance the pressure, the capillary pressure difference across the wick increases. Accordingly, the temperature of the evaporator increases. Consequently, the liquid in the condenser is pushed out of the condenser exit firstly and then the vaporeliquid interface is pushed out, which accompany the rapid change of the temperatures of RTD2 and liquid line. To and fro, the normal circulation flow resumes finally. However, there exist temperature fluctuations at the liquid line exit and in the condenser at 7 g although there is no fluctuation under terrestrial gravity. Compared Fig. 16 with Fig. 8(c), it can be found that the temperature of RTD12 fluctuates with high frequency and that of RTD1 and RTD13 fluctuate slightly in both circumstances. It is believed that the vaporeliquid interface traverses at RTD12 location by the effect of the acceleration. The vapor may be extruded from the compensation chamber or generated inside the evaporator core. In addition, the condensation process is also impacted and the temperature of RTD4 or RTD2 fluctuates. In general, complex combinations of the acceleration force, inertial force, gravitational force, capillary force and surface tension lead to the liquidevapor interface traversing. For the regimes of the temperature fluctuation of the DCCLHP under acceleration conditions, further studies should be carried out. 5. Conclusions Experimental investigations on the operating characteristics of the DCCLHP under terrestrial gravity and elevated acceleration

Fig. 16. Transient temperature profiles at 250 W under configuration C.

311

conditions were conducted in a specifically designed acceleration test apparatus. The loop temperature evolutions were recorded for the purpose of comparison and the experimental results were presented and analyzed in detail. Overall, different operating temperature may be induced even at the same work conditions for the DCCLHP. In most cases, the operating temperature decreases as the acceleration acts. Especially as the effect of the acceleration improves the liquid returning, the loop is easy to reach a steady state. The operation is sensitive to the direction of the acceleration at small heat load and insensitive at large heat load. The effect of the acceleration can be regarded either as an additive heat load or as a cold load. The transition of the operation mode is the function of the acceleration direction, magnitude and heat load. The variation of the thermal conductance is small as the heat load is large. Temperature fluctuation and reverse flow phenomena are observed under elevated acceleration fields. The effect of the acceleration can suppress temperature fluctuations and also can provoke reverse flow. Acknowledgement The authors acknowledge the financial supports from the Fundamental Research Funds for the Central Universities of China (YWF-14-HKXY-019) and the National Natural Science Foundation of China (No.51406009). References [1] M.C. Zaghdoudi, C. Sarno, Investigation on the effects of body force environment on flat heat pipe, J. Thermophys. Heat Transfer 15 (4) (2001) 384e394. [2] Y.W. Wang, J.W. Cen, F.M. Jiang, et al., An experimental study on the performance of a stainless steel-water loop heat pipe under natural cooling condition, J. Therm. Sci. 23 (1) (2014) 91e95. [3] L.Y. Xie, Y.Q. Xie, H.W. Wu, J.Z. Yu, High gravity influence on boiling heat transfer in helical coils, Int. J. Heat Mass Transfer 73 (2014) 706e715. [4] C.S. Sharma, S. Zimmermann, M.K. Tiwari, et al., Optimal thermal operation of liquid- cooled electronic chips, Int. J. Heat Mass Transfer 55 (7e8) (2012) 1957e1969. [5] Y.F. Maydanik, Loop heat pipes, Appl. Therm. Eng. 25 (2005) 635e657. [6] A. Ambirajan, A.A. Adoni, J.S. Vaidya, et al., Loop heat pipes: a review of fundamentals, operation, and design, Heat Transfer Eng. 33 (4e5) (2012) 387e405. [7] C.L. Baker, W.B. Bienert, A.M. Ducao, Loop Heat Pipe Experiment, SAE Paper No.981580, 1998. [8] Y.F. Maydanik, Y.G. Fershtater, V.G. Pastukhov, et al., Some results of loop heat pipes development, tests, and application in engineering, in: Proceedings of the 5th International Heat Pipe Symposium, Melbourne, Australia, 1996, pp. 406e412. [9] T.D. Swanson, Thermal control technologies for complex spacecraft, in: Proceedings of the 13th International Heat Pipe Conference, China Astronautic Publishing House, 2003, pp. 3e11. [10] J. Ku, Operating Characteristics of Loop Heat Pipes, SAE Paper No.1999-012007, 1999. [11] J. Ku, L. Ottenstein, M. Kobel, et al., Temperature oscillations in loop heat pipe operation, AIP Conf. Proc. 552 (2001) 255e262. [12] L.K. Yerkes, D.S. James, L.C. David, et al., An Experimental Investigation into the Transient Performance of a Titanium-Water Loop Heat Pipe Subjected to a Steady-Periodic Acceleration Field, AIAA Paper No.2012-1009, 2012. [13] A.L. Philips, J.E. Fale, N.J. Gernert, et al., Loop Heat Pipe Qualification for High Vibration and High-g Environment, AIAA Paper No. 98e0885, 1998. [14] J. Baumann, B. Cullimore, J. Ambrose, et al., A Methodology for Enveloping Reliable Start-up of LHPs, AIAA Paper No. 2000-2285, 2000. [15] H.X. Zhang, G.P. Lin, T. Ding, et al., Investigation on start-up behaviors of a loop heat pipe, AIAA J. Thermophys. Heat Transfer 19 (2) (2005) 558e565. [16] J. Ku, L. Ottenstein, T. Kaya, et al., Testing of a Loop Heat Pipe Subjected to Variable Accelerating Forces, Part 1: Start-up, SAE Paper No. 2000-01-2488, 2000. [17] J. Ku, L. Ottenstein, T. Kaya, et al., Testing of a Loop Heat Pipe Subjected to Variable Accelerating Forces, Part2: Temperature Stability, SAE Paper No. 2000-01-2489, 2000. [18] A.J. Fleming, S.K. Thomas, K.L. Yerkes, et al., Titanium-water loop heat pipe operating characteristics under standard and elevated acceleration fields, J. Thermophys. Heat Transfer 24 (1) (2010) 184e198. [19] D.A. Wolf, W.B. Bienert, Investigation of Temperature Control Characteristics of Loop Heat Pipes, SAE Paper No.941576, 1994.

312

Y. Xie et al. / Applied Thermal Engineering 81 (2015) 297e312

[20] C. Gerhart, D.F. Gluck, Summary of operating characteristics of a dual compensation chamber loop heat pipe in gravity, in: Proceedings of the 11th International Heat Pipe Conference, Japan Association for Heat Pipes and Seikei University, Tokyo, Japan, September 1999. [21] D.F. Gluck, C. Gerhart, S. Stanley, Characterization of a high capacity, dual compensation chamber loop heat pipe, AIP Conf. Proc. 458 (1) (1999) 943e948. [22] G.P. Lin, H.X. Zhang, X.G. Shao, et al., Development and test results of a dual compensation chamber loop heat pipe, J. Thermophys. Heat Transfer 20 (4) (2006) 825e834. [23] L.Z. Bai, G.P. Lin, D.S. Wen, et al., Experimental investigation of startup behaviors of a dual compensation chamber loop heat pipe with insufficient fluid inventory, Appl. Therm. Eng. 29 (2009) 1447e1456.

[24] G.P. Lin, N. Li, L.Z. Bai, et al., Experimental investigation of a dual compensation chamber loop heat pipe, Int. J. Heat Mass Transfer 53 (2010) 3231e3240. [25] J.T. Feng, G.P. Lin, L.Z. Bai, Experimental investigation on operating instability of a dual compensation chamber loop heat pipe, Sci. China Ser. E-Technol. Sci. 52 (8) (2009) 2316e2322. [26] J.Z. Yu, Principle and Design of Heat Exchanger, first ed., Beihang University Publication, Beijing, 2006, p. 239. [27] P.A. Chuang, An Improved Steady-state Model of Loop Heat Pipes Based on Experimental and Theoretical Analyses, Ph.d thesis, Pennsylvania State University, PA, USA, 2003. [28] A. Okamoto, R. Hatakenaka, Y. Mase, et al., Visualization of a Loop Heat Pipe Using Neutron Radiography, AIAA Paper No.2010-6031, 2010.