University of Nevada, Reno
Modeling of Heat Transfer and Flow Patterns in a Porous Wick of a Mechanically Pumped Loop Heat Pipe: Parametric Study Using ANSYS Fluent
A thesis submitted in partial fulfillment of the requirements for the degree of
Master of Science in Mechanical Engineering
by Md Shujan Ali
Dr. Miles Greiner, Thesis Advisor
May, 2017
THE GRADUATE SCHOOL
We recommend that the thesis prepared under our supervision by
MD SHUJAN ALI
Entitled
Modeling of Heat Transfer And Flow Patterns In A Porous Wick of A Mechanically Pumped Loop Heat Pipe: Parametric Study Using ANSYS Fluent
be accepted in partial fulfillment of the requirements for the degree of
MASTER OF SCIENCE
Miles Greiner, Ph. D., Advisor
Mustafa Hadj-Nacer, Ph. D., Co-advisor and Committee Member
Sage R. Hiibel, Ph. D., Graduate School Representative
David W. Zeh, Ph.D., Dean, Graduate School
May, 2017 i
ABSTRACT
In recent years, NASA space exploration has achieved new feats due to advancement in aerodynamics, propulsion, and other related technologies. Future missions, including but not limited to manned mission to Mars, deep space exploratory missions, and orbit transfer vehicles, require advanced thermal management system. Current state-of-the-art for spacecrafts is a mechanically pumped single phase cooling loop that are not enough to meet thermal-related challenges for future space missions. Loop heat pipes (LHP) are the solution for the required thermal management system that is compact, light-weight, reliable, precise, and energy efficient. These are two-phase systems that employ capillary forces instead of pumps to circulate the coolant. In these devices, the coolant evaporates and condenses in the evaporator and condenser, respectively. The condensed coolant liquid is driven toward the evaporator by capillary action in a wick structure located inside the evaporator. A mechanical pump is added to the liquid line of the loop to reach the distributed heat loads while controlling the temperature to produce an isothermal surface. In this work, flow patterns and heat transfer in the LHP evaporator wick is studied for various flow rates of the working fluid, wick thermal conductivity, porosity and permeability of wick, heat flux, and gravity condition. A CFD model has been developed to predict the performance of LHP due to the change in these parameters. The Volume of Fluid (VOF) model in ANSYS Fluent was modified using a User Defined Function (UDF) to calculate mass transfer between the liquid and vapor phases at the interface. The Lee phase change model was used to calculate the mass flux due to evaporation and condensation.
ii
Acknowledgement
First of all, I would like to express my grateful to the creator ALLAH s.w.t. for the blessing given to me to complete my thesis.
I thank my advisor Dr. Miles Greiner for his great supports, advice, guidance and encouragement. I would like to give special thanks to my supervisor Dr. Mustafa Hadj-Nacer for his continuous supports and mentorship. He guided me in every step of my works. I would also like to say thanks to my committee member Dr. Sage R. Hiibel.
I would say thanks to my colleague Nishan Pandey, who helped me write the user defined function, and colleague Corey with whom I shared interesting discussions several times about my research.
I would acknowledge my family for their supports. Especially, I want to thank my wife for her continuous supports, understanding, and trust.
Finally, I would say thanks to NASA EPSCOR Nevada to provide necessary fund for this research project.
iii
Table of Contents 1 Introduction ...... 1 2 Literature review on Loop Heat Pipe ...... 3 2.1 LHP operating Principles ...... 3 2.1.1 Classification of capillary two-phase loops ...... 5 2.1.2 Classification of LHP ...... 6 2.1.3 Loop Heat Pipes with flat evaporators ...... 6 2.1.4 Mechanically Pumped LHP ...... 8 2.1.5 The thermodynamic cycle of an LHP ...... 9 2.1.6 LHP serviceability ...... 10 2.1.7 Steady-state operating performance ...... 12 2.2 LHP parametric study ...... 13 2.2.1 Effect of fluid charge ...... 13 2.2.2 Effect of the porous wick characteristics ...... 13 2.2.3 Effect of gravity ...... 15 3 CFD Model ...... 17 3.1 Governing Equations ...... 17 3.1.1 Porous media model ...... 18 3.1.2 Volume of fluid model ...... 20 3.1.3 Evaporation-Condensation model ...... 22 3.2 Geometry and boundary conditions ...... 25 3.3 Mesh ...... 26 3.4 Solver Settings ...... 27 4 Results and Discussion ...... 29 4.1 Benchmark problems ...... 29 4.1.1 One-Dimensional Stefan problem ...... 30 4.2 Liquid mass flow rate at inlet ...... 32 4.3 Permeability ...... 33 4.4 Porosity ...... 35 4.5 Heat flux...... 36 4.6 Thermal conductivity ...... 39 4.7 Gravity ...... 41 5 Conclusions ...... 43 BIBLIOGRAPHY ...... 44
iv
List of Tables
Table 2.1: Classification of Loop Heat Pipe (Maydanik [7]) ...... 6 Table 3.1. Construction of mass and energy sources...... 24 Table 3.2. Results for mesh sensitivity test...... 27 Table 4.1. Base case parameters ...... 29 Table 4.2. Properties of working fluid ...... 29 Table 4.3. Relation between vapor content and Pressure Difference...... 38
v
List of figures
Figure 2.1. Flow Schematic of an LHP (Ku [6])...... 3 Figure 2.2. Schematic of (a) a traditional LHP and (b) a traditional CPL (Butler et al. [13].) ...... 5 Figure 2.3. Schematic of (a) active zone of an evaporator with opposite replenishment, (b) active zone of an evaporator with longitudinal replenishment, (c) an evaporator with opposite replenishment, and (d) an evaporator with longitudinal replenishment ( Maydanik [8]) ...... 7 Figure 2.4. Schematic of a Pumped LHP (Jiang et al. (2014) [52])...... 8 Figure 2.5. Diagram of: (a) the LHP working cycle; (b) location of the characteristics points in the LHP (Maydanik [7])...... 9 Figure 2.6. Typical LHP operating curves (Launay et al [24])...... 12 Figure 3.1. Flow chart summarizing the numerical methodology...... 17 Figure 3.2. Mechanically Pumped Loop Heat Pipe (LHP)...... 25 Figure 3.3. Schematic of the numerical model used in ANSYS/FLUENT to simulate heat transfer in a section of the evaporator wick with the boundary conditions...... 25 Figure 3.4. Schematic of the mesh of the computational domain...... 26 Figure 4.1. One-dimensional Stefan problem: (a) Schematic of the problem; (b) Temperature and steam volume fraction contours from our model...... 30 Figure 4.2. Result for 1D Stefan problem from Dongliang Sun et al. (2014) ...... 31
Figure 4.5. Liquid-vapor fronts as a function of liquid mass flow rate 풎̇ 풊풏...... 32 Figure 4.6. Temperature and Pressure difference in wick as functions of liquid mass flow rate 풎̇ 풊풏...... 33 Figure 4.7. Liquid-vapor fronts as a function of Permeability...... 34 Figure 4.8. Temperature and Pressure difference in wick as functions of Permeability. 34 Figure 4.9. Liquid-vapor fronts as a function of Porosity...... 35 Figure 4.10. Temperature and Pressure difference in wick as function of porosity...... 35 Figure 4.11. Liquid-vapor fronts as a function of Heat Flux...... 37 Figure 4.12. Temperature and Pressure difference in wick as functions of Heat flux. .... 38 Figure 4.13. Temperature contour for: (a) Titanium wick; (b) Steel wick; (c) Aluminum wick; (d) Copper wick...... 40 Figure 4.14. Temperature and Pressure difference in wick as functions of material thermal conductivity...... 40 Figure 4.15. Temperature field for different value of gravitational acceleration...... 41 Figure 4.16. Pressure field for different value of gravitational acceleration...... 41 Figure 4.17. Temperature and Pressure difference in wick as functions of the value of the gravitational acceleration...... 41 1
1 Introduction
Thermal control of life-support, power electronics, and other systems is a key enabler for a variety of NASA space, air-flight, and monitoring technologies. It is essential to develop new light-weight, energy efficient, reliable and compact heat transfer technologies for use on near-earth satellites. An example of a system that requires such advanced thermal management devices is the Surface Water Ocean Topography (SWOT) project, with a baseline mission launch in 2020. The primary SWOT payload will be the Ka-band Rader
Interferometer (KaRIn). The allowable KaRIn temperature rate of change is 20 times smaller than typical spacecraft instruments and must be achieved without isolation. NASA thermal control systems typically employ liquid coolant loops that remove heat from heat-generating components and transport it to cooler modules, which either utilize the heat or reject it to space. NASA currently relies on single-phase coolant loops (in which the liquid does not evaporate as it absorbs heat). Loop heat pipes (LHPs), which are passive two-phase cooling systems, are a good alternative compared to single-phase system to reduce the size, weight, temperature variations, and pumping power requirements of thermal control systems, without loss of reliability.
In LHPs, heat is removed from high temperature external components by the device’s evaporator section (that contains a porous wick to deliver cooling liquid to surfaces where it is vaporized). The vapor is transported to a condenser section where heat is absorbed by a cooler component or rejected to space. This condenses the vapor and the resulting liquid returns to the evaporator. Capillary pressure that develops in the evaporator’s wick drives the liquid and vapor flows. LHPs are capable to transfer the removed heat over long distances with minimal temperature gradient between the evaporator and the condenser, and they also provide efficient spatial and temporal temperature stability due to use of latent heat [58, 59, 49-57]. 2
To increase the controllability and reliability of thermal management systems, NASA is considering system in which a pump is placed in the liquid line [58, 59]. The added mechanical pump facilitates distribution of working fluid to discrete heat loads [58], decreases temperature oscillations, increases the heat transport distance, and also minimizes the heat leak from the evaporator to the compensation chamber [52]. Pumped two-phased loops (P2φL) are currently used in some military, aerospace, and electronics cooling applications. One of the objectives of the current work is to apply them to low-gravity space applications.
In a P2φL system, thermal and hydrodynamic mechanisms between its various components are strongly coupled in a complex manner. Pressure dynamics and temperature instabilities, are sometimes found experimentally after changes in operational conditions.
These instabilities can lead to various types of failure, like evaporator wick dry-out, temperature oscillations, inferior performance, which should be avoided. The performance of
LHP is affected by working fluid, liquid flow rate, wick material thermal conductivity, porosity and permeability of the wick, and other parameters. In this work, a parametric study is conducted using ANSYS/Fluent computational fluid dynamics (CFD) code to predict the performance of a section of the evaporator wick for a range of parameters.
3
2 Literature review on Loop Heat Pipe
This chapter presents an extensive review about Loop Heat Pipes (LHP). First LHP operating principles are described. Then different components of LHP and their contribution to the overall system operation are studied. An emphasis is put on the effect of a single operating parameter or design constraint on the overall LHP performance.
2.1 LHP Operating Principles
Figure 2.1. Flow Schematic of an LHP (Ku [6]).
Flow schematic of an LHP is shown in Fig 2.1. It consists of an evaporator, a compensation chamber, and vapor and liquid transport lines. Wicks are present only in the evaporator and compensation chamber, while the rest of the loop is made of smooth wall tubing. The evaporator contains wick with fine pores so that it can develop a capillary pressure to circulate the working fluid around the loop. The compensation chamber has wick with larger pores for the purpose of managing fluid ingress and egress. Heat is applied to the evaporator, as a consequence liquid is vaporized and menisci are formed in the evaporator wick, which develops the capillary forces to push the vapor through the vapor line to the condenser. Due to finite thermal resistance of the wick, vapor temperature and pressure in the vapor grooves, which are in contact which the heated evaporator wall, become higher than 4
temperature and pressure in the compensation chamber. Thus, the wick serves as a “thermal lock”. At the same time, hotter vapor cannot penetrate into the compensation chamber through the saturated wick because of the capillary forces which hold the liquid in it. In that way, the wick acts as a “hydraulic lock”. In the condenser, vapor is condensed, and the capillary forces continue to push liquid back to the evaporator. Here no external power is required because the heat source that needs to be cooled provides the driving force to circulate the working fluid. The two-phase compensation chamber stores surplus liquid and controls the operating temperature of the loop. The total pressure drop of the system must not exceed the maximum capillary pressure that the wick in the evaporator can develop. The meniscus in the liquid-vapor interface tunes its radius of curvature to match the capillary pressure generated by the wick to the total pressure drop of the system. Frictional pressure drops in the evaporator grooves, the vapor line, the condenser, the liquid line, the evaporator wick, and hydrostatic pressure imposed by gravity constitute the total pressure drop across the system.
∆푃푡표푡 = ∆푃푔푟표표푣푒 + ∆푃푣푎푝 + ∆푃푐표푛 + ∆푃푙푖푞 + ∆푃푤 + ∆푃푔 (1)
The capillary pressure rise that the wick can develop is given by
∆푃푐푎푝 = 2𝜎 cos 휃/푅 (2) where 𝜎 is the surface tension of the working fluid, R is the radius of curvature of the meniscus in the liquid-vapor interface, and 휃 is the contact angle between the liquid and the wick. The mass flow rate and total pressure drop in the system increase when the heat load is increased. Then the radius of curvature of the meniscus decreases so that capillary pressure generated in the wick match to the total pressure drop in the system. The radius of curvature of the liquid-vapor interface can decrease in response to increased heat load up to the pore radius of the wick 푅푝. The wick reaches its maximum capillary pumping capability at that point:
∆푃푐푎푝, 푚푎푥 = 2𝜎 cos 휃/푅푝 (3) 5
Vapor penetrates through the wick and the system fails when the heat load is further increased.
2.1.1 Classification of capillary two-phase loops
(a) (b) Figure 2.2. Schematic of (a) a traditional LHP and (b) a traditional CPL (Butler et al. [13].)
Capillary Pumped Loops (CPL) and Loop Heat Pipes (LHPs) are two types of capillary two-phase loops that were developed about the same time. Figure 2.2 schematically shows that both loops consist of an evaporator, a condenser, a vapor line, a liquid line, and a hydro- accumulator. Both loops share many similarities in operating principles and performance characteristics. A significant difference between the two loops is the thermo-hydraulic link between the evaporator and the hydro-accumulator. The compensation chamber of an LHP is made as an integral part of the evaporator. It is connected to the evaporator by a secondary wick, and is located directly in the path of the liquid line. The reservoir of a CPL is placed distantly from the evaporator and is outside the route of fluid circulation. The reservoir usually has no thermal connection with the evaporator, and it needs to be heated before start- up to collapse any vapor bubbles in the evaporator core and to ensure that the wick is wetted.
The LHP technology is rapidly gaining recognition in the aerospace community. LHPs use 6
metal wicks that have fine pores, and are acknowledged for their high pumping capability and robust operation.
2.1.2 Classification of LHP
LHPs differ based on their application and geometry. Although both LHPs and CPLs were developed as satellite thermal management units, they have many other applications nowadays such as electronics cooling, solar water heating, cryogenic telescope cooling.
Maydanik [7] suggests the classification presented in Table 2.1 to categorize the LHP designs that can be found in the literature and the industry.
Table 2.1: Classification of Loop Heat Pipe (Maydanik [7]) LHP design LHP dimensions Evaporator shape Evaporator design Cylindrical Miniature Cylindrical One butt-end (diode) All the rest Flat-disk shaped compensation Conventional Flat rectangular chambers Reversible Two butt-end Flexible compensation Ramified chambers Coaxial Condenser design Number of Temperature range Operating evaporators & temperature control condensers Pipe-in-pipe One Cryogenic With active Flat coil Two or more Low-temperature control Collector High- Without active temperature control
2.1.3 Loop Heat Pipes with flat evaporators
LHP with a flat evaporator is more suitable for heat removal since most of the heat sources or devices that require cooling have a flat contact surface. Existence of a flat surface itself is one of the main benefits of flat evaporators, it does not require any additional
“cylinder-plane” reducers for thermal contact with flat heat dissipating devices. This decreases the thermal resistance and the weight of evaporators. The flat surface is highly deformative to the vapor pressure in LHPs, and this is the main drawback of LHPS with flat evaporators. The vapor pressure should not exceed the ambient pressure. This situation 7
requires an increase in the wall thickness, or limits the choice of working fluids, or restricts the operating temperatures range.
(a) (b)
(c) (d) Figure 2.3. Schematic of (a) active zone of an evaporator with opposite replenishment, (b) active zone of an evaporator with longitudinal replenishment, (c) an evaporator with opposite replenishment, and (d) an evaporator with longitudinal replenishment ( Maydanik [8])
At present, there are mainly two types of flat evaporators: “evaporators with opposite replenishment” (EOR), and “evaporators with longitudinal replenishment” (ELR) [8]. Fig 2.3 shows their schematic. In EOR, liquid flows towards the heated surface of the evaporator, and the compensation chamber is located on the opposite side of the heated surface. In ELR, the main liquid flows in the direction along the evaporator heated wall, and the compensation chamber is located aside from the active zone. The thickness of such evaporators may be reduced to 3-8 mm. At the same time, in consequence of the decrease of thickness and the lateral location of the compensation chamber, the length or width of such evaporators increases significantly. 8
2.1.4 Mechanically Pumped LHP
Figure 2.4. Schematic of a pumped LHP (Jiang et al. (2014) [52]).
Many researchers have experimentally studied mechanically pumped capillary phase changed loop [52-58, 64]. The use of a mechanical pump significantly increases the capillary limits of the passive capillary driven systems, although the total pressure head from the capillary (passive) and mechanical pumping (active) must be still greater than overall pressure drop in the system [54]. Jiang et al. proposed a novel pump assisted capillary phase change loop. As shown in Fig 2.4 their system loop consists of an evaporator, a mechanical pump, a reservoir, an ejector, and a condenser. In pump assisted phase change loops, working fluid is circulated by both capillary and mechanical pumping. Thus, added pump facilitates circulation of working fluid to reach it at discrete heat loads. The evaporator sometimes can handle the supply of sub-cooled liquid, and in this case, both sensitive heat and latent heat of the working fluid is utilized to transfer heat [52]. Mechanical pump forces liquid to circulate in the loop, thereby providing liquid for wick boiling, a liquid is also flowing through the compensation chamber, and thus heat leak from the evaporator to the compensation chamber is taken away [52, 54]. The mechanical pumping in the loop significantly increases heat flux dissipation, heat transfer distance, and also decreases the fluctuation of operating temperature 9
[52-58, 64]. When there is no separate lane for liquid and vapor phase liquid, and vapor mixes at the end of the evaporator (no bypass line is added to the compensation chamber), an enormous pressure drop is observed through the evaporator wick and the evaporator (vapor grooves) is always flooded with water [54, 56-57]. Jiang et al. [52] used two lines for liquid and vapor in their mechanically pumped phase change loop. This arrangement helped to prevent a large pressure drop across the wick, but still they observed that the evaporator is flooded with liquid when the pressure in the vapor side is lower than that of the liquid side.
2.1.5 The thermodynamic cycle of an LHP
(a) (b) Figure 2.5. Diagram of: (a) the LHP working cycle; (b) location of the characteristics points in the LHP (Maydanik [7]).
The typical Pressure-Temperature diagram of the thermodynamic cycle of an LHP in steady-state operation is illustrated in Fig. 2.5a, and Fig 2.5b. These figures show the location of the characteristic points. In the evaporator, the liquid evaporates at the liquid-vapor interface (1) at temperature 푇1 (푇푔푟표표푣푒). It is often supposed that the evaporation happens at the wick-groove interface. However, several authors reported the transposition of the vapor front inside the porous wick. The section 1–2 represents flow of vapor through the vapor- 10
removal channels into the vapor line. Here the vapor experiences a pressure drop along with a slight superheat because it proceeds along the heated surface of the evaporator. The vapor motion in the vapor line (section 2–3) can be considered as isothermal. In the LHP condenser
Pressure losses are usually negligible. Section 3-4 represents condensation of the working fluid in the condenser, and section 4-5 represents super-cooling of the condensed liquid inside the condenser. Further its motion through the liquid line in the diagram is shown as isothermal, though in many actual cases it may be accompanied by considerable heating or cooling in consequence of the heat transfer with the ambient. The liquid arrives the compensation chamber with parameters 푇6, 푃6. Due to heat leak from the evaporator to the compensation chamber liquid temperature at the compensation chamber rises to 푇7 (line 6-7).
Section 7-8 denotes the liquid filtration through the wick into the evaporation zone. On this way the liquid may prove to be superheated, but it does not boil because it exists in that state for a short period of time. The point 8 defines the state of the working fluid adjacent to the evaporating menisci, and the pressure drop ∆푃1−8 resembles the total pressure losses in all the sections of the loop.
2.1.6 LHP serviceability
The first condition of LHP serviceability is determined in the same way as for conventional heat pipes [7] :
∆푃푐 ≥ ∆푃푣 + ∆푃푙 + ∆푃푔 (4)
∆푃푐 is the capillary pressure created in the wick; ∆푃푣, ∆푃푙 are pressure drops due to the motion of the working fluid in the vapor and the liquid phase; ∆푃푔 are hydrostatic pressure of a liquid column. The values of ∆푃푣 and ∆푃푙, can be evaluated by use of the established equations [1-4] for the motion of vapor and liquid in different LHP sections, including the capillary structure.
The value of ∆푃푔may be determined by the formula: 11
∆푃푔 = (𝜌푙 − 𝜌푣)푔푙 sin 휑 (5) where 𝜌푙 and 𝜌푣 are the liquid and the vapor densities respectively; g is the gravitational acceleration; 푙 is the LHP effective length; 휑 is the LHP slope with the horizontal. During operation against gravity, if the value of 푙 is sufficiently large, and the value of 휑 is close or
o equal to 90 , the hydrostatic resistance ∆푃푔become prevalent. In such situations one may consider to use a fine-pored wick, or add a mechanical pump in the liquid line.
The second requirement for serviceability is the obligation of creating an adequate amount of temperature and pressure drop between the evaporating surface of the wick and the compensation chamber. This pressure drop is equivalent to ∆푃퐸푋, which is the sum of pressure losses due to working fluid circulation in all the sections except the wick. This condition, with the use of the terms in Fig. 2.5, may be written as follows:
휕푃 ∆푇1−7 = ∆푃퐸푋 (6) 휕푇푇̅̅푣̅
푑푃 where is the derivative determined by the slope of the saturation line at the point with 푑푇 temperature 푇̅푣, average between 푇1 and 푇7.
∆푇1−7 is the motive temperature head required for pushing the liquid into the compensation chamber during an LHP startup. The diagram (Fig. 2.5) also illustrates that the working fluid should be sufficiently super-cooled for avoiding its boiling-up in the liquid line as an outcome of pressure losses and heating due to some external heat inflows. The value of super-cooling may be determined as:
휕푃 ⌉ ∆푇 = ∆푃 (7) 휕푇 4−6 5−6 푇푣 This is the third condition of LHP serviceability. If the liquid line contains some vapor bubbles, they can freely breach into the compensation chamber. They do not interfere with the normal operation of the device, but they cause an increase in the operating temperature. 12
2.1.7 Steady-state operating performance
Steady-state operating performance of an LHP is generally measured by expressing its temperature or thermal resistance as a function of the heat load. The characteristic temperature of an LHP is frequently defined as the vapor groove temperature 푇푣 [27], but it can also be defined by evaporator temperature, 푇푒 [27], or even by the reservoir temperature
푇푟 [24]. In electronic cooling the LHP performance is mainly measured by 푇푒, because the system requirement often concerns the utmost allowable temperature of the electronic device.
Figure 2.6. Typical LHP operating curves (Launay et al [24]).
Two different LHP operating modes are reported in the literature, the variable conductance mode (VCM) and the fixed conductance mode (FCM) (Fig. 2.6). In LHP VCM operation, the part of the condenser where liquid is sub-cooled is large enough, the liquid leaves the condenser at a temperature near to the heat sink temperature, and the operating curve has a U-shape [24]. In FCM the condenser is almost filled with two-phase fluid and the liquid enters the liquid line at a temperature near to the saturation. An LHP operates in FCM for higher heat loads, and operating temperature increases quasi-linearly with the heat flux. 13
Launay and Vall´ee [28] state that according to experimental results, the prediction of the operating curve shape is difficult due to the fact that they are greatly influenced by heat transfer in the evaporator, and therefore delicate to the evaporator design and thermal characteristics.
2.2 LHP Parametric Study
The LHP is a complex system, and a parametric study on the LHP operation is difficult due to the strongly coupled physical mechanisms involved in LHPs.
2.2.1 Effect of fluid charge
The compensation chamber volume, 푉푐푐, must be able to accommodate at least the liquid volume fluctuation (and density changes) between the hot case and the cold case of the loop operation [5,6]. Lee et al. [39] found the best heat transfer performance at 51.3% volume filling ratio based on experimental study at a horizontal position. The amount of fluid charge, and the fluid distribution greatly affect startup behavior of an LHP.
2.2.2 Effect of the porous wick characteristics
The porous wick characteristics, such as the effective thermal conductivity, the pore diameter, the porosity, and the permeability, all may have a momentous effect on the LHP performance [5].
2.2.2.1 Wick material
Plastic, metal, ceramic, as well as metal foams are suitable materials for the LHP porous structure. Plastic wicks, such as polypropylene, PTFE, polyethylene have a very low thermal conductivity. These wicks reduce the heat leak form the evaporator to the compensation chamber, thus improving the LHP performances. However, they have low value of porosity
(about 50 %). They are suitable for low to moderate heat loads due to constraint in operating temperatures. Ceramic wicks are very stable in thermal and chemical environments, 14
compatible with metals and their microstructure has a wide variety of size and shapes. Most of the experimental studies are on the performance of metallic wicks. Nickel is the broadly used material for metal wicks, but stainless steel, titanium, brass and iron oxide are also used as wick materials. Sintered powder is the wide spread manufacturing process, providing high porosity and low pore size characteristics.
2.2.2.2 Wick thermal conductivity
The thermal conductivity, 푘푒푓푓, regulates the radial heat leak and therefore, the LHP operating temperature. Mishkinis et al. [42] found that the radial heat leak through the wick is proportional to the effective thermal conductivity for low Peclet numbers, which is the ratio of the convective to the conductive heat transfers. The effective thermal conductivity of the wick is not a prevailing factor for high Peclet numbers. According to Figus et al. [43], the wick thermal conductivity does not significantly alter the position of the liquid-vapor interface, but greatly affects the fin superheat, the temperature distribution in the wick, and the axial heat leak.
2.2.2.3 Wick permeability
The capillary and boiling limits depend on wick permeability. The heat transfer coefficient is strongly linked to the location of the liquid-vapor interface in the wick. Figus et al. [43] numerically studied phase change and flow patterns in a porous wick for various parameters including wick permeability. The Authors found that the use of a relatively high permeable wick ease the vapor flow. Generally wick capillary pressure decreases with increase of permeability which is not desirable owing to the fact that capillary pressure head is the driving force. In this case the development of a layered wick can be a potential solution.
In [44] a system having a two layered wick was studied where the 1st layer placed near the fin possesses relatively high permeability and thermal conductivity, while another layer has comparatively smaller pores and a low thermal conductivity. 15
2.2.2.4 Wick porosity and pore size
Boo and Chung [47] have experimentally studied a number of polypropylene wicks, whose pore diameters varied from 0.5 to 25 µm. The authors found that when the pore size was changed to 0.5 µm from 25 µm, there was a 45% increase in maximum thermal load
(푄푚푎푥) and a 33% decrease in the evaporator wick thermal resistance (푅푡ℎ). When the LHP operates at nearly the capillary limit one should use wicks having small pores. In other cases, one should use an optimum pore size so that the wick permeability does not impair the evaporator heat transfer coefficient. Hamdan et al. [37] studied the effect of the wick characteristics on the LHP operating limits by means of numerical simulation for a coherent porous silicon wick. They concluded that reducing the pore size 푅푝 increases the capillary
2 pressure as 1/푅푝 but simultaneously the pressure losses due to friction increases as Q/푅푝 . So, there exists an optimum value of the pressure build-up through the wick that depends on the heat load. The nucleate boiling limit is associated with the liquid superheat. A reduction of the pore radius increases the boiling limit because the superheat required for nucleation is inversely proportional to the pore radius.
2.2.3 Effect of gravity
Gravity has significant effects in case of LHP elevation and tilt. Elevation indicates the location of the evaporator with respect to the condenser, and tilt denotes location of the evaporator compared that of the compensation chamber. At an adverse elevation the evaporator is situated above the condenser, and at an adverse tilt the evaporator is situated above the compensation chamber. In 1 gravity conditions, fluid distribution inside the loop, especially between the evaporator core and compensation chamber alters with LHP tilt, and the LHP operating temperature is increased. At low power applications the LHP operating temperatures at adverse tilts are greater than those at positive tilts [50]. At low power applications, adverse elevation increases LHP operating temperature, while positive elevation 16
decreases LHP operating temperature. At larger heat flux the elevation generally has no effect on the LHP operating characteristics. 17
3 CFD Model
ANSYS/Fluent computational fluid dynamics (CFD) code is used to model a section of the evaporator wick for a range of parameters. The porous media model and the volume of fluid (VOF) multiphase model are used in our calculation. A user defined function based on
Lee phase change model is incorporated in VOF model to calculate evaporation-condensation mass transfer at the liquid-vapor interface, and also to track the interface. The main assumptions of this CFD model are: (1) the porous wick is isotropic; (2) liquid and vapor both are in thermal equilibrium in wick, grooves, and transport lines; (3) fluid flow through porous wick is laminar; (4) both vapor & liquid phases are incompressible. This chapter describes all governing equations, discretization and iteration schemes. Figure 3.1 shows overview of the numerical methodology.
Figure 3.1. Flow chart summarizing the numerical methodology. 3.1 Governing Equations
Continuity, momentum and energy equations solves velocity components, pressure and temperature. The continuity equation is defined as [59]:
휕휌 + ∇. (𝜌푣⃑) = 푆 (3.1) 휕푡 푚 18
where 푆푚 is the mass source. It represents any type of mass addition in the system for the fluid or phase for which the continuity equation is solved, and it also included any user- defined mass source.
The momentum equation, also known as Navier-Stokes Equation, is defined as [59]:
휕 (𝜌푣⃑) + ∇. (𝜌푣⃑푣⃑) = −∇푝 + ∇. (휏̿) + 𝜌푔⃑ + 퐹⃑ (3.2) 휕푡
where 푝 is the static pressure, 휏̿ is the stress tensor, 𝜌푔⃑ is gravitational body force, and
퐹⃑ is the external body force. 퐹⃑ accommodates any source term including user defined source terms and model specific source terms.
The energy equation is defined as [59]:
휕 (𝜌퐸) + ∇. (푣⃑(𝜌퐸 + 푝)) = ∇. (퐾 ∇푇 − ∑ ℎ 퐽⃑⃑⃑ + (휏̿ . 푣⃑)) + 푆 (3.3) 휕푡 푒푓푓 푗 푗 푗 푒푓푓 ℎ
where 퐾푒푓푓 is the effective conductivity, 퐽⃑⃑푗⃑ is the diffusion flux for species 퐽, and 푆ℎ is any volumetric source term including user-defined energy source.
Equations (3.1), (3.2) and (3.3) are basic governing equations for any CFD model. They are modified for incorporation of other models. In this study these equations were modified for incorporation of porous media, multiphase flow, and evaporation-condensation phase change.
3.1.1 Porous media model
Different types of single phase and multiphase problems involving porous media can be simulated using the ANSYS Fluent porous media model. Porous media conditions calculate the pressure loss in the cell zone by solving the momentum and energy equations for porous media [60]. 19
3.1.1.1 Momentum equation for porous media
ANSYS Fluent porous media model use the Superficial Velocity Porous Formulation.
The superficial phase or mixture velocities are calculated based on the volumetric flow rate in a porous region. Standard fluid flow equations are modified by addition of a momentum source term [60].
1 푆 = − (∑3 퐷 휇푣 + ∑3 퐶 𝜌|푣|푣 ) (3.4) 푖 푗=1 푖푗 푗 푗=3 푖푗 2 푗
Here 푆푖 is the source term for the 𝑖th (x, y, or z) momentum equation, |푣| is the velocity magnitude, and D and C are prescribed matrices. In equation (3.4), the first term represents viscous loss and the second term represents inertial loss. Momentum sink 푆푖 generates a pressure drop in the porous cell that is proportional to the fluid velocity in the cell [60].
3.1.1.1.1 Darcy’s law in porous media
In laminar flow, convective acceleration and diffusion can be ignored, and the matrix 퐶푖푗 can be considered zero. Then the porous media model reduces to Darcy’s Law [60]:
휇 ∇푝 = − 푣⃑ (3.5) 훼 where 훼 is the volume fraction, and 휇 is the viscosity of the fluid.
3.1.1.1.2 Inertial losses in porous media
In high flow velocities, the viscous loss term in Equation (3.4) can be ignored, and the matrix C gives a correction for inertial losses in the porous medium. Here the constant C represents the pressure drop as a function of dynamic head of the flow [60]. As an instant, in case of perforated plate or tube bank model, the viscous loss term is negligible and the equation for porous media model reduces to [60]:
1 ∇푝 = − ∑3 퐶 ( 𝜌푣 |푣|) (3.6) 푗=1 푖푗 2 푗 20
3.1.1.2 Energy equation for porous media
For thermal equilibrium between the porous media and the fluid flow, the energy equation becomes [60]:
휕 (훾𝜌 퐸 (1 − 훾)𝜌 퐸 ) + ∇. (푣⃑(𝜌 퐸 + 푝)) = 푆ℎ + ∇. [퐾 ∇푇 − (∑ ℎ 퐽 ) + (휏̿. 푣⃑)] (3.7) 휕푡 푓 푓+ 푠 푠 푓 푓 푓 푒푓푓 푖 푖 푖
Where,
퐸푓 = total energy of fluid
퐸푠 = total energy of solid medium
𝜌푓 = density of fluid
𝜌푠 = density of solid medium
훾 = porosity of the medium
퐾푒푓푓= effective thermal conductivity of the medium
ℎ S푓 = fluid enthalpy source term
ANSYS Fluent calculates the effective thermal conductivity of the porous medium as the volume average of the fluid conductivity and solid conductivity [60]:
퐾푒푓푓 = 훾퐾푓 + (1 − 훾)퐾푠 (3.8)
where,
퐾푓= thermal conductivity of fluid
퐾푠= thermal conductivity of solid
Equation (3.8) is update of equation (3.3) due to porous media consideration.
3.1.2 Volume of fluid model
Two or more immiscible fluids are be modeled by employing the Volume of fluid (VOF) model. This model based on the fact that fluids are not interpenetrating. For each phase the model introduces its volume fraction in the computational cell, and the summation of volume fraction of all phases is unity [59]. The calculation involves a single momentum for the 21
mixture phase, and the resulting velocity field is shared among the phases. The energy equation is also solved for the mixture phase and is shared among the phases [59]. VOF model is not applicable if interface length is small compared to a computational grid.
Accuracy of VOF model decreases when the interface dimension is close to the computational grid scale [59, 61].
3.1.2.1 Volume fraction equation
The VOF tracks the interface(s) by solving a continuity equation for the volume fraction of one (or more) of the phases. For the 푞푡ℎ phase, this volume fraction equation is [59]:
1 휕 푛 [ (훼푞𝜌푞) + ∇. (훼푞𝜌푞푣⃑⃑⃑⃑푞⃑) = 푆훼 + ∑푝=1(푚̇ 푝푞 − 푚̇ 푞푝)] (3.9) 휌푞 휕푡 푞 where 푚̇ 푝푞 is the mass transfer from phase q to phase p, and 푚̇ 푞푝 is the mass transfer from
푡ℎ phase p to phase q. 푆훼푞 is the mass source for 푞 phase. The volume fraction equation is solved for the secondary phases and the primary phase volume fraction is calculated based on the following fact [59]:
푛 ∑푞=1 훼푞 = 1 (3.10)
Due to implication of VOF multiphase model equation (3.10) and (3.11) would be used in calculation instead of continuity equation (3.1). The volume fraction equation can be solved either through implicit or explicit formulation.
3.1.2.1.1 The implicit formulation of volume fraction equation
For implicit formulation volume fraction equation is discretized as follows:
훼푛+1휌푛+1−훼푛휌푛 푞 푞 푞 푞 푉 + ∑ (𝜌푛+1푈푛+1훼푛+1) = [푆 + ∑푛 (푚̇ − 푚̇ )] 푉 (3.11) ∆푡 푓 푞 푓 푞,푓 훼푞 푝=1 푝푞 푞푝
Where,
n+1= index for current time step
n= index for previous time step
푛+1 훼푞 =cell value of volume fraction at current time step 22
푛 훼푞 = cell value of volume fraction at previous time step
푛 th 훼푞,푓 = face value of the q volume fraction at current time step
푛+1 푈푓 = volume flux through the face at previous time step
푉 = cell volume
3.1.3 Evaporation-Condensation model
In our simulation, we considered evaporation and condensation simultaneously. Fluent can simulate evaporation process alone. But, It doesn’t have ability to simulate condensation process. To solve this problem, a user-defined function (UDF) has been compiled to complete the existing FLUENT code. This UDF calculate the mass and heat transfer in the liquid-vapor interface during the evaporation and condensation processes, calculated by the source terms in the governing equations.
3.1.3.1 User defined function
Table 3.1 shows the source terms used to calculate the mass and energy transfer. Mass sources, 푆푀 is plugged in the volume fraction equation and energy sources, 푆퐸 is plugged in the energy equation, where 푇푚푖푥 and 푇푠푎푡 are the mixture temperature and saturation temperature, respectively, and LH is the latent heat of vaporization.
Mass and energy sources in Table 3.1 have been employed in the UDF and connected to the governing equations in FLUENT. The volume fraction for each phase in the cell has been defined by the VOF model. Consequently, two mass sources are essential to calculate mass transfer in evaporation, eqn. (3.12) denotes subtraction of mass from the liquid phase and eqn.
(3.13) denotes addition of mass to the vapor phase. Mass transfer in condensation is evaluated in similar way, eqn. (3.14) and eqn. (3.15) represents the quantity of mass taken from vapor to liquid phase. A single energy source is needed for both phases to calculate heat transfer involved in the evaporation or condensation process. Latent heat for evaporation or condensation is multiplied with the mass source to calculate the energy source or heat transfer 23
for evaporation or condensation, eqn. (3.16) and eqn. (3.17), respectively. Mass sources described here are based on Lee phase change model.
24
Table 3.1. Construction of mass and energy sources Thermal Phase change Temperature Phase Source term energy process condition
Mass Evaporation 푇 > 푇 Liquid 푇푚푖푥−푇푠푎푡 푚푖푥 푠푎푡 푆푀 = −0.1𝜌퐿훼퐿 (3.12) transfer 푇푠푎푡
Vapor 푇푚푖푥−푇푠푎푡 푆푀 = +0.1𝜌퐿훼퐿 (3.13) 푇푠푎푡
Condensation 푇 < 푇 Liquid 푇푚푖푥−푇푠푎푡 푚푖푥 푠푎푡 푆푀 = +50𝜌퐿훼퐿 (3.14) 푇푠푎푡
Vapor 푇푚푖푥−푇푠푎푡 푆푀 = −50𝜌퐿훼퐿 (3.15) 푇푠푎푡
Heat Evaporation 푇 > 푇 Liquid 푇푚푖푥−푇푠푎푡 푚푖푥 푠푎푡 푆퐸 = −0.1𝜌퐿훼퐿 퐿퐻 (3.16) transfer 푇푠푎푡
Condensation 푇 < 푇 Vapor 푇푚푖푥−푇푠푎푡 푚푖푥 푠푎푡 푆퐸 = +50𝜌퐿훼퐿 퐿퐻 (3.17) 푇푠푎푡 3.1.3.2 Lee phase change model
The Lee model is a mechanistic model that has a physical basis [59]. In this model, the liquid-vapor mass transfer (evaporation and condensation) is based on the following vapor transport equation [59]:
휕 (훼 𝜌 ) + ∇. (훼 𝜌 푉⃑⃑ ) = 푚̇ − 푚̇ (3.18) 휕푡 푣 푣 푣 푣 푣 푙푣 푣푙
Where,
훼푣 = vapor volume fraction
𝜌푣 = vapor density
푉⃑⃑푣 = Vapor phase velocity
The mass transfer can be expressed as follows [59]:
If 푇푙 > 푇푠푎푡 (evaporation):
푇푙−푇푠푎푡 푚̇ 푙푣 = 푐표푒푓 푓 ∗ 훼푙𝜌푙 (3.19) 푇푠푎푡
If 푇푣 < 푇푠푎푡 (condensation):
푇푠푎푡−푇푣 푚̇ 푣푙 = 푐표푒푓 푓 ∗ 훼푣𝜌푣 (3.20) 푇푠푎푡 25
the coefficient 푐표푒푓 푓 is like a relaxation time and usually is finely tuned to match experimental data [59].
3.2 Geometry and boundary conditions
Figure 3.2. Mechanically Pumped Loop Heat Pipe (LHP).
(a) (b) Figure 3.3. Schematic of the numerical model used in ANSYS/FLUENT to simulate heat transfer in a section of the evaporator wick with the boundary conditions.
Pumped two phase loop (P2φL) is shown in Fig. 3.2, with the evaporator being the most significant and sophisticated component. Figure 3.3a shows a cross sectional view of the evaporator. Figure 3.3b shows the two dimensional model of a section of the wick that is 26
considered for simulations in this work. Symmetry boundary conditions are employed on the left and right sides of the small section to simplify the problem. The evaporator wick is modeled as a 1×1 cm2 domain. Initially, at t=0 the section of the wick is filled with liquid phase (water). At the top boundary, the mass flow rate 푚̇ 푖푛, temperature 푇푖푛 and pressure 푃푖푛 of the working fluid (water) at saturation (푠=1) is specified. A heat flux (푄ℎ) is supplied to the porous wick from half of the bottom boundary that represents half of a pillar. The liquid flow evaporates at the heated boundary and the mixture of liquid and vapor (푠 ≤ 1) leaves the domain from the outlet boundary. It is desirable that the evaporator would be able to handle some amount of sub-cooled liquid at the wick inlet. For this reason, temperature at the wick
표 inlet 푇푖푛 = 95 퐶, is specified for all the simulation. So, heat transfer in the wick utilizes both sensitive heat and latent heat.
3.3 Mesh
Figure 3.4. Schematic of the mesh of the computational domain.
The mesh of the domain is shown in figure 3.4. Face meshing, edge sizing, and refinement was used to create a fine mesh. VOF model is sensitive to mesh size. It is not suitable when the interface length is smaller than the mesh size. Mesh sensitivity test was conducted. The result of mesh sensitivity test is given in table 3.2. The mesh with 40000 27
elements gives acceptable results. So, this mesh is used in all the simulations to reduce the computational costs. This mesh have 40401 nodes. The average orthogonal quality is 1 due to use of face mapped meshing, and the average skewness of the mesh elements is 1.3058.
Table 3.2. Results for mesh sensitivity test. No. of Temperature Pressure Max. Max. Vapor inside Continuity Elements at heated difference Vapor Velocity wick residual wall [o C] [Pa] VOF [m/s] [10-9 Kg] 5200 101.888 13185.9 0.3685 0.0007 3.349 1e-6
10000 101.886 23166.867 0.3685 0.0007 3.344 1e-6
22000 101.630 21550.127 0.3632 0.00074 2.994 1e-6
40000 101.487 20888.707 0.3575 0.00073 2.744 1e-6
90000 101.324 20012.918 0.3478 0.00072 2.409 1e-5
3.4 Solver Settings
A transient simulation with a time step of 0.01 second is carried out to model the dynamic behavior of the two-phase flow. The simulations reach a steady state after around 60-90 seconds. A combination of the SIMPLE algorithm for pressure-velocity coupling and Quick schemes for the determination of momentum and energy is employed in the model. Volume fraction interpolation is done by use of Modified HRIC discretization, and pressure interpolation is done by use of Body Force Weighted discretization scheme. Water vapor is defined as the primary (vapor) phase and water liquid is defined as the secondary (liquid) phase. For the calculation of the mass and heat transfer during the evaporation and condensation processes, a temperature of 100 표퐶 is used as the saturation temperature and the latent heat in the UDF code is 2455.134 KJ/Kg. When the simulation is started, the liquid 28
pool in the wick is heated first. Once the saturation temperature (100 표퐶) is reached, evaporation starts and phase change occurs. When the scaled residual of the mass is smaller than 10-4, scaled residual of the temperature is smaller than 10-12, and scaled residual of the velocity components is smaller than 10-6, the numerical computation is considered to have converged. Convergence behavior of the fin temperature, vapor volume in the wick, shape of liquid-vapor interface, and mass convergence are also noted the determine solution convergence. 29
4 Results and Discussion
THE CFD model developed in this work, was first used to simulate some benchmark problems to verify accuracy of our model. Next it was applied to other simulations in this study.
Table 4.1 shows the parameters of the wick and flow employed for the base case considered in this study. Based on this case, each of the following parameters; liquid flow rate, wick material thermal conductivity, porosity and permeability of the wick, heat flux, and orientation of the evaporator, will be changed once at a time to quantify its effect on the wick performance. Table 4.2 shows properties of working fluid at different phases.
Table 4.1. Base case parameters Material Steel Liquid mass flow rate 10-6 Kg/s Thermal conductivity 16 Wm-1K-1 Inclination (gravity) -9.81 m/s-2 Density of wick material 8030 Kg/m3 Heat flux 10 W/cm2 Porosity 0.4 Liquid permeability 1x10-13 m2
Table 4.2. Properties of working fluid Thermal cond. (water) 0.6 W/m-k Thermal cond. (vapor) 0.0261 W/m-k Density (water) 998.2 Kg/m3 Density (vapor) 0.5542 Kg/m3 Specific heat (water) 4182 J/Kg-k Specific heat (vapor) 2014 J/Kg-k Viscosity (water) 1.003x10-3 Kg/m-s Viscosity (vapor) 1.34x10-5 Kg/m-s
4.1 Benchmark problems
Our model was used to simulate Stefan problem to validate its accuracy.
30
4.1.1 One-Dimensional Stefan problem
(a)
Steam volume fraction
(b) Figure 4.1. One-dimensional Stefan problem: (a) Schematic of the problem; (b) Temperature and steam volume fraction contours from our model. 31
Figure 4.2. Result for 1D Stefan problem from Dongliang Sun et al. (2014) [62].
Stefan problem is a well-known benchmark problem for testing new models [62, 66].
Figure 4.1a shows schematic of one-dimensional Stefan problem, Fig. 4.1b shows temperature and steam volume fraction contours of this problem obtained from our model. In
3 our model, we use water as the working fluid. Liquid phase has density 𝜌푙= 998.2 Kg/m and
3 vapor phase (steam) has a density 𝜌푣= 0.01 Kg/m . Here, the saturation temperature is
표 표 100 퐶, latent heat ℎ푓푔= 2455.134 KJ/Kg, and the wall temperature is kept at 10 퐶 superheat. Dongliang Sun et al. [62] simulate similar problem with different values for the
3 3 parameters. In their model, liquid density 𝜌푙= 1 Kg/m , vapor density 𝜌푣= 0.1 Kg/m , latent
4 표 heat of vaporization ℎ푓푔= 10 J/Kg, and the wall was at 10 퐶 superheat. As illustrated in Fig.
4.1 and Fig. 4.2, numerical values of temperature and steam volume fraction at a particular cell is different for two models. But for both models the system behaves in a similar way. The temperature of vapor near the wall increases and becomes superheated. Mass transfer initiates in the interface, and at the liquid-vapor interface the temperature is equal to the saturation temperature [66].
32
4.2 Liquid mass flow rate at inlet
Effect of the inlet liquid flow rate on the evaporator performance is shown in Fig. 4.5. For the cases shown in this figure, the water flow rate was increased from 10-6 to 10-4 kg/s while the other parameters are kept constant. Water is forced through the top boundary of the wick.
No capillary pressure is considered in the simulations. For each of the flow rate considered, it can be seen that the liquid water near the heated surface evaporates, and a mixture of vapor and liquid leave through the outlet. Figure 4.5 also shows that when the liquid flow rate is increased by 10 times, the amount of vapor created is significantly reduced. Thus increase of liquid flow rate extends the boiling limit (delays dry-out process) of the porous wick.
-6 -5 -4 (a) 푚̇ 푖푛 = 10 Kg/s (b) 푚̇ 푖푛 = 10 Kg/s (c) 푚̇ 푖푛 = 10 Kg/s
Figure 4.3. Liquid-vapor fronts as a function of liquid mass flow rate 풎̇ 풊풏.
The variation of the heated surface temperature and pressure drop across the wick as function of the mass flow rate is plotted in Fig. 4.6. As the mass flow rate increases, the pressure drop across the wick increases and the temperature of the heated surface decreases.
There is only 1 degree of temperature change due to an increase of the mass flow rate by hundred times. This decrease in temperature is due to the reduction in the amount of vapor in the porous wick. The vapor has a lower thermal conductivity comparing to the liquid water.
As the flow rate is increased, there less thermal resistance caused by the vapor, which causes the temperature of the heated surface to decrease. 33
103.2 16.0 Temperature 103.0
14.0 x 10000 x 102.8 12.0
102.6 ]
C 10.0 [o 102.4 Pressure Difference 102.2 8.0
102.0
6.0 Temperature
101.8 [Pa] Difference Pressure 4.0 101.6 2.0 101.4
101.2 0.0 1 10 50 100 Liquid Flow Rate * 106 [Kg/s]
Figure 4.4. Temperature and Pressure difference in wick as functions of liquid mass flow rate 풎̇ 풊풏. 4.3 Permeability
The effect of wick permeability on the P2φL performance is shown in Fig. 4.7 and
Fig. 4.8. In this case, the wick permeability, 퐾, was changed from 10-12 to 10-14 m2. The change in size and shape of the vapor bubble is not significant, as shown in Fig 4.7. This caused the evaporator heated wall temperature to be constant for all values of permeability considered in this study (see Fig. 4.8). However, the decrease in the wick permeability has caused the pressure drop across the wick to significantly increase, because of the higher resistance presented by the wick to the flow of water. This result suggests that a higher permeability of the wick will lead to a low resistance to the flow of water, so minimize the power consumption of the mechanical pump; however, it will reduce the capillary pressure developed by the wick. Therefore, including the effect of the wick capillary pressure will help determine an optimum value for the wick permeability that will generate enough capillary pressure to assist the pump and reduce the power consumption of the pump. 34
Permeability 1x10-12 m2 Permeability 1x10-13 m2 Permeability 1x10-14 m2 Figure 4.5. Liquid-vapor fronts as a function of Permeability.
103.5 25.00
103 Thousands 20.00
102.5 Temperature ]
C 15.00 [o 102 Pressure Dierence
101.5
10.00
Temperature Pressure Difference [Pa] Difference Pressure 101
5.00 100.5
100 0.00 1.00E-12 1.00E-13 1.00E-14 Permeability [m2] Figure 4.6. Temperature and Pressure difference in wick as functions of Permeability. 35
4.4 Porosity
휀 = 0.3 휀 = 0.4 휀 = 0.5
휀 = 0.6 휀 = 0.7 휀 = 0.8 Figure 4.7. Liquid-vapor fronts as a function of Porosity.
Pressure Difference 2250 114 Temperature
2200 ]
C 112
[o 2150 110 2100 108 2050
106 2000 Temperature
104 1950 Pressure Difference [Pa] Difference Pressure 102 1900 30% 40% 50% 60% 70% 80% Porosity
Figure 4.8. Temperature and Pressure difference in wick as function of porosity.
The effect of wick porosity on the P2φL performance is shown in Fig. 4.9 and Fig. 4.10.
In this case, the wick porosity,휀, was changed from 0.4 to 0.8. The increase of porosity reduces the relative resistance through the wick to the flow of vapor. So, amount of vapor inside the wick increased when porosity is increased (Fig 4.9). Due to low thermal conductivity of vapor, thermal resistance of the wick increases. Temperature of the heated surface increased significantly when porosity has increased. Increased amount of vapor affect 36
liquid-vapor pressure equilibrium in the liquid-vapor interface. Increased amount of vapor increases pressure in the vapor side of liquid-vapor interface, and we observed increased pressure drop across the wick except for porosity in the range of 0.5 to 0.6 (Fig 4.10). At the outlet we observed a mixture of liquid and vapor phase for porosity ranging from 0.3 to 0.5, and this 2-phase mixture contribute to step up the pressure drop. At porosity 0.6 outlet is almost filled with vapor which reduces the pressure drop. At porosity higher than 0.6 the amount of vapor increased more, and pressure at the vapor side of the liquid-vapor interface increased. This increased the pressure drop across the wick with further increase of porosity for porosity higher than 0.6. Thus in our case, optimum value of porosity ranges from 0.5 to
0.6. Optimum value of porosity can be different for different wick configuration.
The results suggest that, use of high permeability wick in the vapor zone may assist removal of vapor through vapor grooves. High permeability wicks usually possess low capillary pressure, which is undesirable. So, development of a layered wick may be the potential solution to fabricate a wick having optimum operating characteristics. The layer adjacent to the fin/heated surface should have high permeability and high thermal conductivity, while the other layer/layers should possess low permeability and low thermal conductivity.
4.5 Heat flux
Figure 4.11 shows evolution of liquid-vapor front as a function of heat load Q. The size of vapor zone increased when Q has increased. Fig 4.12 shows that the temperature of the heated surface increased quasi-linearly with heat flux Q. The operating temperature (temperature of the heated surface) of P2φL may be U-shaped, flattened, or may increase linearly with heat flux. In a typical P2φL, inlet mass flow rate increases with heat flux Q, the system operates in
Variable Conductance mode (VCM), and the operating temperature curve with heat flux forms U shape. 37
Q=4W/cm2 Q=6W/cm2 Q=8W/cm2
Q=10W/cm2 Q=12W/cm2 Q=14W/cm2
Q=16W/cm2 Q=18W/cm2 Q=20W/cm2 Figure 4.9. Liquid-vapor fronts as a function of Heat Flux.
In our simulation, boundary condition is imposed by fixed water mass flow rate at inlet. So, it is similar to a P2φL system operating in Fixed Conductance Mode (FCM) where the condenser remains fully utilized, or in other words, the condenser is almost filled with two- phase zone, and the liquid enters liquid line at a temperature close to saturation. Like any other P2φL system operating in FCM, we obtained that in our system heated surface temperature increased linearly with heat flux Q. In the plot of pressure difference as a function of heat flux, we first observed increase of pressure difference with increase in heat flux, then there is a sharp decrease in pressure difference, and then it increased again. As 38
shown in Fig. 4.11, for heat flux less than 12 W/cm2, at the outlet the flow is a 2-phase mixture which contributes to a higher pressure drop, and we observed pressure drop across the wick is increasing with heat flux. At 푄=12 W/cm2 the outlet is almost filled with vapor that reduces the pressure drop across the wick. When the heat flux is further increased, the amount of vapor inside the wick increased further (Table 4.3) pushing the vapor front further inside the wick, altered pressure at the vapor side of the liquid-vapor interface. As a consequence pressure drop increased with heat load when heat flux is larger than 12W/cm2.
111 2400 Temperature 109 Pressure Difference 2200
] 107
C 2000 [o 105 1800
103 Temperature 1600 101 [Pa] Difference Pressure
99 1400
97 1200 4 6 8 10 12 14 16 18 20 Heat Flux [W/cm2]
Figure 4.10. Temperature and Pressure difference in wick as functions of Heat flux.
Table 4.3. Relation between vapor content and Pressure Difference. Q[W/cm2] 4 6 8 10 12 14 16 18 20 Vapor 0 0 4.28 11.78 21.75 30.08 36.29 41.19 45.1 inside wick*108 [Kg] Pressure 1415 1415 1570 2174 1974 2018 2088 2102 2426 Difference [pa] 39
4.6 Thermal conductivity
Fig 4.13 shows the effect of wick material thermal conductivity on the temperature distribution. Here, Titanium (휆 = 1.2 W/m-k), Steel (휆 = 16 W/m-k), Aluminum (휆 = 203
W/m-k), and copper (휆 =387.6 W/m-k) was used as wick materials. Vapor front position is nearly independent of wick thermal conductivity. This is logical, because the boundary conditions that were imposed on the simulations, the heat flux, were identical. Temperature distribution is significantly different as shown in Fig. 4.13. Fin superheat (Region A) was higher for low conductivity wick as shown in Fig. 4.13 and Fig. 4.14. We got highest evaporative cooling for Copper wick. In Copper wick, amount of vapor inside wick and pressure drop (Fig 4.14) across the wick, both are higher in value. The wick material itself takes heat from the heated wall, and Copper wick takes more heat compared to other materials due to its high thermal conductivity. Wick material thermal conductivity also alters the pressure drop across the wick as shown in Fig 4.14. The results show that the use of high conductivity wick material limits the temperature of the heated surface. On the other hand, using a high conductivity wick material may favor boiling of liquid at the fin-wick interface for small heat loads. High conductivity wick material would surge effective thermal conductivity of the wick, which would surge heat leak to the liquid reservoir and consequently, it will increase P2φL operating temperature. So, it is recommended to use low thermal conductivity wick material. However, several authors [52, 54] reports that adding a mechanical pump in a LHP takes away heat leak from the evaporator to the compensation chamber due to liquid flow in the compensation chamber. In our model, fixed liquid flow rate is prescribed in the inlet which impose same effect as having a continuous liquid flow in the compensation chamber.
40
Vapor front Vapor front
A A
(a) (b)
Vapor front Vapor front
A A
(c) (d) Figure 4.11. Temperature contour for: (a) Titanium wick; (b) Steel wick; (c) Aluminum wick; (d) Copper wick.
107 2320 2300 106 Pressure 2280
] 105 Difference 2260 C
[o 2240 104 Temperature 2220 103 2200 2180
Temperature 102 2160
2140 [Pa] Difference Pressure 101 2120 100 2100 1.2 16 202 387.6 Thermal Conductivity [W/m.K]
Figure 4.12. Temperature and Pressure difference in wick as functions of material thermal conductivity. 41
4.7 Gravity
g= -9.81 ms-2 g= 0 ms-2 Figure 4.13. Temperature field for different value of gravitational acceleration.
g= -9.81 ms-2 g= 0 ms-2 Figure 4.14. Pressure field for different value of gravitational acceleration.
400 2600
350 2500 Temperature 300
] Pressure Difference
C 2400
[o 250
200 2300
150
Temperature 2200
100 [Pa] Difference Pressure 2100 50
0 2000 gravity No gravity
Figure 4.15. Temperature and Pressure difference in wick as functions of the value of the gravitational acceleration. 42
The effect of value of gravitational acceleration is shown in Figs. 4.14, 4.15 and 4.16.
Here, no elevation between the evaporator and condenser is considered. The evaporator is simulated only for different values of gravitational acceleration. It was observed that the temperature distribution and temperature of heated surface is quiet independent of gravitational acceleration, as shown in Fig. 4.14 and Fig. 4.17. In our simulation boundary condition is imposed by fixed water mass flow rate at inlet, and the system is driven by capillary pressure and mechanical pumping. So, it is an inertia driven system, and it is not surprising that temperature distribution and heated surface temperature is independent of gravitational acceleration. Pressure drop across the wick is lower in presence of gravity, because in this orientation, gravity assists the flow to leave the domain by imposing hydrostatic pressure. Due to the same reason, we got more vapor accumulated inside the wick in absence of gravity. In a wick having heterogeneous porosity, increase of pressure drop due to absence of gravity may be more significant. The increased pressure drop across the wick due to absence of gravity can be balanced in cost of using wick with more fine pores, or by decreasing porosity. In our system, added mechanical pump will be sufficient to counteract increased pressure drop across the wick due to absence of gravity or any other reason. In 1g condition, elevation between evaporator and condenser alters total pressure drop across the system due to additional hydrostatic pressure, and thus P2φL operating temperature
(temperature of heated surface) would also change. In microgravity condition, effect of elevation between evaporator and condenser would be weaker, because zero or low gravitational acceleration will produce small hydrostatic pressure due to elevation between evaporator and condenser.
43
5 Conclusions
In this thesis, ANSYS Fluent CFD model was used to simulate heat transfer, phase change and flow patterns in a porous wick of a mechanically-pumped Loop Heat Pipe.
Results from the parametric study illustrated the effect of each parameters on the overall performance of the wick. To design precision thermal management systems such as LHPs for use in NASA future space missions, the porous wick must be optimized. Increase of liquid flow rate at inlet increases boiling limits and decreases operating temperature, but the pressure drop across the wick also increases significantly. There is an optimum value for porosity that can be determined by CFD model of that wick structure. High permeability favors to keep the pressure drop in minimal range. High thermal conductivity wick is not suitable for applications with moderate to high heat flux, although added pump minimizes the heat leak from the vapor fin to the compensation chamber. At low heat flux applications of
LHPs, use of high thermal conductivity wicks can be beneficial due to its ability to achieve higher fin superheat. As heat flux is increased the vapor front retreats inside the wick. Vapor grooves are engulfed with liquid for low heat flux. Multi-layers wicks can be a good improvement for high precision applications such as advanced thermal systems for NASA future space systems. However, these systems need further modeling and experimental investigation to learn more about these options and to justify their capabilities. 44
BIBLIOGRAPHY [1] Faghri, A., 1995, Heat Pipe Science and Technology, 1st ed., Taylor & Francis, Washington, D.C. [2] David Reay, Ryan McGlen and Peter Kew, 2013, Heat Pipes: Theory, Design and Applications, 6th ed., Elsevier, UK. [3] Calvin C. Silverstein, Design and Technology of Heat Pipes for Cooling and Heat Exchange, Taylor & Francis, Washington, D.C. [4] S.W. Chi, 1976, Heat Pipe Theory and Practice, Hemisphere Publishing Corporation, Washington. [5] Stéphane Launay, Valérie Sartre and Jocelyn Bonjour, 2006, Parametric analysis of loop heat pipe operation: a literature review, International Journal of Thermal Sciences 46 (2007) 621–636. [6] Jentung Ku, 1999, Operating Characteristics of Loop Heat Pipes, 29th International Conference on Environmental System, July 12-15, 1999, Denver, Colorado. [7] Yu.F. Maydanik, 2004, Review Loop heat pipes, Applied Thermal Engineering 25 (2005) 635–657. [8] Yu.F. Maydanik, M.A. Chernysheva and V.G. Pastukhov, 2014, Review: Loop heat pipes with flat evaporators, Applied Thermal Engineering 67 (2014) 294e307. [9] Amir Faghri, 2014, HEAT PIPES: REVIEW, OPPORTUNITIES AND CHALLENGES, Frontiers in Heat Pipes (FHP), 5, 1 (2014) [10] Donatas Mishkinis, G. Wang, D. Nikanpour and E. MacDonald, 2005, Advances in Two- Phase Loop with Capillary Pump Technology and Space Applications. [11] Amrit Ambirajan, Abhijit A. Adoni, Jasvanth S. Vaidya, Anand A. Rajendran, Dinesh Kumar & Pradip Dutta, 2011, Loop Heat Pipes: A Review of Fundamentals, Operation,and Design, Heat Transfer Engineering, 33(4–5):387–405, 2012. [12] Benjamin Siedel. Analysis of heat transfer and flow patterns in a loop heat pipe : Modelling by analytical and numerical approaches and experimental observations. Other [cond-mat.other]. INSA de Lyon, 2014. English.
[22] Benjamin Siedel, Valérie Sartre and Frédéric Lefèvre, 2014, Complete analytical model of a loop heat pipe with a flat evaporator, International Journal of Thermal Sciences 89 (2015) 372e386. [23] Benjamin Siedel, Valérie Sartre and Frédéric Lefèvre, 2013, Numerical investigation of the thermohydraulic behaviour of a complete loop heat pipe, Applied Thermal Engineering 61 (2013). [24] Stéphane Launay, Valérie Sartre and Jocelyn Bonjour, 2008, Analytical Model for Characterization of Loop Heat Pipes, Journal of thermophysics and heat transfer, Vol. 22, No. 4, October–December 2008. [25] Yu. Maydanik, S. Vershinin, M. Chernysheva and S. Yushakova, 2011, Investigation of a compact copper-water loop heap pipe with a flat evaporator, Applied Thermal Engineering 31 (2011). [26] Chernysheva, M., Vershinin, S., and Maydanik, Y. F., 2007, Operating temperature and distribution of a working fluid in LHP. International Journal of Heat and Mass Transfer. [27] Chen, Y., Groll, M., Mertz, R., Maydanik, Y. F., and Vershinin, S. Steady-state and transient performance of a miniature loop heat pipe. International Journal of Thermal Sciences, 45:1084–1090, 2006. [28] Launay, S. and Vall´ee, M. State-of-the-art experimental studies on loop heat pipes. Frontiers in Heat Pipes, 2(1), 2011. [29] J. Ku, 1994, Thermodynamic Aspects of Capillary Pumped Loop Operation, 6th AIAA/ASME Joint Thermophysics and Heat Transfer Conference, June 20-23, 1994 / Colorado Springs, CO. [30] Meng-Chang Tsai, 2005, Flat Plate Loop Heat Pipe with a Novel Evaporator Structure, 21st IEEE SEMI-THERM Symposium. [31] Jeehoon Choi, Wataru Sano, Weijie Zhang and Diana-Andra Borca-Tasciuc, 2013, Experimental investigation on sintered porous wicks for miniature loop heat pipe applications, Experimental Thermal and Fluid Science 51 (2013) 271–278. [32] Kaya, T. and Goldak, J. Numerical analysis of heat and mass transfer in the capillary structure of a loop heat pipe. International Journal of Heat and Mass Transfer, 49: 3211–3220, 2006. [33] V. Dhir, S. Kandlikar, Y. Fujita, Y. Iida, R. Heist, Nucleate boiling, in: S. Kandlikar, M. Shoji, V. Dhir (Eds.), Handbook of Phase Change: Boiling and Condensation, 1999, pp. 71– 120 (Chapter 4). [34] Wu, S.-C., Huang, C.-J., Chen, S.-H., and Chen, Y.-M. Manufacturing and testing of the double-layer wick structure in a loop heat pipe. International Journal of Heat and Mass Transfer, 56(1–2):709–714, 2013. [35] Incropera, F. P., Dewit, D. P., Bergman, T. L., and Lavine, A. S., Introduction to Heat Transfer, 5th ed., Wiley, New York, 2007. [36] P.-Y.A. Chuang, An improved steady-state model of loop heat pipes based on experimental and theoretical analyses, PhD thesis, The Pennsylvania State University, 2003, 271 pp. [37] M.O. Hamdan, Loop heat pipe (LHP) modelling and development by utilizing coherent porous silicon (CPS) wicks, PhD thesis, University of Cincinnati, 2003. [38] J. Ku, P. Ottenstein Rogers, K. Cheung, Investigation of low power operation in a loop heat pipe, in: 31st International Conference On Environmental Systems, Orlando, July 2001 (SAE paper 2001-01-2192). [39] W.-H. Lee, K.-W. Lee, K.-H. Park, K.-J. Lee, Study on working characteristics of loop heat pipe using a sintered metal wick, in: 13th IHPC, Shanghai, China, 21–25 September 2004. [40] D. Mishkinis and J.M. Ochterbeck, 2003, Loop Heat Pipes: Criteria Approach, V Minsk International Seminar “Heat pipes, Heat pumps, Refrigerators” Minsk, Belarus, September 8- 11, 2003. 46
[41] Shuang-Ying Wu, authorBao-Xi Cao, Lan Xiao and You-Rong Li, 2011, Parametric study on flow and heat transfer characteristics of porous wick evaporator based on AMTEC, Journal of Mechanical Science and Technology 26 (3) (2012) 973~981. [42] D. Mishkinis, J.M. Ochterbeck, C. Sodtke, J. Ku, D. Butler, Nondimensional analysis and scaling issues in loop heat pipes in: AIAA Aerospace Sciences Meeting and Exhibit, 41st, Reno, Nevada, 6–9 January 2003 (AIAA 2003-0341). [43] C. Figus, Y. Le Bray, S. Bories, M. Prat, Heat and mass transfer with phase change in a porous structure partially heated: continuum model and pore network simulations, Int. J. Heat Mass Transfer 42 (1999) 2557–2569. [44] C. Figus, Y. Le Bray, S. Bories, M. Prat, Heat transfer in porous media considering phase change, capillarity and gravity: Application to capillary evaporators, Proceedings of the 11th International Heat Transfer Conference, Kyongju, 0887, vol. 4. [45] W. Yao, J. Miao, X. Shao, Parametric analysis on LHP/CPL evaporator performance and critical heat flux by two-dimensional calculation, in: 13th IHPC, Shanghai, China, 21–25 September 2004. [46] V. Platel, O. Fudym, C. Butto, P. Briend, Coefficient de transfert, à l’interface de vaporisation, d’une boucle de fluide diphasique à pompage capillaire, Rev. Gén. Therm. 35 (1996) 592–598. [47] Joon Hong Boo and Won Bok Chung, 2005, Experimental Study on the Thermal Performance of a Small-scale Loop Heat Pipe with Polypropylene Wick, Journal of Mechanical Science and Technology (KSME Int. J.), VoL 19, No. 4. [48] A.A.M. Delil, V. Baturkin, Yu. Friedrikhson, Yu. Khmelev, S. Zhuk, Experimental results on heat transfer phenomena in miniature loop heat pipe with a flat evaporator, in: 12th IHPC, Moscow, 19–24 May 2002, 8 pp. [49] J. Baumann, S. Rawal, Viability of loop heat pipes for space solar power applications, in: AIAA Thermophysics Conference, 35th, Anaheim, CA, 11–14 June 2001 (AIAA 2001-3078). [50] T. Kaya, J. Ku, Ground testing of loop heat pipes for spacecraft thermal control, in: 33rd Thermophysics Conference, Norfolk, VA, 28 June – 1 July 1999. [51] Y. Chen, M. Groll, R. Mertz, Y.F. Maydanik, S.V. Vershinin, Steady-state and transient performance of a miniature loop heat pipe, International Journal of Thermal Sciences 45 (11) (2006). [52] C. Jiang, W. Liu, H.C. Wang, D.D. Wang, J.G. Yang, J.Y. Li, Z.C. Liu, Experimental investigation of pump-assisted capillary phase change loop, Applied Thermal Engineering 71 (2014). [53] N. Schweizer, P. Stephan, R. Schlitt, A Concept for a Miniature, Mechanically Pumped Two-phase Cooling Loop, 2008. SAE Paper No. 2008-01-1953. [54] M. Crepinsek, C. Park, Experimental analysis of pump-assisted and capillary driven dual-evaporators two-phase cooling loop, Appl. Therm. Eng. 38 (2012). [55] T. Hoang, R. Baldayff, K. Cheung, Hybrid Two-phase Mechanical Capillary Pumped Loop for High-capacity Heat Transport, 2007. SAE Paper No. 2007-01-3198. [56] C. Park, M. Crepinsek, Effect of operational conditions on cooling performance of pump-assisted and capillary-driven two-phase loop, J. Thermophys. Heat Transfer 25 (2011). [57] C. Park, A. Vallury, J. Zuo, Performance evaluation of a pump-assisted, capillary two- phase cooling loop, J. Therm. Sci. Eng. Appl. 1 (2009) 022004. [58] Eric Sunada, Benjamin Furst, Pradeep Bhandari, Brian Carroll, Gajanana C. Birur and Terry Hendricks, 2016, A Two-Phase Mechanically Pumped Fluid Loop for Thermal Control of Deep Space Science Missions, 46th International Conference on Environmental Systems, 10-14 July 2016, Vienna, Austria. [59] ANSYS Fluent Theory Guide, release 17.0, January 2016. [60] ANSYS Fluent User’s Guide, release 17.0, January 2016. [61] Multiphase flow tutorial, lesson 3, release 15.0, ANSYS INC. 47
[62] Dongliang Sun and Jinliang Xu, 2014, Modeling of the evaporation and condensation phase change problems with Fluent, Numerical Heat Transfer, Part B, 66: 326–342, 2014. [63] Alain Lamy and Vincent Albouys, MISSION DESIGN FOR THE “SWOT” MISSION, CNES, 18 avenue Edouard Belin, 31401 Toulouse Cedex 9, France. [64] Jordan Mizerak, Eric Sunada1, A.J. Mastropietro, Ben Furst, Hiroki Nagai, and Brian Carroll, 2016, Evaporator Design for a Mechanically Pumped Two-phase Cooling Loop in Microgravity. [65] B. Fadhl, L. C. Wrobel & H. Jouhara, Modelling of the thermal behavior of heat pipes, Advanced Computational Methods and Experiments in Heat Transfer XIII. [66] Dongliang Sun and Jinliang Xu, 2012, Development of a vapor-liquid phase change model for volume of fluid method in Fluent, International Communications in Heat and Mass Transfer 39 (2012) 1101–1106. [67] Zhangyuan Wang and Wansheng Yang, 2014, A review on loop heat pipe for use in solar water heating, Energy and Buildings 79 (2014) 143–154. [68 ] V.G. Pastukhov, Yu.F. Maidanik, C.V. Vershinin, M.A. Korukov, Miniature loop heat pipes for electronics cooling, Appl. Therm. Eng. 23 (9) (2003) 1125e1135.