Microfluid Nanofluid (2005) 1: 137–145 DOI 10.1007/s10404-004-0021-8 R ES E AR C H P A PE R Abhijit Mukherjee Æ Satish G. Kandlikar Numerical simulation of growth of a vapor bubble during flow boiling of water in a microchannel Received: 20 July 2004 / Accepted: 5 October 2004 / Published online: 31 March 2005 Springer-Verlag 2005 Abstract The present study is performed to numerically analyze the growth of a vapor bubble during flow of water through a microchannel. The complete Navier– Stokes equations, along with continuity and energy equations, are solved using the SIMPLER (semi-implicit method for pressure-linked equations revised) method. The liquid–vapor interface is captured using the level set technique. The microchannel is 200-lm square in crosssection and the bubble is placed at the center of the channel with superheated liquid around it. The results show steady initial bubble growth followed by a rapid axial expansion after the bubble fills the channel crosssection. A trapped liquid layer is observed between the bubble and the channel as it elongates. The bubble growth rate increased with the incoming liquid superheat, but decreased with Reynolds number. The formation of a vapor patch at the walls is found to be dependent on the time the bubble takes to fill up the channel. The upstream interface of the bubble is found to exhibit both forward and reverse movement during bubble growth. The results show little effect of gravity on the bubble growth rate under the specified conditions. The bubble growth features obtained from the numerical results are found to be qualitatively similar to experimental observations. l0 m p Re r T DT t t0 u u0 v w x y z bT j l m q r s / u Length scale Mass transfer rate at interface Pressure Reynolds number Radius Temperature Temperature difference, TwTsat Time Characteristic time x direction velocity Characteristic velocity y direction velocity z direction velocity Distance in x direction Distance in y direction Distance in z direction Coefficient of thermal expansion Interfacial curvature Dynamic viscosity Kinematic viscosity Density Surface tension Time period Level set function Contact angle Keywords Bubble Æ Microchannel Æ Flow boiling List Cp d g H hfg k of symbols Specific heat at constant pressure Grid spacing Gravity vector Heaviside function Latent heat of evaporation Thermal conductivity A. Mukherjee (&) Æ S. G. Kandlikar Department of Mechanical Engineering, Rochester Institute of Technology (RIT), Rochester, NY, USA E-mail: [email protected] Tel.: +1-585-4755839 Fax: +1-585-4757710 Subscripts evp Evaporation in Inlet l Liquid sat Saturation v Vapor w Wall x ¶/¶x y ¶/¶y z ¶/¶z Superscripts * Non-dimensional quantity fi Vector quantity 138 1 Introduction Two-phase flow through microchannels has been studied extensively in the last decade, since it has the potential of providing very high heat transfer rates in miniature heat exchangers. The studies, however, were primarily confined to experiments. But, experiments at the microscale have their own limitations, and a small inaccuracy in a measurement can lead to a large error. Hence, there is a need to numerically model two-phase flow through microchannels, which would help us better interpret the available experimental data and also explain the underlying physics. When boiling takes place in a microchannel, bubbles nucleate at the wall, but soon grow large enough to fill the entire channel. Thus, the behavior of individual bubbles determines the flow field in the microchannel. The present work is undertaken to numerically simulate the growth of a vapor bubble in a microchannel and analyze its different characteristics. 2 Literature review Jacobi and Thome (2002) developed a heat transfer model for elongated bubbles in a microchannel. They assumed that thin-film evaporation into elongated bubbles is the dominant heat transfer mechanism in microchannel flows. The model required the knowledge of two parameters; the effective nucleation superheat and the initial film thickness. The dependence of heat transfer coefficient on heat flux was found to be low at small heat fluxes, but high at large heat fluxes. The model predicted the heat transfer coefficient to be insensitive to mass flux. Steinke and Kandlikar (2003) experimentally studied flow boiling and pressure drop characteristics in parallel microchannels. They used six parallel microchannels of 207-lm hydraulic diameter and observed conventional flow boiling patterns. The flow patterns observed were bubbly flow, slug flow, and annular flow, similar to conventional flow boiling. They could maintain heat flux up to 930 kW/m2 in the microchannel. One of the major differences noted by them was the reversed flow during rapid growth of a bubble. Hetsroni et al. (2003) studied convective boiling in parallel microchannels using pure water and surfactants. They observed two flow regimes in steam–water flow. The low heat flux regime was characterized by the presence of a liquid phase in part of the parallel microchannels. The high heat flux regime was characterized by convective boiling, accompanied by quasi-periodical rewetting and refilling of the microchannels. They recommended that the boiling of surfactant solutions in microchannels may be used to provide a nearly isothermal heat sink. Peles (2003) studied two-phase boiling in microchannels and obtained flow regime maps. He used 16mm long parallel triangular microchannels with hydraulic diameters ranging from 50 lm to 200 lm. He also observed rapid bubble growth around a nucleating bubble. He concluded that two-phase flow instabilities were of primary importance at the microscale and should be comprehensively addressed. Fogg et al. (2003) numerically studied transient boiling in microchannels. They used a homogeneous model using mass-weighted averages of the local properties of the liquid and the vapor. The two phases were considered to be uniformly distributed within each grid element, precluding any models formulated for the bubble or slug structures. In conjunction, they also solved the one-dimensional transient heat conduction equation at the wall. They concluded that future work needs to model bubbles growing into slugs and annular flow. Ajaev and Homsy (2003) developed a mathematical model of constrained vapor bubbles. They assumed that the shape of the bubble is dominated by capillary forces away from the wall. A lubrication-type analysis was used to find the local vapor–liquid interface shapes and mass fluxes near the wall. The microscopic adsorbed film on the constraining walls was assumed to be in thermodynamic equilibrium with the vapor phase, due to the action of London–van-der-Waals forces. Their solutions indicate large values of mass flux near the contact line. Longer bubbles were obtained for higher heater temperatures. Son et al. (1999) developed a two-dimensional numerical model of growth and departure of single vapor bubbles during nucleate pool boiling. They used the level set technique to implicitly capture the liquid–vapor interface. Mukherjee and Dhir (2003) extended the model to three-dimensional cases and studied the merger and departure of multiple bubbles during nucleate pool boiling. The present analysis is done using a similar model to study the growth of a vapor bubble inside a microchannel. 3 Objectives The objectives of the present work are as follows: – Obtain the flow and thermal fields around a growing vapor bubble inside a microchannel through numerical simulation – Study the behavior of the liquid–vapor interface and its interaction with the constraining walls – Determine the effect of inlet liquid superheat, Reynolds number, and gravity on the bubble growth rate 4 Numerical model 4.1 Method The complete incompressible Navier–Stokes equations are solved using the SIMPLER method (Patankar 1980), 139 which stands for semi-implicit method for pressurelinked equations revised. The continuity equation is turned into an equation for the pressure correction. A pressure field is extracted from the given velocity field. At each iteration, the velocities are corrected using velocity-correction formulas. The computations proceed to convergence via a series of continuity-satisfying velocity fields. The algebraic equations are solved using the line-by-line technique, which uses the tri-diagonal matrix algorithm (TDMA) as the basic unit. The speed of convergence of the line-by-line technique is further increased by supplementing it with the block-correction procedure (Patankar 1981). The multi-grid technique is employed to solve the pressure equations. Sussman et al. (1994) developed a level set approach where the interface was captured implicitly as the zero level set of a smooth function. The level set function was typically a smooth function, denoted as /. This formulation eliminated the problems of adding/subtracting points to a moving grid and automatically took care of the merging and breaking of the interface. Furthermore, the level set formulation generalized easily to three dimensions. The present analysis is done using this level set technique. The liquid–vapor interface is identified as the zero level set of a smooth distance function /. The level set function / is negative inside the bubble and positive outside the bubble. The interface is located by solving the level set equation. A fifth-order weighted, essentially non-oscillatory (WENO) scheme is used for the left- and right-sided discretization of / (Fedkiw et al. 1998). While / is initially a distance function, it will not remain so after solving the level set equation. Maintaining / as a distance function is essential for providing the interface with a width fixed in time. This is achieved by the reinitialization of /. A modification of Godunov’s method is used to determine the upwind directions. The reinitialization equation is solved in fictitious time after each fully complete time step. With Ds ¼ 2ud 0 ; ten s steps are taken with a third-order total variation diminishing (TVD) Runge–Kutta method. The curvature of the interface is defined as: r/ j ð /Þ ¼ r jr/j ð4Þ The mass flux of liquid evaporating at the interface is given by: ~¼ m kl rT hfg ð5Þ The vapor velocity at the interface due to evaporation is: ~ uevp ¼ ~ kl rT m ¼ qv qv hfg ð6Þ To prevent instabilities at the interface, the density and viscosity are defined as: q ¼ qv þ ðql qv ÞH ð7Þ l ¼ lv þ ðll lv ÞH ð8Þ H is the Heaviside function, given by: H ¼1 H ¼0 / H ¼ 0:5 þ 3d þ if />1:5d if /6 1:5d sin2p/ 3d 2p ð9Þ if j/j61:5d where d is the grid spacing. Since the vapor is assumed to remain at saturation temperature, the thermal conductivity is given by: k ¼ kl H 1 ð10Þ The level set equation is solved as: @/ þ ~ u þ~ uevp r/ ¼ 0 @t ð11Þ After every time step, the level set function / is reinitialized as: @/ @t ¼ S ð/0 Þð1 jr/jÞu0 /ðx; 0Þ ¼ /0 ð xÞ ð12Þ where S is the sign function, which is calculated as: 4.2 Governing equations The momentum equation is written as: @~ u þ~ u r~ u ¼ rp þ q~ g qbT ðT Tsat Þ~ q g rjrH @t þ r lr~ u þ rlr~ uT The energy equation is written as: u rT ¼ r krT for / > 0 qCp @T @t þ ~ T ¼ Tsat for /60 ð1Þ ð2Þ The continuity equation is written as: r ~ u¼ ~ m rq q2 ð3Þ /0 S ð/0 Þ ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi /20 þ d 2 ð13Þ 4.3 Scaling factors The governing equations are made non-dimensional using a length scale and a time scale. The length scale l0 given by the channel width is equal to 200 lm. Thus, for the case of water at 100C and Re equal to 100, the velocity scale u0 is calculated as 0.146 m/s. The corresponding time scale t0 is 1.373 ms. All non-dimensional quantities hereafter are indicated with a * superscript. The non-dimensional temperature is defined as: 140 T ¼ T Tsat Tw Tsat ð14Þ 4.6 Boundary conditions The boundary conditions are as follows: – At the inlet (x*=0): 4.4 Computational domain u ¼ u0 ; Figure 1 shows the typical computational domain. The full domain is 4.95·0.99·0.99 non-dimensional units in size. Cartesian coordinates are used with the uniform grid. Flow takes place in the x direction, with liquid entering the domain at x*=0. To take advantage of symmetry and reduce computation time, the bubble is placed at the center of the microchannel cross-section, equidistant from the walls in the y and the z directions. The gravity vector acts in the negative y direction. It is shown later in this paper that gravity has little effect on the bubble growth under the given conditions in this study. Thus, when neglecting gravity, calculations are done only for one quarter of the domain shown in Fig. 1. Otherwise, when gravity is considered, computations are done using one half of the domain (in the z direction) shown in Fig. 1. – At the plane of symmetry, (y*=0, if neglecting gravity): v ¼ w ¼ 0; uy ¼ v ¼ wy ¼ Ty ¼ 0; T ¼ Tin ; /x ¼ 0 /y ¼ 0 ð15Þ ð16Þ – At the plane of symmetry (z*=0): uz ¼ vz ¼ w ¼ Tz ¼ 0; /z ¼ 0 * ð17Þ * – At the walls (y =0.495, y =0.495): u ¼ v ¼ w ¼ 0; T ¼ Tw ; /y ¼ cos u ð18Þ where u is the contact angle. – At the walls (z*=0.495, z*=0.495): u ¼ v ¼ w ¼ 0; T ¼ Tw ; /z ¼ cos u ð19Þ – At the outlet (x*=4.95): ux ¼ vx ¼ wx ¼ Tx ¼ 0; /x ¼ 0 ð20Þ 4.5 Initial conditions 4.7 Effect of grid size The bubble is placed at x*=0.99, y*=0, and z*=0, with radius 0.1l0 in the domain shown in Fig. 1. All velocities in the internal grid points are set to zero. The wall temperatures are set to 107C (T*=1). The vapor inside the bubble is set to a saturation temperature of 100C (T*=0). The liquid temperature inside the domain is set equal to the inlet liquid temperature. All physical properties are taken at 100C. The inlet velocity is specified as u0 and the inlet liquid temperatures Tin are varied as 102C, 104C, and 107C. The Reynolds number is varied as 50, 100, and 150. The contact angle at the walls is specified as 50 for the situation when the bubble touches the walls. We have compared the bubble growth rates for five different grid sizes. Figure 2 plots the time taken for the downstream interface of the vapor bubble to cross fourfifths of the channel length for different grids. The differences in growth time decreases as the grid is refined. To optimize the computation costs and numerical accuracy, 96 grids are used per 0.99l0 for all subsequent calculations. Thus, the number of computational cells in the half-domain used for the calculations when considering gravity are 480·96·48. Fig. 1 Computational domain Fig. 2 Grid independence check 141 5 Results First, the growth of a vapor bubble is considered at a liquid inlet temperature of 102C, with no gravity. The flow and thermal fields around the bubble is analyzed and its interaction with the wall is studied. Next, the effects of the inlet liquid superheat, Reynolds number, and gravity on the bubble growth rates are determined. Finally, the numerical results are compared with experimental observations. 5.1 Bubble growth with Tin=102C, Re=100, g=0 Figure 3 shows the growth of the vapor bubble inside the microchannel. The wall superheat is 7C and the liquid inlet temperature is 102C. The Reynolds number is specified as 100. Gravity is neglected in this case, specifying g=0 in Eq. 1. The time corresponding to each frame is shown at the lower right corner. The frame at 0.00 ms shows the initial condition with a bubble of radius 0.1l0 placed at the center of the channel equidistant from the walls. At 0.082 ms, we see that the bubble has grown bigger due to evaporation, but it still retains its spherical shape. The bubble continues to grow and, at the same time, moves downstream in the direction of flow, as seen at 0.158 ms. At 0.232 ms, the bubble starts to elongate in the direction of flow. Here, the bubble has almost filled the channel cross-section and cannot expand any further in the y–z plane, due to resistance from the wall. At 0.256 ms, the bubble elongates further and turns into a plug. A thin layer of liquid separates the bubble from the walls on all sides. At 0.278 ms, the bubble is seen to form vapor patches at the walls near its upstream end. Our calculations were stopped when the downstream interface crossed a length of 4l0 of the channel, which corresponds to the last frame at 0.278 ms. Figure 4 shows the velocity vectors at a vertical x–y plane through the center of the domain. The first frame at 0.000 ms shows no velocity vectors since before the start of the calculations, all internal velocities are set to zero. Thereafter, as the bubble grows, the liquid is pushed towards downstream of the bubble, which can be seen at 0.082 ms. The velocity vectors are much larger downstream of the bubble compared to upstream. The reference vector shown in the frame corresponds to 10u0. At 0.158 ms, as the bubble interface approaches the walls, vapor jets are seen inside the bubble due to evaporation at the interface near the walls. At 0.232 ms, the bubble has started to elongate along the channel axis and more evaporation takes place from its interface near the walls, causing the bubble to expand rapidly. Large velocity vectors can be seen across the downstream interface at 0.256 ms, which is pushing the liquid out of the channel. Fig. 3 Growth of vapor bubble in microchannel; Tin=102C, Re=100, g=0 142 Fig. 4 Velocity vectors at the central vertical plane; Tin=102C, Re=100, g=0 Fig. 5 Temperature field at the central vertical plane; Tin=102C, Re=100, g=0 The bubble touches the wall as it expands and forms vapor patches, as seen at 0.278 ms. A thin layer of liquid is trapped between the bubble and the wall in the downstream of the vapor patch. Due to the presence of saturated vapor on one side, and superheated wall on the other side, very high rate of evaporation takes place in this liquid layer. Figure 5 shows the temperature field around the bubble at 0.278 ms. Isotherms are plotted in ten intervals between 0 and 1. The wall is at T*=1 and the vapor inside the bubble is at T*=0. Formation of a thermal boundary layer can be seen in the liquid near the walls. Crowding of isotherms between the bubble and the walls at downstream of the vapor patch indicates very high heat transfer rates in these regions. Figure 6 shows the plot of the bubble diameter and the bubble length in the axial direction against time. It also shows the locations of the upstream and downstream interfaces of the bubble in the channel measured from the channel inlet as a function of time. The bubble equivalent diameter is calculated assuming a sphere of equal volume. The bubble length is the length of the bubble at the x axis through the center of the domain. It is calculated as the difference between the upstream and the downstream end locations. The bubble growth rate, indicated by the equivalent diameter, is constant initially, but increases as the bubble equivalent diameter approaches the channel width at around 0.2 ms. The initial bubble growth is linear, as it is inertia controlled. The constant bubble growth rate (d r/d t) before 0.1 ms is found to be around 0.38 m/s. The bubble growth rate is found to decrease slightly after 0.27 ms, due to the formation of the vapor patches. The bubble length is equal to the equivalent diameter initially, as the small bubble grows spherically with superheated liquid around it. However, as the bubble equivalent diameter approaches the channel diameter, it faces resistance from the wall and grows rapidly in the axial direction. This is evident as the bubble length becomes greater than its equivalent diameter after 0.1 ms. As the bubble elongates, more surface area becomes available for evaporation, and the bubble length increases exponentially between 0.1 ms and 0.28 ms. The bubble upstream interface location is seen to decrease slightly until 0.1 ms, as the bubble grows uniformly at the initial stages. Thereafter, as the bubble starts to elongate after 0.1 ms, the upstream interface 143 Fig. 6 Bubble growth rate; Tin=102C, Re=100, g=0 location increases, indicating that the entire bubble is moving downstream with the flow. The bubble downstream interface location increases throughout the entire growth period. It initially increases linearly, similar to the bubble diameter, until 0.1 ms, and afterwards, exponentially as the bubble elongates into a plug. The velocity of the bubble downstream interface is around 6 m/s when the bubble length is 0.5 mm. This is much higher than the inlet velocity of the liquid, which is around 0.146 m/s. Thus, the liquid is pushed out of the channel at a very high rate, due to evaporation and the bubble growth. inlet superheats of 2C, 4C, and 7C. Gravity is neglected in all these cases. The results indicate that the bubble grows considerably faster with increasing liquid inlet temperature, thereby, decreasing the time taken by the bubble to fill the channel. The upstream interface is found to move back initially in all cases, causing reversed flow, indicating the dominance of the bubble growth over the incoming liquid flow. The amount by which the upstream interface receded increased with the liquid inlet temperature, due to higher growth rates. This reversed flow is, however, different to that observed by investigators in parallel microchannels. In the present case of a single channel, the flow at the inlet is constant, and, hence, the backward movement of the upstream interface increases the downstream liquid flow between the bubble and the walls. After some time, the location of the upstream interface becomes constant in all the three cases. In the case of 102C inlet temperature, which has the lowest bubble growth rate, the upstream interface is found to advance considerably in the direction of flow after 0.16 ms. The downstream interface moves forward with bubble growth in all the cases and displays similar growth patterns. Formation of vapor patches at the walls was observed for an inlet superheat of 2C (Fig. 2). Interestingly, it was not observed for the cases of 4C and 7C inlet superheats. This could be due to the fact that the bubbles grew too quickly in the latter two cases. Thus, there was insufficient time for the thin layer of liquid between the bubbles and the walls to evaporate, thereby, preventing the formation of vapor patches. 5.2 Effect of liquid inlet temperature 5.3 Effect of Reynolds number Figure 7 compares the effect of the inlet liquid temperature on the bubble growth. The upstream and downstream interface locations are plotted against time for Figure 8 compares the effect of the Reynolds number on the bubble growth. The upstream and downstream Fig. 7 Effect of inlet liquid temperature on bubble growth; Re=100, g=0 Fig. 8 Effect of Re on bubble growth; Tin=102C, g=0 144 interface locations are plotted against time for a Reynolds number of 50, 100, and 150. Gravity is neglected in all the cases and the liquid inlet temperature is 102C. A higher Reynolds number signifies a higher liquid inlet velocity that pushes the bubble downstream. Hence, the bubble upstream interface is found to move forwards faster with increasing Reynolds number. The bubble downstream interface, however, exhibits the opposite behavior. The case with a Reynolds number equal to 50 shows the highest downstream interface velocity after 0.2 ms. This is due to the fact that, when the bubble interface approaches the walls, the bubble growth rate depends on the thickness of the wall thermal boundary layer. An increased Reynolds number causes the thermal boundary layer at the walls to be thinner. Therefore, a comparatively smaller area of the bubble interface gets exposed to the superheated thermal boundary layer near the walls. An increased Reynolds number also causes the layer of liquid between the bubble and the walls to be thicker due to an increased supply of liquid. This, in turn, decreases the temperature gradient across the liquid layer and the subsequent evaporation rate. As a result, when the bubble has turned into a plug, the bubble growth rate and the downstream interface velocity is less for an increased Reynolds number. Formation of vapor patches at the walls was observed for all the above three cases. 5.4 Effect of gravity All the results so far were obtained neglecting gravity. Figure 9 shows the effect of gravity on the bubble growth rate for the case with Tin=107C. In the case with gravity, calculations are carried out for half the domain shown in Fig. 1. The gravity acts in the negative y direction. The plot shows the bubble equivalent diameter against time, until the bubble diameter reaches 0.2 mm, which is the channel width. It indicates an insignificant Fig. 9 Effect of gravity; Tin=107C, Re=100 difference in the bubble growth rates for the two cases. This justifies the assumption of neglecting gravity for the present channel geometry and Reynolds number. It helps to take advantage of symmetry in the y direction and reduces computation time. The above finding is in agreement with the experimental observations made by Kandlikar and Balasubramanian (2004). They found similar flow patterns for horizontal flows, vertical up flows, and vertical down flows during flow boiling of water in parallel minichannels. 5.5 Comparison with experimental observations The numerical results qualitatively agree with the experimental data obtained by Balasubramanian and Kandlikar (2004). The experiments show an increase in bubble/plug length growth rate with time, similar to present calculations. The plug growth rate in the experiments with subcooled liquid at the inlet was found to be 0.3 m/s at the initial stages of bubble growth and around 3.5 m/s at the later stages. This is comparable to the respective values of 0.38 m/s and 6 m/s obtained from numerical calculations at Tin=102C. However, the above experimental study was undertaken in 1-mm wide parallel channels, 63.5 mm in length. Figure 10 compares the bubble shapes between numerical calculations and experimental observations. This comparison is only qualitative. The times indicated on the frames for numerical results do not correspond to the experimental observations. Fig. 10 Comparison of bubble shapes 145 Both the experiments and the numerical calculations show the formation of vapor patches at the wall and thin layers of liquid between the bubble and channel corners. The bubbles are seen to move downstream with flow as it turns into a plug and fills up the entire channel crosssection. The upstream and downstream interfaces of the plug have similar shapes in both of the cases. 6 Conclusions 1. A vapor bubble growing in superheated liquid inside a microchannel is modeled using the level set technique. 2. The bubble initially grows at a constant rate, but its length increases rapidly when it fills the channel cross-section and expands in a longitudinal direction. This increase in growth rate is due to the thin layer of liquid between the walls and its interface, where a high rate of evaporation takes place. 3. The upstream interface of the bubble is found to exhibit both forward and reverse movement during bubble growth. 4. The bubble growth rate is found to increase with the incoming liquid superheat, but decrease with Reynolds number. 5. Vapor patch formation at the walls is found to be dependent on the time taken by the bubble to fill up the channel length. 6. The effect of gravity is found to be negligible on the bubble growth rate under the specified conditions. 7. The bubble growth features obtained from numerical calculations are found to be qualitatively similar to the experimental observations. Acknowledgments The work was conducted in the Thermal Analysis and Microfluidics Laboratory at the Rochester Institute of Technology (RIT), Rochester, NY, USA. References Ajaev VS, Homsy GM (2003) Mathematical modeling of constrained vapor bubbles. In: Proceedings of the 1st international conference on microchannels and minichannels, Rochester, New York, April 2003, pp 589–594 Balasubramanian P, Kandlikar SG (2004) An experimental study of flow patterns, pressure drop and flow instabilities in parallel rectangular minichannels. In: Proceedings of the 2nd international conference on microchannels and minichannels, Rochester, New York, June 2004, ICMM2004–2371, pp 475– 481 Fedkiw RP, Aslam T, Merriman B, Osher S (1998) A non-oscillatory Eulerian approach to interfaces in multimaterial flows (the ghost fluid method). Department of Mathematics, UCLA, CAM Report 98–17, Los Angeles Fogg DW, Koo JM, Jiang L, Goodson KE (2003) Numerical simulation of transient boiling convection in microchannels. In: Proceedings of the ASME summer heat transfer conference, Las Vegas, Nevada, July 2003, HT2003–47300 Hetsroni G, Klein D, Mosyak A, Segal Z, Pogrebnyak E (2003) Convective boiling in parallel micro-channels. In: Proceedings of the 1st international conference on microchannels and minichannels, Rochester, New York, April 2003, pp 59–67 Jacobi AM, Thome JR (2002) Heat transfer model for evaporation of elongated bubble flows in microchannels. J Heat Trans 124:1131–1136 Kandlikar SG, Balasubramanian P (2004) Effect of gravitational orientation on flow boiling of water in 1054·197 lm parallel minichannels. In: Proceedings of the 2nd international conference on microchannels and minichannels, Rochester, New York, June 2004, ICMM2004-2379, pp 539–550 Mukherjee A, Dhir VK (2003) Numerical study of lateral merger of vapor bubbles during nucleate pool boiling. In: Proceedings of the ASME summer heat transfer conference, Las Vegas, Nevada, July 2003, HT2003-47203 Patankar SV (1980) Numerical heat transfer and fluid flow. Hemisphere Publishing Company, Washington, DC Patankar SV (1981) A calculation procedure for two-dimensional elliptic situations. Num Heat Trans 4:409–425 Peles Y (2003) Two-phase boiling flow in microchannels—instabilities and flow regime mapping. In: Proceedings of the 1st international conference on microchannels and minichannels, Rochester, New York, April 2003, pp 559–566 Son G, Dhir VK, Ramanujapu NK (1999) Dynamics and heat transfer associated with a single bubble during nucleate boiling on a horizontal surface. J Heat Trans 121:623–631 Steinke ME, Kandlikar SG (2003) Flow boiling and pressure drop in parallel flow microchannels. In: Proceedings of the 1st international conference on microchannels and minichannels, Rochester, New York, April 2003, pp 567–579 Sussman M, Smereka P, Osher S (1994) A level set approach for computing solutions to incompressible two-phase flow. J Comput Phys 114:146–159
© Copyright 2026 Paperzz