Modeling of Subcooled Boiling in a Nuclear Reactor Core by Matthew P. Wilcox A Thesis Submitted to the Graduate Faculty of Rensselaer Polytechnic Institute in Partial Fulfillment of the Requirements for the degree of MASTER OF SCIENCE Major Subject: MECHANICAL ENGINEERING Approved: _________________________________________ Ernesto Gutierrez-Miravete, Thesis Adviser Rensselaer Polytechnic Institute Troy, New York December, 2012 © Copyright 2012 By Matthew P. Wilcox All Rights Reserved ii CONTENTS Modeling of Subcooled Boiling in a Nuclear Reactor Core ............................................... i LIST OF TABLES ............................................................................................................. v LIST OF FIGURES .......................................................................................................... vi ACKNOWLEDGMENT ..................................................Error! Bookmark not defined. ABSTRACT .................................................................................................................... vii 1. Introduction.................................................................................................................. 1 2. Background .................................................................................................................. 2 3. Methodology ................................................................................................................ 4 3.1 Turbulence Models............................................................................................. 4 3.2 Two Phase Models (ANSYS Fluent Theory Guide MultiPhase Flow) ............. 5 3.3 3.2.1 Volume of Fluid ..................................................................................... 5 3.2.2 Mixture ................................................................................................... 5 3.2.3 Eulearian ................................................................................................ 6 Solution Models ................................................................................................. 6 4. Equations ....................................................................Error! Bookmark not defined. 5. Numerical Methods ..................................................................................................... 9 6. Natural Convection .................................................................................................... 11 6.1 Introduction ...................................................................................................... 11 6.2 Theory .............................................................................................................. 12 6.3 Examples of Natural Convection ..................................................................... 14 6.3.1 Modeling of a Horizontal Cylinder ...................................................... 14 6.3.2 Modeling of a Vertical Plate ................................................................ 18 7. Laminar Flow with Heat Transfer ............................................................................. 22 8. Turbulence ................................................................................................................. 26 8.1 Calculating Turbulence Parameters ................................................................. 29 9. Turbulence with Heat Transfer .................................................................................. 35 iii 10. Two-Phase Flow ........................................................................................................ 38 10.1 Flow Regimes .................................................................................................. 41 11. Gas Mixing Tank ....................................................................................................... 43 12. Bubble Column .......................................................................................................... 46 13. Population Balance Equation ..................................................................................... 50 13.1 Background ...................................................................................................... 50 13.2 Equation Formulation ...................................................................................... 51 13.2.1 Particle State Vector ............................................................................. 51 13.2.2 Population Balance Equation ............................................................... 52 13.2.3 Particle Growth and Dissolution .......................................................... 52 13.2.4 Particle Birth and Death Due to Breakage and Aggregation ............... 52 13.2.5 Particle Birth by Nucleation ................................................................. 53 13.3 Solution Method ............................................................................................... 54 14. Bubble Column with Population Balance Model ...................................................... 55 15. Pool Boiling ............................................................................................................... 58 16. Subcooled Boiling ..................................................................................................... 61 17. Subcooled Boiling with Population Balance Model .................................................. 62 18. References.................................................................................................................. 63 iv LIST OF TABLES Table 11-1: Mesh Independent Quantities ....................................................................... 45 v LIST OF FIGURES Figure 12-1: Bubble Column ........................................................................................... 46 Figure 12-2: Instantaneous Gas Volume Fraction ........................................................... 47 Figure 12.3: Instantaneous Liquid Velocity Vectors ....................................................... 48 Figure 12-4: Instantaneous Gas Velocity Vectors ........................................................... 49 Figure 13-1: Particle Size Distribution (ANSYS Fluent PBE Guide Figure 2.1) ........... 54 Figure 14-1: Instantaneous Gas Volume Fraction with PBM ......................................... 56 Figure 14-2: Bubble Column Liquid Vector Velocity with PBM ................................... 57 Figure 15.1: Instantaneous Gas Volume Fraction ........................................................... 59 Figure 15.1: Instantaneous Liquid Velocity Vectors ....................................................... 59 Figure 15.3: Volume Fraction of Vapor on Heated Surface ............................................ 60 vi ABSTRACT vii Nomenclature viii 1. Introduction Nuclear reactors have been used for commercial electricity production since 1958. They provide roughly 20% of the electricity in the United States and about 13% world-wide. There are two distinct types of nuclear reactors, Pressurizer Water Reactors (PWR) and Boiling Water Reactors (BWR). The more common of the two types is the PWR which heats water that flows past nuclear fuel rods to produce steam. The benefit of having a pressurized system is that the water can be heated to very high temperatures without bulk boiling occurring. This helps to increase efficiency and prevent fuel failure. Energy is removed from the nuclear fuel rods through an efficient heat transfer process known as nucleate boiling. During nucleate boiling, the heated surface temperature is hotter than the saturation temperature of the fluid causing localized boiling in the subcooled bulk fluid. The amount of subcooled boiling is heavily impacted by fluid inlet temperature, pressure, mass flow and heat flux. Because there are so many factors that play a role in determining the level of subcooled boiling, the amount when it occurs in a nuclear reactor is generally unknown. The power in a reactor core is held constant by keeping the reactivity balance at zero. Positive reactivity leads to a power increase while negative reactivity leads to a power decrease. Some of the components that make up the reactivity balance equation are water temperature, water density, fuel temperature and voiding in the core. The accuracy at which each parameter can be measured impacts the ability to calculate core power during a transient. Being able to accurately measure the core power during a transient removes uncertainty in the safety analyses performed. Currently there are methods to accurately calculate the reactivity components listed above except for voiding in the core. If a more accurate method was developed to calculate voiding under varying conditions, there would be less uncertainty in the power calculation and more safety analysis margin could be gained thus allowing plants to increase power and produce more electricity. The purpose of this thesis is to develop a better understanding of subcooled boiling and generate a more accurate method to measure voiding at different axial locations in the core. 2. Background Electricity is one of the greatest discoveries of the 19th century and its use has greatly increased the world’s standard of living. It is generated by converting thermal energy, from a fuel source, into electrical energy. This is done generally through the Rankine Cycle where fuel is burned to heat water and form steam. The steam is then used to turn a turbine which spins an electric generator. Electricity production involves numerous engineering processes but primarily based around heat transfer and fluid flow. There are many different fuel sources available to electrical power plants, the one in focus here will be nuclear fuel. Nuclear power plants harness the energy released during the fission process to heat the surrounding water called the Reactor Coolant System (RCS) which is then used to produce steam. The heat transfer mechanisms at work within a nuclear reactor core are extremely complex. All three major forms of heat transfer are at work, conduction, convection and radiation. The fluid flow through the reactor is also complex due to the extreme energy transfer and phase change. The RCS is prevented from bulk boiling due to the highly pressurized system; however, a small amount of localized boiling does occur. This is also known as subcooled boiling. This thesis will focus on the convective heat transfer that occurs in the core, and more specifically, subcooled boiling. Subcooled boiling occurs when a fluid comes into contact with a surface that is hotter than its saturation temperature. Small bubbles form on the heated surface in locations called nucleation sites. Eventually the bubbles detach and enter the bulk fluid. At this point there is saturated steam in a subcooled liquid. The bubbles have three options, they can coalesce with other bubbles, they can grow in size or they can shrink in size. The idea of using PBEs for subcooled boiling is a relatively new idea and has shown potential recently to provide great insights into this regime of boiling heat transfer. The amount of voiding that occurs in a nuclear reactor core has a direct impact on fission power due to reactivity feedbacks caused by voiding. If a better understanding of the level of voiding due to subcooled boiling was developed, the accuracy at which fission power is calculated during a transient could be improved and the amount of uncertainty reduced. Fluid properties such as temperature, pressure, mass 2 flux and heat flux will be varied and their impact on the amount of subcooled boiling at different axial locations in a nuclear reactor will be determined. 3 3. Methodology A portion of a fuel bundle will be modeled using ANSYS Fluent. The traditional models available (energy equation, turbulence, two-phase, etc.) in Fluent will be implemented along with a Population Balance Equation (PBE) model. Population balance equations have been introduced in several branches of modern science, mainly in branches with particulate entities. Population balance equations define how populations of separate entities develop in specific properties over time. They are nothing more than a balance on the number of particles in a particular state. The PBE model will be used to determine the number of steam bubbles in the core, reveal how they develop over time and decide if the bubbles shrink and collapse or coalesce and grow in size. Ten models will be created, each more advanced than the previous. The final model will be three-dimensional, use multiple heated rods, allow for turbulent, two-phase flow and have the PBE model implemented. For more information about the model progression and development, see the Model Development section. After each model is developed, it will be compared to known experimental data whenever possible in order to validate the information generated by ANSYS Fluent. After the models have been validated and the final model developed, the initial conditions, temperature, pressure, mass flux and heat flux, will be altered so that the voiding at different axial locations can be determined for the various initial conditions. Once the data has been collected, it will be analyzed to produce either a set of equations or a set of tables that will allow the user to quickly determine how much voiding occurs based on the known conditions. To determine how the bubbles will react within the subcooled fluid, Population Balance Equations (PBE) will be used. Population Balance Equations allow Fluent to calculate probabilities instead of keeping track of each individual steam bubble which would require large amounts of computing power and significant amount of time. 3.1 Turbulence Models 4 3.2 Two Phase Models (ANSYS Fluent Theory Guide MultiPhase Flow) A large number of flows encountered in nature and technology are a mixture of phases. Physical phases of matter are gas, liquid, and solid, but the concept of phase in a multiphase flow system is applied in a broader sense. In multiphase flow, a phase can be defined as an identifiable class of material that has a particular inertial response to and interaction with the flow and the potential field in which it is immersed. For example, different-sized solid particles of the same material can be treated as different phases because each collection of particles with the same size will have a similar dynamical response to the flow field. 3.2.1 Volume of Fluid The VOF model can model two or more immiscible fluids by solving a single set of momentum equations and tracking the volume fraction of each of the fluids throughout the domain. Typical applications include the prediction of jet breakup, the motion of large bubbles in a liquid, the motion of liquid after a dam break, and the steady or transient tracking of any liquid-gas interface. 3.2.2 Mixture The mixture model is a simplified multiphase model that can be used in different ways. It can be used to model multiphase flows where the phases move at different velocities, but assume local equilibrium over short spatial length scales. It can be used to model homogeneous multiphase flows with very strong coupling and phases moving at the same velocity and lastly, the mixture models are used to calculate non-Newtonian viscosity. The mixture model can model multiple phases (fluid or particulate) by solving the momentum, continuity, and energy equations for the mixture, the volume fraction equations for the secondary phases, and algebraic expressions for the relative velocities. Typical applications include sedimentation, cyclone separators, particle-laden flows with low loading, and bubbly flows where the gas volume fraction remains low. 5 The mixture model is a good substitute for the full Eulerian multiphase model in several cases. A full multiphase model may not be feasible when there is a wide distribution of the particulate phase or when the interphase laws are unknown or their reliability can be questioned. A simpler model like the mixture model can perform as well as a full multiphase model while solving a smaller number of variables than the full multiphase model. 3.2.3 Eulearian The Eulerian multiphase model in ANSYS FLUENT allows for the modeling of multiple separate, yet interacting phases. The phases can be liquids, gases, or solids in nearly any combination. An Eulerian treatment is used for each phase, in contrast to the EulerianLagrangian treatment that is used for the discrete phase model. With the Eulerian multiphase model, the number of secondary phases is limited only by memory requirements and convergence behavior. Any number of secondary phases can be modeled, provided that sufficient memory is available. For complex multiphase flows, however, you may find that your solution is limited by convergence behavior. See Eulerian Model in the User's Guide for multiphase modeling strategies. ANSYS FLUENT’s Eulerian multiphase model does not distinguish between fluid-fluid and fluid-solid (granular) multiphase flows. The volume of fluid model solves a single set of momentum equations for two or more fluids and tracks the volume fraction of each fluid throughout the domain. The mixture model solves for the momentum equation of the mixture and prescribes relative velocities to describe the dispersed phases. The Eulerian model solves momentum and continuity equations for each of the phases, and the equations are coupled through pressure and exchange coefficients. 3.3 Solution Models 6 4. Mathematical Formulation A continuity equation in physics is an equation that describes the transport of a conserved quantity. Since mass, energy, momentum and other natural quantities are conserved under their respective appropriate conditions; a variety of physical phenomena may be described using continuity equations. Continuity equations are a stronger, local form of conservation laws. In fluid dynamics, two important continuity equations are the conservation of mass and the conservation of momentum. [Transport Phenomenon] Conservation of Mass in Vector Form: ππ β β πv + (∇ β)= 0 ππ‘ Conservation of Mass in Cartesian Form: ππ π π π (ππ£π₯ ) + (ππ£π ) + (ππ£π§ ) = 0 + ππ‘ ππ₯ ππ ππ§ Conservation of Momentum in Vector Form: π π·v β β π + π∇ β 2v = −∇ β + ππ π·π‘ Conservation of Momentum in Cartesian Form: ππ£π₯ ππ£π₯ ππ£π₯ ππ£π₯ ππ π 2 π£π₯ π 2 π£π₯ π 2 π£π₯ π( + π£π₯ + π£π¦ + π£π§ )=− +π( 2 + + ) + πππ₯ ππ‘ ππ₯ ππ¦ ππ§ ππ₯ ππ₯ ππ¦ 2 ππ§ 2 π( ππ£π¦ ππ£π¦ ππ£π¦ ππ£π¦ π 2 π£π¦ π 2 π£π¦ π 2 π£π¦ ππ + π£π₯ + π£π¦ + π£π§ )=− +π( 2 + + ) + πππ¦ ππ‘ ππ₯ ππ¦ ππ§ ππ¦ ππ₯ ππ¦ 2 ππ§ 2 ππ£π§ ππ£π§ ππ£π§ ππ£π§ ππ π 2 π£π§ π 2 π£π§ π 2 π£π§ π( + π£π₯ + π£π¦ + π£π§ )=− +π( 2 + + ) + πππ§ ππ‘ ππ₯ ππ¦ ππ§ ππ§ ππ₯ ππ¦ 2 ππ§ 2 In many instances of fluid dynamics, energy is being added or removed from the system. In this situation, the conservation of energy equation is important. 7 Conservation of Energy in Vector Form: ππΆΜπ π·π π ln π π·π β β π) − ( = −(∇ ) π·π‘ π ln π π π·π‘ Conservation of Energy in Cartesian Form: ππ ππ ππ ππ πππ₯ πππ¦ πππ§ π ln π π·π ππΆΜπ ( + π£π₯ + π£π¦ + π£π§ ) = − ( + + )−( ) ππ‘ ππ₯ ππ¦ ππ§ ππ₯ ππ¦ ππ§ π ln π π π·π‘ 8 5. Numerical Methods “The numerical solution of heat transfer, fluid flow, and other related processes can begin when the laws governing these processes have been expressed in mathematical form, generally in terms of differential equations. The individual differential equations that are encountered express a certain conservation principle. Each equation employs a certain physical quantity as its dependent variable and implies that there must be a balance among the various factors that influence it.” Some examples of differential equations that may be solved through numerical methods are the conservation of energy, conservation of momentum and time averaged equation for turbulent flow (see Section 4). “A numerical solution of a differential equation consists of a set of number from which the distribution of the dependent variable can be constructed.” The goal of computational fluid dynamics (CFD) is to calculate the temperature, velocity, etc. of a fluid at a particular location within a control volume and thus the independent variable in the differential equations is a physical location. Due to computational limitations, the number of locations (also known as grid points or nodes) is finite. ‘By only focusing on the solution of the differential equations at the grid points, the need to find the exact solution of the differential equation has been replaced. The algebraic equations (also known as discretization equations) involving the unknown values of the independent variable at chosen grid points are derived from the differential equations governing the independent variable. In this derivation, assumptions about the value of the independent variable between grid points must be made.’ This concept is known as discretization. “A discretization equation is an algebraic relation connecting the values of the independent variable for a group of grid points. Such an equation is derived from the differential equation governing the independent variable and thus expresses the same physical information as the differential equation. The fact that only a few grid points participate in a given discretization equation is a consequence of the piecewise nature of the profiles chose. The value of the independent variable at a grid point thereby influences the distribution of the independent variable only in its immediate neighborhood. As the number of grid points becomes very large, the, solution of the discretization equations is expected to approach the exact solution of the corresponding differential equation. This follows from the consideration that, as the grid points get 9 closer together, the change in phi between neighboring grid points becomes small and the actual details of the profile assumption become unimportant.” This is where the term “mesh independent” comes from. If there are too few grid points, the profile assumptions impact the solution and the solution is not mesh independent. To ensure that the results are not dependent on the profile assumptions, the solution should be checked for mesh independence. “The most common procedure for deriving the discretization equations is the through a truncated Taylor series.” Other methods for deriving the discretization equations include variational formulation, method of weighted residuals and control volume formulation. “In the iterative solution of the algebraic equation or in the overall iterative scheme employed for handling nonlinearity, it is often desirable to speed up or to slow down the changes, from iteration to iteration, in the values of the dependent variable. This process is called overrelaxation or underrelaxation, depending on whether the variable changes are accelerated or slowed down. Underrelaxation is a very useful device for nonlinear problems. It is often employed to avoid divergence in the iterative solution of strongly nonlinear equations.” ANSYS Fluent offers numerous spatial discretization solvers for the various independent variables such as gradients, pressure, flow, momentum, turbulence, energy, etc. One of the most common spatial discretization solvers is the upwind scheme which was first proposed by Courant, Isaacson, and Rees in 1952. “It is part of a class of numerical discretization methods for solving hyperbolic partial differential equations. Upwind schemes use an adaptive or solution-sensitive finite difference stencil to numerically simulate the direction of propagation of information in a flow field. The upwind schemes attempt to discretize partial differential equations by using differencing biased in the direction determined by the sign of the characteristic speeds (Wikipedia).” Other methods are available in Fluent however this is the most commonly used solver. Other options include QUICK, power law and third-order MUSCL. 10 6. Natural Convection 6.1 Introduction Convection can be defined as the transport of mass and energy by potential gradients and by bulk fluid motion. If the fluid motion is induced by some external force, it is generally referred to as forced convection (Convective mass and heat transfer). Natural convection is a mechanism, or type of heat transport, in which the fluid motion is not generated by any external source (like a pump, fan, suction device, etc.) but driven by buoyancy-induced motion resulting from body forces acting on density gradients, which, in turn, arise from mass concentrations and or temperature gradients in the fluid. (Convective mass and heat transfer). In natural convection, fluid surrounding a heated surface absorbs energy, becomes less dense, and rises. Then, the surrounding, cooler fluid moves in to take its place. The cooler fluid is then heated and the process continues, forming a convection current that continuously removes energy from the heated surface. Figure 5.1-1: Natural Convection Currents in a Pot In nature, natural convection cells occur everywhere from oceanic currents to air rising above sunlight-warmed land. Most weather patterns are created by natural convection. Natural convection also takes place in many engineering applications such as home heating radiators that use fins to distribute heat and computer chips. 11 6.2 Theory Like forced convection, natural convection also builds boundary layers on the surfaces of solid bodies. The equations that govern forced convection are essentially the same as those for external laminar flow free convection. For natural convection problems, it is easiest to assume that the fluid is Newtonian, meaning a fluid whose stress vs. strain rate curve is linear, and the flow field is two dimensional and steady. The continuity equation for two-dimensional natural convection is as follows: πΏ(ππ’) πΏ(ππ£) + =0 πΏπ₯ πΏπ¦ The momentum equation is similar to that shown in Section 4 except it contains the gravitational body force term (-ρg): ππ’ πΏπ’ πΏπ’ ππ πΏ πΏπ’ + ππ£ =− − ππ + (π ) πΏπ¦ πΏπ¦ ππ₯ πΏπ¦ πΏπ¦ The pressure gradient term is due to the hydrostatic pressure field outside the boundary layer and can be written as: ππ = −π∞ π ππ₯ Therefore the momentum equation becomes: ππ’ πΏπ’ πΏπ’ πΏ πΏπ’ + ππ£ = π(π∞ − π) + (π ) πΏπ¦ πΏπ¦ πΏπ¦ πΏπ¦ The energy equation is: πππ’ πΏπ πΏπ πΏ πΏπ + πππ£ = (π ) πΏπ₯ πΏπ¦ πΏπ¦ πΏπ¦ A more practical way to measure the amount of natural convection occurring is to use the Grashof number. The Grashof number, Gr, is a dimensionless number that approximates the ratio of the buoyancy to viscous forces acting on a fluid. The Grashof number can be calculated by: πΊπ = ππ½(ππ − π∞ )πΏ3 π2 where β is the thermal expansion coefficient: 1 πΏπ π½=− ( ) π πΏπ π 12 The importance of buoyancy forces in a mixed convection flow (i.e., forced and natural convection are occurring simultaneously) can be measured by the ratio of the Grashof and Reynolds numbers: πΊπ ππ½ΔππΏ = 2 π π π2 When this number approaches or exceeds unity, there are strong buoyancy contributions to the flow. Conversely, if the ratio is very small, buoyancy forces may be ignored. In pure natural convection, the strength of the buoyancy-induced flow is measured by the Rayleigh number: π π = πΊπππ The Prandtl number, Pr, is a dimensionless number that approximates the ratio of momentum diffusivity (kinematic viscosity) to thermal diffusivity. ππ = π£ πΆπ π = πΌ π the thermal diffusivity (α) is: πΌ= π πππ Rayleigh numbers less than 108 indicate a buoyancy-induced laminar flow, with transition to turbulence occurring over the range of 108 < Ra < 1010. (ANSYS Help Theory Guide) 13 6.3 Examples of Natural Convection Two of the simplest forms of natural convection are a fluid surrounding a horizontal cylinder and fluid surrounding a vertical plate. Both will be examined in the following subsections. 6.3.1 Modeling of a Horizontal Cylinder In the first example, a uniformly heated cylinder is submerged in an infinite pool. The cylinder is slightly warmer than the surrounding fluid and therefore energy passes from the cylinder to the surrounding fluid. As the fluid absorbs the energy, its temperature begins to increase. The fluid temperature gradient as calculated by Fluent is shown in Figure 5.3.1-1. Figure 5.3.1-1: Horizontal Cylinder Temperature Plot When the temperature increases, the fluid expands and the density decreases. The hottest fluid is that in direct contact with the cylinder. As the density decreases, buoyancy forces take affect and the warmer, less dense fluid begins to rise. The density changes can be seen in Figure 5.3.1-2. Notice that even some distance away from the cylinder there is a density change. This is caused by small amounts of conduction within 14 the fluid which causes small density changes in the fluid not even in contact with the heated cylinder. Figure 5.3.1-2: Horizontal Cylinder Density Plot As the fluid rises, it separates from the cylinder and new, colder fluid takes its place. When the warm fluid rises, it loses energy to the surrounding, cooler bulk fluid. As this heat transfer process occurs the buoyancy driving head diminishes causing the fluid to climb more slowly until it eventually stops. At this point it is pushed to the side by the fluid travelling upwards below it and begins to sink since it is denser than the fluid that surrounds it. This motion begins to forms a large convection cell which eventually returns the fluid to the bottom where it is reheated by the cylinder. This can be seen in the velocity vector plot shown in Figure 5.3.1-3 15 Figure 5.3.1-3: Horizontal Cylinder Velocity Vector Plot Figure 5.3.1-3 shows the magnitude and direction of the fluid in the pool with a submerged horizontal cylinder. It can be seen that the as the cylinder heats the surrounding fluid, it travels around the cylinder, eventually separating and rising. An interesting point is when the fluid loses heat to its surrounding it begins to fall. Some of that falling liquid is heated by the warmer fluid rising and it begins to rise without being heated by the cylinder. This creates a small convection cell about two diameters above the heated cylinder. This process continues ad infinitum as long as there is a temperature gradient (i.e., buoyancy driving head). To ensure that the ANSYS Fluent generated figures are correct, the results were compared to experimental results. The following figure shows isotherms surrounding a horizontal tube in natural convection flow as revealed by an interference photograph. 16 Figure 5.3.1-4: Horizontal Cylinder Experimental Isotherm Plot Show Figure 93, page 167 from Introduction to the Transfer of Heat and Mass The ANSYS Fluent calculated isotherms for a horizontal cylinder is shown in Figure 5.3.1-5: Figure 5.3.1-5: Horizontal Cylinder Isotherm Plot The modeling of a horizontal cylinder submerged in an infinite pool using ANSYS Fluent 14.0 shows good resemblance to experimental data. Figure 5-5 and 5-6 show comparable results. Both have isotherms that extend away from the plate and grow in distance away from one another as they get farther from the plate. Because the isotherms calculated by ANSYS Fluent are similar to those found experimentally, it can confidently be stated that the results for a uniformly heated horizontal cylinder submerged in a pool are reasonable. 17 6.3.2 Modeling of a Vertical Plate The second example is similar to the first, except it involves a uniformly heated vertical flat plate submerged in an infinite pool. Like the cylinder, the plate is also slightly warmer than the surrounding fluid and therefore energy passes from the plate to the surrounding fluid. The main difference between the flat plate and the cylinder example is that in the flat plate example, the fluid has more time in contact with the plate as it rises and is therefore able to absorb more energy. As the fluid absorbs the energy, its temperature begins to increase. Because the fluid is heated as it rises along the plate, the fluid temperature is greater in the flat plate example than in the horizontal cylinder example. Figure 5.3.2-1: Vertical Plate Temperature Plot Since heat is exchanged between the plate and the fluid, a thermal boundary layer is created. Thermodynamic equilibrium demands that the fluid in contact with plate is equal to the temperature of the plate. The region in which the temperature changes from the plate surface temperature to that of the bulk fluid is known as the thermal boundary layer. Notice how the thermal boundary layer is small at the bottom of the plate and much larger at the top. The thermal boundary layer expands as the momentum boundary layer expands which helps pull warm fluid away from the hot plate. 18 For more information on thermal and momentum boundary layers, see Convective Heat and Mass Transfer (Reference XX). Figure 5.3.2-2: Vertical Plate Velocity Vector Plot Figure 5.3.2-2 shows how the fluid velocity is highly vertical and increases as it travels up the plate. This is caused by the fluid having lengthy contact time with the heated surface creating a greater temperature gradient and therefore a large buoyancy force. Comparing the vertical flat plate to the horizontal cylinder, it is expected that the vertical plate would have a greater maximum fluid velocity because the fluid is in contact with the heated surface longer. The maximum fluid velocity for the vertical plate is 0.0149 m/s while the maximum fluid velocity for the horizontal cylinder is 0.0177 m/s. This is quite counterintuitive. The reason why the horizontal cylinder actually has a larger maximum velocity is because the buoyancy driving head is allowed to work freely without any drag from the plate. Although the plate is continuing to heat the fluid as it travels up the plate, the velocity is limited due to friction. For this reason, the plate actually has a smaller maximum velocity. To ensure that the ANSYS Fluent generated figures are correct, the results were compared to experimental results. The following figure shows isotherms surrounding a vertical plate in natural convection flow as revealed by an interference photograph. 19 Figure 5.3.2-3: Vertical Plate Experimental Isotherm Plot Show Figure 94, page 168 from Introduction to the Transfer of Heat and Mass The ANSYS Fluent calculated isotherms for a heated vertical plate is shown in the Figure 5.3.2-4. Figure 5.3.2-4: Vertical Plate Isotherm Plot Figure 5.3.2-3 and Figure 5.3.2-4 show comparable results. Both have isotherms that extend away from the plate and grow in distance away from one another as they get farther from the heated surface. Not only do the isotherms match expectations, the momentum boundary layer calculated by ANSYS Fluent also matches experimental expectations. Figure 5.3.2-5: Experimental Momentum Boundary Layer Show Figure 17-2, page 371 from Convective Heat and Mass Transfer 20 Figure 5.3.2-6: Momentum Boundary Layer Plot The trends shown in Figure 5.3.2-5 and 5.3.2-6 are comparable. Having similar isotherms and momentum boundary layers, it can confidently be stated that the results determined by ANSYS Fluent for a uniformly heated vertical plate submerged in a pool are reasonable. 21 7. Laminar Flow with Heat Transfer “Single-phase fluid flow can be characterized into two categories, laminar or turbulent flow. Laminar flow implies that the fluid moves in sheets, or “laminae,” that slip relative to each other (Introduction to Thermal and Fluids Engineering). Laminar flow occurs at very low velocities where there are only small disturbances and little to no local velocity variations. At low velocities, the fluid tends to flow without lateral mixing, and adjacent layers slide past one another easily. There are no cross currents perpendicular to the direction of flow, nor eddies or swirling of fluid. In laminar flow, the motion of the fluid particles is very orderly. In fluid dynamics, laminar flow is a flow regime characterized by high momentum diffusion and low momentum convection (Wikipedia)”. Turbulent flow, the other category of single-phase flow, is discussed in Section 7. The Reynolds number is used to characterize the flow regime. The Reynolds number, Re, is a dimensionless number that gives a measure of the ratio of inertial forces (fluid resists change in motion) to viscous forces. This helps to quantify the relative importance of these two types of forces for given flow conditions. (Wikipedia) The Reynolds number can be calculated using the following equation: Re = ρVA μ For internal flow, such as within a pipe, laminar flow is characterized by flow with a Reynolds number less than 2300 whereas turbulent flow is characterized by a flow with a Reynolds number greater than 4000. For flow with a Reynolds number between 2300 and 4000, both laminar and turbulent flows are possible. This is called transition flow. The velocity of laminar flow in a pipe is can be calculated by: π’= ππ 2 ππ π2 (− ) (1 − 2 ) 4π ππ₯ ππ However, it is friendlier to express the velocity in terms of the mean velocity, V. π’ = 2π (1 − π2 ) ππ 2 The energy equation for flow through a circular pipe assuming symmetric heat transfer, fully developed flow and constant fluid properties is: 22 πΏπ 1πΏ πΏπ πΏ 2π π’ = πΌ[ (π ) + 2 ] πΏπ₯ π πΏπ πΏπ πΏπ₯ Using ANSYS Fluent, a simple laminar flow problem through a cylindrical pipe with uniform surface termperature was developed. The Reynolds number for the problem was 352 which means the flow in the laminar regime. One of the most notable characteristics of laminar flow is the velocity profile parabolic shape which can be calculated using the equation for velocity above. Figure 6-1 shows the velocity magnitude versus distance from the pipe centerline for various distances from the pipe entrance. The distance from the pipe entrance is given in the legend. For example, “line-10cm” shows the velocity profile 10 cm from the pipe entrance. As the flow developes, i.e., the entrance effects dissipate, the velocity profile becomes more and more parabolic until it reaches a steady state 45 cm from the entrance. This is proven by the fact that the outlet, 50 cm from the entrance, and the velocity profile 45 cm from the entrance are basically the same. 23 Figure 6-1: Laminar Flow Velocity Profile Another characteristic of laminar flow is the lack of mixing that occurs within the fluid as it travels through the pipe. Laminar flow is considered to move in “sheets” and each fluid molecule or atom tends to stay about the same distance from the centerline as it travels through the pipe. This is shown by viewing the temperature profile below. Diffusion and conduction are the primary forms of heat transfer in laminar flow. Notice how the fluid thermal boundary layer grows slowly as it travels down the pipe towards the centerline. Figure 6-2: Laminar Flow Temperature Plot Figure 6-3 shows the radial flow velocity. The greatest radial velocity occurs at the entrance and exit of the pipe. However, as expected, because laminar flow moves in 24 “sheets,” the radial velocity for the middle of the pipe is near zero meaning there is little mixing. Figure 6-4: Laminar Flow Radial Velocity Plot Laminar flow also tends to create large momentum boundary layers which cause frictional force on the wall. The drag force on the wall is shown in Figure 6-5. Figure 6-5: Laminar Flow Wall Shear Stress The wall stress is much larger in the first 5 cm due to entrance effects. Once the entrance effects dissipate, the wall shear stress slowly decreases as the flow becomes more and more parabolic. At the very end, around 49 cm, the wall shear stress begins to increase due to the pipe exit. 25 8. Turbulence In fluid dynamics, turbulence or turbulent flow is a flow regime characterized by chaotic and stochastic property changes. It exists everywhere in nature from the jet stream to the oceanic currents. Giving an exact definition of turbulence is rather difficult. Instead, it is easier to describe a turbulent flow. Turbulent flows are highly irregular or random which makes a deterministic approach to turbulence problems impossible. They have high diffusivity, meaning there is rapid mixing and increased rates of momentum and heat and mass transfer. They all have large Reynolds numbers and contain three-dimensional vorticity fluctuations. The unsteady vortices appear on many scales and interact with each other generating high levels of mixing. Finally, turbulent flows are dissipative. Viscous shear stresses perform deformation work which increases the internal energy of the fluid at the expense of kinetic energy. Because turbulence cannot maintain itself, it depends on its environment to obtain energy. A common source of energy for turbulent velocity fluctuations is shear in the mean flow; other sources, such as buoyancy, exist too. If turbulence arrives in an environment where there is no shear or other maintenance mechanisms, the turbulence decays. The Reynolds number decreases and the flow tends to become laminar again (A First Course in Turbulence). In flows that are originally laminar, turbulence arises from instabilities at large Reynolds numbers. As stated previously, laminar flow starts to transition to turbulent at Reynolds numbers around 2100 (A First Course in Turbulence). Although laminarturbulent transition is not governed by Reynolds number, the same transition occurs if the size of the object is gradually increased, or the viscosity of the fluid is decreased, or if the density of the fluid is increased (Wikidpedia). One of the most common examples of the transition of laminar flow to turbulent flow is smoke rising from a cigarette. 26 http://askphysics.com/wp-content/uploads/2012/01/cigarette.gif As the smoke leaves the cigarette it travels upward in a laminar fashion as shown by the single stream of smoke. At a certain distance, the Reynolds number becomes too large and the flow begins to transition into the turbulent regime. When this happens, the flow becomes more random and mixes with the air causing the smoke to dissipate. Because of its irregular nature, modeling turbulence can be difficult. It requires the solution to the Navier-Stokes equations (see conservation of momentum equation in Section 4). The Reynolds-averaged Navier–Stokes equations (or RANS equations) are timeaveraged equations of motion for fluid flow. The idea behind the equations is Reynolds decomposition, whereby an instantaneous quantity is decomposed into its time-averaged and fluctuating quantities, an idea first proposed by Osborne Reynolds. The RANS equations are primarily used to describe turbulent flows. These equations can be used with approximations based on knowledge of the properties of flow turbulence to give approximate time-averaged solutions to the Navier–Stokes equations (Wikipedia). In Reynolds averaging, the solution variables in the instantaneous (exact) NavierStokes equations are decomposed into the mean (ensemble-averaged or time-averaged) and fluctuating components. For the velocity components: π’π = π’Μ π + π’π′ where π’Μ π and π’π′ are the mean and fluctuating velocity components (i = 1, 2, 3). Likewise, for pressure and other scalar quantities: π = πΜ + π ′ 27 where π denotes a scalar such as pressure, energy, or species concentration. Substituting expressions of this form for the flow variables into the instantaneous continuity and momentum equations and taking a time (or ensemble) average (and dropping the overbar on the mean velocity, π’Μ ) yields the ensemble-averaged momentum equations. They can be written in Cartesian tensor form as: πΏπ πΏ (ππ’π ) = 0 + πΏπ‘ πΏπ₯π πΏ πΏ πΏπ πΏ πΏπ’π πΏπ’π 2 πΏπ’π πΏ ′ ′ Μ Μ Μ Μ Μ Μ (ππ’π ) + (ππ’π π’π ) = − + [π ( + − πππ )] + (−ππ’ π π’π ) πΏπ‘ πΏπ₯π πΏπ₯π πΏπ₯π πΏπ₯π πΏπ₯π 3 πΏπ₯π πΏπ₯π The two equations above are called the Reynolds-averaged Navier-Stokes (RANS) equations. They have the same general form as the instantaneous Navier-Stokes equations, with the velocities and other solution variables now representing ensembleaveraged (or time-averaged) values. Additional terms now appear that represent the ′ ′ Μ Μ Μ Μ Μ Μ effects of turbulence. These Reynolds stresses,−ππ’ π π’π , must be modeled in order to close the second equation. One way that the Reynolds stress is modeled is using the k-Ο΅ turbulence model. The k-Ο΅ model was first introduced by Harlow and Nakayama in 1968. (F. H. Harlow and P. I. Nakayama, Transport of turbulence energy decay rate, Los Alamos Sci. Lab., LA-3854, 1968). The k-Ο΅ model has become the most widely used model for industrial applications because of its overall accuracy and small computational demand. The k in the k-Ο΅ model stands for the turbulent kinetic energy and the Ο΅ stands for its dissipation rate. Turbulent kinetic energy is the mean kinetic energy per unit mass associated with eddies in turbulent flow. Physically, the turbulence kinetic energy is characterized by measured root-mean-square (RMS) velocity fluctuations (Wikipedia). Epsilon (Ο΅) is the rate of dissipation of the turbulence energy per unit mass. Two-equation turbulence models allow the determination of both, a turbulent length and time scale by solving two separate transport equations. The standard k-Ο΅ model in falls within this class of models and has become the workhorse of practical engineering flow calculations in the time since it was proposed by Launder and Spalding. Robustness, economy, and reasonable accuracy for a wide range of turbulent flows explain its popularity in industrial flow and heat transfer simulations. It is a semi- 28 empirical model, and the derivation of the model equations relies on phenomenological considerations and empiricism. The standard k-Ο΅ model is a model based on model transport equations for the turbulence kinetic energy (k) and its dissipation rate (Ο΅). The model transport equation for k is derived from the exact equation, while the model transport equation for Ο΅ was obtained using physical reasoning and bears little resemblance to its mathematically exact counterpart. In the derivation of the k-Ο΅ model, the assumption is that the flow is fully turbulent, and the effects of molecular viscosity are negligible. The standard kΟ΅ model is therefore valid only for fully turbulent flows. As the strengths and weaknesses of the standard k-Ο΅ model have become known, modifications have been introduced to improve its performance. Over the years improvements have been made to the Standard k-Ο΅ model. The improvements helped to create a new model known as the Realizable k-Ο΅ model. The Realizable k-Ο΅ model differs from the Standard k-Ο΅ model in two important ways. First, realizable model contains an alternative formulation for the turbulent viscosity. Second, a modified transport equation for the dissipation rate, Ο΅, has been derived from an exact equation for the transport of the mean-square vorticity fluctuation. The term “realizable” means that the model satisfies certain mathematical constraints on the Reynolds stresses, consistent with the physics of turbulent flows. 8.1 Calculating Turbulence Parameters When using the Realizable k-Ο΅ in ANSYS Fluent certain parameters need to be established to properly set the initial and boundary conditions of the problem. The following calculations were performed to determine the boundary condition and initial condition inputs for the turbulence model. Mass Flow Rate: 0.5 kg/s (randomly chosen flow rate that will give turbulent flow) Pipe Diameter (D): 0.03 m Viscosity (μ): 0.001003 kg/m-s Density (ρ): 998.2 kg/m3 29 Turbulence Empirical Constant (Cμ) = 0.09 [1] Hydraulic Diameter (Dh): π· 2 4 ∗ π΄ π ∗ (2) π·β = = = π· = 0.03 π π 4∗π∗π· Flow Area (A): π· 2 0.03 π 2 π΄ = π∗( ) =π∗( ) = 0.00070686 π2 2 2 Average Flow Velocity (uavg): π’ππ£π = πΜ = π∗π΄ 0.5 ππ/π 998.2 ππ ∗ 0.00070686 π2 π3 = 0.708631 π π Reynolds Number (ReDh): π ππ·β πΜπ·β = = ππ΄ ππ 0.5 π ∗ 0.03 m = 21157 ππ 2 0.001003 π − π ∗ 0.00070686 π Turbulence Length Scale (l): π = 0.07 ∗ π·β = 0.07 ∗ 0.03 π = 0.0021 π Turbulent Intensity (I): − 1 8 1 πΌ = 0.16 ∗ π ππ·β = 0.16 ∗ 21157−8 = 0.0460721 Turbulent Kinetic Energy (k): π= 2 3 3 π π2 2 (π’ππ£π ∗ πΌ) = (0.708631 ∗ 0.0460721) = 0.00159885 2 2 2 π π Dissipation Rate (Ο΅): ε= 3/4 k Cπ 3/2 3/4 π = 0.09 0.001598853/2 0.0021 Using the above calculated turbulent parameters, the following was determined. Below is a figure of the velocity magnitude versus distance from the centerline for various distances from the pipe entrance. The distance from the pipe entrance is given in the legend. For example, “line-10cm” shows the velocity profile 10 cm from the pipe entrance. The figure shown below is very difference from the similar figure for laminar 30 flow. First the velocity profiles are much flatter. The parabolic shape seen in the laminar flow plot is not as clear. Second, the flow reaches the steady state velocity profile quicker for turbulent flow. The flow is stable at about 40 cm where it takes until about 45 cm to become stable for the laminar flow for the same pipe geometry. Below is the wall shear stress versus distance from the entrance. Notice how the shear stress is very large in the beginning and quickly reduces by about 3 cm. The large shear stress at the beginning is due to entrance effects. 31 Also note that the distance needed to dissipate increased shear stress due to entrance effects is much shorter than for laminar flow. It took the laminar flow about 5 cm to dissipate the impact of the entrance on the shear stress. Once the entrance effects dissipate the wall shear stress slowly decreases as the flow becomes more and more stable. At about 22 cm the shear stress reduces again. This can be explained by boundary layer separation. The structure and location of boundary layer separation often changes, sometimes resulting in a reduction of overall drag (Wikipeida). At the very end of the pipe, around 49 cm, the wall shear stress begins to increase. This is caused by the pipe exit. Another very different feature of turbulent flow compared with laminar flow is the mixing that occurs radially. The figure below shows the radial flow velocity. The greatest radial velocity occurs at the entrance and exit of the pipe. This is also true of laminar flow. However, the radial velocity is non-zero in the middle of the pipe meaning that mixing is occurring. 32 The radial velocity figure for turbulent flow is very different than that for the laminar flow. That is because laminar flow is characterized by moving in “sheets” while turbulent flow is characterized by random motion. The next two figures show the turbulent kinetic energy and the production of turbulent kinetic energy. Notice how all of the turbulent kinetic energy is near the wall. This is because the wall helps generate turbulent kinetic energy. The shape of the production of turbulent kinetic energy versus distance from the entrance has the same trend as that of the wall shear stress. This makes sense because shear stress, caused by the wall, produces turbulent kinetic energy. 33 The entrance effects also help to produce turbulent kinetic energy which is why the most amount of turbulent kinetic energy is produced at the wall near the entrance. 34 9. Turbulence with Heat Transfer The problem of turbulent flow can be further complicated by adding energy transfer. The flow components with heat transfer are basically the same as that for solely turbulent flow however temperature changes and therefore density changes do have a slight impact. The radial energy equation given in Section 4 can be applied for turbulent flow as well as laminar flow. The ANSYS Fluent problem discussed in Section 7 was expanded to include a constant surface heat flux. The figure below shows the fluid temperature change caused by the constant surface heat flux. The turbulent kinetic energy is the same between the two cases which is expected since turbulent kinetic energy is momentum based not thermal based. However, if the heat transfer was large enough that buoyance effects began to influence the flow, then there would be an increase in turbulent kinetic energy in the example with heat transfer. The radial velocity is the same between the two cases, which is expected since the radial velocity is caused by turbulent flow. 35 The velocity profiles for each distance from the entrance except that of the entrance is shifted up a small amount. The heat transfer that occurs causes the density of the fluid to decrease. To keep a constant mass flow through the pipe, the velocity increases slightly. The shear stress is the same which is expected since it is based solely on the flow and momentum boundary layer. 36 37 10.Two-Phase Flow A large number of flows encountered in nature and technology are a mixture of phases. Physical phases of matter are gas, liquid, and solid, but the concept of phase in a multiphase flow system is applied in a broader sense. In multiphase flow, a phase can be defined as an identifiable class of material that has a particular inertial response to and interaction with the flow and the potential field in which it is immersed. For example, different-sized solid particles of the same material can be treated as different phases because each collection of particles with the same size will have a similar dynamical response to the flow field. Multiphase flow regimes can be grouped into four categories: gas-liquid or liquid-liquid flows; gas-solid flows; liquid-solid flows; and three-phase flows. The area that will be focused on is gas-liquid flow and more specifically gasliquid flow caused by boiling heat transfer. (Theory Guide). Boiling heat transfer is defined as a mode of heat transfer that occurs with a change in phase from liquid to gas. There are two basic types of boiling, pool boiling and flow boiling. Flow boiling is boiling in a flowing stream of fluid, where the heating surface may be the channel wall confining the flow. Since the heat transfer rate in boiling is generally very high, it has been used to cool devices requiring high heat transfer rates such as nuclear reactors. There are four regimes of boiling shown in the figure below. http://www.thermalfluidscentral.org/e-resources/download.php?id=1 The first regime of boiling is natural convection boiling, up to point A. During this regime, no bubbles form. Instead, heat is transferred from the surface to the bulk 38 fluid by natural convection (http://www.thermalfluidscentral.org/e5/4 resources/download.php?id=1). The heat transfer rate is proportional to π₯ππ ππ‘ (Tong). The second regime of boiling, from point A to point C, is called nucleate boiling. During this stage vapor bubbles are generated at certain preferred locations on the heated surface called nucleation sites. Nucleation sites are often microscopic cavities or cracks in the surface (http://www.thermalfluidscentral.org/e-resources/download.php?id=1). When the liquid near the wall superheats it evaporates, forming bubbles at the various nucleation sites. The bubbles transport the latent heat of the vaporization and also increase the convective heat transfer by agitating the liquid water near the heated surface. There are two subregimes of nucleate boiling. The first is local boiling which is boiling in a subcooled liquid, where the bubbles form at the heating surface tend to condense locally. Subcooled boiling is discussed further in Section 11. Bulk boiling is nucleate boiling in a saturate liquid; in this case, the bubbles do not collapse (Tong). Nucleate boiling has very high heat transfer rates for only small temperature difference between the bulk fluid and the heated surface. For this reason it is considered the most efficient boiling regime for heat transfer. As the heated surface increases in temperature, more and more nucleation sites become active. The bubbles begin to merge together and form columns and slugs of gas, thus decreasing the contact area between the bulk fluid and the heated surface. The decrease in contact area causes the slope of the line to decrease until a maximum is reached (point C). Point C is referred to as the critical heat flux. The vapor begins to form an insulating blanket over the heated surface and thereby dramatically increases the surface temperature. This is called the boiling crisis or departure from nucleate boiling. As the temperature delta increases past the critical heat flux, the rate of bubble generation exceeds the rate of bubble separation. Bubbles at the different nucleation sites begin to merge together and boiling becomes unstable. The surface is alternately covered with a vapor blanket and a liquid layer, resulting in oscillating surface temperatures. This regime of boiling is known as partial film boiling or transition boiling (Tong). If the temperature difference between the surface and the fluid continues to increase, stable film boiling is achieved. At this point, there is a continuous vapor blanket that 39 covers the heated surface and phase change occurs at the liquid-vapor interface instead of the heated surface. During this regime, most heat transfer is carried out by radiation (http://www.thermalfluidscentral.org/e-resources/download.php?id=1) (Advanced Heat and Mass Transfer by Amir Faghri, Yuwen Zhang, and John R. Howell). To know the flow pattern of a two-phase flow is as important as to know whether the flow is laminar or turbulent in a single-phase flow. The number of characteristic flow patterns and the name used for them vary somewhat from investigator to investigator. Some of them include, in order of increasing quality from liquid to gas, are bubbly flow, plug flow, stratification, wave flow, slug flow, annular flow, dispersed flow and fog or mist flow. An image of each of these flow types can be seen in the figure below. Add Tong, Figure 3-2, page 50 The flow patterns stated above can be further classified into three categories, bubbly flow, slug flow and annular flow. Bubbly flow is when the liquid phase in continuous and the vapor phase is discontinuous. The vapor is distributed in the liquid in the form of bubbles. The flow pattern occurs at low void fractions. This is the flow pattern of subcooled boiling. Slug flow is when there are relatively large liquid slugs in the flow as a result of the beginning of agglomeration of vapor bubbles. This flow occurs at moderate void fractions and relatively low flow velocities. Annular flow is when the liquid phase is continuous in an annulus along the wall and the vapor phase is continuous in the core. This flow pattern occurs at high void fractions and high flow velocities. (TONG) The following figure depicts the three flow patterns described above as a fluid is heated while travelling in a pipe. Add Tong, Figure 3-6, page 56 40 10.1 Flow Regimes The first step in solving any multiphase problem is to determine which of the regimes described in Multiphase Flow Regimes in the Theory Guide best represents your flow. Model Comparisons in the Theory Guide provides some broad guidelines for determining appropriate models for each regime, and Detailed Guidelines provides details and how to determine the degree of interphase coupling for flows involving bubbles, droplets, or particles, and the appropriate model for different amounts of coupling. 41 The gas volume fraction and the ratio of liquid velocity to gas velocity determine the flow regime that describes a given flow. Show Figure 1.2 from One-Dimensional Two Phase Flow (p.9) Show Figure 3-4 from TONG (p. 54) The following regimes are gas-liquid: ο· ο· ο· ο· Bubbly flow: This is the flow of discrete gaseous or fluid bubbles in a continuous fluid. Droplet flow: This is the flow of discrete fluid droplets in a continuous gas. Slug flow: This is the flow of large bubbles in a continuous fluid. Stratified/free-surface flow: This is the flow of immiscible fluids separated by a clearly-defined interface. 42 11.Gas Mixing Tank In steelmaking processes, gas injection techniques have been extensively utilized to enhance chemical reaction rates, homogenize temperature and chemical compositions, and remove impurities and nonmetallic inclusions. The advancements made in mixing have increased the amount on control available over the steelmaking process which has led to improved the quality of steel produced. Gas injection uses a porous plug that is located at the bottom of a tank that hold molten metal. The porous plug allows the addition of a gas into the mixing tank at a given velocity and an approximate bubble diameter. The gas travels through the liquid due to buoyancy forces and causes mixing of the liquid due to drag forces. In this model, the Schiller-Nauman drag model is used drag model will be used which is generally acceptable for all multiphase calculations. To better understand how gas injection impacts fluid movement and therefore mixing, CFD modeling is used. This model implements the Eulerian multiphase model with the standard k-Ο΅ turbulence model, a bubble diameter of 0.00005 m and an inlet velocity of 0.05 m/s. 43 44 Mesh Independence: Table 11-1: Mesh Independent Quantities Nodes 2701 Elements 2592 Max Liquid Velocity 0.37243 Max Gas Velocity 0.37249 Max Liquid Dynamic Pressure 72.6646 Max Gas Dynamic Pressure 0.08920 Max Liquid Volume Fraction 0.94933 Min Gas Volume Fraction 0.05067 45 12.Bubble Column A bubble column reactor is an apparatus used for gas-liquid reactions. A bubble column is a vertical column of liquid with gas introduced continuously at the bottom through a sparger. Bubbles form and travel upwards through the column due to the inlet gas velocity and buoyancy. The mixing is done by the gas sparging and it requires less energy than mechanical stirring. Bubble column reactors are characterized by a high liquid content and a moderate phase boundary surface. The bubble column is particularly useful in reactions where the gas-liquid reaction is slow in relation to the absorption rate. Bubble column reactors are often used in industry to develop and produce chemicals and fuels for use in chemical, biotechnology, and pharmaceutical processes. The reactors involve gas-liquid flows where the gas is dispersed as bubbles in a continuous volume of liquid. When reactants are introduced into the flow, the interaction of gas and liquid can cause chemical reactions. Figure 12-1: Bubble Column http://upload.wikimedia.org/wikipedia/commons/thumb/9/93/Bubble_column.svg/200px -Bubble_column.svg.png In all gas-liquid flows, the bubbles can increase and decrease in size due to coalescence and breakup. Coalescence is two or more bubbles colliding, whereby the thin liquid barrier between ruptures to form a larger bubble. Breakup of bubbles is 46 caused by collisions with turbulent eddies, approximately equal in size to the bubbles (BC Master Thesis). The ability to calculate the change in bubble size due to turbulent eddies is discussed in Section 14. The results shown in this section implement the Eulerian multiphase model with the standard k-Ο΅ turbulence model, a bubble diameter of 0.0048 m and an inlet velocity of 0.05 m/s. Figure 12-2 shows a comparison between gas volume fraction at 1 second and 5 seconds after gas has started flowing through the bubble column. Note that in both time points the gas tends to flow in slugs. After 5 seconds the gas has reached the top of the liquid and interaction between the liquid-gas interface has occurred causing it to change shape. It can also be seen that the liquid level after 5 seconds is higher than that after 1 seconds. This shows that the gas flowing through the bubble column, through drag forces and displacement pushes the liquid level higher. Figure 12-2: Instantaneous Gas Volume Fraction (left image: 1 second; right image: 5 seconds) 47 Figure 12.3 shows a comparison between liquid velocity vectors fraction at 1 second and 5 seconds after gas has started flowing through the bubble column. Distinct paths of liquid movement can be seen in both 1 second and 5 seconds. Most of the liquid seems to be pushed along the wall and the center of the column. Figure 12.3: Instantaneous Liquid Velocity Vectors (left image: 1 second; right image: 5 seconds) 48 Figure 12-4 shows a comparison between gas velocity vectors fraction at 1 second and 5 seconds after gas has started flowing through the bubble column. It can be easily observed from Figure 12-4 where the gas particles are after 1 second in the bubble column. It is also interesting that the original gas-liquid interface is not flat but two parabolas. This is most likely due to bubble coalescence and wall drag. The gas velocity after 5 seconds shows that the greatest gas velocity occurs near the walls which is surprising due to friction effects caused by the walls. The greatest velocity is shown at the liquid-gas interface. Figure 12-4: Instantaneous Gas Velocity Vectors (left image: 1 second; right image: 5 seconds) 49 13.Population Balance Equation 13.1 Background “Several industrial fluid flow applications involve a secondary phase with a size distribution. The size distribution of particles, including solid particles, bubbles, or droplets, can evolve in conjunction with transport and chemical reaction in a multiphase system. The evolutionary processes can be a combination of different phenomena like nucleation, growth, dispersion, dissolution, aggregation, and breakage producing the dispersion. Thus in multiphase flows involving a size distribution, a balance equation is required to describe the changes in the particle population, in addition to momentum, mass, and energy balances. This balance is generally referred to as the population balance. Cases in which a population balance could apply include crystallization, precipitative reactions from a gas or liquid phase, bubble columns, gas sparging, sprays, fluidized bed polymerization, granulation, liquid-liquid emulsion and separation, and aerosol flows. To make use of this modeling concept, a number density function is introduced to account for the particle population. With the aid of particle properties (for example, particle size, porosity, composition, and so on), different particles in the population can be distinguished and their behavior can be described.” (ANSYS Fluent PBE Guide, Section 1.0) The population balance model gives the ability to track steam bubbles on a particle size basis after they have detached from a heated wall. The fate of a steam bubble traveling in a subcooled bulk fluid is not well understood. There are a number of possibilities that can occur which include breakup into smaller steam bubbles, coalescence of multiple bubbles into one larger bubble or shrinkage due to transfer of energy from the bubble to the surrounding fluid. The population balance models allows for a more accurate calculation of void fraction during subcooled nucleate boiling. In this analysis the Discrete method is used. “In the discrete method, the particle population is discretized into a finite number of size intervals. This approach has the advantage of computing the particle size distribution (PSD) directly. This approach is also particularly useful when the range of particle sizes is known a priori and does not 50 span more than two or three orders of magnitude. In this case, the population can be discretized with a relatively small number of size intervals and the size distribution that is coupled with fluid dynamics can be computed. The disadvantage of the discrete method is that it is computationally expensive if a large number of intervals is needed.” (ANSYS Fluent PBE Guide Section 1.1) 13.2 Equation Formulation The goal of this section is to present an overview of the theory and governing equations for the methods used to calculate particle growth and nucleation. 13.2.1 Particle State Vector The particle state vector is characterized by a set of external coordinates (π₯), which denote the spatial position of the particle and “internal coordinates” (φ), which could include particle size, composition, and temperature. From these coordinates, a number density function π(π₯, φ, t) can be postulated where φ Ο΅ πΊπ , π₯ π πΊπ₯ . Therefore, the average number of particles in the infinitesimal volume πππ₯ πππ is π(π₯, φ, t) πππ₯ πππ . The total number of particles in the entire system is ∫ ∫ ππππ₯ πππ ππ₯ β ππ The local average number density in physical space ( that is, the total number of particles per unit volume is given by π(π₯, π‘) = ∫ ππππ πΊπ The total volume fraction of all particles is given by πΌ(π₯, π‘) = ∫ π π(π) πππ πΊπ Where π(π) is the volume of a particle in state φ. 51 13.2.2 Population Balance Equation Assuming that φ is the particle volume, the transport equation for the number density function is given as π ππ‘ [π(π, π‘)] + ∇ β [π’ β π(π, π‘)] + ∇π β [πΊπ π(π, π‘)] = π ∫ π 2 0 1 (π − π ′ , π ′ ) π (π − π ′ , π‘) π (π ′ , π‘) ππ ′ ∞ Birth due to Aggregation − ∫0 π (π, π ′ ) π (π, π‘) π (π ′ , π‘) ππ ′ Death due to Aggregation + ∫πΊ ππ (π ′ ) π½ (π|π ′ ) π (π ′ , π‘) ππ ′ Birth due to Breakage −π (π) π (π, π‘) Death due to Breakage π The boundary and initial conditions are given by π (π, π‘ = 0) = ππ ; π(π = 0, π‘) πΊπ = πΜ 0 Where πΜ 0 is the nucleation rate in particles / m3-s. 13.2.3 Particle Growth and Dissolution In the population balance equation given in Section 12.2, ∇π β [πΊπ π(π, π‘)] is the particle growth term. The growth rate is based on particle volume, πΊπ , and therefore surface area. In nucleate boiling, the bulk fluid is subcooled. When steam bubbles form on the heated surface and eventually detach, they travel within the subcooled bulk fluid loosing energy through the steam-liquid interface. Because of this, the growth rate is set to a negative value. 13.2.4 Particle Birth and Death Due to Breakage and Aggregation The birth and death of particles occur due to breakage and aggregation processes. In the case of subcooled nucleate boiling, turbulence plays an important role in the birth and death of steam bubbles. During mixing processes, mechanical energy is supplied to the fluid. This energy creates turbulence within the fluid. The turbulence creates eddies, which in turn help dissipate the energy. The energy is transferred from the largest eddies to the smallest eddies in which it is dissipated through viscous interactions. Particle 52 birth is caused by the breakage of a single large bubble into multiple smaller bubbles due to liquid turbulence eddies. Particle death is due to the coalescence of multiple small bubbles into one larger bubble. The Luo model is used in this analysis because it encompasses both the breakage frequency and the PDF of breaking particles and only requires the specification of surface tension. 13.2.5 Particle Birth by Nucleation Depending on the application, spontaneous nucleation of particles can occur due to the transfer of molecules from the primary phase. In boiling applications, the creation of the first vapor bubbles is a nucleation process referred to as nucleate boiling. There are two types of nucleation sites. The first is formed in a pure liquid and can either be a high energy molecular group or a cavity resulting from a local pressure reduction such as in accelerated flow (cavitation). The other type forms on a foreign object such as a cavity on a wall or a suspended foreign material. In subcooled nucleate boiling, the nucleation sites are created at the cavities of the heated surface. The number of potential nucleation sites is dependent on the surface condition of the heated wall. A very smooth surface has a low number of cavities and therefore a low number of potential nucleation sites. A rough surface has a large number of cavities and therefore a large number of potential nucleation sites. However, just because a heated surface has a high number of potential nucleation sites it does not mean that they are all active nucleation sites. The population of active sites was found to be Μ = π0 exp (− π πΎ 3 ππ€πππ ) Where N0 and K represent the liquid an surface conditions. There is no possible way to predict N0 and K for a particular boiling system. However, it can be seen that the population of active sites is a strong function of wall temperature and therefore heat flux. (TONG) 53 13.3 Solution Method The discrete method (also known as the classes or sectional method) was developed by Hounslow [10] (p. 65), Litster [16] (p. 65), and Ramkrishna [25] (p. 66). It is based on representing the continuous particle size distribution (PSD) in terms of a set of discrete size classes or bins, as illustrated in Figure 12.3-1. The advantages of this method are its robust numerics and that it gives the PSD directly. The disadvantages are that the bins must be defined a priori and that a large number of classes may be required. (ANSYS Fluent PBE Guide Section 2.3.1) Figure 13-1: Particle Size Distribution (ANSYS Fluent PBE Guide Figure 2.1) 54 14.Bubble Column with Population Balance Model The work discussed in Section 12 was extended to include a population balance model. The entire bubble population may not be a single size formed from the inlet at the sparger due to growth, coalescence, and breakup. The implementation of a population balance model allows for the direct calculation of the growth, breakup, and coalescence of bubbles as they travel up the bubble column. The population balance model also does not necessitate estimating the proper characteristic bubble size but rather permits a specified bubble size distribution. The same model used in Section 12 was utilized with the addition of a population balance model with 3 discrete bubble sizes (0.0030 m, 0.0048 m and 0.0076 m). The inlet uses a gas bubble diameter distribution of 25% 0.0030 m, 50% 0.0048 m and 25% 0.0076 m. The coalescence and breakage of bubbles is determined using the luo model with a surface tension of 0.072 N/m Figure 14.1 shows a comparison between gas volume fraction at 1 second and 5 seconds after gas has started flowing through the bubble column. Note that at both time points the gas tends to flow in slugs. After 5 seconds the gas has reached the top of the liquid and interaction between the liquid-gas interface has occurred causing it to change shape. It can also be seen that the liquid level after 5 seconds is higher than that after 1 seconds. This shows that the gas flowing through the bubble column, through drag forces and displacement pushes the liquid level higher. When comparing this to Figure 12.2, there are significant differences. One of the more obvious differences is the distribution of the phases at both 1 second and 5 seconds. With the PBM implemented the phase distribution seems to be much more uniform without any large areas with high gas volume. This is especially noticeable at the bottom. Another difference is the liquid-gas interface. At 5 seconds, the liquid-gas interface appears to have increased in elevation only slightly due to displacement and drag forces compared with the bubble column results without PBM. Another major difference is that after 1 second the gas seems to have traveled farther into the bubble column compared with Figure 12.2. 55 Figure 14-1: Instantaneous Gas Volume Fraction with PBM (left image: 1 second; right image: 5 seconds) Figure 14-2 shows a comparison between liquid velocity vectors at 1 second and 5 seconds after gas has started flowing through the bubble column. Similar to Figure 12-3, there are distinct paths of liquid movement can be seen in both 1 second and 5 seconds. Figure 14-2a shows that the liquid velocity in the upper portion of the column is greater than zero which matches the gas volume fraction results in Figure 14-1. Figure 14-2 shows a more uniform mixing of liquid velocity throughout the bubble column, no sections of little to no movement. The maximum velocity in Figure 14-2 is also less than that of Figure 12-3. 56 Figure 14-2: Bubble Column Liquid Vector Velocity with PBM (left image: 1 second; right image: 5 seconds) Table 14-1 shows the bubble population at different heights in the bubble column. Table 14-1: PBM Bubble Size Bin-0 (0.0076 m) Bin-1 (0.0048 m) Bin-2 (0.0030 m) Inlet (Fraction) Outlet (Fraction) Net (Fraction) 0.250 0.500 0.250 0.865 0.117 0.018 +0.557 -0.308 -0.134 Weighted Average (Fraction) 0.776 0.174 0.051 This shows that as the bubbles travel up the column, the smaller bubbles primarily coalesced into larger bubbles. This means that the turbulent forces were weak and did not break apart the bubbles. 57 15.Pool Boiling “Pool boiling is the type of boiling that occurs when a heater is submerged in a pool of initially stagnant liquid. When the surface temperature of the heater sufficiently exceeds the saturation temperature of the liquid, vapor bubbles nucleate on the heater surface. The bubbles grow rapidly in the superheated liquid layer next to the surface until they depart and move out into the bulk liquid. While rising is the result of buoyancy, the bubbles either collapse or continue their grown, depending upon whether the liquid is locally subcooled or superheated. Thus, in pool boiling, a complex fluid motion around the hater is initiated and maintained by the nucleation growth, departure, and collapse of bubbles; and by natural convection” (TONG p. 5). This model implements the mixture multiphase model with laminar flow. Laminar flow was chosen sine most of the fluid is not moving and there is only an outlet boundary condition (no inlet). The fluid starts 1 degree subcooled and is heated by a wall (at the bottom) that is constant temperature of 10 degrees above saturation. Figure 15.1 shows the instantaneous gas volume fraction after 1 second and 2 seconds of heating. After 1 second, the entire bottom of the control volume is heated and some steam begins to form. Figure 15.2 shows that there is no movement of the liquid and therefore all heat transfer is occurring via conduction. However, just one second later, enough energy has been absorbed by the fluid that buoyancy effects have take affect and the fluid begins to move. The steam created on the bottom surface moves upward and cooler liquid moves down to take its place creating eddies which can also be seen in Figure 15.2. Figure 15.1 shows four nucleation sites on the heated wall where steam is being formed. From these sites a phase changes occurs where bubbles nucleate, grow and detach from the heated surface. 58 Figure 15.1: Instantaneous Gas Volume Fraction (left image: 1 second; right image: 2 seconds) Figure 15.1: Instantaneous Liquid Velocity Vectors 59 (left image: 1 second; right image: 2 seconds) To better understand the vapor production on the heated surface and graphically see the location of nucleation sites, Figure 15.3 was created. Figure 15.3 shows the volume fraction of vapor on the heated surface at 10 seconds. Most of the vapor production occurs on the left half of the heated surface, as can be seen by the spikes in vapor volume fraction from 0 m to 0.15 m. Figure 15.3: Volume Fraction of Vapor on Heated Surface 60 16.Subcooled Boiling Show Figure 3-8 from TONG (p. 60) 61 17.Subcooled Boiling with Population Balance Model 62 18.References [1] ANSYS, Inc., 2012, ANSYS FLUENT 13.0 Theory Guide. [2] 63