Aps Division of Fluid Dynamics Meeting Abstracts, Nov 1, 1997
The lattice Boltzmann method (LBM) has been considered as an alternative numerical method for sol... more The lattice Boltzmann method (LBM) has been considered as an alternative numerical method for solving fluid flows. The standard lattice Boltzmann method uses regularly spaced lattices as grids; thus it cannot handle curved boundaries with desirable accuracy. A finite difference-based lattice Boltzmann method (FDLBM) in curvilinear coordinates is explored using body-fitted coordinates with non-uniform grids. The impulsively started cylindrical Couette flow, steady state cylindrical Couette flow, steady flow over flat plates, and steady flow over a circular cylinder are used to examine various fundamental issues of FDLBM. Those aspects investigated include effects of boundary conditions for the distribution functions on the solution, the merits between second-order central difference and upwind schemes for advection terms, and the effect of the Reynolds number. Favorable results are obtained using FDLBM in curviliner coordinates, indicating that the method is potentially capable of solving finite Reynolds number flow problems in complex geometries.
The method of lattice Boltzmann equation (LBE) is an alternative, kinetic-based method for solvin... more The method of lattice Boltzmann equation (LBE) is an alternative, kinetic-based method for solving fluid flow problems. The standard explicit LBGK scheme simplifies the computation for the particle distribution function, and is suitable for unsteady flow computations. For steady flows, the computation often takes long time to converge due to the explicitness of the scheme. To expedite the convergence rate, it is found that a careful implementation of nested loop approach can significantly improve the convergence to steady state. The idea is to start the LBGK solution on a very coarse grid system. The steady state can be quickly reached due to the use of larger time step and less number of grids. This steady state solution for the particle distribution functions must be carefully transferred to the next fine grid system to ensure that the density and the stresses are the same in the two grids before and after the transfer. For flow inside a 2-D lid driven cavity and flow over a column of cylinders, the use of three level grid systems can result in the speed up by a factor of 4 to 5 comparing with the time it takes to reach the same steady state solution in the finest grids.
Aps Division of Fluid Dynamics Meeting Abstracts, Nov 1, 2000
In the lattice Boltzmann method(LBM), the probability density function(pdf) in the phase space is... more In the lattice Boltzmann method(LBM), the probability density function(pdf) in the phase space is used as the primary variable. In general, the velocity or pressure inlet boundary condition is prescribed as part of problem formulation. Then these conditions are converted to the corresponding pdf there. Two methods have been employed to calculate the pdf with given velocity profiles. One is based on the bounce-back idea, while the other is based on the local state of the open boundary. With the bounce-back scheme, the pdf is affected not only by the inlet velocity but also by the fluid density elsewhere. In LBM, the fluid density is computed based on the pdf, and is not held constant even if physically the fluid density is fixed. Accordingly, the fluctuation of the estimated fluid density in the course of computation will cause the boundary condition to fluctuate also. It is found that, with the bounce-back treatment, the convergence rate can be slow, numerical instabilities may appear, and the situation worsens when the Reynolds number increases. On the other hand, using the local state of the macroscopic condition, the boundary treatment is not sensitive to the fluctuation in the main domain, and the convergence rate becomes more favorable. Examples using flows surrounding moving cylinders and airfoils will be used to help demonstrate the characteristics of these alternative boundary treatments.
Http Dx Doi Org 10 1080 10407780590948927, Sep 2, 2006
ABSTRACT The effect of conjugate heat transfer resulting from a microelectromechanical systems (M... more ABSTRACT The effect of conjugate heat transfer resulting from a microelectromechanical systems (MEMS)-based thermal shear stress is investigated. Due to the length-scale disparity and large solid–fluid thermal conductivity ratio, a two-level computation is used to examine the relevant physical mechanisms and their influences on wall shear stress. The substantial variations in transport properties between the fluid and solid phases and their interplay with regard to heat transfer and near-wall fluid flow structures are investigated. It is demonstrated that for state-of-the-art sensor design, the buoyancy effect can noticeably affect the accuracy of the shear stress measurement.
ABSTRACT We report our investigation on electroosmotic flow (EOF) in a wavy channel between a pla... more ABSTRACT We report our investigation on electroosmotic flow (EOF) in a wavy channel between a plane wall and a sinusoidal wall. An exact solution is obtained by using complex function formulation and boundary integral method. The effects of the channel width and wave amplitude on the electric field, streamline pattern, and flow field are studied. When a pressure gradient of sufficient strength in the opposite direction is added to an EOF in the wavy channel, various patterns of recirculation regions are observed. Experimental results are presented to validate qualitatively the theoretical description. The solution is further exploited to determine the onset condition of flow recirculation and the size of the recirculation region. It is found that they are dependent on one dimensionless parameter related to forces (K, the ratio of the pressure force to the electrokinetic force) and two dimensionless parameters related to the channel geometry (α, the ratio of the wave amplitude to the wavelength, and h, the ratio of the channel width to the wavelength).
The lattice Boltzmann equation (LBE) is an alternative kinetic method capable of solving hydrodyn... more The lattice Boltzmann equation (LBE) is an alternative kinetic method capable of solving hydrodynamics for various systems. Major advantages of the method are due to the fact that the solution for the particle distribution functions is explicit, easy to implement, and natural to parallelize. Because the method often uses uniform regular Cartesian lattices in space, curved boundaries are often approximated by a series of stairs that leads to reduction in computational accuracy. In this work, a second-order accurate treatment of the boundary condition in the LBE method is developed for a curved boundary. The proposed treatment of the curved boundaries is an improvement of a scheme due to O. Filippova and D. Hänel (1998, J. Comput. Phys.147, 219). The proposed treatment for curved boundaries is tested against several flow problems: 2-D channel flows with constant and oscillating pressure gradients for which analytic solutions are known, flow due to an impulsively started wall, lid-driven square cavity flow, and uniform flow over a column of circular cylinders. The second-order accuracy is observed with a solid boundary arbitrarily placed between lattice nodes. The proposed boundary condition has well-behaved stability characteristics when the relaxation time is close to 1/2, the zero limit of viscosity. The improvement can make a substantial contribution toward simulating practical fluid flow problems using the lattice Boltzmann method.
Aps Division of Fluid Dynamics Meeting Abstracts, Nov 1, 1998
A discrete element simulation (DES) for cohesive powder in a simple shear flow is carried out. Th... more A discrete element simulation (DES) for cohesive powder in a simple shear flow is carried out. The effect of the Van der Waals force is characterized by the surface energy and the JKR force model is used for the contact force. Transition of flow regime occurs when the powder concentration becomes nonuniform and the normalized powder flow stresses experience sudden changes as the the shear rate gradually decreases in the simple shear flow simulation. An energy-based cohesion number is derived to characterize the flow transition of the powder. The transition occurs over a quite narrow range of the cohesion number. The effects of the shear rate, bulk concentration, particle material density, particle Young's modulus, particle size, and particle surface energy on the transition are captured using this energy-based cohesion number. The transition results from the insufficent kinetic ernrgy to overcome the cohesive energy upon the collision of particles. In the fluid-like regime, the continuum model based on the kinetic theory of gas for cohesionless granula particles is applicable.
In the lattice Boltzmann method, fluid viscosity is represented by using a relaxation time. For v... more In the lattice Boltzmann method, fluid viscosity is represented by using a relaxation time. For variable viscosity, such as caused by variable temperature, non-Newtonian behavior, or turbulence, local relaxation time has often been used to simulate the local viscosity. The present study attempts to investigate the error in the velocity profile associated with the use of the local relaxation time for the local viscosity. For the pressure driven channel flow, a method is devised to separate the error associated with the effect of the inexact boundary condition from the error caused by variable viscosity. The effect of the variable viscosity is then quantified over a range of computational parameters. For an order-one change in the viscosity across the channel caused by temperature variation, the error associated with the variable viscosity is much higher than the machine error but significantly lower than the error caused by the boundary condition. For a variable viscosity profile that is similar to turbulent eddy viscosity, the error in the velocity profile associated with the variable viscosity can be much higher than the error caused by the boundary condition. Furthermore, an inadequate resolution to represent the rapid near-wall viscosity variation can result in a significant under-prediction in the velocity profile.
A discrete element simulation is carried out to study the behavior of the powder in Couette flow ... more A discrete element simulation is carried out to study the behavior of the powder in Couette flow via the modeling on the microscopic level (particle scale). The cohesive force (due to the van der Waals force and accounted for by using the surface energy) and Hertzian contact force upon particle collision are incorporated in the force-displacement model. The dependence of the powder flow velocity and concentration profiles and the stresses on the average powder concentration, shearing velocity and surface energy is studied. The powder concentration in Couette flow at higher shearing velocity is highly nonuniform due to shear induced migration. Particle mean velocity profile is highly nonlinear. Regions of solid- like and fluid-like behavior can be clearly identified by examining the time-area-averaged particle velocity on the macroscale and the powder cluster structure on the microscale. Thus extreme caution must be exercised in using the nominal shear rate based on the shearing velocity and cell height in powder flow studies. The surface energy has weak (strong) effect on the powder velocity profile at high (low) shearing velocity. It has minor effect on the particle concentration distribution but affects strongly the formation and sustenance of cluster structure.
Solar thermochemical energy storage has enormous potential for enabling cost-effective concentrat... more Solar thermochemical energy storage has enormous potential for enabling cost-effective concentrated solar power (CSP). A thermochemical storage system based on a SrO/SrCO3 carbonation cycle offers the ability to store and release high temperature (≈1200 °C) heat. The energy density of SrCO3 /SrO systems supported by zirconia-based sintering inhibitors was investigated for 15 cycles of exothermic carbonation at 1150 °C followed by decomposition at 1235 °C. A sample with 40 wt % of SrO supported by yttria-stabilized zirconia (YSZ) shows good energy storage stability at 1450 MJ m(-3) over fifteen cycles at the same cycling temperatures. After further testing over 45 cycles, a decrease in energy storage capacity to 1260 MJ m(-3) is observed during the final cycle. The decrease is due to slowing carbonation kinetics, and the original value of energy density may be obtained by lengthening the carbonation steps.
Aps Division of Fluid Dynamics Meeting Abstracts, Nov 1, 1997
The lattice Boltzmann method (LBM) has been considered as an alternative numerical method for sol... more The lattice Boltzmann method (LBM) has been considered as an alternative numerical method for solving fluid flows. The standard lattice Boltzmann method uses regularly spaced lattices as grids; thus it cannot handle curved boundaries with desirable accuracy. A finite difference-based lattice Boltzmann method (FDLBM) in curvilinear coordinates is explored using body-fitted coordinates with non-uniform grids. The impulsively started cylindrical Couette flow, steady state cylindrical Couette flow, steady flow over flat plates, and steady flow over a circular cylinder are used to examine various fundamental issues of FDLBM. Those aspects investigated include effects of boundary conditions for the distribution functions on the solution, the merits between second-order central difference and upwind schemes for advection terms, and the effect of the Reynolds number. Favorable results are obtained using FDLBM in curviliner coordinates, indicating that the method is potentially capable of solving finite Reynolds number flow problems in complex geometries.
The method of lattice Boltzmann equation (LBE) is an alternative, kinetic-based method for solvin... more The method of lattice Boltzmann equation (LBE) is an alternative, kinetic-based method for solving fluid flow problems. The standard explicit LBGK scheme simplifies the computation for the particle distribution function, and is suitable for unsteady flow computations. For steady flows, the computation often takes long time to converge due to the explicitness of the scheme. To expedite the convergence rate, it is found that a careful implementation of nested loop approach can significantly improve the convergence to steady state. The idea is to start the LBGK solution on a very coarse grid system. The steady state can be quickly reached due to the use of larger time step and less number of grids. This steady state solution for the particle distribution functions must be carefully transferred to the next fine grid system to ensure that the density and the stresses are the same in the two grids before and after the transfer. For flow inside a 2-D lid driven cavity and flow over a column of cylinders, the use of three level grid systems can result in the speed up by a factor of 4 to 5 comparing with the time it takes to reach the same steady state solution in the finest grids.
Aps Division of Fluid Dynamics Meeting Abstracts, Nov 1, 2000
In the lattice Boltzmann method(LBM), the probability density function(pdf) in the phase space is... more In the lattice Boltzmann method(LBM), the probability density function(pdf) in the phase space is used as the primary variable. In general, the velocity or pressure inlet boundary condition is prescribed as part of problem formulation. Then these conditions are converted to the corresponding pdf there. Two methods have been employed to calculate the pdf with given velocity profiles. One is based on the bounce-back idea, while the other is based on the local state of the open boundary. With the bounce-back scheme, the pdf is affected not only by the inlet velocity but also by the fluid density elsewhere. In LBM, the fluid density is computed based on the pdf, and is not held constant even if physically the fluid density is fixed. Accordingly, the fluctuation of the estimated fluid density in the course of computation will cause the boundary condition to fluctuate also. It is found that, with the bounce-back treatment, the convergence rate can be slow, numerical instabilities may appear, and the situation worsens when the Reynolds number increases. On the other hand, using the local state of the macroscopic condition, the boundary treatment is not sensitive to the fluctuation in the main domain, and the convergence rate becomes more favorable. Examples using flows surrounding moving cylinders and airfoils will be used to help demonstrate the characteristics of these alternative boundary treatments.
Http Dx Doi Org 10 1080 10407780590948927, Sep 2, 2006
ABSTRACT The effect of conjugate heat transfer resulting from a microelectromechanical systems (M... more ABSTRACT The effect of conjugate heat transfer resulting from a microelectromechanical systems (MEMS)-based thermal shear stress is investigated. Due to the length-scale disparity and large solid–fluid thermal conductivity ratio, a two-level computation is used to examine the relevant physical mechanisms and their influences on wall shear stress. The substantial variations in transport properties between the fluid and solid phases and their interplay with regard to heat transfer and near-wall fluid flow structures are investigated. It is demonstrated that for state-of-the-art sensor design, the buoyancy effect can noticeably affect the accuracy of the shear stress measurement.
ABSTRACT We report our investigation on electroosmotic flow (EOF) in a wavy channel between a pla... more ABSTRACT We report our investigation on electroosmotic flow (EOF) in a wavy channel between a plane wall and a sinusoidal wall. An exact solution is obtained by using complex function formulation and boundary integral method. The effects of the channel width and wave amplitude on the electric field, streamline pattern, and flow field are studied. When a pressure gradient of sufficient strength in the opposite direction is added to an EOF in the wavy channel, various patterns of recirculation regions are observed. Experimental results are presented to validate qualitatively the theoretical description. The solution is further exploited to determine the onset condition of flow recirculation and the size of the recirculation region. It is found that they are dependent on one dimensionless parameter related to forces (K, the ratio of the pressure force to the electrokinetic force) and two dimensionless parameters related to the channel geometry (α, the ratio of the wave amplitude to the wavelength, and h, the ratio of the channel width to the wavelength).
The lattice Boltzmann equation (LBE) is an alternative kinetic method capable of solving hydrodyn... more The lattice Boltzmann equation (LBE) is an alternative kinetic method capable of solving hydrodynamics for various systems. Major advantages of the method are due to the fact that the solution for the particle distribution functions is explicit, easy to implement, and natural to parallelize. Because the method often uses uniform regular Cartesian lattices in space, curved boundaries are often approximated by a series of stairs that leads to reduction in computational accuracy. In this work, a second-order accurate treatment of the boundary condition in the LBE method is developed for a curved boundary. The proposed treatment of the curved boundaries is an improvement of a scheme due to O. Filippova and D. Hänel (1998, J. Comput. Phys.147, 219). The proposed treatment for curved boundaries is tested against several flow problems: 2-D channel flows with constant and oscillating pressure gradients for which analytic solutions are known, flow due to an impulsively started wall, lid-driven square cavity flow, and uniform flow over a column of circular cylinders. The second-order accuracy is observed with a solid boundary arbitrarily placed between lattice nodes. The proposed boundary condition has well-behaved stability characteristics when the relaxation time is close to 1/2, the zero limit of viscosity. The improvement can make a substantial contribution toward simulating practical fluid flow problems using the lattice Boltzmann method.
Aps Division of Fluid Dynamics Meeting Abstracts, Nov 1, 1998
A discrete element simulation (DES) for cohesive powder in a simple shear flow is carried out. Th... more A discrete element simulation (DES) for cohesive powder in a simple shear flow is carried out. The effect of the Van der Waals force is characterized by the surface energy and the JKR force model is used for the contact force. Transition of flow regime occurs when the powder concentration becomes nonuniform and the normalized powder flow stresses experience sudden changes as the the shear rate gradually decreases in the simple shear flow simulation. An energy-based cohesion number is derived to characterize the flow transition of the powder. The transition occurs over a quite narrow range of the cohesion number. The effects of the shear rate, bulk concentration, particle material density, particle Young's modulus, particle size, and particle surface energy on the transition are captured using this energy-based cohesion number. The transition results from the insufficent kinetic ernrgy to overcome the cohesive energy upon the collision of particles. In the fluid-like regime, the continuum model based on the kinetic theory of gas for cohesionless granula particles is applicable.
In the lattice Boltzmann method, fluid viscosity is represented by using a relaxation time. For v... more In the lattice Boltzmann method, fluid viscosity is represented by using a relaxation time. For variable viscosity, such as caused by variable temperature, non-Newtonian behavior, or turbulence, local relaxation time has often been used to simulate the local viscosity. The present study attempts to investigate the error in the velocity profile associated with the use of the local relaxation time for the local viscosity. For the pressure driven channel flow, a method is devised to separate the error associated with the effect of the inexact boundary condition from the error caused by variable viscosity. The effect of the variable viscosity is then quantified over a range of computational parameters. For an order-one change in the viscosity across the channel caused by temperature variation, the error associated with the variable viscosity is much higher than the machine error but significantly lower than the error caused by the boundary condition. For a variable viscosity profile that is similar to turbulent eddy viscosity, the error in the velocity profile associated with the variable viscosity can be much higher than the error caused by the boundary condition. Furthermore, an inadequate resolution to represent the rapid near-wall viscosity variation can result in a significant under-prediction in the velocity profile.
A discrete element simulation is carried out to study the behavior of the powder in Couette flow ... more A discrete element simulation is carried out to study the behavior of the powder in Couette flow via the modeling on the microscopic level (particle scale). The cohesive force (due to the van der Waals force and accounted for by using the surface energy) and Hertzian contact force upon particle collision are incorporated in the force-displacement model. The dependence of the powder flow velocity and concentration profiles and the stresses on the average powder concentration, shearing velocity and surface energy is studied. The powder concentration in Couette flow at higher shearing velocity is highly nonuniform due to shear induced migration. Particle mean velocity profile is highly nonlinear. Regions of solid- like and fluid-like behavior can be clearly identified by examining the time-area-averaged particle velocity on the macroscale and the powder cluster structure on the microscale. Thus extreme caution must be exercised in using the nominal shear rate based on the shearing velocity and cell height in powder flow studies. The surface energy has weak (strong) effect on the powder velocity profile at high (low) shearing velocity. It has minor effect on the particle concentration distribution but affects strongly the formation and sustenance of cluster structure.
Solar thermochemical energy storage has enormous potential for enabling cost-effective concentrat... more Solar thermochemical energy storage has enormous potential for enabling cost-effective concentrated solar power (CSP). A thermochemical storage system based on a SrO/SrCO3 carbonation cycle offers the ability to store and release high temperature (≈1200 °C) heat. The energy density of SrCO3 /SrO systems supported by zirconia-based sintering inhibitors was investigated for 15 cycles of exothermic carbonation at 1150 °C followed by decomposition at 1235 °C. A sample with 40 wt % of SrO supported by yttria-stabilized zirconia (YSZ) shows good energy storage stability at 1450 MJ m(-3) over fifteen cycles at the same cycling temperatures. After further testing over 45 cycles, a decrease in energy storage capacity to 1260 MJ m(-3) is observed during the final cycle. The decrease is due to slowing carbonation kinetics, and the original value of energy density may be obtained by lengthening the carbonation steps.
Uploads