Particulate Modeling with ANSYS CFD · Particulate Modeling with ANSYS CFD ... alpha-alumina...
Transcript of Particulate Modeling with ANSYS CFD · Particulate Modeling with ANSYS CFD ... alpha-alumina...
© 2011 ANSYS, Inc. September 7, 20111
Particulate Modeling with ANSYS CFD
Mohan Srinivasa
© 2011 ANSYS, Inc. September 7, 20112
Introduction
Review models for particulate flows
Models for dense particulate flows
Example applications
Questions
Agenda
© 2011 ANSYS, Inc. September 7, 20113
Introduction
© 2011 ANSYS, Inc. September 7, 20114
Spans wide range of
• Length scales
• Time scales
Physics
• Particulate physics
• Fluid particle interaction
• Particle size distribution
• Homogenous and heterogeneous reaction
• Particle structure interaction
Particulate Flows: Issues and Physics
From: Fundamental of Multiphase Flow, C. E. Brennen
© 2011 ANSYS, Inc. September 7, 20115
A Platform for Modeling Particulate Flows
© 2011 ANSYS, Inc. September 7, 20116
From: E Loth, Particles, drops and bubblesFluid dynamics and numericalmethods
Fluid Solid Interactions
© 2011 ANSYS, Inc. September 7, 20117
For Dense Phase Flows …
Effect of particle volume on gas phase
• Excluded volume
© 2011 ANSYS, Inc. September 7, 20118
Fluid phase
• Almost always Eulerian!
Particle phase
• Eulerian or Lagrangian
Size of particles with respect to fluid grid
• Sub-grid or super grid
• Particle fluid interaction
– Modeled or resolved
Particle-particle interactions
• Modeled or resolved
Models for Dense Particulate Flows
© 2011 ANSYS, Inc. September 7, 20119
http://www.deakin.edu.au/itri/cmfi/research/areas/cfd.php
For Example …
© 2011 ANSYS, Inc. September 7, 201110
Particle phase• Treated in a multi-fluid framework
– Ensemble and time averaged over particles to arrive at PDE
– Maximum packing
• Cell based
– Volume fraction, velocity, temperature
Particle-particle interactions• Modeled. Kinetic theory, frictional
models
• Granular pressure, granular viscosity
Particle fluid interactions• Empirical models
Euler – Granular model
© 2011 ANSYS, Inc. September 7, 201111
Prediction of Solid Inversion
Hollow char Hollow charGlass beads Glass beads
Char and glass beads are well mixed atthe low liquid velocity of 2.5mm/s
Inversion of char and glass beadsat a high liquid velocity of 9mm/s
© 2011 ANSYS, Inc. September 7, 201112
Conversion of methane to synthesis gas is an important chemical process
Modeling of partial oxidation of methane in presence of Ni-alpha-alumina catalyst is carried out in a bubbling fluidized bed reactor.
Simulation results were compared to experimental data obtained from literature
Catalytic Oxidation of Methane
© 2011 ANSYS, Inc. September 7, 201113
Simulation Results
Figure 3 Effect of feed composition on methane conversion
40
50
60
70
80
90
100
1 1.5 2 2.5 3 3.5 4 4.5
pCH4/pO2
XC
H4
, %
CFD prediction
Experimental data (Mleczko & Wurzel, 1997)
Figure 4 Variation of species concentration with height above the gas distributor
0
10
20
30
40
50
60
70
-0.01 0.04 0.09 0.14 0.19 0.24
Height (m)V
ol %
CFD-CH4
CFD-CO
CFD-H2
Expt-CH4
Expt-CO
Expt-H2
© 2011 ANSYS, Inc. September 7, 201114
NETL Fluidization Challenge
© 2011 ANSYS, Inc. September 7, 201115
Framework to track a dispersed phase in a Lagrangian framework
• Account for the presence of the dispersed phase in the equations of motion for the primary phase
• Treat fluid-particle interactions in an efficient manner
Particle-particle interactions
• Collision breakup model for droplets
• Expression for solid pressure – KTGF
• Explicit particle interaction – DEM
• Other models …
Particle fluid interactions
• Empirical models
DDPM
KTGF based collision
© 2011 ANSYS, Inc. September 7, 201116
• Deeper bed has a tendency to exhibit gas streaming (larger fluctuations in DP) compared to shallow bed (Issangya et. al. 2007). DDPM model qualitatively predicts the trend as observed in experiments.
NETL Challenge:Effect of Depth on Fluidization
© 2011 ANSYS, Inc. September 7, 201117
Axial Pressure Gradient Profile
Effect of Fines on Fluidization
© 2011 ANSYS, Inc. September 7, 201118
© 2011 ANSYS, Inc. September 7, 201119
DEM
Particle-particle interactions – DEM
• Resolved using a spring dashpot model
• Can model size dependent phenomenon
– Brazil nut effect
• Model Particles as parcels or individual particles
• Accurate modeling of particle-particle interaction
• Independent of fluid mesh!
© 2011 ANSYS, Inc. September 7, 201120
© 2011 ANSYS, Inc. September 7, 201121
Hard sphere analogy
Particle phase
• Lagrangian representation of BIG particles
• Particles ideally span several cells
Particle-particle interactions
• Resolved
Particle fluid interactions
• Determined as part of the solution
Macroscopic Particle ModelAvailable as a UDF
© 2011 ANSYS, Inc. September 7, 201122
© 2011 ANSYS, Inc. September 7, 201123
Summary: Models for Particulate Flows
Model Numerical approach Particle fluid interaction
Particle-Particle interaction
Particle size distribution
DPM Fluid – Eulerian Particles – Lagrangian
Empirical models for sub-grid particles
Particles are treated as points
Easy to include PSD because of Lagrangian description
DDPM - KTGF Fluid – Eulerian Particles – Lagrangian
Empirical; sub-gridparticles
Approximate P-P interactions determined by granular models
Easy to include PSD because of Lagrangian description
DDPM - DEM Fluid – Eulerian Particles – Lagrangian
Empirical; sub-grid particles
Accurate determinationof P-P interactions.
Can account for all PSD physics accurately including geometric effects
Euler Granular model
Fluid – EulerianParticles – Eulerian
Empirical; sub-grid particles
P-P interactions modeled by fluidproperties, such as granular pressure, viscosity, drag etc.
Different phases to account for a PSD; when size change operations happen use population balance models
MacroscopicParticle Model
Fluid – Eulerian Particles – Lagrangian
Interactions determined as part of solution; particles span many fluid cells
Accurate determinationof P-P interactions.
Easy to include PSD; if particles become smaller than the mesh, uses an empiricial model