|
National Science FoundationCombined Research and Curriculum DevelopmentMultiphase Transport Phenomena
Charles Petty (PI), Mei Zhuang (co-PI), George Chase (co-PI), Ram Mohan (co-PI), Marilyn Amey (project evaluator) |
Logo by Michael Skeggs |
||||||
 Figure A.1. Axial (top) and Radial (bottom) Velocity Contours in the Longitudinal Symmetry Plane (units: m/s). Click figure to enlarge. |
Pharmacia Design Team A: Michael Skeggs (Akron) and Andrew Yoder (MSU)
Academic Mentor: Charles Petty (MSU)
Industrial Mentor: Mark Widman (Pharmacia Corporation)
CFD Mentors: Chinh Nguyen (MSU), Steven Parks (MSU), and Shiwei Shao (MSU)
This CFD design project examines the flow patterns of an aqueous suspension in a mixing tank that uses a VibromixerŽ. The continuous phase contains surfactants, a suspending agent, and preservatives. Micronized drug particles are suspended in an aqueous phase. The dispersed particles have a mean diameter of ten microns and the largest particles have diameters below twenty microns. The goal is to develop an understanding between the operating characteristics of the mixer and the spatial uniformity of the solid/liquid suspension. Maintaining a spatially uniform liquid suspension of solids in a batch tank during the filling stage and the withdrawal stage is a challenging unit operation in the pharmaceutical industry (Widman and Jonas, 2000).
The VibromixerŽ consists of a shaft and a perforated plate; the perforations in the plate are roughly conical in shape. As the plate is rapidly vibrated up and down, the suspension is pumped through the perforations forming multiple jetlets. The direction of the pumping action depends on the orientation of the plate. Figure A.1 shows the axial and radial components of the velocity predicted by the RANS-equation with a k-e model for the Reynolds stress. For the results illustrated, the vibrator produces a pumping rate equivalent to an average mass flux of 38 kg/m2-s above the surface of the disk vibrator.
 |
|
From Left: Michael Skeggo (Akron), Chinh Nguyen (MSU),
Luis Gomez (Tulsa), Charles Petty (MSU), Mark Hoyack (Krebs), Tim Olson
(Krebs), Andy Yoder (MSU)
|
Design Team B: Dina El-dein (MSU), Nicholas Lynn (MSU), and Michael Schafer
(MSU)
Academic Mentors: Andre Benard (MSU) and Mei Zhuang (MSU)
Industrial Mentor: Ray Rite (The Trane Company)
CFD Mentors: Figen Lacin (MSU) and Dilip Mandel (MSU)
Good distribution of a two-phase refrigerant in heating, ventilating, and air conditioning (HVAC) systems is critical. In an HVAC system, the role of the header is to deliver an equal amount of refrigerant to each heat exchanger circuit without changing the ratio of vapor to liquid (Rite, 2000). This project examines the flow patterns within a vertical, cylindrical header with a single axial influent tube and multiple radial effluent tubes evenly spaced along the header. Figure B.1 illustrates the flow domain for an end-of-header service. The computational grid for the simulation is shown in Figures B.2a and 2b. Figure B.3 shows the streamlines in the longitudinal symmetry plane predicted for a single-phase fluid (i.e., air at 283 oK and 100 kPa) by the RANS-equation with a k-e closure for the Reynolds stress.
|
|
|
|
|
Figure B.1. Definition of Header Design Scales and Sequel Problem. |
Figure B.2a. Computational Grid: Side View. Click figure to enlarge |
 |
 |
|
|
Figure B.2b. Computational Grid: Top View. Click figure to enlarge |
Figure B.3. Streamlines in the Longitudinal Symmetry
Plane. |
 |
|
Back Row, From Left: Dina El-dein (MSU), Figen Lacin
(MSU), Dilip Mandel (MSU)
Front Row, From Left: Michael Schafer (MSU), Nick Lynn (MSU) |
Design Team C: Joshua Herron (Akron), Gregory McColley (MSU), Jin Wang (Tulsa)
Academic Mentor: George Chase (Akron)
Industrial Mentor: Kevin Fontenot (Eastman Chemical)
CFD Mentors: Dilip Mandal (MSU) and Brian Raber (Akron)
Many processes within the chemical and biochemical industries utilize three-phase slurry bubble columns. In some cases, these columns may be sixty feet high and only six feet in diameter (Fontenot, 2000). Mixing the contents of these large-scale reactors at relatively low power, albeit challenging, can be accomplished by natural convection. For example, a low-density fluid can be created in the lower portion of a column reactor by mixing air with the continuous phase to form a bubbly fluid that rises due to buoyancy. In the upper regions of the column, transport of gas across the gas/liquid interface decreases the local concentration of the light dispersed phase. This unstable situation (i.e., heavy fluid on top of light fluid) supports a convective flow between the bottom and the top of the column. Gaslift columns with large aspect ratios (i.e., H/D >> 1) may contain an internal annulus or some other means to guide the natural recirculation flow.
The objective of this project is to model a slurry bubble column to determine the effect of height/diameter ratio and gas load on the spatial distribution of the solid/gas/liquid suspension. Figure C.2 illustrates the natural convective currents for an analogous thermal column of water with H/D = 5 and ÷DT˝=80 oK predicted by the RANS-equation and the Reynolds averaged internal energy balance with a k-e closure for the Reynolds stress and the turbulent energy flux.
 |
 |
|
|
Figure C.1 Gaslift Column. Click figure to enlarge. |
Figure C.2 Magnitude of the Velocity in the Symmetry
Plane (m/s). |
 |
|
From Left: Brian Raber (Akron), Josh Herron (Akron),
George Chase (Akron), Jin Wang (Tulsa), Paul Gillis (Dow Chemical), Kevin
Fontenot (Eastman Chemical), Greg McColley (MSU)
|
|
|
Design Team D: Radu Dunca (MSU), Seth Jentner (Akron), and Jose Severino (Tulsa)
Academic Mentor: Ram Mohan (Tulsa)
Industrial Mentor: Gene Kouba (Chevron)
CFD Mentor: Ferhat Erdal (Tulsa) and Steve Parks (MSU)
A large (80,000 barrels) wet crude oil gravity separation tank is used to separate water from crude oil so the water can be safely returned to the environment (see Figure D.1). The feedstream may contain more than 20% water by volume. As more production wells are added and existing wells mature, the total flow rate and the ratio of water to oil in the feed stream increase. The oil/water dispersion is introduced into the tank through an array of pipes located near the bottom, as illustrated by Figure D.2 (Kouba, 2000).
A major challenge in the design is the prediction of the distribution of the oil and water dispersion over the cross sectional area of the tank near the inlet manifold. The goal of this project is to estimate the performance of the separator as the flow rate and water cut increase. Figure D.3 shows the streamlines of the continuous phase (density = 875 kg/m3; viscosity = 0.00823 kg/m-s) at the first split. The RANS -equation with a k-e closure for the Reynolds stress was used in the simulation.
 |
 |
|
|
Figure D.2 Overhead Diagram of Inlet Spreader Configuration. |
Figure D.3 Streamlines in the Longitudinal Symmetry Plane (streamlines are color-coded based on the local speed of the fluid, m/s ; flow in main line is from right to left). Click figure to enlarge. |
 |
|
Back Row, From Left: Ram Mohan (Tulsa), Ferhat Erdal
(Tulsa), Radu Dunca (MSU), Seth Jentner (Akron)
Front Row, From Left: Jose Severino (Tulsa), Jeff Henning (AEA) |
Design Team E: Luis Gomez (Tulsa), Hongmin Li (Akron), and Steve Parks (MSU)
Academic Mentor: Charles Petty (MSU)
Industrial Mentors: Mark Hoyack and Tim Olson (Krebs Engineers)
The geometry of a hydrocyclone determines the internal flow behavior of the continuous phase and the separation of the dispersed phase. The objective of this study is to simulate the internal flows within two different hydrocyclones to understand the influence of shape on solid/liquid separation (Hoyack and Olson, 2000). The study focuses on the conical section of the hydrocyclone. Figure E.1 shows the relative shapes of the two hydrocyclones and Figure E.2 gives a top view of the computational grid (58,000 computational elements for the 20o cyclone and 66,000 for the 10.5o cyclone). Figure E.3 shows the tangential velocity of the continuous phase (water) predicted by the RANS-equation with an RNG k-e closure for the Reynolds stress. Figure E.4 shows the axial velocity for the same calculation.
 |
 |
|
|
Figure E.1. Definition of Geometric Scales (top: 20o
cone, bottom: 10.5 o cone). Click figure to enlarge.
|
Figure E.2. Computational Grid, Top View (75 axial divisions
for the 20o cyclone, 85 divisions for the 10.5 o cone). Click figure to
enlarge.
|
 |
 |
|
|
Figure E.3. Magnitude of the Tangential Component of
the Velocity in the Longitudinal Symmetry Plane (top: 20o cone, bottom:
10.5 o cone). Click figure to enlarge.
|
Figure E.4 Magnitude of the Axial Component of the Velocity
in the Longitudinal Symmetry Plane (top: 20o cone, bottom: 10.5 o cone).
Click figure to enlarge.
|
|
|
|
From Left: Michael Skeggo (Akron), Chinh Nguyen (MSU),
Luis Gomez (Tulsa), Charles Petty (MSU), Mark Hoyack (Krebs), Tim Olson
(Krebs), Andy Yoder (MSU)
|
