Tutorial 5 Fluent

Tutorial 5.
Modeling Radiation and Natural Convection
Introduction: In this tutorial, combined radiation and natural convection are solved in a two-dimensional square box on a mesh consisting of quadrilateral elements. In this tutorial you will learn how to: Use the radiation models in FLUENT (Rosseland, P-1, DTRM, discrete ordinates (DO), and surface-to-surface (S2S)) and understand their ranges of application Use the Boussinesq model for density Set the boundary conditions for a heat transfer problem involving natural convection and radiation Separate a single wall zone into multiple wall zones Change the properties of an existing uid material Calculate a solution using the segregated solver Display velocity vectors and contours of stream function and temperature for ow visualization Prerequisites: This tutorial assumes that you are familiar with the menu structure in FLUENT, and that you have solved Tutorial 1. Some steps in the setup and solution procedure will not be shown explicitly. Problem Description: The problem to be considered is shown schematically in Figure 5.1. A square box of side L has a hot right wall at T = 2000 K, a cold left wall at T = 1000 K, and adiabatic top and bottom walls. Gravity points downwards. A buoyant ow develops because of thermally-induced density gradients. The medium 5-1
c Fluent Inc. November 29, 2001
contained in the box is assumed to be absorbing and emitting, so that the radiant exchange between the walls is attenuated by absorption and augmented by emission in the medium. All walls are black. The objective is to compute the ow and temperature patterns in the box, as well as the wall heat ux, using the radiation models available in FLUENT, and to compare their performance for dierent values of the optical thickness aL. The working uid has a Prandtl number of approximately 0.71, and the Rayleigh number based on L is 5 105 . This means the ow is inherently laminar. The Boussinesq assumption is used to model 3 ) is 0.02, and measures buoyancy. The Planck number k/(4LT0 the relative importance of conduction to radiation; here T0 = (Th + Tc )/2. Three values for the optical thickness are considered: aL = 0, aL = 0.2, and aL = 5. Note that the values of physical properties and operating conditions (e.g., gravitational acceleration) have been adjusted to yield the desired Prandtl, Rayleigh, and Planck numbers.
Adiabatic
= 1000 kg/m3 cp= 1.1030x10

4
J/kgK
y x
k = 15.309 W/mK -3 = 10 kg/ms -5 = 10 1/K -5 g = -6.96 x 10 m/s 2 a = 0, 0.2, 5 1/m L=1m Ra = 5 x 10 Pr = 0.71 Pl = 0.02
5
Tc = 1000 K
T = 2000 K
= 0.2, 5
Figure 5.1: Schematic of the Problem
5-2
Preparation
1. Copy the le rad/rad.msh from the FLUENT documentation CD to your working directory (as described in Tutorial 1). 2. Start the 2D version of FLUENT.
Step 1: Grid
1. Read the mesh le rad.msh. File Read Case... As the mesh is read in, messages will appear in the console window reporting the progress of the reading. The mesh size will be reported as 2500 cells. 2. Check the grid. Grid Check FLUENT performs various checks on the mesh and reports the progress in the console window. Pay particular attention to the minimum volume. Make sure this is a positive number.
5-3
3. Display the grid (Figure 5.2). Display Grid...
Note: All the walls are currently contained in a single wall zone, wall-4. You will need to separate them out into four dierent walls so that you can specify dierent boundary conditions for each wall.
5-4
Grid
Jul 25, 2001 FLUENT 6.0 (2d, segregated, lam)
Figure 5.2: Graphics Display of Grid
4. Separate the single wall zone into four wall zones. Grid Separate Faces...
5-5
(a) Select the Angle separation method (the default) under Options. (b) Select wall-4 in the Zones list. (c) Specify 89 as the signicant Angle. (d) Click on the Separate button. Faces with normal vectors that dier by more than 89 will be placed in separate zones. Since the four wall zones are perpendicular (angle = 90 ), wall-4 will be separated into four zones. 5. Display the grid again. (a) Select all Surfaces and click on Display. Notice that you now have four dierent wall zones instead of only one. Extra: You can use the right mouse button to check which wall zone number corresponds to each wall boundary. If you click the right mouse button on one of the boundaries in the graphics window, its zone number, name, and type will be printed in the FLUENT console window. This feature is especially useful when you have several zones of the same type and you want to distinguish between them quickly. In some cases, you may want to disable the display of the interior grid so as to more accurately select the boundaries for identication.
5-6
Step 2: Models
As discussed earlier, in this tutorial you will enable each radiation model in turn, obtain a solution, and postprocess the results. You will start with the Rosseland model, then use the P-1 model, the discrete transfer radiation model (DTRM), and the discrete ordinates (DO) model. At the end of the tutorial, you will use the surface-to-surface (S2S) model. 1. Keep the default solver settings. Dene Models Solver...
5-7
2. Turn on the Rosseland radiation model. Dene Models Radiation...
When you click OK in the Radiation Model panel, FLUENT will present an Information dialog box telling you that new material properties have been added for the radiation model. You will be setting properties later, so you can simply click OK in the dialog box to acknowledge this information. Note: FLUENT will automatically enable the energy calculation when you enable a radiation model, so you need not visit the Energy panel.
5-8
3. Add the eect of gravity on the model. Dene Operating Conditions...
(a) Turn on Gravity. The panel will expand to show additional inputs. (b) Set the Gravitational Acceleration in the Y direction to -6.94e-5 m/s2 . As mentioned earlier, the gravitational acceleration has been adjusted to yield the correct dimensionless quantities (Prandtl, Rayleigh, and Planck numbers). See Figure 5.1 and the associated comments. (c) Set the Operating Temperature to 1000 K. The operating temperature will be used by the Boussinesq model, which you will enable in the next step.
5-9
Step 3: Materials
The default uid material is air, which is the working uid in this problem. However, since you are working with a ctitious uid whose properties have been adjusted to give the desired values of the dimensionless parameters, you must change the default properties for air. You will use an optical thickness aL of 0.2 for this calculation. (Since L = 1, the absorption coecient a will be set to 0.2.) Later in the tutorial, results for an optically thick medium with aL = 5 and non-participating medium with aL = 0 are computed to show how the dierent radiation models behave for dierent optical thicknesses. Dene Materials...
5-10
1. Select boussinesq in the drop-down list next to Density, and then set the Density to 1000 kg/m3 . For details about the Boussinesq model, see the Users Guide. 2. Set the specic heat, Cp, to 1.103e4 J/kg-K. 3. Set the Thermal Conductivity to 15.309 W/m-K. 4. Set the Viscosity to 0.001 kg/m-s. 5. Set the Absorption Coecient to 0.2 m1 . Hint: Use the scroll bar to access the properties that are not initially visible in the panel. 6. Keep the default settings for the Scattering Coecient and the Scattering Phase Function, since there is no scattering in this problem. 7. Set the Thermal Expansion Coecient (used by the Boussinesq model) to 1e-5 K1 . 8. Click on Change/Create and close the Materials panel.
5-11
Step 4: Boundary Conditions

Dene Boundary Conditions... 1. Set the boundary conditions for the bottom wall (wall-4.006). Note: The bottom wall should be called wall-4.006, but to be sure that you have the correct wall, use your right mouse button to click on the bottom wall in the graphics window. When you do this, the corresponding zone will be selected automatically in the Zone list in the Boundary Conditions panel. You can do this when you set boundary conditions for the other walls as well, to be sure that you are dening the correct conditions.
(a) Change the Zone Name to bottom. (b) Retain the default thermal conditions (heat ux of 0) to specify an adiabatic wall.
5-12
Note: The Rosseland model does not require you to set a wall emissivity. Later in the tutorial, you will need to dene the wall emissivity for the other radiation models. 2. Set the boundary conditions for the left wall, wall-4. (a) Change the Zone Name to left. (b) Select Temperature under Thermal Conditions and set the Temperature to 1000 K. 3. Set the boundary conditions for the right wall, wall-4:007. (a) Change the Zone Name to right. (b) Select Temperature under Thermal Conditions and set the Temperature to 2000 K. 4. Set the boundary conditions for the top wall, wall-4:005. (a) Change the Zone Name to top. (b) Retain the default thermal conditions (heat ux of 0) to specify an adiabatic wall.
5-13
Step 5: Solution for the Rosseland Model

1. Set the parameters that control the solution. Solve Controls Solution...
(a) Retain the default selected Equations (all of them) and UnderRelaxation Factors. (b) Under Discretization, select PRESTO! for Pressure, and Second Order Upwind for Momentum and Energy.
5-14
2. Initialize the ow eld. Solve Initialize Initialize...
(a) Set the Temperature to 1500 K and click on Init.
5-15
3. Enable the plotting of residuals during the calculation. Solve Monitors Residual...
(a) Under Options, select Plot. (b) Click OK. Note: There is no extra residual for the radiation heat transfer because the Rosseland model does not solve extra transport equations for radiation; instead, it augments the thermal conductivity in the energy equation. When you use the P-1 and DO radiation models, which both solve additional transport equations, you will see additional residuals for radiation.
5-16
4. Save the case le (rad ross.cas). File Write Case... 5. Start the calculation by requesting 200 iterations. Solve Iterate... The solution will converge in about 180 iterations. 6. Save the data le (rad ross.dat). File Write Data...
5-17
Step 6: Postprocessing for the Rosseland Model

1. Display velocity vectors (Figure 5.3). Display Vectors...
5-18
2.11e-04 1.90e-04 1.68e-04 1.47e-04 1.26e-04 1.05e-04 8.42e-05 6.32e-05 4.21e-05 2.11e-05 2.61e-09
Velocity Vectors Colored By Velocity Magnitude (m/s)
Figure 5.3: Velocity Vectors for the Rosseland Model
5-19
2. Display contours of stream function (Figure 5.4). Display Contours...
The recirculatory patterns observed are due to the natural convection in the box. At a low optical thickness (0.2), radiation should not have a large inuence on the ow. The ow pattern is expected to be similar to that obtained with no radiation (Figure 5.5). However, the Rosseland model predicts a ow pattern that is very symmetric (Figure 5.4), and quite dierent from the pure natural convection case. This discrepancy occurs because the Rosseland model is not appropriate for small optical thickness.
5-20
6.95e-02 6.26e-02 5.56e-02 4.87e-02 4.17e-02 3.48e-02 2.78e-02 2.09e-02 1.39e-02 6.95e-03 0.00e+00
Contours of Stream Function (kg/s)
Figure 5.4: Contours of Stream Function for the Rosseland Model
Extra: If you want to compute the results without radiation yourself, turn o all the radiation models in the Radiation Model panel, set the under-relaxation factor for energy to 0.8, and calculate until convergence. (Remember to reset the underrelaxation factor to 1 before continuing with the tutorial).
5-21
1.97e-02 1.77e-02 1.57e-02 1.38e-02 1.18e-02 9.84e-03 7.87e-03 5.90e-03 3.94e-03 1.97e-03 0.00e+00
Contours of Stream Function (kg/s)
Figure 5.5: Contours of Stream Function with No Radiation
5-22
3. Display lled contours of temperature (Figure 5.6). Display Contours...
The Rosseland model predicts a temperature eld (Figure 5.6) very dierent from that obtained without radiation (Figure 5.7). For the low optical thickness in this problem, the temperature eld predicted by the Rosseland model is not physical.
5-23
2.00e+03 1.90e+03 1.80e+03 1.70e+03 1.60e+03 1.50e+03 1.40e+03 1.30e+03 1.20e+03 1.10e+03 1.00e+03
Contours of Static Temperature (k)
Figure 5.6: Contours of Temperature for the Rosseland Model
2.00e+03 1.90e+03 1.80e+03 1.70e+03 1.60e+03 1.50e+03 1.40e+03 1.30e+03 1.20e+03 1.10e+03 1.00e+03
Contours of Static Temperature (k)
Nov 28, 2001 FLUENT 6.0 (2d, segregated, lam)
Figure 5.7: Contours of Temperature with No Radiation
5-24
4. Plot the y velocity along the horizontal centerline of the box. (a) Create an isosurface at y = 0.5, the horizontal line through the center of the box. Surface Iso-Surface...
i. Select Grid... in the Surface of Constant drop-down list and select Y-Coordinate from the list below. ii. Click on Compute to see the extents of the domain. iii. Set a value of 0.5 in the Iso-Values eld, and change the New Surface Name to y=0.5. iv. Click on Create to create a surface at y = 0.5.
5-25
(b) Create an XY plot of y velocity on the isosurface. Plot XY Plot...
i. Check that the Plot Direction for X is 1, and the Plot Direction for Y is 0. With a Plot Direction vector of (1,0), FLUENT will plot the selected variable as a function of x. Since you are plotting the velocity prole on a cross-section of constant y , the x direction is the one in which the velocity varies. ii. Select Velocity... and Y Velocity under Y Axis Function. iii. Select y=0.5 in the Surfaces list. iv. Click on Plot.
5-26
y=0.5
2.50e-04 2.00e-04 1.50e-04 1.00e-04 5.00e-05
Y Velocity (m/s)
0.00e+00 -5.00e-05 -1.00e-04 -1.50e-04 -2.00e-04 -2.50e-04 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position (m)
Y Velocity
Figure 5.8: XY Plot of Centerline y Velocity for the Rosseland Model The velocity prole reects the rising plume at the hot right wall, and the falling plume at the cold left wall. Compared to the case with no radiation, the prole predicted by the Rosseland model exhibits thicker wall layers. As discussed before, the expected prole for aL = 0.2 is similar to the case with no radiation. (c) Save the plot data to a le. i. Select the Write to File option, and click the Write... push button. ii. In the resulting Select File dialog box, specify rad ross.xy in the XY File text entry box and click OK.
5-27
5. Compute the total wall heat ux on each lateral wall. Report Fluxes...
(a) Select Total Heat Transfer Rate under Options. (b) Select right and left under Boundaries. (c) Click the Compute button. The total wall heat transfer rate is reported for the hot and cold walls as approximately 7.43 105 W. The sum of the heat uxes on the lateral walls is a negligible imbalance. 6. Save the case and data les (rad ross.cas and rad ross.dat). File Write Case & Data... Thus far in this tutorial, you have learned how to set up a natural convection problem using the Rosseland model to compute radiation. You have also learned to postprocess the results. You will now turn on the P-1 model and compare the results so computed with those of the Rosseland model. 5-28
Step 7: P-1 Model Denition, Solution, and Postprocessing

You will now repeat the above calculation using the P-1 radiation model. The main steps are identical to the procedure described above for the Rosseland model. 1. Enable the P-1 model. Dene Models Radiation... 2. Conrm that the wall emissivity is 1 for all walls. Dene Boundary Conditions... For each wall boundary, there will be a new entry, Internal Emissivity, in the Thermal section of the Wall panel. Retain the default value of 1. 3. Modify the under-relaxation parameters. Solve Controls Solution... (a) Under Under-Relaxation Factors, set the factor for P1 to 1.0, and retain the default factors for Pressure, Momentum, and Energy (0.3, 0.7, and 1.0). Note that an additional equation, P1, appears because the P-1 model solves an additional radiation transport equation. This problem is relatively easy to converge for the P-1 model since there is not much coupling between the radiation and temperature equations at low optical thicknesses. Consequently a high under-relaxation factor can be used for P-1. 4. Save the case le (rad p1.cas). File Write Case... 5. Continue the calculation by requesting another 200 iterations. Solve Iterate... The P-1 model reaches convergence after about 115 additional iterations.
5-29
6. Save the data le (rad p1.dat). File Write Data... 7. Examine the results of the P-1 model calculation. Note: The steps below do not include detailed instructions because the procedure is the same one that you followed for the Rosseland model postprocessing. See Step 6: Postprocessing for the Rosseland Model if you need more detailed instructions. (a) Display velocity vectors (Figure 5.9). Display Vectors...
2.87e-04 2.58e-04 2.29e-04 2.01e-04 1.72e-04 1.43e-04 1.15e-04 8.61e-05 5.75e-05 2.89e-05 2.27e-07
Figure 5.9: Velocity Vectors for the P-1 Model (b) Plot the y velocity along the horizontal centerline (Figure 5.10), and save the plot data to a le called rad p1.xy. Plot XY Plot... ! You will need to reselect Y Velocity under Y Axis Function. Also, remember to turn o the Write to File option so that you can access the Plot button to generate the plot.
5-30
y=0.5
2.50e-04 2.00e-04 1.50e-04 1.00e-04 5.00e-05 0.00e+00
Y Velocity (m/s)
-5.00e-05 -1.00e-04 -1.50e-04 -2.00e-04 -2.50e-04 -3.00e-04 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position (m)
Y Velocity
Figure 5.10: XY Plot of Centerline y Velocity for the P-1 Model (c) Compute the total wall heat transfer rate. Report Fluxes ... The total heat transfer rate reported on the right wall is 8.47 105 W. The heat imbalance at the lateral walls is negligibly small. You will see later that the Rosseland and P-1 wall heat transfer rates are substantially dierent from those obtained by the DTRM and the DO model. Notice how dierent the velocity vectors and y -velocity prole are from those obtained using the Rosseland model. The P-1 velocity proles show a clear momentum boundary layer along the hot and cold walls. These proles are much closer to those obtained from the non-radiating case (Figures 5.11 and 5.12). Though the P-1 model is not appropriate for this optically thin limit, it yields the correct velocity proles since the radiation source in the energy equation, which is proportional to the absorption coecient, is small. The Rosseland model uses an eective conductivity to account for radiation, and yields the wrong temperature eld, which in turn results in an erroneous velocity eld.
5-31
2.16e-04 1.94e-04 1.73e-04 1.51e-04 1.29e-04 1.08e-04 8.63e-05 6.47e-05 4.31e-05 2.16e-05 8.78e-09
Figure 5.11: Velocity Vectors with No Radiation
y=0.5
2.50e-04 2.00e-04 1.50e-04 1.00e-04 5.00e-05
Y Velocity (m/s)
0.00e+00 -5.00e-05 -1.00e-04 -1.50e-04 -2.00e-04 -2.50e-04 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position (m)
Y Velocity
Figure 5.12: XY Plot of Centerline y Velocity with No Radiation
5-32
Step 8: DTRM Denition, Solution, and Postprocessing

1. Turn on the discrete transfer radiation model (DTRM) and dene the ray tracing. Dene Models Radiation...
(a) Select Discrete Transfer under Model. The panel will expand to show additional inputs.
5-33
(b) Accept the defaults by clicking OK. The Ray Tracing panel will open automatically.
(c) Accept the default settings for Clustering and Angular Discretization by clicking OK. When you click OK, FLUENT will open a Select File dialog box so you can specify a name for the ray le used by the DTRM. A detailed description of the ray tracing procedure can be found in the Users Guide. In brief, the number of Cells Per Volume Cluster and Faces Per Surface Cluster control the total number of radiating surfaces and absorbing cells. For a small 2D problem, the default number of 1 is acceptable. For a large problem, however, you will want to increase these numbers to reduce the ray tracing expense. The Theta Divisions and Phi Divisions control the number of rays being created from each surface cluster. For most practical problems, the default settings will suce. (d) In the Ray File text entry box in the Select File dialog box, enter rad dtrm.ray for the name of the ray le. Then click OK. FLUENT will print an informational message describing the progress of the ray tracing procedure.
5-34
2. Retain the current under-relaxation factors for pressure, momentum, and energy (0.3, 0.7, and 1.0). Solve Controls Solution... 3. Save the case le (rad dtrm.cas). File Write Case... 4. Continue the calculation by requesting another 100 iterations. Solve Iterate... The solution will converge after about 80 additional iterations. 5. Save the data le (rad dtrm.dat). File Write Data... 6. Examine the results of the DTRM calculation. Note: The steps below do not include detailed instructions because the procedure is the same one that you followed for the Rosseland model postprocessing. See Step 6: Postprocessing for the Rosseland Model if you need more detailed instructions. (a) Display velocity vectors (Figure 5.13). Display Vectors...
5-35
2.88e-04 2.59e-04 2.30e-04 2.02e-04 1.73e-04 1.44e-04 1.15e-04 8.65e-05 5.77e-05 2.90e-05 2.08e-07
Figure 5.13: Velocity Vectors for the DTRM
5-36
(b) Plot the y velocity along the horizontal centerline (Figure 5.14), and save the plot data to a le called rad dtrm.xy. Plot XY Plot... ! You will need to reselect Y Velocity under Y Axis Function. Also, remember to turn o the Write to File option so that you can access the Plot button to generate the plot.
y=0.5
2.50e-04 2.00e-04 1.50e-04 1.00e-04 5.00e-05 0.00e+00
Y Velocity (m/s)
-5.00e-05 -1.00e-04 -1.50e-04 -2.00e-04 -2.50e-04 -3.00e-04 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position (m)
Y Velocity
Figure 5.14: XY Plot of Centerline y Velocity for the DTRM (c) Compute the total wall heat transfer rate. Report Fluxes ... The total heat transfer rate reported on the right wall is 6.06 105 W. Note that this is substantially lower than the values predicted by the Rosseland and P-1 models.
5-37
Step 9: DO Model Denition, Solution, and Postprocessing

1. Turn on the discrete ordinates (DO) radiation model and dene the angular discretization. Dene Models Radiation...
(a) Select Discrete Ordinates under Model. The panel will expand to show additional inputs for the DO model. (b) Set the number of Flow Iterations Per Radiation Iteration to 1. This is a relatively simple ow problem, and will converge easily. Consequently it is useful to do the DO calculation every iteration of the ow solution. For problems that are dicult to 5-38
converge, it is sometimes useful to allow the ow solution to establish itself between radiation calculations. In such cases, it may be useful to set Flow Iterations Per Radiation Iteration to a higher value, such as 10. (c) Retain the default settings for Angular Discretization and NonGray Model. For details about the angular discretization used by the DO model, see the Users Guide. The Number of Bands for the Non-Gray Model is zero because only gray radiation is being modeled in this tutorial. Note: When you click OK in the Radiation Model panel, FLUENT will present an Information dialog box telling you that new material properties have been added for the radiation model. The property that is new for the DO model is the refractive index, which is relevant only when you are modeling semi-transparent media. Since you are not modeling semi-transparent media here, you can simply click OK in the dialog box to acknowledge this information. 2. Retain the current under-relaxation factors for pressure, momentum, and energy (0.3, 0.7, and 1.0), as well as the default underrelaxation of 1 for the discrete ordinates transport equation. Solve Controls Solution... 3. Save the case le (rad do.cas). File Write Case... 4. Continue the calculation by requesting another 100 iterations. Solve Iterate... The solution will converge after about 25 additional iterations. 5. Save the data le (rad do.dat). File Write Data...
5-39
6. Examine the results of the DO calculation. Note: The steps below do not include detailed instructions because the procedure is the same one that you followed for the Rosseland model postprocessing. See Step 6: Postprocessing for the Rosseland Model if you need more detailed instructions. (a) Display velocity vectors (Figure 5.15). Display Vectors...
2.89e-04 2.61e-04 2.32e-04 2.03e-04 1.74e-04 1.45e-04 1.16e-04 8.70e-05 5.80e-05 2.91e-05 1.91e-07
Figure 5.15: Velocity Vectors for the DO Model
5-40
(b) Plot the y velocity along the horizontal centerline (Figure 5.16), and save the plot data to a le called rad do.xy. Plot XY Plot... ! You will need to reselect Y Velocity under Y Axis Function. Also, remember to turn o the Write to File option so that you can access the Plot button to generate the plot.
y=0.5
3.00e-04
2.00e-04
1.00e-04
Y Velocity (m/s)
0.00e+00
-1.00e-04
-2.00e-04
-3.00e-04 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position (m)
Y Velocity
Figure 5.16: XY Plot of Centerline y Velocity for the DO Model (c) Compute the total wall heat transfer rate. Report Fluxes ... The total heat transfer rate reported on the right wall is 6.12 105 W. Note that this is about 1.5% higher than that predicted by the DTRM. The DO and DTRM values are comparable to each other, while the Rosseland and P-1 values are both substantially dierent. The DTRM and DO models are valid across the range of optical thickness, and the heat transfer rates computed using them are expected to be closer to the correct heat transfer rate.
5-41
Step 10: Comparison of y -Velocity Plots

In this step, you will read the plot les you saved for all the solutions and compare them in a single plot. Plot File...
1. Read in all the XY plot les. (a) Click on the Add... button. (b) In the resulting Select File dialog box, select rad do.xy, rad dtrm.xy, rad p1.xy, and rad ross.xy in the Files list. They will be added to the XY File(s) list. If you accidentally add an incorrect le, you can select it in this list and click Remove. (c) Click OK to load the 4 les. 2. Click on Plot. Extra: You can click Curves... in the File XY Plot panel to open the Curves panel, where you can dene dierent styles for different plot curves. In Figure 5.17, dierent symbols have been selected for each curve. 5-42
3. Resize and move the legend box so that you can read the information inside it. (a) To resize the box, press any mouse button on a corner and drag the mouse to the desired position. (b) To move the legend box, press any mouse button anywhere else on the box and drag it to the desired location.
3.00e-04
2.00e-04
Y Velocity Y Velocity Y Velocity (rad_p1.xy) Y Velocity (rad_dtrm.xy) Y Velocity (rad_do.xy)
1.00e-04
Y Velocity
0.00e+00
-1.00e-04
-2.00e-04
-3.00e-04 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position
Y Velocity
Figure 5.17: Comparison of Computed y Velocities for aL = 0.2 Notice in Figure 5.17 that the velocity proles for the P-1 model, DTRM, and DO model are nearly identical even though the reported wall heat transfer rates are dierent. This is because in an optically thin problem, the velocity eld is essentially independent of the radiation eld, and all three models give a ow solution very close to the non-radiating case. The Rosseland model gives substantially erroneous solutions for an optically thin case.
5-43
Step 11: Comparison of Radiation Models for an Optically Thick Medium

In the previous steps, you compared the results of four radiation models for an optically thin (aL = 0.2) medium. It was found that, as a result of the low optical thickness, the velocity elds predicted by the P-1, DTRM, and DO models were very similar and close to that obtained in the nonradiating case. The wall heat transfer rates for DO and DTRM were very close to each other, and substantially dierent from those obtained with the Rosseland and P-1 models. In this step, you will recalculate a solution (using each radiation model) for an optically thick (aL = 5) medium. This is accomplished by increasing the value of the absorption coecient from 0.2 to 5. You will repeat the process outlined below for each set of case and data les that you saved earlier in the tutorial. 1. For each radiation model, calculate a new solution for aL = 5. (a) Read in the case and data le saved earlier (e.g., rad ross.cas and rad ross.dat). File Read Case & Data... (b) Set the absorption coecient to 5. This will result in an optical thickness aL of 5, since L = 1. Dene Materials... (c) Calculate until the new solution converges. Solve Iterate... ! For the DTRM calculation, you may need to click the Iterate button repeatedly until the radiation eld is updated. Since the number of Flow Iterations Per Radiation Iteration in the Radiation Model panel is 10, it is possible that the radiation eld will not be updated for as many as 9 iterations, although FLUENT will report that the solution is converged. If this happens, keep clicking the Iterate button until the radiation eld is updated and the solution proceeds for multiple iterations.
5-44
(d) Save the new case and data les using a dierent le name (e.g., rad ros5.cas and rad ros5.dat). File Write Case & Data... (e) Compute the total wall heat transfer rate. Report Fluxes... (f) Plot the y velocity along the horizontal centerline, and save the plot data to a le (e.g., rad ros5.xy). Plot XY Plot... 2. Compare the computed heat transfer rates for the four models. The wall heat transfer rates predicted by the four radiation models range from 3.50 105 to 3.97 105 W. 3. Compare the y -velocity proles in a single plot (Figure 5.18). Plot File... Note: Use the Delete button in the File XY Plot panel to remove the old XY plot data les. The XY plots of y velocity are nearly identical for the P-1 model, DO model, and DTRM. The Rosseland model gives somewhat different velocities, but is still within 10% of the other results. The Rosseland and P-1 models are suitable for the optically thick limit; the DTRM and DO models are valid across the range of optical thicknesses. Consequently, they yield similar answers at aL = 5. For many applications with large optical thicknesses, the Rosseland and P-1 models provide a simple low-cost alternative.
5-45
5.00e-04 4.00e-04 3.00e-04 2.00e-04 1.00e-04
Y Velocity Y Velocity Y Velocity (rad_p15.xy) Y Velocity (rad_dtr5.xy) Y Velocity (rad_do5.xy)
Y Velocity
0.00e+00 -1.00e-04 -2.00e-04 -3.00e-04 -4.00e-04 -5.00e-04 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position
Y Velocity
Figure 5.18: Comparison of Computed y Velocities for aL = 5
5-46
Step 12: S2S Model Denition, Solution and Postprocessing for a Non-Participating Medium
In the previous steps, you compared the results of four radiation models for optically thin (aL = 0.2) and optically thick (aL = 5) media. The surface-to-surface (S2S) radiation model cannot be used to model participating radiation problems, but it is suitable for modeling the enclosure radiative transfer without participating media. The S2S model assumes that all surfaces are gray and diuse. Thus, according to the gray-body model, if a certain amount of radiation is incident on a surface, a fraction is reected, a fraction is absorbed, and a fraction is transmitted. For most applications the surfaces in question are opaque to thermal radiation (in the infrared spectrum), so the surfaces can be considered opaque. The transmissivity, therefore, can be neglected. Eectively, for the S2S model the absorption coecient can be considered to be zero. In this step, you will calculate a solution for aL = 0 using the S2S radiation model. In the next step, you will use the DTRM and DO models for aL = 0, and compare the results of the three models. The Rosseland and P-1 models are not considered here as they have been shown (earlier in the tutorial) to be inappropriate for optically thin media. 1. Turn on the surface-to-surface (S2S) radiation model and dene the view factor and cluster parameters. Dene Models Radiation...
5-47
(a) Select Surface to Surface under Model. The panel will expand to show additional inputs for the S2S model.
5-48
(b) Set the view factor and cluster parameters. i. Click Set... under Parameters. The View Factor and Cluster Parameters panel will open automatically.
ii. Click OK to accept the default settings. The S2S radiation model is computationally very expensive when there are a large number of radiating surfaces. The number of radiating surfaces is reduced by clustering surfaces into surface clusters. The surface clusters are made by starting from a face and adding its neighbors and their neighbors until a specied number of faces per surface cluster is collected. For a small 2D problem, the default value of 1 for Faces Per Surface Cluster is acceptable. For a large problem, you can increase this number
5-49
to reduce the memory requirement for the view factor le that is saved in a later step. This may also lead to some reduction in the computational expense. However, this is at the cost of some accuracy. Using the Blocking option ensures that any additional surface that is blocking the view between two opposite surfaces is considered in the view factor calculation. In this case, there is no obstructing surface between the opposite walls, so selecting either the Blocking or the Nonblocking option will produce the same result. The default setting for Smoothing is None, which is appropriate for small problems. The Least Square option is more accurate, but also more computationally expensive. See the Users Guide for details about view factors and clusters for the S2S model. (c) Compute the view factors for the S2S model. This step is required only if the problem is being solved for the rst time. For subsequent calculations, you can read the view factor and cluster information from an existing le (by clicking Read... instead of Compute/Write...). i. Click Compute/Write... under Methods in the Radiation Model panel. FLUENT will open a Select File dialog box so you can specify a name for the le where the cluster and view factor parameters are stored. ii. In the S2S File text entry box in the Select File dialog box, enter rad s2s.s2s for the name of the S2S le. Then click OK. FLUENT will print an informational message describing the progress of the view factor calculation.
5-50
2. Retain the current under-relaxation factors for pressure, momentum, and energy (0.3, 0.7, and 1.0). Solve Controls Solution... 3. Save the case le (rad s2s.cas). File Write Case... 4. Continue the calculation by requesting another 100 iterations. Solve Iterate... The solution will converge after about 80 additional iterations. 5. Save the data le (rad s2s.dat). File Write Data... 6. Examine the results of the S2S calculation. Note: The steps below do not include detailed instructions because the procedure is the same one that you followed for the Rosseland model postprocessing. See Step 6: Postprocessing for the Rosseland Model if you need more detailed instructions. (a) Display velocity vectors (Figure 5.19). Display Vectors...
5-51
2.47e-04 2.22e-04 1.98e-04 1.73e-04 1.48e-04 1.24e-04 9.89e-05 7.43e-05 4.96e-05 2.49e-05 1.92e-07
Figure 5.19: Velocity Vectors for the S2S Model
5-52
(b) Plot the y velocity along the horizontal centerline (Figure 5.20), and save the plot data to a le called rad s2s.xy. Plot XY Plot... ! You will have to reselect Y Velocity under Y Axis Function. Also, remember to turn o the Write to File option to access the Plot button to generate the plot.
y=0.5
2.50e-04 2.00e-04 1.50e-04 1.00e-04 5.00e-05
Y Velocity (m/s)
0.00e+00 -5.00e-05 -1.00e-04 -1.50e-04 -2.00e-04 -2.50e-04 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position (m)
Y Velocity
Figure 5.20: XY Plot of Centerline y Velocity for the S2S Model (c) Compute the total wall heat transfer rate. Report Fluxes ... The total heat transfer rate on the right wall is 6.77 105 W.
5-53
Step 13: Comparison of Radiation Models for a Non-Participating Medium

In this step, you will calculate a solution for the aL = 0 case, using the DTRM and DO models, and then compare the results with the S2S results. 1. For the DTRM and DO models, calculate a new solution for aL = 0. (a) Read in the case and data les saved earlier (e.g., rad dtrm.cas and rad dtrm.dat). File Read Case & Data... (b) Set the absorption coecient to 0. This will result in an optical thickness aL of 0. Dene Materials... (c) Calculate until the new solution converges. Solve Iterate... (d) Save the new case and data les using a dierent le name (e.g., rad dtr0.cas and rad dtr0.dat). File Write Case & Data... (e) Compute the total wall heat transfer rate. Report Fluxes... (f) Plot the y velocity along the horizontal centerline, and save the plot data to a le (e.g., rad dtr0.xy) Plot XY Plot... 2. Compare the computed heat transfer rates for the three models. For the S2S model, the total heat transfer rate on the right wall was 6.77 105 W. This is about 5% higher than that predicted by the DTRM and 1.5% higher than DO. Although the S2S, DO, and DTRM values are comparable to each other, this problem involves enclosure radiative transfer without participating media. Therefore, the S2S model provides the most accurate solution. 5-54
3. Compare the y -velocity proles in a single plot (Figure 5.21) Plot File... (a) Use the Delete button in the File XY Plot panel to remove the old XY plot data les. (b) Read in all the XY plot les you saved for the S2S, DTRM, and DO models. (c) Click on Plot.
2.50e-04 2.00e-04 1.50e-04 1.00e-04 5.00e-05 0.00e+00 -5.00e-05 -1.00e-04 -1.50e-04 -2.00e-04 -2.50e-04 0 0.1
Y Velocity Y Velocity Y Velocity (rad_dtr0.xy) Y Velocity (rad_do0.xy)
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Y Velocity
Aug 02, 2001 FLUENT 6.0 (2d, segregated, lam)
Figure 5.21: Comparison of Computed y Velocities for aL = 0 In Figure 5.21, the velocity proles for the DTRM, DO, and S2S models are almost identical even though the wall heat transfer rates are dierent.
5-55
Summary: In this tutorial, you studied combined natural convection and radiation in a square box and compared the performance of four radiation models in FLUENT for optically thin and optically thick cases, and the performance of three radiation models for a non-participating medium. For the optically thin case, the Rosseland and P-1 models are not appropriate; the DTRM and the DO model are applicable, and yield similar results. In the optically thick limit, all four models are appropriate and yield similar results. In this limit, the less computationallyexpensive Rosseland and P-1 models may be adequate for many engineering applications. The S2S radiation model is appropriate for modeling the enclosure radiative transfer without participating media, where the methods for participating radiation may not always be ecient. For more information about the applicability of the dierent radiation models, see the Users Guide.
5-56

Tutorial 5 Fluent

Uploaded by

Copyright:

Available Formats

Tutorial 5 Fluent

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Tutorial 5 Fluent

Uploaded by

Copyright:

Available Formats

Tutorial 5.

Modeling Radiation and Natural Convection

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

= 1000 kg/m3 cp= 1.1030x10

Figure 5.1: Schematic of the Problem

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

3. Display the grid (Figure 5.2). Display Grid...

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

Jul 25, 2001 FLUENT 6.0 (2d, segregated, lam)

Figure 5.2: Graphics Display of Grid

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

2. Turn on the Rosseland radiation model. Dene Models Radiation...

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

3. Add the eect of gravity on the model. Dene Operating Conditions...

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

Step 4: Boundary Conditions

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

Step 5: Solution for the Rosseland Model

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

2. Initialize the ow eld. Solve Initialize Initialize...

(a) Set the Temperature to 1500 K and click on Init.

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

Step 6: Postprocessing for the Rosseland Model

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

Velocity Vectors Colored By Velocity Magnitude (m/s)

Jul 25, 2001 FLUENT 6.0 (2d, segregated, lam)

Figure 5.3: Velocity Vectors for the Rosseland Model

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

2. Display contours of stream function (Figure 5.4). Display Contours...

c Fluent Inc. November 29, 2001

Modeling Radiation and Natural Convection

Contours of Stream Function (kg/s)

Jul 25, 2001 FLUENT 6.0 (2d, segregated, lam)