TUTORIAL CFD-STUDIO

NATURAL CONVECTION IN A SQUARE CAVITY

Processor (Numerical Simulator)

Numerical Parameters :

In the main window click in Processor and after click in Numerical Scheme to define the interpolation Scheme, the P-V Coupling method, the time step, the maximum number of iterations in time and the tolerances in the solver (Figure 29).


Figure 29 – CFD-Studio main Console

After clicking a window will appear (Figure 30) 



Figure 30 – Numeric Parameters Console

In this problem we will use the CDS interpolation scheme and the SIMPLEC Method of
(For more details about interpolation scheme, see the Scientific Manual)


Figure 31 – Interpolation Scheme
pressure-velocity coupling. (For more details about PV Coupling methods, see Scientific Manual). On this box you can also set a sub-relaxation on the pressure. If the problem has some boundary condition where the velocity is unknown, like “open wall” and “outlet”, you can facilitate the convergence setting a Reference Pressure in the domain. This ensure that will be just one solution for the pressure field. 


Figure 32 – PV Coupling Console
We continue, setting a step time of 0,05s,

If you desire to estimate the time interval size to be used in a certain problem you can take as base the maximum step in the allowed time if the explicit formulation was used. More details can be found in the chapter 5 of Maliska, 2004.



Figure 33 – Time Settings Console
with 20 iterations in the PV Coupling looping, and 2 in the temperature loop.




Figure 34 – Maximum Iterations Console
We will use 10-5 the tolerance for the steady state for all variables. The verification is done in the two central lines of the mesh (joining S to N and W to E).



Figure 35 – Steady State Errors Console.

Therefore, for our problem you will have to configure the window as in Figure 36:



Figure 36 – Numeric Parameters Console.

Now click OK.

Choice of the Linear System Solver:


The next step is to define the solver to be used in the solution of the linear systems. In the main window click in Processor and after in Solver Settings, (Figure 37).


Figure 37 - CFD-Studio main Console

and a window as the Figure 38 will appear;



Figure 38 – Solver Console
The solver choice is an essential step. The direct solvers are much more precise but they need a very big memory space. The iterative solvers just work with the non null elements of the matrix, saving like this a lot of memory in the computer, but they are subject to fail with unstable linear systems (evil-conditioned matrix), because they spread mistakes according to the direction of the sweeping in the matrix The behavior of the several linear solution system methods can be found in the linear algebra text.

For our problem we will use the TDMA solver. The tolerance and the maximum number of iterations must should be defined as in Figure 38.

Simulator


IN the main window click in Processor and then in Simulator, as shown in Figure 39.



Figure 39 – CFD-Studio main Console
You will see a window like the one in Figure 40



Fig. 40 – Console do terminal de controle.

In this window, you will see in the right side, the residues of the simulation. The plot shows the residues against the number of iterations. After the end of the simulation it is necessary to click in Exit to see the numerical results. You can see the images of the simulation (such as the iso-lines, pressure fields) connecting the 3D window, where you can see the simulation in real time. Only enable this function if you have a 3D video board with OpenGL support.


Simulation Progress :

In the main window clicking in Processor and after in Simulator Progress and you will see a window like the one in Figure 41.



Figure 41 – Numeric Engine Console
(It is not necessary to open this window now, it will open automatically as soon as the simulation begins, the same will happen with the 3D window)
Images

Now we are ready to begin the simulation. Clicking in Init Simulation the simulation will start. Now you will see a sequence of images (Illustrations 42-55) from the beginning to the end of the simulation. After this we will enter in the final part, the post-processing.
Figure 42


 Figure 43
Figure 44


 Figure 45
Figure 46


 Figure 47
Figure 48


 Figure 49
Figure 50


 Figure 51
Figure 52 Figure 53
Figure 54 Figure 55
 
© SINMEC/EMC/UFSC, 2001. All rights reserved.