Question from a beginner
 

Hello all, being a newcomer to cfd, I have many unanswered questions on which you could enlighten me: 1. i do not have access to any commercial cfd package like fluent etc. So, would it be possible to progress QUICKLY as far as the lid driven cavity problem using C or MATLAB or FORTRAN? What is the approximate time required to code the above problem with each of these three languages and would commercial CFD codes help in reducing the time required to code? 2. What are the computational resources required to install FLUENT? Thank you. ~AK

Re: Question from a beginner
 

check the below website and I am pretty sure they have an example test problem for the lid driven cavity. What's more, it has detailed text book related to its code.

http://wissrech.iam.uni-bonn.de/rese...3DGP/index.htm

Junseok

Re: Question from a beginner
 

Hi Arun,

First, if you are a student, you can ask your supervisor to buy an academic licence. It is fearly cheap compared to the price they for industrial licenses. All major CFD codes are now avalaible for PCs.

Second of all, if you go with your home made solver, you do not "code a problem". What you code is the numerical method that will solve the NS equations. The time depends on the method you chose and the number of applications you want to solve. Let say it can go from 1week to a life time...(depending on the number of bugs...).

I a gave a quick look to NAST 3D, it seems verywell documented and is very easy to build. Moreover, it comes with the cavity flow.

Good luck

grid size
 

Hi, Can any one tel me (in more details) how can I see the the influence of the grid size on the computed solution. I'm traying to use a vof code. thanks.

Re: grid size
 

1) Solve the problem with coarse grid and then solve it again with systematic grid refinement on 2 (or 3) other grids. Let say the mean size with the coarse grid is h0, then the size for the other grids can be chosen like this:

1D) h1=0.5*h0, h2=0.5*h1 and so forth. 2D) h1=0.7*h0, h2=0.7*h1 and so forth. 3D) h1=0.8*h0, h2=0.8*h1 and so forth. This choice will most likely double the number of grid points from one grid to the other one.

When the exact solution of the problem is known (which I doubt if you are using VOF), the rate of convergence of the scheme can be measured for the problem beign solved.

Then you can compare the solutions and compute reference quantities that apply to your problem. (Such as the drag).

2) You can also try to refine the mesh in region of interest (like the boundary layers). But it is not as straithforward as the other method.