0
$\begingroup$

I would like to incorporate a CFD topic in a project I currently have. The deadline is a little less than a month from now. Basically, what I would like to do is solve some equations and simulate fluid behavior. The programming language I will use to accomplish this in is C, however I intend to rely on MATLAB for later visualizations and better flexibility - after getting my solutions for the PDEs from my C code.

I am currently a beginner in CFD and would like to boost my understanding by getting myself involved in such an interesting project - at least from my perspective!

Are there any interesting yet simple examples out there to help me accomplish this? I so far have not one idea in mind, especially since most of the topics circulating the web are kind of 'advanced' for someone in my situation as of yet. Any references, tutorials, ideas or the like would be greatly appreciated !!

$\endgroup$
  • $\begingroup$ What equation(s) are you looking to solve? $\endgroup$ – spektr Mar 6 '16 at 15:18
  • $\begingroup$ No particular idea in mind. Anything that is probably 'hot' in this field and of-course easy to implement by someone like me with such a relatively small time frame... $\endgroup$ – Inquisitor101 Mar 6 '16 at 15:22
  • $\begingroup$ Do yourself a favor and find an opensource code with some examples codes to modify to your specific needs. An opensource code such as Openfoam comes to mind. Learn from the documentation which equations are solved; in my opinion as a beginner you are simply not equipped to write a fullfledged CFD solver in less than a month. $\endgroup$ – nluigi Mar 6 '16 at 15:25
  • $\begingroup$ I suspect a lot of the "hot" areas in the field are going to be out of reach for a beginner to build within a month, unfortunately. However, this document might give you enough to try making something yourself if you really have your heart set on it: www3.nd.edu/~gtryggva/CFD-Course/2011-Lecture-17.pdf $\endgroup$ – spektr Mar 6 '16 at 15:28
  • 1
    $\begingroup$ Well in that case I can also suggest looking into Lattice Boltzmann methods. While understanding the full theory can have a bit of a learning curve (to be honest equally difficult to work through as flux limited Euler equations), the code to accomplish some simple flow systems is only at most 200 lines of code in MATLAB. You could look into taylor-green vortices easily with some periodic boundary conditions. Easily doable in a month. $\endgroup$ – nluigi Mar 6 '16 at 16:22
4
$\begingroup$

There are many simple cases that you could look at, e.g.

2D-flow over a backward-facing step.

Lid-driven cavity flow.

(Inviscid) wave resistance of thin ships or travelling pressure distributions.

They should be easy to do in a month. There are also hidden depths (e.g. 3D-turbulence effects, non-linearity, etc) to the problems that could give years of interest, fascination and frustration.

$\endgroup$
  • $\begingroup$ Thanks. I see there are lots of papers on those issues. Can you recommend a good starting place on the '2D-flow over a backward-facing step' ? $\endgroup$ – Inquisitor101 Mar 6 '16 at 19:34
  • $\begingroup$ I think you would get a lot from... Fletcher, Clive A. J., "Computational Techniques for Fluid Dynamics", Vol. II, Specific Techniques for Different Flow Categories, Springer-Verlag, 2nd edition, 1991, 5th printing 2003. $\endgroup$ – Lysistrata Mar 6 '16 at 20:10
1
$\begingroup$

In my opinion, the simplest way to start is with some 1D compressible flow. e.g. Sod's shock tube problem. Then maybe extend that to 2D for visuals. Incompressible flow is "trickier" than compressible flow.

Density vs Time in Sod's shock tube

$\endgroup$
0
$\begingroup$

You can go through the following MOOC.

http://openedx.seas.gwu.edu/courses/GW/MAE6286/2014_fall/about

Going through the lessons and coding the examples and exercises will help you understand and build the tools you require. With that you can try out more substantial problems.

$\endgroup$
  • $\begingroup$ Great. Thanks a lot, this seems a very useful tool for afterwards. Alas however it addresses a more general audience, the scientific computing community, rather than specific areas such as CFD. $\endgroup$ – Inquisitor101 Mar 7 '16 at 10:52
  • $\begingroup$ Actually most of CFD is done for either the incompressible problem or the compressible problem. Compressiblity implies hyperbolicity which is the wave problem. Incompressibility would be the elliptic problem which is the heat equation type of problem. Bot are covered in the course. What is not covered is the modelling and I don't think that is in any way "simple". $\endgroup$ – Vikram Mar 8 '16 at 9:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.