CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   New to CFD with some basic question (http://www.cfd-online.com/Forums/main/12451-new-cfd-some-basic-question.html)

Newbie October 23, 2006 04:47

New to CFD with some basic question
 
Hi all, I'm new to CFD but not to simulation codes (rigid bodies and robotics). Due to my personal interests I approached the CFD theory and techniques.

Recently I've seen a simulation consisting in a ship subject to a train of waves that interact with it. The waves broke on ship and moving it along the liquid surface. It was amazing! :)

Nevertheless I have a question (maybe due to my poor understanding of FD simulations): in CFD the approaches to simulation are based on solution of PDE problems. Is it possible to simulate that kind of model I've seen or such a model could be better simulated using a 'rigid bodies' modeling of fluid particles?

I apologize if my question is very near to be 'stupid'!


Mani October 23, 2006 09:51

Re: New to CFD with some basic question
 
When you say 'rigid body fluid particles' I am suspecting you imagine a molecule traveling through space. What fluid dynamicists call 'particle' is usually different. Lagrangian experts out there will correct me if I am wrong, but my understanding of a fluid particle is a certain amount of fluid (certainly more than one molecule) that is small enough to stay well defined in a closed boundary, and to have a more or less common velocity, but large enough to be treated as a continuum. These particles can change shape and/or density and are not 'rigid bodies'.

It's possible to consider molecules (describing fluid flow based on molecular collision models), but there are so many of them that this is typically done by statistical means (Boltzmann equation for kinetic theory of gases), and even as such is extremely computing intensive!

In some methods 'particle' could also mean something completely different than described above. For example, vortex methods track the position and strength of singular vortices in an incompressible flow to solve a particular form of the flow equations.

But the short answer is: Yes, there are method out there to solve flow equations with Lagrangian (as opposed to Eulerian) methods.

Newbie October 23, 2006 12:20

Re: New to CFD with some basic question
 
Thanks for your response. Effectively my question concerns the possibility to mix eulerian and lagrangian metohds to 'manage' both rigid bodies and fluids interacting. Look at

- 'Wave and body interaction' at

http://www.nextlimit.com/xflow/seakeeping_1.mov

- 'Body interaction' at

http://www.nextlimit.com/xflow/solidcube.avi

http://www.nextlimit.com/xflow/sphere_floating.mov

It seems that there is a level of simulation in which the fluid is treated via an eulerian method, and a level of simulation in which particles (not molecules) are treted via a lagrangian method (to interact with ship, cube or sphere). And the two levels are (should be) coupled!

p.s. The 'particle' I wrote about was the molecule of a fluid.


Newbie October 23, 2006 12:22

Re: New to CFD with some basic question
 
Thanks for your response. Effectively my question concerns the possibility to mix eulerian and lagrangian metohds to 'manage' both rigid bodies and fluids interac


All times are GMT -4. The time now is 21:36.