Software for modeling fluid flow?

[jwaggs]

Hey everyone,

This is my first-ever post on a forum, so my apologies if this is the wrong section or if this topic has been covered...

Anyways, I'm trying to design an optimized propeller/generator that will be in a closed pipe, but I have no idea what kind of software to use as I've never worked with this kind of stuff before. The water will obviously be driving the system, so I want something that will help me get the best design for a propeller or generator mechanism. I looked up a couple programs, Fluent and Solidworks, but I don't know if they can do what I want. So, what software would you suggest? If you can, include cost and computing requirements.

Thanks!

 

[MasterX]

If the propeller/generator is rotating then you need a software that can solve FSI (Fluid-Structure Interaction) problems. These are fairly complicated problems, but I think there is a software named ADINA, that can solve this type of problems. I have not used it, however.

If, on the other hand, the propeller/generator does not rotate, then there are many software that you could use (e.g. Fluent, FemLab, openfoam, etc). Fluent and Femlab, are the easiest to use, as the provide a graphical interface to draw your geometry. But, they are not free.

 

[minger]

That's not entirely true. FSI is only needed if he's concerned with vibratory stresses in the blades, thermal issues, or more generally, any issue within the solid structure.

Many techniques are employed now in CFD to allow moving and rotating boundaries. However, for a problem like this, I think you would be well-served using a simple 1D or meanline code. These are codes which basically solve velocity triangles to determine a basic design.

 

[MasterX]

That's not entirely true. FSI is only needed if he's concerned with vibratory stresses in the blades, thermal issues, or more generally, any issue within the solid structure.

Many techniques are employed now in CFD to allow moving and rotating boundaries. However, for a problem like this, I think you would be well-served using a simple 1D or meanline code. These are codes which basically solve velocity triangles to determine a basic design.

By the term FSI I meant that jwaggs would have to modify the NS eqns in order to solve this problem. Yes, you are right, if the blades do not deform, you could use other techniques, but you also need to modify the eqns, if these blades are rotating.
For example you could consider a mesh that rotates along with the blades. This is NOT FSI, but you need to have an extra term in the NS eqn, due to the mesh velocity, like you do when you are solving FSI problems.

 

[minger]

The N-S equations already include mesh motion. When you expand them into a form which actually can be solved, they can be written (in non-conservative form) in one dimension as:
$$\frac{\partial Q}{\partial\tau} + \frac{\partia\xi}{\partial t}\frac{\partial Q}{\partial \xi} + \frac{\partial\xi}{\partial x} \frac{\partial F}{\partial \xi}$$
Where Q is the vector of conserved variables, and F is the flux vector. You'll notice that the de/dt term is the grid metrics, or essentially how the mesh is moving in time. The governing equations are already elegant enough to include these effects.

*note that in order to convert these to conservative form a little algebra is done, although it's not too bad.

Either way, for a relatively simple problem like this, the OP would be well decided to open a turbomachinery textbook and start reading. He'll find that many of his problems can be solved analytically for about $70 (used of course) rather than investing time/money in learning a numerical solution, which in this case to do it properly would take significantly longer.

 

[nathanlee52]

Solidworks is not up to the task. fwiw. Not unless you're a super-duper user.

 

[AngieD]

Waggs,

Not sure if this would work for your application, but I am just learning to use a pipe flow modeling software that may be functional for your analysis:

www.flowmaster.com

You may want to check it out to see if it would be applicable to your situation...
Good luck!