OT: Pro/e models and CFD

I was asked today if I'd ever prepped a Pro/e model for analysis by CFD. I say yes and told them basically how you prep a model for Mechanica analysis (suppress rounds, holes, chamfers) to avoid meshing small features. Not hearing any gasps or depressed sighs in the background, I figured I was pretty close. But, the more I thought about this, the more I realized that fluids running through pipes under pressure have little to do with the external features like rounds and chamfers. And least likely that you'd eliminate holes where you're looking for cavitation in simulation. So, what kind of prep is done to models for CFD?

David Janes

Reply to
Janes
Loading thread data ...

I say yes and told them basically how you prep a model for Mechanica analys= is (suppress rounds, holes, chamfers) to avoid meshing small features. Not = hearing any gasps or depressed sighs in the background, I figured I was pre= tty close. But, the more I thought about this, the more I realized that flu= ids running through pipes under pressure have little to do with the externa= l features like rounds and chamfers. And least likely that you'd eliminate = holes where you're looking for cavitation in simulation. So, what kind of p= rep is done to models for CFD?

As you state, avoiding extremely small features that aren't likely to affect the flow helps a lot in meshing the geometry in the same way as meshing an FEA model. Tiny surfaces, gaps caused by poor modeling (geom checks, tolerance settings) will also cause problems.

On the other hand if they are large relative to the size of the flow field being modeled, rounds, chamfers, leak gaps, holes, bolt heads, etc can have a big effect on the flow results so they may be necessary to get accurate CFD results.

Real assemblies may have small clearances or gaps that in reality leak flow. If these leak gaps are small relative to the flow field of interest it may be desirable to eliminate them from the model. This requires some judgment or simple 1D calculation to determine whether the leakage is important to the flow results of interest.

It depends on the meshing tool used, but often a solid model of the fluid must be constructed. This can be done by using the solid model of the part to cut a fluid volume.

Finally, it is usually desirable to extend the fluid zone inlet and exit length so that the flow boundary conditions at the inlet and exit are separated from the region of interest. This requires either adding actual upstream and downstream geometry to the model, or creating some simplified geometry that mimics the effective area and perhaps the flowpath curvature of the actual geometry.

Dave

Reply to
dgp

As David points out, you'd be delivering wetted surfaces to the analyst, typically with the 'in' and 'out' surfaces coloured to identify them. Depending on the CFD system, and the product type, how you present that is a matter of whatthe meshe monkey wants, For instance, if you're using Star, they'll want a .stl file with a chord height of 0.02 (so make sure your model accuracies are appropriate or you won't get it) whereas for fluent, they'll probabaly want .igs

My perennial comment about not using relative accuracy applies here. If your models are full of holes, the mesh monkey will have to fill them in, and may hideously misinterpret your intent.

Make sure you sit the mesher down in front of your screen & explain what's going on, as they are social retards almost without exeption, and won't necessarily tell you if they get stuck. They'll just make some crap up & blow your budget on something you can't use, then blame you. Killing them is frowned upon, even though by any rational measure, they aren't properly human, as they aren't designers.

The bitter voice of experience.

Reply to
John.R.Wade

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.