Does anyone have any experience setting up this type of problem? I am trying to determine the leakage rate through a clearance type diameter fit of two parts.

I am setting up the problem in this manner:

Part is assembled into the tube. Negative of the part and tube is made within ProE. Step file of the negative is exported. I can clearly see the connected crack between the two volumes. The crack has very tiny mesh compared to the connected volumes. Pressure pulse of 50 psi is applied to the "inlet". Vacuum pressure of 1 psi is applied to the "outlet".

When veiwing the results, I can see the pressure differential form around the "crack" but there is no velocity streamlines forming through it. I have even tried using up to 10,000 seed points thinking maybe it was not high enough but to no avail!

Does this just mean that there will be no leakage or have I set up the problem incorrectly?

It depends on the size of the clearance. How small is the clearance? Also, if the streamlines are visible on both sides of the crack, but not in the crack, then it's a visual artifact. Does the leakage flow rate seem reasonable? Is the solution converged?

Bore fit clearance is about .60/.37 so the crack is about .37/2 mm thick (.185 mm). The fluid is air and the mesh size is about .01 min so I think it is okay.

If you are able to see the results does this not mean that the solution has converged? If not, how do you check?

Where are you seeding the streamlines? If the crack is very small, then it will be very hard to find a streamline that goes through. Try seeding the streamline from within the crack by seeding them from a slice plane that cuts through them, for instance. Calculate the streamline forward and backward to see where they come from and where they are going.

By the way, if the crack is very small relative to the domain, you should probably run the solver in double precision.

-CycLone

Yes, the dimensions and the relative mesh size sounds alright. If you are able to see the results, that does not mean that the solution has converged. You can see the results for a non-converged solution also if the solver finishes the number of iterations specified. If the error residuals reach the minimum specified target, then the solutions is considered to be converged.

That worked! I used double precision in the solver and created a plane through the crack. Then by making a plane in the middle of the crack and seeding from the crack in both directions the leak was revealed!