[ybcfd]

Hi guys,

Recently i have been dealing with an air induction system of a passenger car. The objective is obtaining pressure loss through the components and system. I need also velocity uniformity data for some sections.

Here is the geometry that I simplified and the mesh that I obtained by Star CCM+.

I have defined stagnation inlet for the sphere's surface and minus mass flow rate for the clean air duct's exit. I used k-epsilon turb. mod. and ideal gas and coupled flow as physics. With this mesh and physics I had reverse flow warnings throughout the analysis and after almost 1000 iterations it couldn't arrive the convergence.

I have substantially simplified original internal volume of the system. I have tried without sphere ( bc: direct mass flow rate inlet from the snorkel face - pressure outlet).

I have been going crazy with this analysis for almost 2 weeks. How can i get it converged ? I can even send my trial .sim files by mail express. Thanks in advance!

[taillanm]

hi
Is there a particular reason why you choose these two boundary conditions?

I don't remember what you give as input for the stagnation inlet: a pressure? a mass flow?
You could try with seg. flow model depending on the mach's involved

[fluid23]

You are getting the reverse flow warning because you applied the stagnation inlet to a large non-planar surface. It really was meant more for duct/pipe openings. You might get better results using Freestream boundary type with M=0. That being said, the reverse flow warnings are not something to worry about for this setup unless they are occurring somewhere other than your inlet boundary.

As for your convergence, it is almost definitely due to your choice of boundary conditions... What you should do is go into solvers > coupled and enable 'temporary storage retained'. Then step the solution once and create a scalar scene with a set of thresholds assigning the various residuals as the field functions. Make sure you set the threshold to show 'between' and put the limits such that you can see the the extreme 10-20% of the residuals. If the cells that show up in the scene are clustered exclusively near the inlet then you are probably OK as long as your pressure drop, mass flow ,etc... have converged. If not, then you may need to refine your mesh in areas with high residuals.

[marmot]

I think pressure for outside the car and a negative (velocity or mass flow inlet) for the air ducts, I believe seg. flow solver is the way to go.

[ybcfd]

taillanm and marmot==> I have tried segregated flow before but i couldn't have converged results by a few different mesh, then i changed it into the coupled flow since it is compressible. AFAIK for compressible flows it is better to use coupled. Thanx for the help

MBdonCFD==> I have simplified the geometry and also changed the bc to the freestream with M=0, now it is better but still doesn't converge (especially Tdr). And unfortunately i have reversed flow warnings for my inlet boundary which is the sphere's surface representing the atmospheric air. I have't tried "temporary storage retained" before. I ll try for sure i hope it directs me to the some point. Thanks for detailed answer and help


If you have any other suggestions please share. I am new at CFD and really sick of that i couldnt manage to complete my first quasicomplex system analysis :S