Low Pressure External Flow (95 Pa)

Jeff P. · July 20, 2016, 13:21

Hi everyone,

I'm performing a compressible external flow analysis in a low pressure / density environment. My problem is that the solver always diverges even after trying many different approaches. Is there anything obvious that i'm missing or should do to start converging?

Thanks!

ghorrocks · July 20, 2016, 21:05

The general answer to this is in the FAQ: http://www.cfd-online.com/Wiki/Ansys...do_about_it.3F

You will need to be more specific if we are to help you more with your case.

Jeff P. · July 21, 2016, 16:54

Sorry about the vagueness. I'm running a simulation of flow around a blunt body in very low pressure / density / temperature conditions at supersonic, i.e. compressible, speeds. I'm using a supersonic velocity and static pressure inlet and supersonic outlet with entrainment as the surrounding atmospheric boundary, which has seemed to work well in the past.

As a result, every solution attempt I try never decreases in residuals but always increases... a major error somewhere. My guess is that the pressures are so low that the wake region of my body pulls the values into a weird, near vacuum state, that causes errors. Any thoughts?

ghorrocks · July 21, 2016, 18:14

If your theory is correct then you should be able to see this in post processing. Save a backup file before it crashes and have a look.

But the usual approach is to include residuals in the results/backup file and save a results/backup file before it crashes. Have a look at the residuals in the post processor and you will probably have a small region of high residuals - and that is the problem area. That is the first step in fixing the problem.

Jeff P. · July 22, 2016, 11:55

I took a look at my results and it seems like nothing is happening. All I see is the constant velocity based initial conditions in the domain. Seems like the solver is failing immediately.

ghorrocks · July 22, 2016, 18:34

Try a very small timestep to get it started. But if it crashes before anything happens then there is obviously a fundamental inconsistency in your simulation.

Jeff P. · July 28, 2016, 11:15

So I've run a few more simulations with some relaxation factors, extra (2) mass loops, and high speed numerics with very small (0.000 000 001 s) and large timesteps and nothing seems to get my mass residuals converging. The only factor that helps is the local timescale factor (6) and has improved convergence of all momentum residuals expect my mass flow which remains high (~0.1) in every case. I've redone my mesh, it seems refined but not exceptionally dense and good quality. It is a tetrahedral mesh. I'm not sure what else I can be missing.

ghorrocks · July 28, 2016, 18:57

Please post an image of your mesh, the geometry and your CCL.

July 20, 2016, 13:21	Low Pressure External Flow (95 Pa)	#1
Jeff P. New Member Jeff Join Date: Jun 2016 Posts: 20 Rep Power: 9	Hi everyone, I'm performing a compressible external flow analysis in a low pressure / density environment. My problem is that the solver always diverges even after trying many different approaches. Is there anything obvious that i'm missing or should do to start converging? Thanks!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Issues on the simulation of high-speed compressible flow within turbomachinery	dowlee	OpenFOAM Running, Solving & CFD	11	August 6, 2021 06:40
porousSimpleFoam: pressure too low (no resistance to flow)	AdriaanDB	OpenFOAM Running, Solving & CFD	0	June 1, 2016 05:59
BC for Actuator Disk - Flow Vel from High Pressure to Low Pressure	vinguva	OpenFOAM Running, Solving & CFD	2	March 7, 2016 23:46
Multicomponent fluid	Andrea	CFX	2	October 11, 2004 05:12
External Flow Pressure	Meri	CFX	2	May 6, 2004 10:36

July 20, 2016, 21:05		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	The general answer to this is in the FAQ: http://www.cfd-online.com/Wiki/Ansys...do_about_it.3F You will need to be more specific if we are to help you more with your case.

July 21, 2016, 16:54		#3
Jeff P. New Member Jeff Join Date: Jun 2016 Posts: 20 Rep Power: 9	Sorry about the vagueness. I'm running a simulation of flow around a blunt body in very low pressure / density / temperature conditions at supersonic, i.e. compressible, speeds. I'm using a supersonic velocity and static pressure inlet and supersonic outlet with entrainment as the surrounding atmospheric boundary, which has seemed to work well in the past. As a result, every solution attempt I try never decreases in residuals but always increases... a major error somewhere. My guess is that the pressures are so low that the wake region of my body pulls the values into a weird, near vacuum state, that causes errors. Any thoughts?

July 21, 2016, 18:14		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	If your theory is correct then you should be able to see this in post processing. Save a backup file before it crashes and have a look. But the usual approach is to include residuals in the results/backup file and save a results/backup file before it crashes. Have a look at the residuals in the post processor and you will probably have a small region of high residuals - and that is the problem area. That is the first step in fixing the problem.

July 22, 2016, 11:55		#5
Jeff P. New Member Jeff Join Date: Jun 2016 Posts: 20 Rep Power: 9	I took a look at my results and it seems like nothing is happening. All I see is the constant velocity based initial conditions in the domain. Seems like the solver is failing immediately.

July 22, 2016, 18:34		#6
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	Try a very small timestep to get it started. But if it crashes before anything happens then there is obviously a fundamental inconsistency in your simulation.

July 28, 2016, 11:15		#7
Jeff P. New Member Jeff Join Date: Jun 2016 Posts: 20 Rep Power: 9	So I've run a few more simulations with some relaxation factors, extra (2) mass loops, and high speed numerics with very small (0.000 000 001 s) and large timesteps and nothing seems to get my mass residuals converging. The only factor that helps is the local timescale factor (6) and has improved convergence of all momentum residuals expect my mass flow which remains high (~0.1) in every case. I've redone my mesh, it seems refined but not exceptionally dense and good quality. It is a tetrahedral mesh. I'm not sure what else I can be missing.

July 28, 2016, 18:57		#8
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	Please post an image of your mesh, the geometry and your CCL.