Problem with convergence

Iman Sengupta · June 23, 2015, 06:18

I am working on CFD simulation of hydrocyclone for steady state condition. For a particular set of dimensions I have used mesh density in the range of 300000. I have used VOF for multiphase model, RSM for turbulence model. I have used QUICK for momentum, PRESTO for pressure field convergence and scheme is SIMPLE. Under relaxation factors are kept as default. I have some problem with convergence. When the residual term of continuity equation approaches the range of 10^-6 the solution is converged but at first residual term of continuity is decreasing and then it increases. After slight decrease again it saturates at the range of 10^-3 of 10^-4. I found out that the some skewed meshes in geometry was causing this problem. By slight modification in geometry I removed all skewed meshes. But still the problem is there. When I am applying first order upwind scheme for momentum, instead of QUICK there is no problem. When I am using QUICK or second order upwind scheme I have trouble with convergence. Can anybody help me please.

jamalf64 · June 23, 2015, 06:55

This can help you :
http://www.cfd-online.com/Forums/flu...2-cyclone.html

Iman Sengupta · June 23, 2015, 07:05

Thank you for the link but that didi not help me to solve my problem.

jamalf64 · June 23, 2015, 07:09

Please describe your problem in detail

Iman Sengupta · June 23, 2015, 07:19

I am working with CFD simulation of a hydrocyclone under steady state condition. I have adopted VOF for multiphase flow model and RSM as turbulence model. For my geometry the mesh density is in the range of 300000. I have used QUICK scheme for momentum field calculation and PRESTO for pressure field calculation. I used SIMPLE for solving the problem. When I am using first order upwind scheme for momentum field computation I get converged results but for better accuracy I am using QUICK scheme and I am getting some problem with convergence.

jamalf64 · June 23, 2015, 07:21

Please describe why VOF?

Iman Sengupta · June 23, 2015, 07:30

I want to check the pressure field and velocity field inside hydrocyclone. I am injecting pure water as feed. Due to rotational motion inside hydrocyclone and high turbulent field, a low pressure zone is formed along the axis of the hydrocyclone. As both the outlets are open to atmosphere air will be sucked into hydrocyclone and an air core will be formed. This air core has great influence on pressure and velocity field of hydrocyclone. To track the shape of air core VOF is used.

jamalf64 · June 23, 2015, 07:36

Ok. You can't use steady state and you must run your problem unsteady. Please read about vof model from fluent user guide.

Iman Sengupta · June 23, 2015, 07:40

If I use explicit VOF model, I must use unsteady state model. But for implicit model I can run steady state model. I adopted implicit VOF.

jamalf64 · June 23, 2015, 07:47

Fluent user guide :
. A steady-state VOF calculation is sensible only when your solution is independent of the initial conditions and there are distinct inflow boundaries for the individual phases. For example, since the shape of the free surface inside a rotating cup depends on the initial level of the fluid, such a problem must be solved using the time-dependent formulation

Iman Sengupta · June 23, 2015, 07:57

Thank you for the suggestion. I will try with transient problem.

jamalf64 · June 23, 2015, 07:59

Be success,
Jamal

Llume · June 23, 2015, 08:44

Sorry I posted on the wrong discussion and I'm trying to delete this

June 23, 2015, 06:18	Problem with convergence	#1
Iman Sengupta New Member Iman Sengupta Join Date: May 2015 Location: India Posts: 14 Rep Power: 11	I am working on CFD simulation of hydrocyclone for steady state condition. For a particular set of dimensions I have used mesh density in the range of 300000. I have used VOF for multiphase model, RSM for turbulence model. I have used QUICK for momentum, PRESTO for pressure field convergence and scheme is SIMPLE. Under relaxation factors are kept as default. I have some problem with convergence. When the residual term of continuity equation approaches the range of 10^-6 the solution is converged but at first residual term of continuity is decreasing and then it increases. After slight decrease again it saturates at the range of 10^-3 of 10^-4. I found out that the some skewed meshes in geometry was causing this problem. By slight modification in geometry I removed all skewed meshes. But still the problem is there. When I am applying first order upwind scheme for momentum, instead of QUICK there is no problem. When I am using QUICK or second order upwind scheme I have trouble with convergence. Can anybody help me please.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Convergence problem in Fluent for quenching process	kaeran	FLUENT	4	December 1, 2014 03:14
Rotate frame reference convergence problem!	wjy-c	CFX	2	September 26, 2014 07:03
Centrifugal pump OpenFOAM, convergence problem, ANSA model	RDD	OpenFOAM Running, Solving & CFD	0	July 5, 2014 10:12
Convergence Problem in Axisymmetric Periodic Flow	atheresia	FLUENT	3	February 10, 2014 04:00
Convergence of CFX field in FSI analysis	nasdak	CFX	2	June 29, 2009 02:17

June 23, 2015, 06:55		#2
jamalf64 Senior Member Jamal Foroozesh Join Date: Oct 2012 Location: Iran Posts: 162 Rep Power: 14	This can help you : http://www.cfd-online.com/Forums/flu...2-cyclone.html

June 23, 2015, 07:05		#3
Iman Sengupta New Member Iman Sengupta Join Date: May 2015 Location: India Posts: 14 Rep Power: 11	Thank you for the link but that didi not help me to solve my problem.

June 23, 2015, 07:09		#4
jamalf64 Senior Member Jamal Foroozesh Join Date: Oct 2012 Location: Iran Posts: 162 Rep Power: 14	Please describe your problem in detail

June 23, 2015, 07:19		#5
Iman Sengupta New Member Iman Sengupta Join Date: May 2015 Location: India Posts: 14 Rep Power: 11	I am working with CFD simulation of a hydrocyclone under steady state condition. I have adopted VOF for multiphase flow model and RSM as turbulence model. For my geometry the mesh density is in the range of 300000. I have used QUICK scheme for momentum field calculation and PRESTO for pressure field calculation. I used SIMPLE for solving the problem. When I am using first order upwind scheme for momentum field computation I get converged results but for better accuracy I am using QUICK scheme and I am getting some problem with convergence.

June 23, 2015, 07:21		#6
jamalf64 Senior Member Jamal Foroozesh Join Date: Oct 2012 Location: Iran Posts: 162 Rep Power: 14	Please describe why VOF?

June 23, 2015, 07:30		#7
Iman Sengupta New Member Iman Sengupta Join Date: May 2015 Location: India Posts: 14 Rep Power: 11	I want to check the pressure field and velocity field inside hydrocyclone. I am injecting pure water as feed. Due to rotational motion inside hydrocyclone and high turbulent field, a low pressure zone is formed along the axis of the hydrocyclone. As both the outlets are open to atmosphere air will be sucked into hydrocyclone and an air core will be formed. This air core has great influence on pressure and velocity field of hydrocyclone. To track the shape of air core VOF is used.

June 23, 2015, 07:36		#8
jamalf64 Senior Member Jamal Foroozesh Join Date: Oct 2012 Location: Iran Posts: 162 Rep Power: 14	Ok. You can't use steady state and you must run your problem unsteady. Please read about vof model from fluent user guide.

June 23, 2015, 07:40		#9
Iman Sengupta New Member Iman Sengupta Join Date: May 2015 Location: India Posts: 14 Rep Power: 11	If I use explicit VOF model, I must use unsteady state model. But for implicit model I can run steady state model. I adopted implicit VOF.

June 23, 2015, 07:47		#10
jamalf64 Senior Member Jamal Foroozesh Join Date: Oct 2012 Location: Iran Posts: 162 Rep Power: 14	Fluent user guide : . A steady-state VOF calculation is sensible only when your solution is independent of the initial conditions and there are distinct inflow boundaries for the individual phases. For example, since the shape of the free surface inside a rotating cup depends on the initial level of the fluid, such a problem must be solved using the time-dependent formulation

June 23, 2015, 07:57		#11
Iman Sengupta New Member Iman Sengupta Join Date: May 2015 Location: India Posts: 14 Rep Power: 11	Thank you for the suggestion. I will try with transient problem.

June 23, 2015, 07:59		#12
jamalf64 Senior Member Jamal Foroozesh Join Date: Oct 2012 Location: Iran Posts: 162 Rep Power: 14	Be success, Jamal

June 23, 2015, 08:44		#13
Llume New Member Join Date: Jun 2015 Posts: 7 Rep Power: 11	Sorry I posted on the wrong discussion and I'm trying to delete this