steady state simulation diverge

Ema40 · October 13, 2015, 20:40

Hi all.

I am simulating an open channel flow with geometry discontinuities (with rapid changes in the cross sections...), in a steady state.

Why does it diverge? In a steady state there is not the temporal discretization, thus the divergence of the solution does not depend on the mesh size, right?
Can the divergence be caused by the boundary conditions?

Thank you

CFDYourself · October 20, 2015, 09:27

a few points, I am also having trouble with divergence of a steady state model currently.

- it does depend on the mesh. If the mesh quality is poor, your simulation will struggle to converge.
- the mesh density near the wall ("inflation layer") needs to be the right size for your turbulence model. y+ of first cell = ca. 30 to 300 for wall functions, y+ = 1 if the boundary layer is being resolved.
- you may have issues solving particular flow features in your model, such as detachment or high-shear regions.
- easy things you can try to aid/investigate convergence issues:
-reduce the URFs, then return them to defaults once solving.
-make material properties constants, not dependent on T or P.
-model isothermal & no chemistry case first.
-make transport eqns 1st order or solve with a simpler model to begin with.
-contour plot -> residuals -> mass imbalance.

Ema40 · October 20, 2015, 16:54

Thank you very much.
I resolved using the QUICK scheme instead of Second Order.

How do you mean with the last sentence (mass unbalance)?

Thank you
Bye

CFDYourself · October 21, 2015, 09:22

on the last point:

it lets you colour the mesh in fluent in terms of "mass imbalance", which can help identify cells that are potentially causing problems with convergence.

Ema40 · October 21, 2015, 11:13

Thank you very much!
Do you know also how to plot water depths in open channel flow? Because in the Contour plot option there is not water depth.

Thank you
Regards

iitpandey@3 · July 21, 2018, 15:09

Creat a plane (vertical), iso-clip it with water phase having min - 0.499 and max -0.501. it will track the water depth for VOF 0.5. now, go to xy plot in post processing portal. and get the water depth..if it does not work ..do let me know

()

October 13, 2015, 20:40	steady state simulation diverge	#1
Ema40 Senior Member Join Date: Oct 2014 Posts: 124 Rep Power: 12	Hi all. I am simulating an open channel flow with geometry discontinuities (with rapid changes in the cross sections...), in a steady state. Why does it diverge? In a steady state there is not the temporal discretization, thus the divergence of the solution does not depend on the mesh size, right? Can the divergence be caused by the boundary conditions? Thank you

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Setting the height of the stream in the free channel	kevinmccartin	CFX	12	October 13, 2022 22:43
mass flow in is not equal to mass flow out	saii	CFX	12	March 19, 2018 06:21
Solver for transonic flow?	Martin Hegedus	OpenFOAM Running, Solving & CFD	22	December 16, 2015 05:59
Steady state simulation with transient partilcle tracking	mali28	FLUENT	2	February 7, 2013 15:25
steady state simulation	manoj	FLUENT	1	March 20, 2004 08:15

October 20, 2015, 09:27		#2
CFDYourself Member Jim Join Date: May 2015 Posts: 47 Rep Power: 11	a few points, I am also having trouble with divergence of a steady state model currently. - it does depend on the mesh. If the mesh quality is poor, your simulation will struggle to converge. - the mesh density near the wall ("inflation layer") needs to be the right size for your turbulence model. y+ of first cell = ca. 30 to 300 for wall functions, y+ = 1 if the boundary layer is being resolved. - you may have issues solving particular flow features in your model, such as detachment or high-shear regions. - easy things you can try to aid/investigate convergence issues: -reduce the URFs, then return them to defaults once solving. -make material properties constants, not dependent on T or P. -model isothermal & no chemistry case first. -make transport eqns 1st order or solve with a simpler model to begin with. -contour plot -> residuals -> mass imbalance.

October 20, 2015, 16:54		#3
Ema40 Senior Member Join Date: Oct 2014 Posts: 124 Rep Power: 12	Thank you very much. I resolved using the QUICK scheme instead of Second Order. How do you mean with the last sentence (mass unbalance)? Thank you Bye

October 21, 2015, 09:22		#4
CFDYourself Member Jim Join Date: May 2015 Posts: 47 Rep Power: 11	on the last point: it lets you colour the mesh in fluent in terms of "mass imbalance", which can help identify cells that are potentially causing problems with convergence.

October 21, 2015, 11:13		#5
Ema40 Senior Member Join Date: Oct 2014 Posts: 124 Rep Power: 12	Thank you very much! Do you know also how to plot water depths in open channel flow? Because in the Contour plot option there is not water depth. Thank you Regards

July 21, 2018, 15:09		#6
iitpandey@3 New Member Gujrat Join Date: Jul 2018 Posts: 2 Rep Power: 0	Creat a plane (vertical), iso-clip it with water phase having min - 0.499 and max -0.501. it will track the water depth for VOF 0.5. now, go to xy plot in post processing portal. and get the water depth..if it does not work ..do let me know ()