Results does not converges at higher flow rates

faiq01 · August 18, 2021, 06:41

Hello Everybody!
I am doing CFD analysis to find blood damage inside double lumen cannula used for ECMO. My inner lumen is very narrow with diameter of 1.3 mm. I mesh my lumen with minimum size of 0.02mm. I simulate the case with
steady, Viscous laminar (b/c Re= 924), pressure based solver,
pressure velocity coupling = SIMPLE,
Discretization = Green Gauss Node based,
flow rate of 0.00026 kg/sec (velocity = 0.55 m/sec)
and everything goes well (I mean results converges upto e-5)

.
Then I change my flow rate to 0.00053 kg/sec (velocity = 1.11 m/sec), my results still converges but take little more time.
Then I change my flow rate to 0.00106 (velocity = 2.22 m/sec) and my results are does not converge.
I have refined my mesh but that does not work for me instead lead to floating point exception error

.
I have also checked my mesh quality with maximum skewness and aspect ratio of 0.44 & 6 respectively.
What I want to know is why my case converge at low flow rate but diverge at high flow rates

.
Any kind of help will be highly appreciated and willing to share more information if anybody wants.
Pic shows cross sectional view.
Thank you
12.jpg

LoGaL · August 18, 2021, 07:12

Maybe you have to underrelax a bit the equations in the first iterations. It is just reasonable that if you double the velocity, the flow is a little bit trickier to converge.

I would anyways suggest you to use the coupled PV rather than simple, it is much more stable. You may even use the pseudo-transient method in the coupled

faiq01 · August 18, 2021, 07:43

Quote:

Originally Posted by LoGaL

Maybe you have to underrelax a bit the equations in the first iterations. It is just reasonable that if you double the velocity, the flow is a little bit trickier to converge.

I would anyways suggest you to use the coupled PV rather than simple, it is much more stable. You may even use the pseudo-transient method in the coupled

Dear Logal, thanks for your reply
but when I switched from SIMPLE to COUPLES PV, my fluent just crashed (closed by itself)

shereez234 · August 19, 2021, 05:49

This might be a way,

500-1000 iterations in steady state with SIMPLEC with first order upwind accuracy, with relaxation factors Energy eqn (0.9), Pressue eqn(0.3), everything else 0.7

Then switch to your schemes (temporal and spatial) of interest?

jwlee1978 · August 23, 2021, 22:49

When flow rate is low, the flow is laminar (Re = 924). Now mass flow rate is four times. Then Re = ~ 3,700, then the flow might be turbulent. If so, it is very difficult to get converged solution by using laminar model.

faiq01 · August 25, 2021, 02:09

Quote:

Originally Posted by jwlee1978

When flow rate is low, the flow is laminar (Re = 924). Now mass flow rate is four times. Then Re = ~ 3,700, then the flow might be turbulent. If so, it is very difficult to get converged solution by using laminar model.

Thank you for you response.

Reynolds Nos for maximum flow rate is 924, which comes under laminar range.

August 18, 2021, 06:41	Results does not converges at higher flow rates	#1
faiq01 New Member Faiq Ahmad Join Date: Mar 2020 Location: Pakistan Posts: 3 Rep Power: 6	Hello Everybody! I am doing CFD analysis to find blood damage inside double lumen cannula used for ECMO. My inner lumen is very narrow with diameter of 1.3 mm. I mesh my lumen with minimum size of 0.02mm. I simulate the case with steady, Viscous laminar (b/c Re= 924), pressure based solver, pressure velocity coupling = SIMPLE, Discretization = Green Gauss Node based, flow rate of 0.00026 kg/sec (velocity = 0.55 m/sec) and everything goes well (I mean results converges upto e-5). Then I change my flow rate to 0.00053 kg/sec (velocity = 1.11 m/sec), my results still converges but take little more time. Then I change my flow rate to 0.00106 (velocity = 2.22 m/sec) and my results are does not converge. I have refined my mesh but that does not work for me instead lead to floating point exception error. I have also checked my mesh quality with maximum skewness and aspect ratio of 0.44 & 6 respectively. What I want to know is why my case converge at low flow rate but diverge at high flow rates. Any kind of help will be highly appreciated and willing to share more information if anybody wants. Pic shows cross sectional view. Thank you 12.jpg

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Very noisy results in the flow around the tube array	Shay_	FLUENT	0	February 22, 2019 06:39
Exit Corrected Mass Flow Rate Mesh Sensitivity Study	s__s__s	CFX	4	July 20, 2016 11:46
inlet has higher pressure than outlet in axial flow fan	sivakumar	Main CFD Forum	37	July 21, 2014 08:34
Gate valve flow simulations...	nikesh	FloEFD, FloWorks & FloTHERM	5	January 28, 2014 01:31
target flow rate: varying results	Ryan	FLUENT	3	January 9, 2008 13:54

August 18, 2021, 07:12		#2
LoGaL Senior Member Lorenzo Galieti Join Date: Mar 2018 Posts: 374 Rep Power: 12	Maybe you have to underrelax a bit the equations in the first iterations. It is just reasonable that if you double the velocity, the flow is a little bit trickier to converge. I would anyways suggest you to use the coupled PV rather than simple, it is much more stable. You may even use the pseudo-transient method in the coupled

August 19, 2021, 05:49		#4
shereez234 Senior Member M Sereez Join Date: Jan 2014 Location: England Posts: 352 Blog Entries: 1 Rep Power: 13	This might be a way, 500-1000 iterations in steady state with SIMPLEC with first order upwind accuracy, with relaxation factors Energy eqn (0.9), Pressue eqn(0.3), everything else 0.7 Then switch to your schemes (temporal and spatial) of interest?

August 23, 2021, 22:49		#5
jwlee1978 New Member J. Lee Join Date: Aug 2021 Posts: 1 Rep Power: 0	When flow rate is low, the flow is laminar (Re = 924). Now mass flow rate is four times. Then Re = ~ 3,700, then the flow might be turbulent. If so, it is very difficult to get converged solution by using laminar model.