transient convergence- PISO/Coupled vs Simplec

Ritah · July 3, 2019, 07:56

Hi,

I am doing a transient simulation with a large time step (0.8s) to compute the flow and turbulence field in a Zn-bath. In the fluent user manual is indicated that the transient simulation should work better with PISO or Coupled (for larger time steps). However when I use PISO or Coupled the maximum Velocity in the bath increases very much and the mean velocity varies also quite a lot in comparison with the Simplec scheme. I also use very low under-relaxation factors of 0.05 for flow and turbulence to get a stable solution.

Does anyone have experience with different schemes for the transient simulation with large time steps?

Thank you

LuckyTran · July 3, 2019, 11:27

PISO does not like large time steps. COUPLED with default settings is closer in behavior to SIMPLE than PISO.

Using a large time-step and small URF is just not optimal (the small URF negates the large time-step).

I think you just have general convergence / stability issues and the problem is not really about which scheme to use. But if you're going to use large time-steps, definitely don't use PISO.

Ritah · July 4, 2019, 07:16

Thank you for your answer, LuckyTran.

I have first run a Steady calculation and reduced the under-relaxation factors to get a stable solution. However the local Velocity field still changed slightly, so I continued with a large time step transient and frozen-flux formulation to further stabilize it. At your suggestion I would try bigger URFs.

We have a complicated geometry/mesh and it is difficult to obtain a stable Steady state flow. I thought that a proper scheme would help the stability...

Best,

July 3, 2019, 07:56	transient convergence- PISO/Coupled vs Simplec	#1
Ritah New Member Andrea Join Date: Jun 2019 Posts: 12 Rep Power: 6	Hi, I am doing a transient simulation with a large time step (0.8s) to compute the flow and turbulence field in a Zn-bath. In the fluent user manual is indicated that the transient simulation should work better with PISO or Coupled (for larger time steps). However when I use PISO or Coupled the maximum Velocity in the bath increases very much and the mean velocity varies also quite a lot in comparison with the Simplec scheme. I also use very low under-relaxation factors of 0.05 for flow and turbulence to get a stable solution. Does anyone have experience with different schemes for the transient simulation with large time steps? Thank you

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Using convergence manager for Transient simulation	lma80	FLUENT	3	February 19, 2018 05:13
Convergence Criteria for first Timestep in transient run	aerost	CFX	3	February 19, 2013 04:01
transient case diverged all of a sudden after already convergence	wc34071209	OpenFOAM Running, Solving & CFD	2	November 12, 2012 06:55
Convergence for a transient problem	jgrisham	FLUENT	0	July 17, 2012 11:47
Transient axial rotor/stator convergence issue?	Nicola Viscanti	CFX	3	March 17, 2010 04:15

July 3, 2019, 11:27		#2
LuckyTran Senior Member Lucky Join Date: Apr 2011 Location: Orlando, FL USA Posts: 5,674 Rep Power: 66	PISO does not like large time steps. COUPLED with default settings is closer in behavior to SIMPLE than PISO. Using a large time-step and small URF is just not optimal (the small URF negates the large time-step). I think you just have general convergence / stability issues and the problem is not really about which scheme to use. But if you're going to use large time-steps, definitely don't use PISO.

July 4, 2019, 07:16		#3
Ritah New Member Andrea Join Date: Jun 2019 Posts: 12 Rep Power: 6	Thank you for your answer, LuckyTran. I have first run a Steady calculation and reduced the under-relaxation factors to get a stable solution. However the local Velocity field still changed slightly, so I continued with a large time step transient and frozen-flux formulation to further stabilize it. At your suggestion I would try bigger URFs. We have a complicated geometry/mesh and it is difficult to obtain a stable Steady state flow. I thought that a proper scheme would help the stability... Best,