Pseudo transient settings

Adithya_11 · August 28, 2021, 09:11

Hello
I have questions about the pseudo transient method and could not find the needed recommendation in other threads or in the manual.

I have a geometry of an aircraft that i am simulating on fluent with the SA model and inlet turbulence conditions set as per the recommendations of spalart and rumsey.

My solution diverges with negative nut for the fine mesh with yplus<1 but works with the coarse mesh with first cell height in the log=law layer.

I run the coupled scheme but i have left the pseudo transient settings as default.

What is the effect of reducing the time scale from 1? Is it recommended to reduce it to 1e-4 for example? Should i leave it at automatic method or specify it as user defined?

Similarly, what are the recommendations for the length scale?

LoGaL · August 28, 2021, 12:17

I wouldn’t go any lower than 0.1. Substantially that number allows you to trade stability vs convergence time. In your case you can try to lower it to 0.1 and see if it converges. It is indeed equivalent to under-relaxing the equations.

Btw are you sure the y+ = 1 mesh is not of bad quality? I would not expect the calculation to blow. Actually when the coupled pseudo transient solver blows and you are using simple physics, in my experience there’s something else wrong

Adithya_11 · August 29, 2021, 05:36

Alright, thanks for your recommendation.

I was also quite skeptical about my mesh.
But, the metric in ANSYS MESHING SEEMs satisfactory and I don't get any warnings when I import to Fluent either.

I really am not sure which cells would be causing the issue, if it is due to bad mesh quality.

What I plan on doing now is running the coarser mesh with SIMPLE and the using that to initialize my finer mesh and run that with coupled.

LoGaL · August 29, 2021, 06:14

Nono run everything with coupled, it’s by far the most stable method xD
Just drop that factor to 0.1 and run it

August 28, 2021, 09:11	Pseudo transient settings	#1
Adithya_11 Member Adithya Join Date: May 2021 Posts: 73 Rep Power: 5	Hello I have questions about the pseudo transient method and could not find the needed recommendation in other threads or in the manual. I have a geometry of an aircraft that i am simulating on fluent with the SA model and inlet turbulence conditions set as per the recommendations of spalart and rumsey. My solution diverges with negative nut for the fine mesh with yplus<1 but works with the coarse mesh with first cell height in the log=law layer. I run the coupled scheme but i have left the pseudo transient settings as default. What is the effect of reducing the time scale from 1? Is it recommended to reduce it to 1e-4 for example? Should i leave it at automatic method or specify it as user defined? Similarly, what are the recommendations for the length scale?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Pseudo Transient vs Time-Averaged solutions in Fluent	yousefaz	FLUENT	6	March 16, 2023 00:41
what is the advantage of using pseudo transient?	harshavardhanps3839	FLUENT	1	December 21, 2020 00:12
regarding using pseudo transient?	harshavardhanps3839	FLUENT	1	December 18, 2020 15:21
problem in transient heat transfer of ground heat exchanger	teguhtf	FLUENT	0	April 23, 2020 06:46
Pseudo Transient method	djerezg	FLUENT	6	April 8, 2020 15:50

August 28, 2021, 12:17		#2
LoGaL Senior Member Lorenzo Galieti Join Date: Mar 2018 Posts: 375 Rep Power: 12	I wouldn’t go any lower than 0.1. Substantially that number allows you to trade stability vs convergence time. In your case you can try to lower it to 0.1 and see if it converges. It is indeed equivalent to under-relaxing the equations. Btw are you sure the y+ = 1 mesh is not of bad quality? I would not expect the calculation to blow. Actually when the coupled pseudo transient solver blows and you are using simple physics, in my experience there’s something else wrong

August 29, 2021, 05:36		#3
Adithya_11 Member Adithya Join Date: May 2021 Posts: 73 Rep Power: 5	Alright, thanks for your recommendation. I was also quite skeptical about my mesh. But, the metric in ANSYS MESHING SEEMs satisfactory and I don't get any warnings when I import to Fluent either. I really am not sure which cells would be causing the issue, if it is due to bad mesh quality. What I plan on doing now is running the coarser mesh with SIMPLE and the using that to initialize my finer mesh and run that with coupled.

August 29, 2021, 06:14		#4
LoGaL Senior Member Lorenzo Galieti Join Date: Mar 2018 Posts: 375 Rep Power: 12	Nono run everything with coupled, it’s by far the most stable method xD Just drop that factor to 0.1 and run it