time step size too small and max. time steps is not enough

czhao86 · September 16, 2013, 11:31

Hi,

I am doing a steady state DPM, and the time step is always in the order of 1e-7s with Brownian force on, and it takes a lot of time steps for them to reach the outlet. I changed the 'Step Length Factor' from 5 to 1, and the time steps I assume only change by 5 times, and it is still too small. Anyone have some idea? Thanks.

flotus1 · September 16, 2013, 12:19

Its the Brownian forces again...

I assume you have rather small particles.
The usual integration mechanisms when Brownian forces are involved only work (because of stability AND accuracy constraints) if the time step size is in the order of or smaller than the particle relaxation time.
For small particle Reynolds numbers (

) the particle relaxation time is
$\tau_p= \frac{d^2 \rho_p}{18 \eta_f}$ .
Here

is the particle diameter, $\rho_p$ is the particle density and $\eta_f$ is the dynamic viscosity of the fluid.
This relaxation time becomes small for small particles, leading to small time step sizes.

czhao86 · September 16, 2013, 12:23

So in another word, if the Brownian motion is on, I can not simulate several hundreds of seconds process, as the time step is 1e9...

flotus1 · September 16, 2013, 12:27

You could do this with implicit integration methods, but this will lead to huge errors in the particle velocity distribution.
I dont know if this is possible in Fluent.

czhao86 · September 16, 2013, 13:27

Thank you very much. Time for me to learn implicit theme then...

September 16, 2013, 11:31	time step size too small and max. time steps is not enough	#1
czhao86 Member Rexxar Join Date: May 2012 Location: Bethlehem, PA Posts: 36 Rep Power: 13	Hi, I am doing a steady state DPM, and the time step is always in the order of 1e-7s with Brownian force on, and it takes a lot of time steps for them to reach the outlet. I changed the 'Step Length Factor' from 5 to 1, and the time steps I assume only change by 5 times, and it is still too small. Anyone have some idea? Thanks. __________________ Best, Rexxar

September 16, 2013, 12:19		#2
flotus1 Super Moderator Alex Join Date: Jun 2012 Location: Germany Posts: 3,399 Rep Power: 46	Its the Brownian forces again... I assume you have rather small particles. The usual integration mechanisms when Brownian forces are involved only work (because of stability AND accuracy constraints) if the time step size is in the order of or smaller than the particle relaxation time. For small particle Reynolds numbers () the particle relaxation time is $\tau_p= \frac{d^2 \rho_p}{18 \eta_f}$ . Here is the particle diameter, $\rho_p$ is the particle density and $\eta_f$ is the dynamic viscosity of the fluid. This relaxation time becomes small for small particles, leading to small time step sizes. czhao86 likes this.

September 16, 2013, 12:23		#3
czhao86 Member Rexxar Join Date: May 2012 Location: Bethlehem, PA Posts: 36 Rep Power: 13	So in another word, if the Brownian motion is on, I can not simulate several hundreds of seconds process, as the time step is 1e9... __________________ Best, Rexxar

September 16, 2013, 13:27		#5
czhao86 Member Rexxar Join Date: May 2012 Location: Bethlehem, PA Posts: 36 Rep Power: 13	Thank you very much. Time for me to learn implicit theme then... __________________ Best, Rexxar

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
AMI speed performance	danny123	OpenFOAM	21	October 24, 2020 04:13
Problem of simulating of small droplet with radius of 2mm	liguifan	OpenFOAM Running, Solving & CFD	5	June 3, 2014 02:53
InterFoam negative alpha	karasa03	OpenFOAM	7	December 12, 2013 03:41
pisoFoam with k-epsilon turb blows up - Some questions	Heroic	OpenFOAM Running, Solving & CFD	26	December 17, 2012 03:34
directMapped problem	panda60	OpenFOAM Bugs	4	July 8, 2010 10:23

September 16, 2013, 12:27		#4
flotus1 Super Moderator Alex Join Date: Jun 2012 Location: Germany Posts: 3,399 Rep Power: 46	You could do this with implicit integration methods, but this will lead to huge errors in the particle velocity distribution. I dont know if this is possible in Fluent.