transient solution

faw · March 12, 2012, 11:03

Hi everyone,

Just a general question, for transient model in fluent, if we use different time step sizes, why does the simulation results different? Is it because of the nature of the software (as we know that using smaller time step size will give more accurate results). What causes the difference in results? Is it because of the algorithm used (generic transport equation) in fluent as explained in this manual : https://www.sharcnet.ca/Software/Flu...th/node375.htm

based on the equation, does the change of dp��/dt (since dt changes when using different time step size) result in different results for different time steps?

Hope there's someone that can explain this in detail, thank you!

robboflea · March 12, 2012, 11:40

One answer can be the fact that if I set a delta_t, so basically a sampling frequence, I set also the maximum frequence of non-stationary phenomena that can be possibliy captured by the solver.There are I guess also numerical aspects to be taken into account (my knowledge of which is really limited to be fair) but this can be surely an explanation (even if only partial) to your answer.

Cheers,

Rob

faw · March 12, 2012, 12:25

Yes, you are right. That's what I've been thinking of actually..Sure, you did answer my question tho, and thanks Rob!

Does anyone else have any other opinions?

March 12, 2012, 11:03	transient solution & generic transport equation	#1
faw New Member Fara Join Date: Feb 2012 Posts: 6 Rep Power: 14	Hi everyone, Just a general question, for transient model in fluent, if we use different time step sizes, why does the simulation results different? Is it because of the nature of the software (as we know that using smaller time step size will give more accurate results). What causes the difference in results? Is it because of the algorithm used (generic transport equation) in fluent as explained in this manual : https://www.sharcnet.ca/Software/Flu...th/node375.htm based on the equation, does the change of dp��/dt (since dt changes when using different time step size) result in different results for different time steps? Hope there's someone that can explain this in detail, thank you! Last edited by faw; March 13, 2012 at 10:27.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Transient conduction possible in fluent?	jlefevre76	FLUENT	2	February 5, 2013 10:53
Best practice for transient simulations?	siw	CFX	5	October 30, 2010 06:45
residual plot after solution reset	plaunie	STAR-CCM+	1	April 3, 2010 07:53
Solution method for transient problems	Roberto Ciardulli	Siemens	10	November 28, 2000 05:38
steady and unsteady flow	shisheng	Main CFD Forum	10	July 23, 1999 07:00

March 12, 2012, 11:40		#2
robboflea Senior Member Join Date: Nov 2011 Posts: 109 Rep Power: 15	One answer can be the fact that if I set a delta_t, so basically a sampling frequence, I set also the maximum frequence of non-stationary phenomena that can be possibliy captured by the solver.There are I guess also numerical aspects to be taken into account (my knowledge of which is really limited to be fair) but this can be surely an explanation (even if only partial) to your answer. Cheers, Rob

March 12, 2012, 12:25		#3
faw New Member Fara Join Date: Feb 2012 Posts: 6 Rep Power: 14	Yes, you are right. That's what I've been thinking of actually..Sure, you did answer my question tho, and thanks Rob! Does anyone else have any other opinions?