CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Main CFD Forum (https://www.cfd-online.com/Forums/main/)
-   -   actual flow time versus simulation time (https://www.cfd-online.com/Forums/main/172997-actual-flow-time-versus-simulation-time.html)

engineer.iman June 10, 2016 15:28

actual flow time versus simulation time
 
i have a question regarding the time of simulation.
if i have an internal flow system that runs for 20 days for example and i'm simulating that flow using CFD (i.e. Fluent) in order to track some changes to the flow which actually takes place in reality (e.g. particles formation and deposition).... how do i cover the 20 days period of actual flow that is in the reality by the simulation time in Fluent set up or simulation work?

engineer.iman June 12, 2016 01:08

i want to change my question:
for simulating a process that takes 20 days, should i specify 1728000 time steps (if assuming the size of each time step is 1 s)????

20days * 24 hours* 60Minutes* 60 seconds =1728000

flotus1 June 12, 2016 03:46

First of all, you have to estimate a suitable time step size for your simulation.
It should be based on the smallest physical time scale relevant for your simulation. From there on it is quite straightforward to calculate the amount of time steps needed to cover the physical time.

engineer.iman June 12, 2016 15:02

Quote:

Originally Posted by flotus1 (Post 604465)
First of all, you have to estimate a suitable time step size for your simulation.
It should be based on the smallest physical time scale relevant for your simulation. From there on it is quite straightforward to calculate the amount of time steps needed to cover the physical time.

so that means that the total size of all time steps is equal to the physical time of the process??

flotus1 June 12, 2016 18:22

Of course it is.

kwwilcox June 13, 2016 09:48

The following blog:
https://caeai.com/blog/selecting-tar...p-cfd-analysis

Gives a good overview of considerations in selecting a timestep. I would suggest starting with the Courant-Fredrichs-Lewy (CFL) condition, keeping it less than one.

When you are in the solution is also important. You may need a smaller timestep at the beginning which can be increased once the solution becomes stable.


All times are GMT -4. The time now is 05:03.