CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   FLUENT (https://www.cfd-online.com/Forums/fluent/)
-   -   Choosing the suitable number of time steps (https://www.cfd-online.com/Forums/fluent/202267-choosing-suitable-number-time-steps.html)

bmaicuong May 25, 2018 12:33

Choosing the suitable number of time steps
 
Hi everybody,

I am very new in Fluent and I have some problems when using this software to do my thesis. Please give me some advice.

I am simulating the flow over a solid sphere and find the drag coefficients on sphere wall, using transient modeling. The radius of sphere is 1m, the length of computational domain is 25m and the sphere is located at 10m from the inlet. The flow velocity is estimated at 2.5m/s.

Based on CFL, I choose the time steps as 0.01s. Please help me choose number of time steps so that Fluent can give me the best results.

Many thanks in advance. Sorry for my bad English.

LuckyTran May 25, 2018 15:27

There's no precise answer.

You need to run long enough for the initial condition to flow out of the domain before collecting any data. So you need to run at least 1 flow-through time, more to be safe. How much is safe enough? 2? 3? 4? It's not known. At 25 m and 2.5 m/s, 1 flow-through time is 10s. During this period, you do not collect any data because the flow is not the flow that you want.

Then you need to average over enough flow cycles of any transient phenomenon. I.e. 10 or 20 or 30 cycles of the vortex shedding. How many is needed? Again there's no precise answer. But you can estimate the error using a normal distribution, i.e. sigma/sqrt(n) and use this to estimate how many cycles you should average over. Then there is the question of what is the vortex shedding frequency? Often you don't know beforehand.


All times are GMT -4. The time now is 17:38.