CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   FLUENT (https://www.cfd-online.com/Forums/fluent/)

 kumarswamy September 28, 2007 08:51

how to decide time step and no. of time steps for a problem

 kb September 28, 2007 15:54

At first try with smaller time step (e.g. 0.0001)if you see your solution converage try to increase it an order of magnitude but if see the solution not converage usually you must decrease it. The number of time steps depend in your experimental data you must know how much time you need from experiments in order your results reach steady state solution.

Cheers kb

 JWilliams September 28, 2007 16:14

Determining your timestep depends on your system. One rule of thumb, is take the characteristic length scale (L) of your system (i.e. length scale of pipe, length scale from inlet to outlet etc.). Next is estimate an average velocity scale (V) of your fluid.

Find the estimated residence time of your fluid by (L/V). If you multiply it by 1/3 you will come up with an estimated timestep. Try this, if it converges in a timely manner you are good to go, if it does not converge, start reducing its value. If it converges but takes forever, start increasing its value.

For a natural convection problem you can estimate the timestep by

SQRT(L/(g Beta DT))

where:

L = characteristic length scale

g = gravity

Beta = thermal expansivity of fluid

DT = temperature difference

 PratiK Mehta October 2, 2007 09:59