how to deceide the time step for unsteady turbulent flows. is there any theoritical procedure avialable to calculate the time step.

plz help me

maybe evaluate the time needed for the flow going through the domain and then choose a resolution? but with chemical kinetics or other phenomena (combustion/explosion....)you may think the time scale for that specific phenomenon.

Hi,

There is also an adaptive time approach to the problem. Instead of fix time step you can use adaptive one ...

PW

If your flow is periodic, obtain the period ie. 1/frequency, then divide by the number of time steps you want say 100. This will be the value of your "Time Step Size".

Example your frequency is 220hz, then you period = 0.00454545s, for 100 time steps, your "Time Step Size" = 4.54545e-5.

The value you use for "Number of Time Steps" depends on the number of cycles you want. if you want 5 cycles then you use 5*100=500.

You can then use the default value for the rest.

thank you....... is time step depends on the property which we require( for example temerature in turbulent flow or pressure in turbulent flow)?. does it depend on the accuracy of the property which we require?