Learning Forum

[murarz]

Hello everyone!

So far, I have been working with steady state simulations. But now I need to perform a transient one. Is there any rule of thumb when relating transient with steady state simulations in terms of the number of required iterations? For instance, my steady state simulation requires more or less 20 000 iterations to converge. In the transient case I would need a 0.001 time step size and because the residence time is 1 sec, the number of time steps would be equal to 1000. Please don't tell me that I would have to set 20 000 iterations/time step. 

 Or can I just run a steady state simulation and use the results from it to run a transient simulation simply by clicking continue from previous solution and then I could use less iterations/time step e.q. 100 ?

Regards

[danbence]

Reformulating your question is helpful. Basically what you are asking is:-n nHow different is your converged steady state simulation compared to the initial guess (initialization algorithm) ?nHow different is the flow field between one time step and the next ?nHow much bigger is ?1? compared to ?2?n nThe correct answer (but not that useful) is that it is impossible to say. It depends on:-n nHow good the initialization is at guessing the flow fieldnHow quicky the flow field in your model is supposed to change (turbulence, deforming topology etc)nWhat timestep you picked.n nHowever, one can safely say that ?the smaller the timestep, the fewer iterations needed per timestep?. With a small timestep, the solver does not have to work so hard, because the flow field of the previous step is so similar.n n

[murarz]

Thanks for the response. But as for your answer 1:How good the initialization is at guessing the flow field. In your experience is it an acccurate way to employ state state results as initializing parameters for the transient case? Or should I initialize it manually?n

[danbence]

The result of a steady state simulation can be an excellent starting point for a transient simulation. n

[murarz]

Thanksn