[mfuchs]

Hi,

I´m trying to modell the following 2d-system. There is a glas plate with a thickness of about 2mm and about 4 cm length. It is covered with a thin metall film (~ 10 microns). The film is periodocally heated from the upper side on one end of the system by a given heat current with an frequency of about 100 mHz. The heat sink is placed on the opposite end of the glas plate and is just a wall with constant temperature (293K)



The modelling and meshing is fine but it turns out that it takes very long to calculate and evaluate the experiment. The problem is that I have to use rather small timesteps (1/1000 sec) because of the evaluation that follows after the simulation. So here is the question:

Is it possible to get a time periodic solution? I have found something similar for ansys fluent:

https://www.sharcnet.ca/Software/Flu...ug/node392.htm

I do not need the start up phase but the start up phase lasts some hundred seconds. When I load the results in Ansys post it takes hours to get graphs.

thanks for your help
mfuchs

[SaschaGH]

Hey mfuchs,

if I understand you correctly all you need to do is manually delete the unnecessary (.trn) of the start up process (check with your timestepsize) or a priori define in cfx-pre when the output of your transient results should begin, which can be done in the output control tab. With the reduced number of timesteps you can then create a transient chart in post to see whether your problem has time periodic character.

If your startup phase lasts some "hundred" seconds but you use 1/1000 sec, how many hours/days does it take to calculate the problem ? Ever thought of using a bigger timestep for the beginning to obtain transient stabilization faster and then switch to the desired timestep for evaluation ?
Greetings

[mfuchs]

ok this works fine. Thank you.