Discussion Forum

Hi

is it a time run with a smoothed (heaviside see functions in V4, index your doc in V3.5) pulse you are looking for, the I would define a function of time and apply a defined load.

hope this answers you question
Ivar

hi Ivar

i am using 3.5a ....i didnt get you ....is it function ? or what ?

Hi

a "Heaviside" function = a smoothed step function with derivable turn on and turn off transitions, check your doc.

Ivar

hi ivar

i founf the function thanks alot , the problem I have 2 phase filed flow and i am doing 2 step tp solve it (first for ohase field then the whole structure) i was wandring should i solve it in saparit ? (just for Axial Symmetry, Stress-Strain) then solve it for the whole structure ?


i realy greatful for u r answers

salim

Hi

I'm not sure I catch everything of your issue, but in general, when you know that two problems can be segregated / separated then its always easier to run them one after the other, (or loop a few times back and forth) as normally each model is simpler and calculates faster. For thsi you have several options in COMSOL: segregated solver set-up so you solve for some variables and then for the others, or the sequencing of the solver manager last sequence tab, whree you perform your solving, step by step, save the set-up and store the intermediate solution, and then you can re-run (3.5a) the sequence automatically next time.

Have fun Comsoling
Ivar

Hi Ivaer
I am greatful for your efforts. In my modeling fSI problem. I am using 2 phase flow. so first i solve it for phi and psi . then solve it for the whole system. As i read in the Documents to solve the heaviside i need to solve the w and Ur so am confused how to solve them seprartly and the then solve the whole system together is that possible ?
another thing have u read about sponge layers, (gradually increasing viscosity) to damp the wave refliction ?


many thanks

Hi

I'm more a structural than a FSI guy, but I remember there were some discussions about PML (perfectly matched Layers) for fluidics on the forum, I believe a few weeks ago. If I remeber well these are not implemented as standard by COMSOL but there are some articles with tricks how to do something.

For me the Heavyside functions are there basically to smoothen rapid transients in time or space coordinates, I do not catchwhy you need for that reason to split up your model.
I was thinking of a multiphyiscs model where you have two rather separate phenomena going on, where you could decouple both in a reasonnable way, i.e. high frequency EM oscillations and slow thermal variations. Then you can separate a frequency scan for the electromagnetics (EM), and take the power average per periode and use that as a scalar or slow varying value for a time or stationary analysis of your heat transfer (HT).

But from what you are sying I suspect it's not directly applicable for you.
Again there are others more experience in fluidics on the forum that could have further ideas.

Good luck
Ivar