[jojoFD]

Hi,

I am simulating with a dynamic mesh the oscillating movement of a piston inside a closed tank, pressurized at 200 bars. The velocity follows a sinus curve, from 0 to 5m/s for one oscillation. The problem is that throughout my simulation, the entropy is building up and never going down, then affecting all my quantities of interest also (pressure, temperature, density, etc) - see enclosed screenshots.

I was using a k-w SST model + ideal gas law for the density in Fluent, but I just read that it works only for low pressure, and that I need to take into account the compressibility factors (https://www.sciencedirect.com/topics.../ideal-gas-law).

Any ideas on how to do that in Fluent ?

Thank you so much in advance for your answer !

entropy.jpg

temperature.jpg

force x rod valve.jpg

[LoGaL]

So instead of ideal gas law you choose Peng-Robinson gas law in the material settings.

[Martin_Sz]

U need to describe Ur problem more precisely
For example do U need to estimate temperature field
In Ur model
In first step make simulation on atmospheric pressure
To check that Ur problem dissapear

[jojoFD]

Thank you both very much for your answers
My main quantity of interest is the pressure losses in the system, and the force on the walls but I also monitor temperature and density.
I'll try both and keep you posted !

[jojoFD]

So the try with peng-robinson was not concluent : I got the same phenomenon

[LoGaL]

If peng robinson didn’t work, your problem is not compressibility/real gas effects

[jojoFD]

Originally Posted by Martin_Sz

U need to describe Ur problem more precisely
For example do U need to estimate temperature field
In Ur model
In first step make simulation on atmospheric pressure
To check that Ur problem dissapear

And at atm pressure it does the same, decreasing/increasing quantities. So it seems to be a numerical error right ?

[jojoFD]

Hi LoGal, thanks for your anwser
Yeah that's what I thought. Peng-robinson seems to affect the results though, but not enough to stabilize it. I tried switching to Spalart-Allmaras model but same thing happens.. I'll refine the mesh to see if there is any difference. The only thing is to know if the result is physical or not... CFD basic problem haha
Let me know if you have some other ideas. Thx !

[LoGaL]

I was just thinking, which thermal boundary condition are you applying at the wall? Adiabatic? If yes, where is all the heat generated by friction going? Could be consistent with the fact that you have entropy building up.

[jojoFD]

Thanks for the idea, I thought about that too and tried : even with non adiabatic walls (heat flux at the walls), I still get the same phenomenon