Learning Forum

[Perrine]

Hi,

I'm trying to model the expansion of air into an initial void box. The inlet is a small hole in one of the walls of the box.

I used a pressure inlet to simulate the imposed pressure of the outside (1 atm) and i initialized the volume's pressure to 1e-9 atm. I'm doing a density based, transient simulation. My goal is to study the temperature of the air during the high speed admission into the box.

But I get an inlet velocity of 1e5 m/s... My theoretical analysis gives a velocity around 200 m/s. Do you know what could be happening to have such a difference ?

Also, I get an error : "divergence detected in AMG solver" which stops my simulation... I'm using a 1e-5 s timestep.

I have also tried to use the pressure based solver but i get a floating point esception from the first iteration.

Please help !!! Thank you !

[Rob]

How are you modelling air density? What are the initial conditions? How well refined is the mesh? How long will the flow take to cross a cell at 200m/s? 

[Perrine]

Thank you for you reply ! I use ideal gas for the air density.

My initial conditions are :

• pressure in the box = 1e-9 atm, pressure outside the box = 1 atm
• temperature = 21°C in and outside the box

My mesh is average and refined around the hole, the cells have a height of 8 cm for a diameter of the hole of 50 cm. So it would take 4e-5 s for the flow to cross a cell at 200 m/s. 

 

[Rob]

And cell length? Ie how good is the aspect ratio?  Did you patch in any speed for the jet region? 

[Perrine]

The cell lenght is the same as the height : 8 cm. My average aspect ratio is 1.8.

I didn't patch any velocity for the inlet, there isn't anything else than the pressure drop making the air enter the box.

[Rob]

You may need to reduce the time step. But also consider the effect on temperature as gas expands from 1atm to near enough vacuum. Speed initially might be somewhat higher than 200m/s because of the ideal gas law too. 

[Perrine]

I've also tried with a timestep of 1e-7 and i still get a floating point exception at the first iteration... 

I have made another atempt at this problem by modeling a sphere filled with air around my box and using a pressure outlet on the surface of the wall. But it doesn't work either...

[Rob]

Try a less extreme low pressure to get some idea of what's going on.