Pressure inlet of a multiphase model
 

Hi,

I am modeling a liquid injection into a cylinder of constant volume. I set the boundary condition of the inlet as pressure inlet (9bar). There's no outflow of my system since I expected the liquid jet to cool and condensate the vapor phase (4bar) in the cylinder.

However, no matter how I changed other parameters, the pressure of the mixture in the cylinder would become 9bar immediately after the first iteration. Back flow was also noticed around the inlet. And there was no phase change.

But when I change the inlet condition into a velocity inlet, everything turns okay (at least until now).

What is wrong with this? Why can't I use the pressure inlet condition?

Thank you in advance.

How big is you time-step and how big the cylinder? The gas is compressible?

I have time steps of 5ms and 50us. The cylinder is 100mm in diameter and 250mm in length. Why do they matter?

Because pressure changes propagate with the velocity of sound. Let's say you have a velocity c=300m/s. If you change the pressure at the inlet, about 83us later, the whole cylinder (250mm) has this pressure. This is what you see.

If you change the massflow at the inlet, the volume will slowly fill up with this velocity. Thus, also the pressure changes only slowly.

You need to decrease the time-step drastically if you need to use a pressure inlet.

The gas is compressible.

Hi, Philipp. Thank you for your input. I tested it by using a very small time step (1us) and found the interesting effect you described. Actually, the pressure oscillated violently around inlet pressure, ranging from 2bar to 20bar. I guess that's why no phase change occured. (Because the pressure went out of range of what I defined.)

So, was the cylinder pressure of 9bar in previous long time step model an mean pressure? Or it just reached a steady state after a few microseconds? If it was the latter, why no phase change then?