Hi,

I am solving a 2d laminar channel flow. I am using the pressure based solver/ mixture model / cavitation model.

I want to model a pressure oscillation within the channel in the mixture of water and vapor (when cavitation happens). both fluids are compressible.

I implement the oscillation in terms of water density (UDF) which gives the correct pressure wave.

Now everything works fine, i can see the pressure wave moving in the channel. But when i increase the amplitude of the wave the pressure reaches vapor pressure of the water and cavitation occurs in some places in the channel.

at this stage i can see in the volume fraction fields that the volume fraction of vapor is increasing accordingly. However, the pressure wave continues so that, a moment later, the pressure again is higher then the vapor pressure.

this is when the problem arise. I would expect that the vapor should condense again to water when the pressure increases. But what i can see from the calculations is that pressure starts to increase unbounded in the points where there previously was cavitation. also i can see that the vapor volume fraction in these areas is INCREASING. After a while the pressure increases in the whole domain to unphysical values. also vapor volume fraction is reaching 1. in the areas originally with cavitation. but also the areas are increasing in size so that after a while half of the domain contains 100 % vapor.

I checked the phase change rate from vapor->liquid and this is also large during this time (it grows uncontrolled). phase change rate from liquid-> vapor is 0 as it should.

During the complete computation all residuals are converging at least 3 orders of magnitude.

i tried to decrease the time step by a factor 10 but the problem still remained.

Any ideas on this? i am at a loss in this problem and have no further ideas on how to approach this so anything is welcome !!!

-mange

I am aware of the cavitation model inside FLUENT, but I've never used it before and I'm not familliar with the theory behind it, what the limitations are etc.

Looking up the FLUENT 6.3 documentation, section 23.11.2, it talks about needing to specify mass transfer rates in the phase interaction panel BEFORE turning on the cavitation model. Have you tried doing this?

Yes i tried doing this also, but it does not seem to make a difference

Further comments anyone?

[congtuha]

Originally Posted by mange
;155854

Yes i tried doing this also, but it does not seem to make a difference

Further comments anyone?

I guess you applied wrong boundary conditions. Is the pressure specified at the outlet? Also, if there are unphysical values in the domains there might be something still wrong. Let's use different cavitation constants.