CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   Cavitation (http://www.cfd-online.com/Forums/main/944-cavitation.html)

tom June 24, 1999 17:45

Cavitation
 
Hy, I'll try to do a caitation-simulation (water in a pump) with fluent. the model is a type of vof-model ? what exact is the volume of fluid model ? are the results good ? is the model ok ? does the model show the changing volume of the bubbles with pressure that gets smaller downstream and the effects of that additional "volume" on the fluid ? thanks

Jim Park June 25, 1999 08:15

Re: Cavitation
 
Go the library, look up

C. W. Hirt and B. D. Nichols, "Volume of Fluid (VOF) Method for the Dynamics of Free Boundaries", Journal of Computational Physics, v. 39, pp. 201-225 (1981).

There are other, later papers on the same subject by these authors.

Jim

Raza Mirza June 25, 1999 12:34

Re: Cavitation
 
I guess your question is more on VOF than cavitation, but since the topic for the thread is cavitation, I though it will of value for those intersted to mention the cavitation model in CFD-ACE+.

To learn more about the "full" cavitation model, visit our website at http://www.cfdrc.com. You can find information on cavitation in the "what's new" section located on the CFD Reserach Corporation (CFDRC) main page.

Those interested in applying it to your problems can contact me at rmm@cfdrc.com.

Sung-Eun Kim June 25, 1999 17:47

Re: Cavitation
 
Hi,

The cavitation model in FLUENT employs a so-called homogeneous mulitphase model wherein the mass transfer between gas and liquid is taken into account. Please read the User's guide that has more details of the model.

jurek July 1, 1999 01:32

Re: Cavitation
 
Dear Mr. Tom,

I also try to simulate cavitation in a centrifugal-pump with FLUENT 5. The bubbles can became bigger and smaller depending on the pressure-difference, but I still have no converged solution. Perhaps we can keep contact because cavitation-simulation seems to be a very difficult problem. jurek

Erik Torbergsen July 1, 1999 02:42

Re: Cavitation
 
Hi,

Have you really published with FLUENT any results for the simulation of cavitation in a centrifugal impeller ? I must say, I haven't seen any real two-phase simulation of cavitation in a rotating passage yet so I would very much like to get references of this. At our shop we have been studying the cavitation zones in high speed impellers using single-phase models and this will of course be wrong when the cavitation zones increase and to a large extent influence the single phase flow. But for smaller zones the results are satisfactory.

Regards,

Erik

Raza Mirza July 1, 1999 11:27

Re: Cavitation
 
Over the past seven years, CFD Reserach Corporation (CFDRC) has assessed many different approaches for simulating cavitative flows. The company has developed and published several new and innovative methods. The result of this effort is CFDRC's new Full Cavitation Model.

This model accounts for the following major effects on Cavitation:
[*]Phase Change
[*]Bubble Dynamics
[*]Turbulence
[*]Presence of non-condensable gases

Each of these effects has been found to be very important. The first two effects,phase change and bubble dynamics,are well known and need no further elaboration.

CFDRC's new model also addresses turbulent pressure fluctuations, which have been commonly observed to play a key role for Cavitation. No other past model has considered this effect.

Another strong feature of the new model is accounting for the presence of non-condensable gases in the liquid (which is commonly the case for water). Several researchers have reported this effect to be very significant.

The final model has been tested for a variety of problems and will be undergoing further and continuous assessment in our own application group.

Selected validation and application cases are:
[*]Hydrofoil
[*]Orifice Flow
[*]Axial Water Pump
[*]Centrifugal Water Pump
[*]Diesel Fuel Injector
[*]Automotive Oil Pump
[*]Cavitation in shock absorber
[*]Aerospace LOX Turbopump

CFDRC's many years of Cavitation modeling experience have led to a very robust new formulation. The model solves reliably, despite the many numerical challenges introduced by the large density ratios encountered in Cavitation problems (up to 1:50,000).

CFDRC's Full Cavitation Model is now available as a part of the CFD-ACE+ software package.

Pleae visit our home on the web, http://www.cfdrc.com and explore our experience in simulating problems with cavitation.

John C. Chien July 1, 1999 22:52

Re: Cavitation
 
(1). Thank you for the invitation. I have visited the home page on the pump cavitation. (2). I think, 3-D pump cavitation prediction is a very difficult problem.

Jurek July 2, 1999 01:28

Re: Cavitation
 
Hello, what do you understand under a single-phase-model ? Do you think that the cavitation-model in fluent is a single-phase-model ? I'm just trying to get a cavitated solution with fluent. But I still don't exact what convergence-criterion I can use. The time-steps converge (res. and surface-monitors (stresses)) but the monitored stresses on the surfaces still change from time-step to time-step. I try to continue the calculations until I get constant stresses on the surfaces. Do you think that is ok ? An other problem: I first calculated a steady-state-solution without cavitation. The delta-p was about 300000 Pa (Pressure-outlet, velocity-inlet). Now I reduced the operational pressure, so that pmin in the whole pressure-field was about 50 Pa under the vaporization-pressure of water (about 2340 Pa) and cavitation should start. The model even realized this fact and started. If I look from time to time the (unconverged ?) solution I see really big areas with vapor. Was the difference of 50 Pa too much or is it normal, that it takes a longer time (not only about 5 time-steps of about 0.0001 as in the fluent-tutorial) until a reasonable solution is reached ? thank you

Erik Torbergsen July 2, 1999 03:29

Re: Cavitation
 
Hi.

I don't know the cavitation model in FLUENT but as I have understood it is a two-phase model. I haven't any experience with the convergence criteria in this case so perhaps the FLUENT support desk can help you with this. I have been working both with CFD problems and experimental studies in pumps and I still think the cavitation phenomenas in a centrifugal pump are a real challenge for CFD. Thats why I am very interested in hearing about others simulating it with a two-phase cavitation model.

The way we have been studying it at our shop is by using ordinary CFD models and using iso-pressure fields at the vapour pressure to study the zones all over the pump performance curve. This is compared to full scale video filming of the inlet of the impeller. One interesting thing is that almost regardless of the inlet pressure, a small, narrow cavitating zone will appear at the leading edge. The pressure gradient in this area is very high so am curious about how a two-phase model will describe this. The ordinary cavitation zone (at reduced suction pressure) in a centrifugal impeller at rated point typically appears as a triangular shaped bubble on the blade suction side with the widest part at the tip, but this will change at off-design condition. Another interesting feature is the interactions with main flow when the bubble increases. Do you have any tests or benchmarks for your cavitation simulations ? If not, I would reccommend at least a qualitative study of typical cavitation shapes and pressure profiles in the presence of cavitation. One of the few papers I have seen on this is:

" Centrifugal Pump Performance Drop due to Leading Edge Cavitation: Numerical Predictions Compared with Model Tests" R. Hirschi et al. ASME FEDSM'97 3342.

They present a cavitation simulation although it is single phase.


Jurek July 12, 1999 16:52

Re: Cavitation
 
Hy Erik, thank you for the literature. I tried up to now one mass-flow and have up to now no reasonable results. For example, I reduced the pressure about 5 Pa (!) under the vap.-pressure, and the pump brought about 15 % more pressure. I think, that's not reasonable. Or: If I change the time-steps, the monitored forces change. I have not a good feeling about the cavitation modell in fluent. Jurek


All times are GMT -4. The time now is 06:06.