Easy question: how to simulate water flow?
 

Hi mates,

I want to simulate in a "wind" tunnel chamber a flow past a flat plate oriented perpendicular to the flow, with dimensions (of the flat plate), but being the flow water, instead of air.

The fact is that I only have the pressure inlet conditions of this water flow (40kPa), but I don't know how to set up this kind of problem, as I am used to set up velocity inlet and outflow as boundary conditions and not pressure inlet & outlet.

Could you please guide me a little in how to achieve this and how to set up this pressure inlet and outlet BC?

Thanks a lot CHeers, Freeman

Re: Easy question: how to simulate water flow?
 

I agree that velocity inlet & pressure outlet is the best way to set a problem. I would model unsteady, setting a velocity inlet (with a UDF) as function of time, from 0 to a value that you consider a maximum, and stopping the velocity change (or controlling with a feedback algorithm) when the requested inlet pressure is reached. After that you can return in steady. I never tried with Fluent, but with other (old, less sophisticated, less robust) CFD codes works. Gian PS: vmax=sqrt(2*p/rho)=8.94 m/sec

Re: Easy question: how to simulate water flow?
 

Hi Gian!

I had great succeed in my simulation: although simulation case was quite rough for the solver, I was able to simulate the flow past the plate parallel to the flow. I had some negative absolute pressures, something not real, but I read here at CFD-online also some other people having the same problem while simulating compressible flow. As I am interested only in structural loads on the plate (that is total pressure over the plate), that's not a big problem for me

Again, turbulent viscosity was quite high due to the very high reynolds problem (10e6, water at 9m/s, strong recirculation zones), but I kept TVR below 1e5 in the 94% of the domain (1.2Mill cells).

Thanks for your advice Gian,

Regards, Freeman

Re: Easy question: how to simulate water flow?
 

At which velocity you obtain the wanted inlet pressure ? Negative pressure is not nice: maybe if you increase the velocity more gently, the problems disappears. The problems is in any case in the solver and in the material properties: the density of water is high, and the compressibility is not negligible if the velocity is normal to a wall: you have like a water hammer problem, where pressure waves travel at the sound speed (1200 m/sec) with pressure over (and under) shooting Gian

Re: Easy question: how to simulate water flow?
 

Yes, you're right Gian: that's the only flaw in my simulation, that negative absolute pressure... but for now, I am only interested in the total pressure over the plate for a structural analysis, so absolute negative pressure and cavitation is not a headache for now, as the gauge pressure distribution seems to be pretty fair over the plate.

I am using pressure inlet BC: I set 42kPa at inlet, but as you said, that is aprox. 9m/s, calculated by Bernuilli's equation... and 9m/s and dealing with water is a great velocity.

I really appreciate your notes, Gian. By the way, now I am trying to learn VOF method, and I think I understand the main proceedings to set up and simualte. But I need more examples, specially 3D in order to get a global and broad picture of it... perhaps you have ever work with this VOF method? Have you got any case studies examples of it?

Thanks in advance,

Warmest regards, Freeman

Re: Easy question: how to simulate water flow?
 

As first attempt with Fluent, I simulated dam break problem using VOF, comparing results with experiments and simulations with a 2D solver of shallow water equations (Riemman solver) Results were not so bad, mainly in the region with bottom steps, where the flow is 3d. For the rest the results were similar. If you like, I can send to you a zip file with the test case. Cheers Gian

Re: Easy question: how to simulate water flow?
 

I would be really grateful to recieve these files, giovanni: I am very interested in VOF method for modelling also my case study.

I send you a private email to manage this...

Cheers,

Freeman