# Emptying tank & Multiphase

 Register Blogs Members List Search Today's Posts Mark Forums Read

 August 15, 2007, 14:00 Emptying tank & Multiphase #1 Ruggero Guest   Posts: n/a I have to simulate an emptying tank with a multiphase liquid-gas inside, in a transient simulation. I'm using a Eulerian Multiphase treatment. My problem is: 1-How to insert a fixed mass flow outlet (eg: exit of 0,1 kg/s)? When i tried an error occurs saying that there is no pressure gradient that allows that mass outflow. But i cannot insert a pressure condition! 2-Isn't possible at all to use a ideal model for the gas phase? It let me use only isobaric model for density. I have some others questions, but for the moment I think it's enought. Thanks to allo of you for answering.

 August 16, 2007, 05:54 Re: Emptying tank & Multiphase #2 groundhog Guest   Posts: n/a Hi Why not try a negative facing inlet. That can be used to suck fluid out of the domain. Please check the manual for details of the density options. Hope this helps.

 August 16, 2007, 07:24 Re: Emptying tank & Multiphase #3 Ruggero Guest   Posts: n/a Ok. I'll try and then i'll be back with the results. I hope.

 August 16, 2007, 07:54 Re: Emptying tank & Multiphase #4 Ruggero Guest   Posts: n/a Here i am. Unluckily it doesn't work. I have back an error: ## error 064: fixed velocity boundary condition outflow exceeds inflow by 5 per cent. Continuity correction not attempted. What shall i do? The fact is that i need to know the emptying pressure with a fixed mass outflow.

 August 16, 2007, 08:07 Re: Emptying tank & Multiphase #5 groundhog Guest   Posts: n/a I guess you need to think about the boundaries. You have, I guess, an inflow inlet, and now an outflow inlet. If the density is constant, then the system is overconstrained.

 August 16, 2007, 08:11 Re: Emptying tank & Multiphase #6 Ruggero Guest   Posts: n/a No! I choose an isobaric model for density. And i have no inlet, since the tank is full at the beginning of simulation end shoud empy little by little.

 August 16, 2007, 14:57 Re: Emptying tank & Multiphase #7 Harry Guest   Posts: n/a What other boundaries do you have in the model?

 August 16, 2007, 15:36 Re: Emptying tank & Multiphase #8 allan Guest   Posts: n/a er...how can you empty the tank with only one outlet. In reality if you had a full tank with one outlet and it was opened, it would not empty. It needs a breather. If you put a put a pressure boundary on a couple of cells at the top of your tank and set it to zero, you should get outflow. You probably need gravity here as well...or have I mis-interpreted what your trying to do! cheers allan

 August 16, 2007, 16:01 Re: Emptying tank & Multiphase #9 Harry Guest   Posts: n/a that was what I was wondering when i asked about boundaries and surely vof would be a better choice than e2p

 August 17, 2007, 03:45 Re: Emptying tank & Multiphase #10 Ruggero Guest   Posts: n/a I've got the wall boundaries and the symmetry boundaries and, of course, the outlet. No other kind of boundaries.

 August 17, 2007, 05:28 Re: Emptying tank & Multiphase #11 Harry Guest   Posts: n/a I think you need to think about the physics and not just the CFD, if there is no inlet (or method of fluid ingress) how is it ever going to empty? Get an old beer bottle fill it to the top with water, wrap your lips around it and try and suck out the water, it's never going to happen.

 August 17, 2007, 05:40 Re: Emptying tank & Multiphase #12 Ruggero Guest   Posts: n/a Yeah, but the tank should have a pressure bigger than outside. The problem should be an incompressible gas-phase, but i used an isobarico model for density, so there should not be. What i have to simulate is how much I can empty such a tank, in which, while emptying, the pressure inside get lower and so the fluid evaporate (so tending to go against the pressure falls). I could use a pressure boundary, but the mass outflow is limited and i cannot impose this.

 August 17, 2007, 05:43 Re: Emptying tank & Multiphase #13 Ruggero Guest   Posts: n/a As I explained in another post, the difference of pressure between inside and outside the tank should let the fluid to exit. But the problem is how to limitate the max mass outflow.

 August 17, 2007, 06:15 Re: Emptying tank & Multiphase #14 allan Guest   Posts: n/a I don't really understand why you're wanting to limit the max mass outflow. The flow area or pressure drop will determine this. allan

 August 17, 2007, 06:22 Re: Emptying tank & Multiphase #15 Ruggero Guest   Posts: n/a The idea is that this tank should feed an engine, so it should guaranted a determined massflow (no more, no less) for 5 seconds. We are assigned to study the tank with an autopressurization sistem.

 August 17, 2007, 06:30 Re: Emptying tank & Multiphase #16 allan Guest   Posts: n/a in that case then, I would use a suction inlet as and outlet suggested previously, but you will still need a pressure inlet to allow air into the tank to let it drain allan

 August 17, 2007, 06:36 Re: Emptying tank & Multiphase #17 Ruggero Guest   Posts: n/a ok! But what pressure use? In my simulation pressure should be found not imposed.

 August 17, 2007, 06:43 Re: Emptying tank & Multiphase #18 Ralphie Guest   Posts: n/a I think the problem is similar to an emptying LPG bottle or a lighter. You have a pressurised liquid-gas-mixture inside your tank. When the valve is opened gas is released, the pressure in the tank drops and more liquid fuel evaporates, i.e. the pressure is buffered.

 October 24, 2007, 13:23 Re: Emptying tank & Multiphase #19 AndyR Guest   Posts: n/a Ruggero, Work backward from the engine. The engine wants X gpm, but it also needs Y psi to flow that amount from your supply through the plumbing, carbeurators, injectors etc etc. In otherwords a requirement of just X gpm is incomplete, it needs to be X gpm at Ypsi. If you cant get that data, make an estimate. You know have a known fixed pressure that you can apply to the "outlet" boundary of your problem. Then you can adjust the tank pressure iteratively to get the required mass flow rate at the required supply pressure. The code should run just fine that way. Andy R