# How to model a "suction"?

Hypothetical problem. Imagine if you had some fluid in a chamber and you wanted to suck it up. How would you model this in DualSPHysics, if possible?

An outflow with an extremly high velocity value at the outlet? Or is there a better way?

Kind regards

• edited December 2020

Are you thinking about something like drinking from a glass with a straw? In that physical situation you have to draw air first and water follows, so if this is a two-phase flow. In a single-phase flow, you cannot apply a pressure difference between the surface of the glass and the top end of the straw because there is nothing (no particles) in the straw. You lack the medium to transmit the pressure wave.

So you have to set conditions at the two surfaces of water, the one in the chamber, the other inside the straw at the same level (except for a meniscus). This is an unsteady process. As you empty the glass, the pressure conditions in the water outside are steady-ish (atmospheric pressure) but inside the straw they obviously change depending on the unsteady discharge. Also, you have flow contraction and recirculation at the intake which causes energy losses.

So you could impose a steady discharge inside the straw, as you suggest. forgetting the start-up stage. But then the straw has to have a sufficient draught inside the glass, so that the entry effects (the head losses) inside the submerged part of the straw can be simulated. I guess you would bother to use to DSPH mainly for modelling those fine-scale processes at the intake. Otherwise, SPH is an overkill, and Bernoulli, century-old empirical formulas and spreadsheet could do -- in that case you have to think of hypothetical device is providing the energy to lift the flow (a pump). Of course modelling the pump in SPH is an also option, but it could take all of your compute power for ever. (Exaggerated; I have no experience there.)

My two cents, without too much of an inner-peer reviewing. I have never looked into that: how do you set the Poiseulle flow into motion? Maybe this gives some ideas.

• Thanks for the detailed response!

I understand that it is a bit more difficult, than I initially anticipated.

I think you are right in, that SPH might be a bit too overkill for this kind of topic, when Bernouilli can give a pretty precise estimate

Kind regards