Hi all, I have following problem. I am trying to model very simple situation: a pipe with the water. The pipe outflows free to air. Since there is very high pressure at the inlet BC the water reaches velocity of approx. 200m/s near the outlet. Everything seems to be ok, however I notice about 25% of overestimation of flow rate comparing to measurements. I guess that there is a mistake in setting pressure to 0.0 at the outlet (which I was thinking is adequate for free outlet). I thing that there should be some pressure loss factor due to fragmentation of the flow into droplets etc. (which one I dont want to model since my interest is in flow rate and behavior inside a pipe). Anybody can help?

Luk

If you're modelling your outlet as the physical area with a velocity boundary condition, you could well cause an overshoot in the flow rate.

When you specify velocity as the boundary condition, CFX calculates the mass flow directly from w = density * velocity * area (setting a velocity condition actually sets the massflux through the boundary (density*velocity) as CFX will interpolate the density based on the surrounding conditions and will measure area directly from the grid). If velocity is specified, and the boundary is modelled as a physical area not accounting for discharge coeffcient, your massflow will have to increase in order to satisfy the velocity requirement, assuming it treats water as an incompressible fluid.

In this case, you can either regrid the outlet assuming some discharge coefficient or you can use your existing grid and apply a loss coeffcient.

Hi John, The outlet I gave is not a velocity but Static Pressure BC. I set 0.0 of relative pressure for outlet and total pressure of 30 MPa in the inlet.

I gave title for this tread "high velocity outlet bc" because now I anticipate that it should be some kind of formula for pressure drop on the outlet. I guess that high speed of the outflowing water, which may be in order of 100m/s causes a serious aerodynamic resistance on the outlet. The problem is that I dont want to calculate a breakup of primary flow since I am not interested in the shower range nor droplets diameter but only in flow in the pipe.

Lukasz

Hi Luk,

As you have suggested, the problem is that the losses are not sufficient to restrict flow. Whether the problem is at the outlet or internal to the domain is unclear. However, if you know what the outlet mass flow rate should be, you should specify the mass flow rate at the outlet.

Note that you can still force a uniform static pressure at this location (to simulate the sudden expansion), just set the "Mass Flow Update" option on the outlet boundary condition to "Shift Pressure" and specify a value of zero. The shift pressure option will use shift the supplied pressure profile up and down to give you the desired mass flow rate. Supply a constant value means you have a constant static pressure profile.

If the outlet static pressure seems too high, you may deduce there are not enough losses within the domain. The other thing to consider is the inlet velocity profile and turbulence level.

-CycLone

Thanks for Your great answer CycLone.

Since I had no a very good understanding of the Help in matter of Shift Pressure I wan wondering how it works. Is that right that it fits static outlet pressure to give a desired outlet mass flow?

Thanks again, Luk

Maybe I describe in a bit more details my arrangement: There is a simple pump (consisted of cilindrical volume and a piston) with a pipe at the outlet. In the end of the pipe is a security high pressure valve that realeases a water free to air. The pressure given on the piston is 56 MPa. The diameter of the piston is about to be 20 cm while the pressure of the pipe and the inlet to the valve is 20 mm. It is real testing arrangement for security valves. The guys handling this has some technical problems and are interested what are the pressure losses in the valve in the connection between the pump of the cylinder and connecting pipe so on.. Right now I have two questions: - a first is given above - does anywant know a head loss coefficient of the pipe (the outlet of the valve) that finishes free to air? For example if pipe enters to cavity there is unity loss factor but I dont know a loses free to air, - a second question is what would be a best BC for the valve in the pump. This valve is pressed and pump out the water from the pump cavity. Is the static pressure is the best?

I would be very gratefull for help.

Luk