Best Practice for weirs/dams with free surface

[siegijank]

I have to do research on the flow capacity of a weir for a hydro power project. So I wanted to use the free surface option to simulate various volume flow rates to get the losses on that weir.

Basically I have only one solid part for the water with a mass flow rate at the inlet and a zero pressure condition at the outlet.

While during the first time steps the (physically instationary) flows seems pretty like that what I expect later the model fills completely with "water" to the top of the geometry. I know form the real live model, that doesn't represents what it should be.

Are there any values for mesh size & time step size for such setups? I want to have a physically "stationary" flow as a result to get the head losses on the weir.

Attached you can see state_02 which looks pretty like as I expect first. state_02 shows a state, which doesn't represent the reality. The water level should be at the head water level approximately in the middle of the geometry....

[siegijank]

I moved a little bit forward: now  I get reasonable results by increasing Inner Iterations from 3 to 10.

I'm still searching for an opportunity to adjust the tailwater level to a fixed height.

[Jon.Wilde]

I have run a similar model. I used an inlet extended downwards at the beginning of the analysis (it was not a particularly large region, not then entire floor before the weir). I had a pressure condition here, using rho.g.h that represented the pressure on the floor of the model. That gave me the inlet at the correct height. Perhaps this might help?

It also looks like you could use a finer mesh - this is especially important where the fluid and air interact.

[siegijank]

Yes, - I thought of such a pressure BC anyway. But in practise I didn't know how to apply that (which should give a sort of triangle pressure distribution). I tried using 10 steps with constant pressure to approach that, - but had problems with solver stability.

The most reasonable results I got using an additional auxiliary fluid path for bending the flow to the horizontal and have then a (constant) pressure condition there. What’s more it should be only static pressure? there because I have to investigate the weir due to a certain tailwater level.

So how did you define rho.g.h in practise then?

thanks, Siegi

[Jon.Wilde]

Could you share a cross section of your latest model at all?

I may need a translation on tailwater level, do you mean you have a set water inlet level?

Kind regards,

Jon

[siegijank]

Attached you will find a sketch on the issue. The tail water level is the static water level downstream, - mostly given from measurements or from 1d or 2d HEC-RAS simulations of the river branch.

So basically it should work like: attaching a static pressure on outlet and corresponding flow at the inlet. What I experienced was, that outflow differs from inflow sometimes more than 30%, - regardless mesh setup. The same setup using also a pressure for the inlet gives quite accurate inlet and outlet flows, - but is not a good way as I want to know the water level upstream due to a given flow.