numerical oscillation when using H-curve to represent discharge through turbine
 

Hi All,

I am using a Head/Discharge curve (as shown) to define the discharge allowed in my model through a turbine.

http://s24.postimg.org/k1mrltxw5/photo_1.jpg

When the head difference either side of the turbine reaches the point on the curve that the discharge begins to fall, numerical oscillations in the model in velocities, water levels, and discharge occur.

If I edit the Head/Discharge curve, such that the discharge does not fall when it reaches high head, and instead is kept constant once it reaches a maximum, the oscillations disappear.

I am interested to hear peoples' thoughts regarding why this may be occurring. I would like to use a realistic Head/Discharge curve for my model.

I suppose it is not a numerical diffusion/dispersion issue as otherwise it would likely still occur when the Head/Discharge curve is levelled off? Reducing the time step does not eliminate the problem either.

Thanks for your thoughts