Problem with suspended sediment concentration (scour model)
 

Hej all

Another question arised during the last weeks of workng with FLOW-3D (v11).
I have a very simple geometry, a straight rectangular channel with a constant bed slope of 0.0127 (=1.27%, quite steep). The width is 4 m, the flow depth is h = 0.4 m, Froude Fr = 2, mean velocity u = 3.96 m/s, hydraulic roughness ks = 0.002 m and discharge Q = 6.34 m3/s. It is a steady-state situation with uniform flow conditions. Turbulence model: RNG

I "added" an suspended sediment concentration at the boundary condition Xmin: particle diameter 0.00045 m(= 0.45 mm), 1 kg/m3. As I am already working on the problem soon described, I already disabled some parameters of the sediment scour model which should not be needed for my case. Richardson-Zaki = 0, Bed Load Coefficient = 0. Drag and entrainment coefficient stayed = 1 (default), density of particles = 2650 kg/m3.

Investigating the flow condition described above, the sediments should stay in suspension and should in no way settle. Now, please have a look at the attached picture (it is the channel from the side, showing the first 40 m; shown is the suspended sediment concentration in a range from 0 to 1 kg/m3). The question is: Why are sediments settling (linearly)? They should be kept in suspension or picked up due to the high flow velocities and turbulence. I already changed parameters like diameter, entrainment and drag coefficient. It changed the results slightly, but they should not be calibration parameters, right?

I would appreciate your help!

The Richardson-Zaki Coefficient Multiplier should be 1, not 0. It multiplies the Richardson-Zaki coefficient, which is a power in the R-Z equation. That equation controls drag and therefore relative separation velocity between particles and liquid. I think you have made the settling velocity very high by setting the multiplier = 0. Try it with 1. Also, since you are using RNG turbulence, make the 'maximum turbulent mixing length' TLEN = 10% of the depth of the flow. The 'dynamic option' is not theoretically reliable for sediment/water mixtures. Do these two changes fix the result?