Average Velocity calculation for Turbine Grid Points > 1

[prith-gs]

Hello,

By default, the default 'cubic-mean' is used to calculate the average velocity for a Turbine. Please correct me if I am wrong.

If I have to use the 'simple-mean', how can I do it?

I can see in floris/core/rotor_velocity.py there are other methods as well.

Best,
Pritheesh

[misi9170]

Hi @prith-gs ,

The "cubic-mean" method is hard-coded into FLORIS as it is the main approach we use. However, a small change in the code will allow you to use the "simple-mean" instead.

If you wish to use the "simple-mean" approach only for calculating the rotor-average wind speed as an output (but maintain the use of the "cubic-mean" for all internal calculations, such as thrust and power calculations), you can change this line in FlorisModel.turbine_average_velocities to instead read method="simple-mean". Calling fmodel.turbine_average_velocities should then return the simple mean over the rotor grid points.

If, instead, you would like to use the "simple-mean" also for all internal calculations, then the best approach will be to change the average_method attribute on TurbineGrid. You'll see there that average_method is hardcoded to "cubic-mean". You can change this line to average_method = "simple-mean". That should propagate throughout the calculations; however, I'm aware that there are at least a few other places where the "cubic-mean" is used implicitly, since it is the default for the average_method keyword argument in some calculations. An example of this is in FlorisModel._get_turbine_powers(), where average_method is not explicitly provided (you could remedy this by passing in average_method=self.core.grid.average_method, as is done in the FlorisModel.get_turbine_ais and FlorisModel.get_turbine_thrust_coefficients---if you do that, please consider submitting a bug fix pull request to incorporate this change into the FLORIS main branch!

[misi9170]

Actually, the first of those options doesn't necessarily require you to change anything on FlorisModel. Alternatively, you can simply import the average_velocity function and use it directly. Here is an example script:

from floris import FlorisModel
from floris.core.rotor_velocity import average_velocity

fmodel = FlorisModel(configuration="defaults") # Requires floris >= 4.3
fmodel.set(
    layout_x=[0, 700],
    layout_y=[0, 0],
    wind_directions=[180, 270],
    wind_speeds=[8, 8],
    turbulence_intensities=[0.06, 0.06]
)
fmodel.show_config() # Requires floris >= 4.3
fmodel.run()

rotor_average_vels = average_velocity(velocities=fmodel.core.flow_field.u, method="simple-mean")

print(rotor_average_vels)

Running the method="simple-mean" produces

[[7.96350938 7.96350938]
 [7.96350938 5.3087649 ]]

while running with method="cubic-mean" produces

 [[7.97368581 7.97368581]
 [7.97368581 5.34410334]]