Bug report: cubature interpolation does not work with CC model #709
Description
Cubature interpolation throws error for CC wake velocity deficit model
How to reproduce
Here is a minimal code that reproduces the error:
# Instantiate FLORIS using either the GCH or CC model
from floris.tools import FlorisInterface
fi = FlorisInterface("inputs/cc.yaml")
# Use turbine cubature grid for rotor interpolation
fi_dict = fi.floris.as_dict()
fi_dict["solver"] = {
"type": "turbine_cubature_grid",
"turbine_grid_points": 3
}
fi = FlorisInterface(fi_dict)
# Calculate wakes and retrieve turbine powers
fi.calculate_wake()
turbine_powers = fi.get_turbine_powers()
print(turbine_powers.flatten())
Relevant output
The example code above results in the following error first:
Exception has occurred: AttributeError
'TurbineCubatureGrid' object has no attribute 'x'
File "/mnt/c/Users/Bart/python_scripts/floris/floris/simulation/grid.py", line 381, in set_grid
self.x = np.take_along_axis(self.x_sorted, self.unsorted_indices, axis=2)
If we then remove those lines from the code (is that correct?), the example script runs fine when you replace cc.yaml with gch.yaml. With cc.yaml, we have a second error come up:
Exception has occurred: ValueError
non-broadcastable output operand with shape (1,1,1,1,1) doesn't match the broadcast shape (1,1,1,1,1,9,3)
File "/mnt/c/Users/Bart/python_scripts/floris/floris/simulation/solver.py", line 564, in cc_solver
effective_yaw_i += added_yaw
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
What I also found is that the velocities variable being passed into the Ct(...) function in turbine.py seems to not be the right dimensions when using the CC model.
- With CC, the shape of
velocitiesis(1, 1, 3), - With GCH, the shape of
velocitiesis(1, 1, 1, 9, 1).
Floris version
Using the current develop branch with latest commit 8ece0f.
System Information
- OS: Ubuntu 22.04
- Python version: Python 3.11
- Library versions:
numpy==1.24.3numexpr==2.8.4scipy==1.10.1