09 - How to solve optimization problem in Pyleecan not useable
Opened this issue · 4 comments
i have many questions for pyleecan and i can go to no one whether documentation or youtube, there is no good explanantion of the Methods Classes etc. and if Tutorials exist there are generally false Methods for Example in the Opimization, we vary geometrical variables without changing the winding so fill factors up to 180% are easily possible.I want set up this right to have something useable. And I do not understand how Objectives like efficiency etc. can be put in the simulation. And then there is the point of how i can add new geometries for me and maybe share it later with the community. There is no useful Documentation. Maybe anyone can help me:(
Hello,
If I understand correctly you want to run an optimization simulation that would change the geometry of a slot while keeping a constant fill factor and add an objective on efficiency. Regarding the first one, there is no function in pyleecan to do that, you will need to write yourself the setter function of the DesignVariable (a function that takes the simulation and a value in argument, the value would be the desired slot dimensions and the function would edit several parameters to adapat the slot and the winding). For the efficiency, it's still a work in progress in pyleecan there is a notebook tuto_Efficiency_map.ipynb but we will need to check if this is still relevant.
Best regards,
Pierre
Hello Bonnel,
Thank you for your support you are great.
The First issue could it look like this:
def set_slot_geometry(simulation, slot_dimensions):
"""
Setter function to adjust the slot dimensions and winding parameters
while keeping the fill factor constant.
Parameters:
- simulation: The simulation object from Pyleecan
- slot_dimensions: The new slot dimensions to be set
"""
# Extract current winding parameters
winding = simulation.machine.stator.winding
# Update slot dimensions
simulation.machine.stator.slot.H0 = slot_dimensions['H0']
simulation.machine.stator.slot.H1 = slot_dimensions['H1']
simulation.machine.stator.slot.H2 = slot_dimensions['H2']
simulation.machine.stator.slot.W0 = slot_dimensions['W0']
simulation.machine.stator.slot.W1 = slot_dimensions['W1']
simulation.machine.stator.slot.W2 = slot_dimensions['W2']
# Compute new slot area
slot_area = compute_slot_area(simulation.machine.stator.slot)
# Update winding parameters to keep fill factor constant
fill_factor = winding.fill_factor
conductor_area = fill_factor * slot_area / winding.Npcpp
winding.conductor.area = conductor_area
# Ensure other necessary parameters are adjusted accordingly
# (e.g., number of turns, layers, etc.)
def compute_slot_area(slot):
"""
Compute the area of the slot based on its dimensions.
Parameters:
- slot: Slot object with dimensions H0, H1, H2, W0, W1, W2
Returns:
- slot_area: The computed slot area
"""
# Implement the computation of the slot area based on its dimensions
# This is a simplified example; adjust as necessary for your specific slot type
slot_area = slot.H2 * (slot.W0 + slot.W1) / 2
return slot_area
Or am I on the false route🫡
Thanks again
Best regards,
Hello,
You can take inspiration from https://github.com/Eomys/pyleecan/blob/master/Tests/Validation/Multisimulation/test_slot_scale.py
Your setter is linked to a Design Variable, so I would suggest setting a float in an interval for the value rather than a dictionary
The slot area is already available with simulation.machine.stator.slot.comp_surface() and even comp_active_surface(). You can also use simulation.machine.stator. comp_fill_factor()
To make sure that the fill factor doesn’t change you can either play on the conductor dimensions or introduce a clever scaling rule on your slot that keeps the active surface constant.
Best regards,
Pierre