Optimizing wind farm design: Leveraging FLORIS for wake steering and flow control

by Polk, Joshua A.
Abstract:
This study explores FLORIS’s layout optimization method and a FLORIS integrated OpenMDAO optimization method used to optimize a wind farm’s turbine layout to maximize the annual energy production of the farm. An initial turbine layout was selected by using general placement rules as well as a boundary for the farm to be used as an initial input for the optimization driver to use as a baseline. The OpenMDAO method integrates FLORIS’s wake calculation tool to predict the AEP for each iteration throughout the optimization process. Upon comparing the improvements in AEP found by each optimizer, FLORIS’s optimization method resulted in a significantly larger increase in AEP. This highlighted the need for more exploration into the optimization methods OpenMDAO offers to find either a better suited method for this optimization problem or the need for increasing the number of turbines and the size of the farm being studied as the optimization method used may be better suited for larger layouts when compared to smaller ones. Further exploration could reveal methods better suited for different sizes of farms to help create optimizers more suited for specific needs of each individual farm.
Reference:
Joshua A. Polk, "Optimizing wind farm design: Leveraging FLORIS for wake steering and flow control", MECH 4391 Mechanical Engineering Project, The University of Memphis, May 2024.
Bibtex Entry:
@techreport{Polk2024REU,
    author = {Polk, Joshua A.},
    title = {Optimizing wind farm design: Leveraging FLORIS for wake steering and flow control},
    institution = {MECH 4391 Mechanical Engineering Project, The University of Memphis},
%    url = {},
    year = {2024},
    month = may,
%    day = {},
%    note = {},
    abstract = {This study explores FLORIS's layout optimization method and a FLORIS integrated OpenMDAO optimization method used to optimize a wind farm's turbine layout to maximize the annual energy production of the farm. An initial turbine layout was selected by using general placement rules as well as a boundary for the farm to be used as an initial input for the optimization driver to use as a baseline. The OpenMDAO method integrates FLORIS's wake calculation tool to predict the AEP for each iteration throughout the optimization process. Upon comparing the improvements in AEP found by each optimizer, FLORIS's optimization method resulted in a significantly larger increase in AEP. This highlighted the need for more exploration into the optimization methods OpenMDAO offers to find either a better suited method for this optimization problem or the need for increasing the number of turbines and the size of the farm being studied as the optimization method used may be better suited for larger layouts when compared to smaller ones. Further exploration could reveal methods better suited for different sizes of farms to help create optimizers more suited for specific needs of each individual farm.},
}