Description: Researchers from BYU and Stanford University are enhancing the design of wind farms through methods such as uncertainty quantification and new wake modeling.
Start: September 1, 2015
End: August 31, 2017
- Sponsor: National Science Foundation
- Principal Investigator: Andrew Ning
- Website: http://flow.byu.edu/
Many of the earth’s energy sources are depleting as we get further into the next generations. Hence, studies regarding energy are continually at the forefront of research. Renewable energy sources are particularly a valued development in the scientific world due to the fact that they are naturally replenished by the environment. Some common examples include energy harnessed from the sun, water, and wind. The use of these renewable resources is not only commonly less expensive, but also provides an alternative to environmentally harmful energy sources such as fossil fuels. Professors Andrew Ning of BYU and Juan Alonso of Stanford University are currently working in collaboration to optimize the energy harnessed from wind power. Their work focuses on the maximization of wind energy produced through the integrated design of wind farms.
Wind farms are made up of many individual wind turbines, each which produce power when its blades are turned by the wind. Through this process, the kinetic energy of the blades rotating is turned into electricity. As the blades spin, the turbine effectively creates a wake behind it where the wind flows, much like how a boat propeller creates ridges and patterns in the water behind it used in wakeboarding. The wake of this one wind turbine often alters the wind patterns for the other wind turbines in the farm. This causes the energy of other wind turbines to be less than optimal than if they were in an isolated condition. Consequently, some wind farms are producing energy only at an 80% efficiency rate than as they were originally designed to do.
Currently, the techniques and tools used to project optimization of wind farm designs are generally simplistic and usually include up to 50 variables. The researchers plan to take on a more complex approach using exact derivatives and gradients to analyze the wakes of wind turbines, and incorporating hundreds of additional variables into the equation. Trials of these efforts will be validated on three separate wind farms. The team will use a method called uncertainty quantification (UQ) to develop an optimized layout design. Since there are many variables that cannot be determined exactly, UQ determines the likelihood of uncertain variables based on statistical analysis and other known pieces of the design. For instance, if a student here at BYU was trying to predict when he would meet and marry his wife, he could consider known values such as the number of girls in his ward and classes, how many of those girls were still single, or how many dates he planned to go on a week. However, no amount of prediction could exactly describe factors such as how many girls would say yes, how many would be taken by other men during the middle of a semester, or how many would display mutual feelings. He would not know all of the parameters exactly and could only look at the statistics of certain possibilities occurring. The researchers are currently developing methods of UQ that can be largely scaled to determine the best methods for designing wind farms to their fullest capacity, despite many unknown variables.
Professors Ning and Alonso plan to publish their wake models on GitHub, a site that warehouses projects and new innovations in software. Through this publication, other researchers and designers of wind farms can access and utilize the models designed in this project. Additionally, the researchers’ development of UQ and optimization methodologies can be applied to many more areas besides wind farms. The techniques they develop will have an impact on multiple industries that wish to integrate designs of their systems with greater optimization. Furthermore, the implications of these advances will not only affect the scientific and industrial communities, but also influence the sphere of education. Throughout the project, undergraduate and graduate students from the engineering and mathematics fields will be directly involved in the development and design of these tools.