In this talk I will present two machine learning approaches for wind farm planning. First I would present a multivariate copula based approach to form predictive distribution for long term site assessment.
Copula models are new to machine learning, and they are able to model multiple variables with arbitrary marginals and model non-linear correlation. This approach lends itself very well to the wind resource estimation where the goal is to estimate the wind resource at a test location given the resource at other neighboring locations. I will present results for multiple sites where we tested this approach.
http://groups.csail.mit.edu/EVO-DesignOpt/groupWebSite/uploads/Site/Veeramachaneni_copula.pdf
In the second part, I will present a new approach to layout optimization. In this, instead of directly searching for coordinates of turbines to minimize energy-cost, we attempt to learn the function that is capable of generating an optimized layout. Once learned, the function can be used to generate layouts in real time for any new site, thus enabling the layout optimization to be interactive.
http://groups.csail.mit.edu/EVO-DesignOpt/groupWebSite/uploads/Site/p745-wilson.pdf
BIO: Kalyan is a Research Scientist in the Computer Science and Artificial Intelligence Laboratory (CSAIL,MIT). His primary research interests are in machine learning and building large scale statistical models that enable discovery from large amounts of data. His research is at the intersection of Big data, machine learning and data science. He co-leads a group called Any Scale learning for all. The group is interested in Big data science and Machine learning, and is comprised of 20 members: postdoctoral fellows, graduate (MEng, S.M., and Ph.D), and undergraduate students.