Your Biggest Fan

Inspiration

As a team of two CS Students, one Physics and one ECE-major we tried to find a topic topic to combine the knowledge of all the fields. The sector of renewable energy was instantly appealing. When talking to PHD students we found out about the challenges that the placement of wind turbines post and how different parameters of the wind as well as the turbine itself effect the power that can be generated. The natural question to pop up was: Where should a specific wind turbine be placed?

What it does

We implemented a windmill-location assistant: it has three modalities which allow a windmill provider to choose the optimal locations among a set of possibilities. The first modality assumes that the windmill provider has statistical knowledge of the wind behavior in each of the possible locations, and then predicts the best one by performing a statistical analysis. The second modality only assumes that the provider has only measured the wind velocity pattern in each location, and then uses gradient descent along with maximum likelihood estimates to fit the best parameters to the wind distribution obtained. The third modality assumes no previous information on the locations, but will only require the latitude and longitude of each point. Machine learning will be used to estimate the parameters, using geolocational methods to estimate the distance between each possible location and the k nearest weather towers, and then using the corresponding wind parameters of each weather tower to predict a result.