By Clark Saben
Faculty mentor: Dr. Desmond Villalba
2:00-2:50pm, HCC 329
Machine learning (ML) in physics has become a powerful tool for big data analysis in the last decade. ML offers an effective way to discover patterns from large input vectors when a suitable architecture is applied to a given data type.
This work investigates the architecture of dense neural networks (DNN) and their viability in wind power generation prediction from an individual wind turbine farm. Specifically, we developed a model that evaluates average ambient wind direction and average nacelle direction in order to predict active power production. The aim of this work was to accumulate model architectures for future applications in research.
