Automated Inference of Vortex Core Physics of Hovering Rotor Wakes Using Machine Learning Techniques

Abstract

Machine learning can significantly enhance the insights gained from large aerodynamic computational data sets. Specifically, 3D rotor wake computations produce detailed vortex structures, but typically, only a small fraction of the data is leveraged. A critical limiting factor is the time needed to process these large 3D data sets and infer the correct physics. The development of machine learning methods to process these data sets reduces the data processing burden on the analyst and also increases the volume of data that can be processed within a reasonable amount of time. Recent efforts in this area have resulted in the development of a novel machine learning-based vortex-core data extraction methodology. The present study demonstrates the application of this method to hovering rotor wakes to compare different blade tip geometries, ground effect, wake grid spacing, and collective changes and discusses the extracted vortex physics for each case.