Speaker: Dixia Fan, Post-doc Associate, MIT Sea Grant

Active and Physics-Informed Learning of Vortex-Induced Vibration (VIV): A confession by a “lazy” fluid mechanist

50 years ago, J.C.R. Licklider, the pioneer of interactive computing hoped that “in not too many years, human brains and computing machines will be coupled together very tightly, and that the resulting partnership will think as no human brain has ever thought and process data in a way not approached by the information-handling machines we know”. Today, twenty years into the twenty-first century, we demonstrate such a vision via the newly-established intelligent towing tank (ITT) at the MIT “lab of the pink coach”, a robotic towing tank capable of learning complex fluid-structure dynamics by itself – the example problem we chose is the canonical vortex-induced vibrations (VIV).
 
In this talk, I will describe in detail the ITT’s active sequential experimentation process by deploying Gaussian process regression as the adaptive learning strategy. This enables for the first time a systematic exploration of the large input parametric space for rigid cylinder VIV problems, and allows us to assess the validity of the underlying strip theory assumption for flexible cylinder VIV modeling. In addition, I will provide a preliminary example of VIV parameter online estimation, and close the talk with a proposal for a potential future research paradigm shift for VIV modeling using a physics-informed (and -informative), data-driven probabilistic active learning approach.