Speaker: Felipe A.C. Viana
Host: Ulisses Braga-Neto
Abstract: Building digital twins for prognosis and health management of industrial assets is a daunting task that requires models to be highly scalable and predictive at an asset level. Therefore, digital twins present a natural application space for machine learning. Dr. Viana will challenge the myth that building prognosis models with machine learning requires large datasets. First, he will address how physics-driven and data-driven kernels can be combined within deep neural networks. This framework pioneered in the Probabilistic Mechanics Lab allows for neural network to directly implement differential equations while accounting for uncertainty in the model form as well as observations. Dr. Viana will give an overview on the theoretical aspects and show engineering applications in prognosis models for main bearing of wind turbines, aircraft fuselage panels, and batteries used to power electric vehicles.
Biography: Dr. Felipe Viana is an Assistant Professor at UCF, where he leads the Probabilistic Mechanics Laboratory. His research focuses on fusing machine learning and probabilistic methods with physics-based models for optimization and uncertainty quantification. Before joining UCF, Dr. Viana was a Sr. Scientist at GE Renewable Energy, where he led the development of computational methods for improving wind turbine performance and reliability. Prior to that role at GE, he spent five years at GE Global Research, where he led and conducted research on design and optimization under uncertainty, probabilistic analysis of engineering systems, and services engineering. Dr. Viana holds a PhD in Aerospace Engineering from the University of Florida and PhD and MSc in Mechanical Engineering from Federal University of Uberlandia (Brazil).