Joar Axås

I am a PhD student at the Institute for Mechanical Systems at the Department of Mechanical and Process Engineering at ETH Zürich under the supervision of George Haller.

I obtained both my Bachelor's and Master's in Mechanical engineering at Chalmers University of Technology in Sweden. During 2019 and 2020, I worked at Fraunhofer Research Center for Industrial Mathematics, where I later wrote my Master's thesis supervised by Dr. Johannes Quist on combined vehicle dynamics and particle mechanics simulation.

Current Research

Data-driven modeling and prediction of nonlinear dynamical systems
The goal of this project is to develop model order reduction methods based on data generated by non-linearizable systems. Such systems exhibit phenomena that are impossible to capture by linearization, such as several steady states, limit cycles, and chaos. Our methods take as input measurements from experiments or simulations, automatically identifies an optimal nonlinear reduced space based on spectral submanifold theory, and returns a sparse, rigorous, low-dimensional model of the dynamics. We refer to these methods as dynamics-based machine learning. My research is focused on development of optimal embedding methods for invariant manifolds in observable spaces in order to analyze high-dimensional datasets from multimodal, internally resonant systems. I am also collaborating with experimentalists in applying dynamics-based machine learning to nonlinear structural and fluid dynamics problems.




See Google Scholar