On October 4, we organized a scientific seminar on “Use of Neural Ordinary Differential Equations for Reduced Order Modeling: A Case Study“. Ján Boldocký, a PhD student at Slovak University of Technology in Bratislava, led the seminar, which was organized in the framework of the FrontSeat project as part of the seminar series on “Research Seminar on Smart Cybernetics.”
Abstract:
This talk presents the implications of using neural ordinary differential equations in the modeling of reduced space dynamics. These are showcased within a case study, where the dynamics of an impinging jet were modeled. The study also touches on a filtering task of the wall shear stress measurements to improve the prediction accuracy. In addition, the performance of neural ordinary differential equations is compared with some of the most commonly used model structures in reduced-order modeling. This work has resulted from a collaboration with Ali Mjalled and Martin Mönnigmann during the presenter’s research stay at Ruhr University Bochum.
This project has received funding from the European Union’s Horizon under grant no. 101079342 (Fostering Opportunities Towards Slovak Excellence in Advanced Control for Smart Industries).
0 Comments