Modelling vortex-induced loads using Recurrent Neural Networks⁎This work was supported by the Flemish fund for scientific research FWO under license number G0068.18N. Issue 20 (2021)

• Record Type: Journal Article
• Authors: Foster, Joshua A.; Troyer, Tim De; Runacres, Mark C.; Decuyper, Jan
• Is Part Of: IFAC-PapersOnLine. Volume 54:Issue 20(2021)
• Page Start: 32
• Page End: 37
• Publication Date: 2021
• Subjects: Machine learning -- Long Short-Term Memory Recurrent Neural Networks -- unsteady fluid dynamics -- vortex-induced vibrations; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2021.11.149
• Languages: English
• ISSNs: 2405-8963

Abstract

Abstract: In this work recurrent neural networks of the LSTM type are used to describe unsteady fluid loading. Indirect load measurements are especially useful for structural health monitoring applications. Relying on discrete spatiotemporal measurements of the velocity field, both the forces and the corresponding displacement of a cylinder subjected to vortex shedding are modelled.