Naslov
Novel methods of describing mechanical behaviour through machine learning

Autori
Zlatić, Martin ; Čanađija, Marko

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

ISBN
978-953-8246-18-0

Skup
4th edition of annual conference for doctoral students of engineering and technology „MY FIRST CONFERENCE“

Mjesto i datum
Rijeka, Hrvatska, 24.09.2020

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Nije recenziran

Ključne riječi
Machine learning, solid mechanics, nonlinear, materials

Sažetak
Traditional methods of modelling mechanical behaviour include matrices that contain material constants and provide a link between stress and strain. Such formulations are well established in solid mechanics and are used in real-world problems for centuries. They are often simplified but still provide a good model for engineers. However new engineering challenges have given rise to a need for new materials and these materials are often poorly described by conventional models or a multitude of models are used to describe the same behaviour for different materials. A possible way of describing new material properties is the use of machine learning that could use raw experiment data to describe the behaviour of materials avoiding the need for specific material models and speeding up the implementation in engineering calculations.

Izvorni jezik
Engleski

Znanstvena područja
Strojarstvo, Interdisciplinarne tehničke znanosti

POVEZANOST RADA