BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:MechE Colloquium: Machine-learning based constitutive modeling
DTSTART:20240312T120000
DTEND:20240312T130000
DTSTAMP:20260502T060631Z
UID:1f3f7f5ddc02864a6e1650b790d4e3d7a766278d3108791c0398d168
CATEGORIES:Conferences - Seminars
DESCRIPTION:Prof Dirk Mohr\, ETHZ  Chair for Artificial Intelligence in M
 echanics and Manufacturing\; Department of Mechanical and Process Engineer
 ing\nAbstract: \n\nRecurrent neural network (RNN) models are emerging as 
 the most promising deep learning technique to describe the three-dimension
 al stress-strain behaviors of elasto-plastic materials under complex loadi
 ng scenarios. At the same time\, there is growing evidence that RNN formul
 ations for language processing (e.g. LSTMs\, GRUs) may lead to erroneous r
 esults when they are used in the context of constitutive modeling. We pres
 ent the development of mechanics-specific RNN formulations that satisfy im
 portant mathematical requirements such as self-consistency by construction
 . The resulting Minimal State Cells (MSCs) are compact networks that allow
  for the choosing the number of state variables independently from the deg
 ree of mathematical flexibility. This decoupling enables the models to det
 ect the optimal number of state variables for a given constitutive modelin
 g problem. We show that MSCs convincingly reproduce a variety of material
  behaviors and recognize important governing mechanisms. In a first applic
 ation\, we make use of RNNs to come up with a computationally efficient su
 rrogate model to represent crystal plasticity. In a second application\, w
 e also demonstrate their ability to capture rate-dependent material behavi
 or. To be able to train RNNs directly from experiments\, we discuss the po
 tential of transfer- and multi-task learning approaches. Furthermore\, we 
 pursue the development of automated mechanical testing systems to generate
  “big” experimental data for the training of data-driven plasticity mo
 dels. \n\n\nBiography: \n\nProfessor Dirk Mohr currently holds the Chai
 r of Artificial Intelligence in Mechanics and Manufacturing at ETH’s De
 partment of Mechanical and Process Engineering. He joined the faculty of E
 TH in 2015 after heading the Experimental Dynamics Group at the Solid Mech
 anics Laboratory at Ecole Polytechnique (France). He was educated in Struc
 tural and Computational Mechanics at the University of Karlsruhe (Germany)
 \, the Ecole Nationale des Ponts et Chaussées (France) and the Massachuse
 tts Institute of Technology (USA) where he received his PhD in Applied Mec
 hanics in 2003. He is Associate Editor of the International Journal of Sol
 ids and Structures (IJSS) and the International Journal of Impact Engineer
 ing (IJIE). He also serves on the editorial boards of the journals Strain\
 , Journal of Manufacturing and Materials Processing and the International 
 Journal of Plasticity. His research focusses on developing experimentally-
 ​validated computational models to enable the optimal design and manufac
 turing of lightweight materials and structures that are subject to extreme
  loading conditions in real-​life applications.\n\n\n\n 
LOCATION:MED 0 1418 https://plan.epfl.ch/?room==MED%200%201418 https://epf
 l.zoom.us/j/61626448592
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR