> Academic opportunities > PHD positions

Neural networks for hyperelastic model reduction in dynamics for virtual reality

Technological challenge: Numerical simulation (learn more)

Department: Département Intelligence Ambiante et Systèmes Interactifs (LIST)

Laboratory: Laboratoire de Simulation Interactive

Start Date: 01-09-2022

Location: Saclay

CEA Code: SL-DRT-22-0661

Contact: anders.thorin@cea.fr

The proposal consists in developing AI methods to train neural networks to learn the dynamics of hyperelastic solids. The aim is to reduce drastically computation time. The first step will consist in getting well versed with the governing equations of hyperelastic dynamic systems, and with state-of-the-art methods of nonlinear model reduction. Both classical methods such as PGD and "IA methods" will be covered by the literature review. The candidate will then select a few reference cases which will be used for benchmarking accuracy, computational cost, robustness and other properties. Simulation methods for the reference solutions will then be chosen. The heart of the thesis will tackle development of neural networks that satisfy the objectives in terms of scope, accuracy and evaluation time.

See all positions Download the offer (.zip)

Email Bookmark and Share