DeepPhysics: a physics aware deep learning framework for real-time simulation

Alban Odot (1) , Ryadh Haferssas (1) , Stéphane Cotin (1)

1 MIMESIS - Computational Anatomy and Simulation for Medicine

Real-time simulation of elastic structures is essential in many applications, from computer-guided surgical interventions to interactive design in mechanical engineering. The Finite Element Method is often used as the numerical method of reference for solving the partial differential equations associated with these problems. Yet, deep learning methods have recently shown that they could represent an alternative strategy to solve physics-based problems 1,2,3. In this paper, we propose a solution to simulate hyper-elastic materials using a data-driven approach, where a neural network is trained to learn the non-linear relationship between boundary conditions and the resulting displacement field. We also introduce a method to guarantee the validity of the solution. In total, we present three contributions: an optimized data set generation algorithm based on modal analysis, a physics-informed loss function, and a Hybrid Newton-Raphson algorithm. The method is applied to two benchmarks: a cantilever beam and a propeller. The results show that our network architecture trained with a limited amount of data can predict the displacement field in less than a millisecond. The predictions on various geometries, topologies, mesh resolutions, and boundary conditions are accurate to a few micrometers for non-linear deformations of several centimeters of amplitude.

Mots clés: Neural Network; Deep Learning; Finite Element Method; Newton-Raphson; Real-Time; Physics Informed Neural Network

Domaines: Intelligence artificielle [cs.AI]

Alban Odot, Ryadh Haferssas, Stéphane Cotin. DeepPhysics: a physics aware deep learning framework for real-time simulation. 2021. ⟨hal-03327818v1⟩