MASSIVE

Unlike fluids, for which a number of numerical models have been proposed, complex bulk materials such as fibrous media remain less studied from a macroscopic perspective: there is currently no widely accepted numerical model capable of predicting the dynamics of entangled materials.

However, industrial demand has grown significantly in recent years, encompassing fields such as cosmetology (precise prediction of hair’s mechanical behavior), virtual entertainment (simulation of hair and fur for special effects), and virtual prototyping in mechanical engineering (cable tangles, composite materials).

   

Coordinators

F. Bertails-Descoubes (Inria-LJK)

Fr. Saramito (LJK)

B. Raffin (Inria-LIG)

Results

The MASSIF exploratory project provided the initial decisive impetus to launch this long-term research project focused on the macroscopic modeling of fibrous materials. In particular, this project gave us the unique opportunity to hire Gilles Daviet as a doctoral student (funded by Persyval from 2013 to 2016) to continue our initial studies on the subject. Gilles Daviet’s thesis began with a comprehensive review of the state of the art regarding existing models for fiber assemblies; we then focused on a simplified yet still relevant challenge: designing a numerical model for macroscopic granular flows capable of capturing the non-smooth stick-slip transitions occurring between grains. The corresponding constitutive law, known as the Drucker-Prager law, has generally been smoothed out by previous approaches, thereby losing certain key emergent effects in the flow dynamics, such as stable stacking. Gilles Daviet’s thesis work demonstrated that it was not only possible to numerically preserve the non-smooth nature of the law, but also to efficiently solve the resulting one-step problem by adapting existing solvers for discrete frictional contact problems.

This multidisciplinary work has resulted in two journal publications [1,2], as well as several talks and posters at various conferences ranging from mechanics and physics to computer graphics. All of the corresponding source code has been released freely under the GNU GPL license. The thesis was awarded the 2017 Thesis Prize by the GdR “Geometric and Computer Graphics, Virtual Reality, and Visualization.”

Selected Publications

[1] G. Daviet, F. Bertails-Descoubes, “A Semi-Implicit Material Point Method for the Continuum Simulation of Granular Materials,” ACM SIGGRAPH 2016.

[2] G. Daviet, F. Bertails-Descoubes, “Nonsmosoth simulation of dense granular flows with pressure-dependent yield stress,” JNNFM 2016.