Skip to content

← ArchivePaper2018

Fast Corotated FEM using Operator Splitting

Tassilo Kugelstadt, Dan Koschier, Jan Bender

CGF41 citesCFX

This paper introduces an operator-splitting approach for corotated FEM simulation by separating the corotated linear deformation energy into a stretching term and a volume-preservation term.

Abstract

This paper introduces an operator-splitting approach for corotated FEM simulation by separating the corotated linear deformation energy into a stretching term and a volume-preservation term. Formulating backward Euler as an optimization shows the stretching term is rotation invariant, so it can be solved accurately with a precomputed Cholesky factorization while the volume term is handled with compliant constraints and Gauss-Seidel iterations. The result is a fast, stable solver for elastic solids and shells that supports large time steps.

How to read this

Category
Method: an operator-splitting solver for corotated FEM
Contributions
  • Splits the corotated linear deformation energy into a stretching term and a volume-preservation term
  • Shows the stretching term is rotation invariant so it is solved accurately with a precomputed Cholesky factorization, while volume is handled via compliant constraints and Gauss-Seidel
  • Yields a fast, stable solver for elastic solids and shells that supports large time steps
Context
Sits in the corotated linear FEM and constraint-based simulation lineage, relating to efficient character-skinning elasticity such as McAdams et al. 2011 (Efficient Elasticity for Character Skinning with Contact and Collisions).Builds on: Efficient Elasticity for Character Skinning with Contact and Collisions
Correctness
Relies on the corotated linear energy model and a precomputed factorization, so accuracy depends on that linearization holding and on the Gauss-Seidel volume iterations; demonstrated on elastic solids and shells, but it is an approximation, not a fully nonlinear hyperelastic solve.
Clarity
Moderately technical; a first pass conveys the splitting intuition, and a second pass is needed to follow the backward-Euler-as-optimization derivation and the rotation-invariance argument.
How to read it
Read for the energy-splitting idea and why the stretching term reuses one Cholesky factorization; do a second pass on the optimization formulation if you implement implicit integration.

Builds on

Built upon by

Nothing yet.

Related work

Keywords

This page summarises the entry and links to its original source. The archive never hosts or redistributes the publication itself.Show it in the full archive list →