A formulation of the linear discrete Coulomb friction problem via convex optimization

This paper presents a new formulation of the dynamical Coulomb friction problem in finite dimension with discretized time. The novelty of our approach is to capture and treat directly the friction model as a parametric quadratic optimization problem with second‐order cone constraints coupled with a fixed point equation. This intrinsic formulation allows a simple existence proof under reasonable assumptions, as well as a variety of solution algorithms. We study mechanical interpretations of these assumptions, showing in particular that they are actually necessary and sufficient for a basic example similar to the so‐called “paradox of Painlevé”. Finally, we present some implementations and experiments to illustrate the practical aspect of our work. This paper presents a new formulation of the dynamical Coulomb friction problem in finite dimension with discretized time. The novelty of this approach is to capture and treat directly the friction model as a parametric quadratic optimization problem with second‐order cone constraints coupled with a fixed point equation. This intrinsic formulation allows a simple existence proof under reasonable assumptions, as well as a variety of solution algorithms. The authors study mechanical interpretations of these assumptions, showing in particular that they are actually necessary and sufficient for a basic example similar to the so‐called “paradox of Painlevé”. Finally, the authors present some implementations and experiments to illustrate the practical aspect of their work.