On the structure preserving high-order approximation of quasistatic poroelasticity

TL;DR

This paper develops high-order, structure-preserving numerical schemes for quasistatic poroelasticity, ensuring stability, energy dissipation, and optimal convergence, with detailed analysis for Taylor-Hood elements and Runge-Kutta methods.

Contribution

It introduces a unified error analysis and a high-order time discretization scheme that preserves the problem's structure and energy properties.

Findings

Achieves high-order convergence rates in space and time.

Ensures stability and energy dissipation in numerical approximations.

Extends applicability to multi-field formulations and various time discretizations.

Abstract

We consider the systematic numerical approximation of Biot's quasistatic model for the consolidation of a poroelastic medium. Various discretization schemes have been analysed for this problem and inf-sup stable finite elements have been found suitable to avoid spurios pressure oscillations in the initial phase of the evolution. In this paper, we first clarify the role of the inf-sup condition for the well-posedness of the continuous problem and discuss the choice of appropriate initial conditions. We then develop an abstract error analysis that allows us to analyse some approximation schemes discussed in the literature in a unified manner. In addition, we propose and analyse the high-order time discretization by a scheme that can be interpreted as a variant of continuous-Galerkin or particular Runge-Kutta methods applied to a modified system. The scheme is designed to preserve both, the underlying differential-algebraic structure and energy-dissipation property of the problem. In summary, we obtain high-order Galerkin approximations with respect to space and time and derive order-optimal convergence rates. The numerical analysis is carried out in detail for the discretization of the two-field formulation by Taylor-Hood elements and a variant of a Runge-Kutta time discretization. Our arguments can however be extended to three- and four field formulations and other time discretization strategies.