Efficient parameter-robust preconditioners for linear poroelasticity and elasticity in the primal formulation
Weizhang Huang, Zhuoran Wang

TL;DR
This paper introduces new nonsingular preconditioners for linear poroelasticity and elasticity problems that are robust against parameter variations and effective in accelerating iterative solutions, especially in locking cases.
Contribution
The authors develop parameter-robust, nonsingular preconditioners that do not require Schur complement computation, improving the efficiency of solving large-scale saddle-point systems in poroelasticity.
Findings
Eigenvalues cluster around 1 with an outlier of order 1/λ.
Preconditioners are robust to mesh size, time step, and locking parameters.
Numerical results confirm effectiveness and robustness.
Abstract
Poroelasticity problems play an important role in various engineering, geophysical, and biological applications. Their full discretization results in a large-scale saddle-point system at each time step that is becoming singular for locking cases and needs effective preconditioners for its fast iterative solution. Instead of constructing spectrally equivalent ones, we develop nonsingular preconditioners so that the eigenvalues of the preconditioned system consist of a cluster around and an outlier in the order of , where is a Lam\'{e} constant that is large for locking cases. It is known that the convergence factor of GMRES is bounded by the radius of the cluster for this type of systems. Both two- and three-field block triangular Schur complement preconditioners are studied. Upper bounds of the radius of the eigenvalue cluster for those systems are obtained and…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsElasticity and Material Modeling · Numerical methods in engineering · Matrix Theory and Algorithms
