Construction of a discrete divergence-free basis through orthogonal factorization in $${\mathcal{H}}$$ -arithmetic

Abstract

In computational fluid dynamics, linear constraints on the fluid velocity lead to challenging indefinite linear systems of equations. In this paper, we propose to compute an approximation to the constrained linear space of divergence-free functions using hierarchical matrix techniques. This approach will yield a data-sparse, well-conditioned basis of the desired subspace in almost optimal computational complexity which is confirmed by numerical tests. The novelty of this paper lies in the application of hierarchical matrix techniques to orthogonal factorization as well as the construction of an explicit approximation to the subspace basis.