From Navier–Stokes to Krylov: A Bridge Between CFD and Numerical Linear Algebra

Abstract

Computational Fluid Dynamics (CFD) has become an essential tool in chemical engineering, enabling detailed simulations of fluid flow, heat transfer, and chemical reactions. Central to CFD is the numerical solution of the Navier–Stokes equations, which often leads to large, sparse, and ill-conditioned linear systems. Efficient solution of these systems is crucial for practical simulations. Krylov subspace methods, a class of iterative solvers from numerical linear algebra, provide a robust framework for solving such systems with reduced computational cost. This work bridges CFD and numerical linear algebra by exploring how discretized Navier–Stokes equations naturally give rise to linear systems suitable for Krylov methods, highlighting solver selection, preconditioning strategies, and convergence considerations. A numerical example demonstrates the effectiveness of these methods in a benchmark incompressible flow problem, emphasizing the practical relevance of this bridge for chemical engineering applications.