Bartosz Protas

Jonathan Gustafsson, Bartosz Protas

In this study we revisit the problem of computing steady Navier-Stokes flows in two-dimensional unbounded domains. Precise quantitative characterization of such flows in the high-Reynolds number limit remains an open problem of theoretical fluid dynamics. Following a review of key mathematical properties of such solutions related to the slow decay of the velocity field at large distances from the obstacle, we develop and carefully validate a spectrally-accurate computational approach which ensures the correct behavior of the solution at infinity. In the proposed method the numerical solution is defined on the entire unbounded domain without the need to truncate this domain to a finite box with some artificial boundary conditions prescribed at its boundaries. Since our approach relies on the streamfunction-vorticity formulation, the main complication is the presence of a discontinuity in the streamfunction field at infinity which is related to the slow decay of this field. We demonstrate how this difficulty can be overcome by reformulating the problem using a suitable background "skeleton" field expressed in terms of the corresponding Oseen flow combined with spectral filtering. The method is thoroughly validated for Reynolds numbers spanning two orders of magnitude with the results comparing favourably against known theoretical predictions and the data available in the literature.

Arian Novruzi, Bartosz Protas

In this paper we introduce a new gradient method which attains quadratic convergence in a certain sense. Applicable to infinite-dimensional unconstrained minimization problems posed in a Hilbert space $H$, the approach consists in finding the energy gradient $g(λ)$ defined with respect to an optimal inner product selected from an infinite family of equivalent inner products $(\cdot,\cdot)_λ$ in the space $H$. The inner products are parameterized by a space-dependent weight function $λ$. At each iteration of the method, where an approximation to the minimizer is given by an element $u\in H$, an optimal weight $\hlambda$ is found as a solution of a nonlinear minimization problem in the space of weights $Λ$. It turns out that the projection of $κg(\hlambda)$, where $0<κ\ll 1$ is a fixed step size, onto a certain finite-dimensional subspace generated by the method is consistent with Newton's step $h$, in the sense that $P_u(κg(\hlambda))=P_u(h)$, where $P_u$ is an operator describing the projection onto the subspace. As demonstrated by rigorous analysis, this property ensures that thus constructed gradient method attains quadratic convergence for error components contained in these subspaces, in addition to the linear convergence typical of the standard gradient method. We propose a numerical implementation of this new approach and analyze its complexity. Computational results obtained based on a simple model problem confirm the theoretically established convergence properties, demonstrating that the proposed approach performs much better than the standard steepest-descent method based on Sobolev gradients. The presented results offer an explanation of a number of earlier empirical observations concerning the convergence of Sobolev-gradient methods.