Paperlane

no. 17 · Wednesday 18 November 2026

Finite element nonlinear nudging for the Navier-Stokes equations

Elizabeth Carlson, Rui Fang and Ali Pakzad

math.NA

Abstract

Nonlinear nudging augments the classical linear nudging (continuous data as- similation) term with a nonlinear feedback that strengthens as the observed error decreases, and yields super-exponential convergence to the true state. In this paper, we study nonlinear nudging for the 3-dimensional incompressible Navier-Stokes equations with the finite element method, which allows complex geometries and physical boundary conditions. We propose an implicit-explicit (IMEX) finite element scheme with inf-sup stable elements, prove its unconditional long-time stability, and show for the spatially semi-discrete scheme that the error reaches a prescribed tolerance in finite time. Our analysis also suggests that the non- linear nudging term requires the mesh resolution to be fine enough (${χ= O(H^{-1})}$, where $χ$ is the nonlinear nudging intensity and ${H}$ is the observation mesh size). Further, our error estimates suggest that a larger nonlinear exponent $γ$ requires finer observational resolution, which is confirmed by numerical tests. Numerical experiments, including two-dimensional forced viscous flow in a periodic box, flow past a cylinder, and flow over a flat plate, show that nonlinear nudging synchronizes with the true solution considerably faster than linear nudging. We also show that observations in a subregion suffice to synchronize the whole flow

fields
Numerical Analysis
pdf ↗ page arXiv:2611.23036 ↗