A flu i d laye r finit e -eleme n t softwar e

Dennis Wörthmüller, Gaetano Ferraro, Pierre Sens, Michele Castellana · 2025

Read A flu i d laye r finit e -eleme n t softwar e by Dennis Wörthmüller, Gaetano Ferraro, Pierre Sens, Michele Castellana (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-17827.txt.

Introduction

Dennis Wörthmüller and three co-authors present IRENE, a software library for simulating a thin layer of viscous fluid spread over a surface in three-dimensional space. Such layers appear everywhere: lipid membranes on the microscopic scale, coatings and films at human scale, and air flows at planetary scale. General finite-element packages struggle with open two-dimensional surfaces, so IRENE is built specifically for them, on top of the FEniCS platform.

The library accepts a wide range of edge conditions and obstacles of any shape sitting between surfaces. It handles both creeping flows and inertia-driven ones, and couples flow within the layer to bending of the layer itself, surface tension, and elastic response. The authors validated it against known solutions and showed physical examples. They offer IRENE as a shared tool for researchers whose problems live on surfaces rather than in volumes.

Catalog record

CREATORDennis Wörthmüller, Gaetano Ferraro, Pierre Sens, Michele Castellana
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE82 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-17827

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A flu i d laye r finit e -eleme n t softwar e Dennis Wörthmüller, Gaetano Ferraro, Pierre Sens, Michele Castellana Abstract We present a finite-element software library, , for numerically solving the steady state and dynamics of a two-dimensional viscous fluid layer embedded in three-dimensional space, on multiple physical scales. The library is designed to handle a wide range of physical regimes—both low-Reynolds-number and turbulent ones—and geometries, capturing the coupling between in-plane flows, out-of-plane deformations, surface tension, and elastic response. In addition, the software can treat complex geometries, including those with intra-layer obstacles or arbitrary layer boundaries. We validate against known analytical and numerical results and demonstrate its capabilities through examples relevant to biological membranes and macroscopic air flows. 1 Introduction Thin fluid layers are central to a wide range of physical systems, from biological membranes [ 1 ] and cell cortices [ 2 ] to soap films [ 3 ] and industrial coatings. These two-dimensional interfaces often exhibit complex dynamics, involving in-plane viscous flow, out-of-plane deformations, and interactions with surrounding media. Modeling such systems presents substantial challenges due to their geometric nonlinearity, the coupling between flows and shape and, in some regimes, the presence of turbulence or flow instabilities. In biological contexts, deformable, curved fluid interfaces such as the cell cortex or lipid bilayers play a key role in morphogenetic processes, cell motility, and signaling. Membrane-bound proteins can locally alter mechanical or transport properties and interact with curvature to generate feedback mechanisms [ 1 , 4 ] . On macroscopic scales, fluid sheets can exhibit instabilities, folding, or turbulent behavior driven by external forcing or intrinsic flow dynamics [ 5 , 6 , 7 ] . On mesoscopic, planetary scales, layers of clouds at low altitude in the atmosphere may exhibit complex, turbulent behavior on spatial extensions so large that the curvature of the layer surface—induced by the the Earth’s curvature—is significant [ 8 ] . A unified numerical framework that captures both in- and out-of-plane deformations across regimes and scales is essential for understanding such phenomena. Several numerical approaches have been proposed to simulate surface hydrodynamics, including boundary-integral methods [ 9 ] , immersed-boundary methods [ 10 , 11 ] , spectral methods [ 12 ] , Monte Carlo methods [ 13 ] , and surface finite elements [ 14 ] . While each of these methods is well suited for specific physical conditions and geometries, there remains a need for a versatile, extensible framework capable of describing surface flows on evolving geometries, with robust control over boundary conditions and material properties. We present a fluId layeR finitE-Element softwarE ( ), which allows to numerically solve for the steady state and dynamics of a two-dimensional viscous fluid layer embedded in three-dimensional space, in a large variety of physical regimes and geometries. Our method captures the coupling between in-plane flows, out-of-plane deformations, surface tension, and elastic response of the fluid layer. ’s modular design allows for a wide range of boundary conditions and can treat a large variety of geometries, including fluid layers with intra-layer obstacles, or layer boundaries with arbitrary shapes. The method is designed to operate across different physical regimes—from low-Reynolds number flows relevant for cellular systems [ 15 ] , to inertia-driven regimes characteristic of macroscopic or mesoscopic air films [ 8 ] . We validate the method against known analytical and numerical results and demonstrate its capabilities through representative examples. These include the steady-state flow of a lipidic membrane with a trans-membrane protein [ 16 ] in a ring geometry, and the dynamics of Poiseuille flow of air [ 17 ] on a curved channel. The modular design of our implementation makes it a promising tool for future extensions, such as the inclusion of active stresses and interactions with surrounding fluids on multiple physical scales [ 18 , 19 ] . The manuscript is organized as follows. In Section 2 we present the partial differential equations which describe the dynamics of the fluid layer. In Section 3 we present ’s solution of these partial differential equations in a few representative examples for the steady state for cell membranes, Section 3.1 , and for the dynamics of macroscopic air flow, Section 3.2 . Finally, Section 4 is devoted to the discussion and interpretation of the results, and future directions. 2 Method Let us consider a two-dimensional fluid layer embedded in three-dimensional space, see Fig. 1 . The elements of the fluid flow tangentially to the layer and the layer itself can deform, exhibiting a velocity normal to the layer surface: we will denote the tangential and normal velocities above by v and w respectively. Geometrically speaking, the surface of the layer constitutes a differential manifold, which we will denote by \mathscr{M} . The velocity v^{i} is a vector field in the tangent bundle of \mathscr{M} , while w is a scalar field on \mathscr{M} [ 20 ] . The surface tension of the layer is denoted by \sigma . Choosing—for example—the Monge parametrization [ 21 ] to parametrize \mathscr{M} , the layer height is given by a function z({\bm{x}}) , where {\bm{x}} are the coordinates on a domain \Omega\in{\mathbb{R}}^{2} , see Fig. 1 . In what follows, we will assume that the elastic response of the layer is described by the Helfrich free energy [ 22 ] . However, the modular structure of allows to implement other elastic models, see Section 4 . The equations which describe the dynamics of the velocities, tension and shape of the layer are [ 23 , 24 , 25 ] \displaystyle\nabla_{i}v^{i}-2Hw \displaystyle=0, (1) \displaystyle\rho(\partial_{t}v^{i}+v^{j}\nabla_{j}v^{i}-2v^{j}wb^{i}_{j}-w\nabla^{i}w) \displaystyle= \displaystyle\nabla^{i}\sigma+\eta\left[-\nabla_{\text{{LB}}}v^{i}-2\left(b^{ij}-2\,H\,g^{ij}\right)\nabla_{j}w+2Kv^{i}\right], (2) \displaystyle\rho\left[\partial_{t}w+v^{i}\left(v^{j}b_{ji}+\nabla_{i}w\right)\right] \displaystyle= \displaystyle 2\kappa\left[\nabla_{\text{{LB}}}H-2H(H^{2}-K)\right]+2\sigma H+2\eta\left[(\nabla^{i}v^{j})b_{ij}-2w(2H^{2}-K)\right] , (3) \displaystyle\partial_{t}z \displaystyle=w\,(\hat{N}^{3}-\hat{N}^{i}\partial_{i}z), (4) are a set of partial differential equations [ 26 ] , defined on a set \Omega with boundary {\partial\Omega} , for the two components v^{i} of the tangent velocity v , the normal velocity w , the surface tension \sigma , and the fluid shape profile z , where each unknown depends on both the coordinates {\bm{x}} on \Omega and time t . In , \nabla is the covariant derivative, g the metric tensor, b the second fundamental form, H and K the mean and gaussian curvatures, respectively, and \nabla_{\text{{LB}}} the Laplace-Beltrami operator [ 27 , 23 ] . The vector \hat{N} is the unit normal vector to \Omega , which lives in the Euclidean, three-dimensional space in which \Omega is embedded, while n is the normal to a boundary of \mathscr{M} , and it belongs to the tangent bundle of \mathscr{M} [ 27 ] , see Fig. 1 . It is important to point out that, despite the fact that in ’s examples we will make use of the Monge coordinate system [ 21 ] , are covariant with respect to a general diffeomorphism x^{i}\rightarrow x^{\prime i} [ 20 ] —which ensures that the physical behavior of the system is independent of the coordinate choice . In what follows, we will consider two geometries for \Omega : A square geometry, shown in Fig. 1 , and a ring geometry. For the ring geometry, we will consider radially symmetric boundary-value problems: Such problems allow for a numerically exact solution, which we will leverage to test ’s solutions. Figure 1: Differential manifold of

Read the complete text