Generalisation of the Total Linearisation Method to Three-dimensional Free-Surface Flows
Tyler Benkley 1, S. Deparis 2, P. Ricci 3, A. Mortensen 1 · 2025
Read Generalisation of the Total Linearisation Method to Three-dimensional Free-Surface Flows by Tyler Benkley 1, S. Deparis 2, P. Ricci 3, A. Mortensen 1 (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-2508-19700.txt.
Introduction
Tyler Benkley and co-authors stretch a lean sum method to full three-space flows in a 2025 paper. The flows have a free top skin, like waves on a tank, and may touch walls at a moving line. Past lean methods worked in flat two-space and at right touch angles only.
Their loop bends the weak form of the sums around a fixed stand-in shape. The bulk of the zone then drops out of the unknowns, near halving the solve size next to stock ways. A fresh prep step keeps the lean solve fast. Flow teams gain a cheaper route to true 3D skin flows, touch lines and all.
Catalog record
| CREATOR | Tyler Benkley 1, S. Deparis 2, P. Ricci 3, A. Mortensen 1 |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 132 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-19700
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Generalisation of the Total Linearisation Method to Three-dimensional Free-Surface Flows Tyler Benkley 1, S. Deparis 2, P. Ricci 3, A. Mortensen 1 Abstract An iterative Finite Element method predicated on a linearisation of the weak form around a reference configuration is derived for general, three-dimensional, free-surface flows, including systems with moving contact lines. The method is a rigorous generalisation of the Total Linearisation Method that was proposed by Kruyt et al. 1988 for two-dimensional flows with contact angles limited to 90 ∘ 90^{\circ} . In contrast to existing numerical methods for free-surface-flow problems, the present linearisation produces a weak form that is devoid of displacement degrees of freedom in the bulk, thus nearly halving the size of the linear system when compared to standard linearised methods. A novel preconditioner, whose implementation is made possible by the size reduction, is employed to solve the large resulting monolithic Jacobian systems with the Generalised Minimum Residual Method. The proposed method and the preconditioner are shown to be effective on two numerical examples of capillary flows, namely (i) the cylindrical die swell problem, solved both in 3D Cartesian coordinates and under the Ansatz of axisymmetry; and (ii) an enclosed 2D thermo-capillary problem. For the die swell, numerical results are validated by existing experimental results and prior simulations, and confirm both the extension to 3D and theoretical convergence rates. For the thermo-capillary problem, simulations verify earlier calculations. Additional simulations are also carried out for a new range of contact angles made possible by the extension. keywords Free-Surface Flows, Computational Fluid Dynamics, Numerical Methods, Newton Method † † affiliation: 1 Laboratory of Mechanical Metallurgy, EPFL, Switzerland 2 MATH, EPFL, Switzerland 3 Swiss Plasma Center, EPFL, Switzerland 1 Introduction Free boundary problems (FBPs) ( Crank 1987 ) are a challenging type of partial differential equation (PDE) where the domain boundary is an unknown that must also be determined as part of the solution. An additional boundary condition (BC) then arises on the free boundary as a result of it being unknown. FBPs have a vast range of applications including, for example, free-surface flows ( Katopodes 2019 ) in micro-fluidic devices ( Stone et al. 2004 ) or ink-jet printers ( Wijshoff 2018 ) , multi-phase fluid flows ( Brennen 2006 ) in secondary petroleum recovery ( Lake 1988 ) or aluminium production by electrolysis ( LaCamera et al. 1992 ; Gerbeau et al. 2004 ) , solidification processes ( Kurz et al. 2023 ) such as casting of metals ( Viswanathan et al. 2008 ) or Laser Powder Bed Fusion in additive manufacturing ( Chowdhury et al. 2022 ) , fluid-structure interaction problems ( Hou et al. 2012 ; Dowell & Hall 2000 ) , relevant for instance in the prevention of blood-vessel aneurysms ( Di Martino et al. 2001 ) or in assessing the structural integrity of long-span bridges ( Elgohary & Attia 2017 ) , or optimal stopping in mathematical finance ( Peskir & Shiryaev 2006 ) such as that of American options ( Rapuch 2005 ) . Typically, the free boundary’s position depends on the value of field variables whose value, in turn, depends on the position of the free boundary, thus, often requiring the need for iteration over the domain. In the last three decades, in the context of fluid dynamics, researchers have focused primarily on time-dependent FBPs. As it stands, the advent of Eulerian diffuse-interface methods has revolutionised this field of study. Techniques such as Volume-of-Fluid ( Hirt & Nichols 1981 ; Mohan & Tomar 2024 ; Popinet 2009 ) , the Level-Set method ( Osher & Sethian 1988 ; Sharma 2015 ) , the Phase-Field method ( Cahn & Hilliard 1958 ; Kim 2012 ) , coupled with the Continuum Surface Model ( Brackbill et al. 1992 ) , have endowed computer simulations with the ability to handle complicated interface dynamics riddled with breakup and coalescence, without the difficulties encountered in mesh-movement steps in Arbitrary Lagrangian-Eulerian (ALE) methods ( Scardovelli & Zaleski 1999 ) . Smoothed-Particle Hydrodynamics ( Lucy 1977 ; Monaghan 2005 ) , another diffuse-interface alternative to ALE, eludes the issues posed by topology change by virtue of its mesh-free nature. While diffuse-interface methods are undoubtedly better equipped to track complex moving interfaces than sharp-interface methods, the latter continue to be needed when precise tracking of a smooth interface is required. Rather than recovering the location of a free surface a posteriori , sharp-interface methods explicitly embed its location into boundary conditions (BCs) like the Young-Laplace law, the no-shear stress condition or the kinematic equation, obviating the need for reconstruction and the errors that can ensue therefrom. In sharp-interface methods, these BCs are all seamlessly integrated as natural BCs into one fully-coupled finite-element weak form. The sharp nature of the interface thus implies that surface-tension effects need not be smeared across the free surface as they are in diffuse methods, where capillary forces are usually approximated via the widely-used Continuum Surface Model ( Brackbill et al. 1992 ) . This can, for example, lead to spurious eddy currents when surface-tension effects are large, as explained by Scardovelli & Zaleski 1999 . Sharp-interface methods are also inherently well-suited to the solution of steady problems. Indeed, zeroing partial derivatives with respect to time in the Navier-Stokes equations and reframing the kinematic equation as an impermeability equation makes it possible to deploy the entire tool-kit of numerical techniques for nonlinear system solution. This opens the door to using, for instance, Newton or quasi-Newton methods and thus achieving quadratic convergence if ensuing linear systems can be inverted. By contrast, in order to incorporate the smoothed free-interface variable, diffuse-interface methods typically rely on special interface-advection equations, whose solution is usually decoupled from the Navier-Stokes equations by virtue of their transient nature. As a result, for the solution of steady problems, the equations in transient form are generally evolved from an initial state with the aim of converging, in time, to a steady solution. Convergence rates are therefore at the mercy of problem dynamics, which may entail a restrictive Courant-Friedrichs-Lewy condition and, in turn, extra computations to reach a solution. We devote here our attention to sharp-interface methods and address problems in need of high-resolution tracking of an unknown smooth interface free of topological change. Previously tractable only in simple configurations, sharp-interface methods for FBPs saw large progress in the latter part of the 20^{\text{th}} century with the advent of increased computing power. In the 1970s, first numerical solutions of FBPs on dam flow, stellar evolution, viscous laminar jet flow were reported ( Cryer 1976 ; Reddy & Tanner 1978 ) . The first technique to be used, known as the Trial Method, involves deliberately relaxing one BC on the free boundary. Then, the original FBP is treated as a fixed-domain PDE in the absence of the relaxed BC, and solved using standard numerical methods designed for PDEs over a fixed domain. Having thus solved the relaxed problem over a guessed domain, using the relaxed BC, one formulates a new guess of the domain so as to execute an ALE scheme. As in a Picard process, one thus iterates on the domain with the aim of converging to the free-boundary solution. The fixed-point nature of