An Embedded Mesh Approach for Isogeometric Boundary Layers in Contact Mechanics

Eugenia Gabriela Loera Villeda, Ivo Steinbrecher, Alexander Popp · 2026

Read An Embedded Mesh Approach for Isogeometric Boundary Layers in Contact Mechanics by Eugenia Gabriela Loera Villeda, Ivo Steinbrecher, Alexander Popp (2026) 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-2603-01857.txt.

Introduction

Eugenia Gabriela Loera Villeda, Ivo Steinbrecher, and Alexander Popp mesh contact problems in two independent parts. Their 2026 work separates the contact surface mesh from the bulk volume mesh of each body. Contact zones and solid interiors need different element types and resolutions, so one shared mesh serves both poorly.

The contact surface uses a smooth boundary layer mesh built from the CAD model boundary curves. NURBS curves give exact shape and smooth joins between elements on that layer. The bulk instead uses a simple structured Cartesian grid, cheap to build and refine. A mortar-type constraint couples the two non-matching meshes so forces pass correctly across the seam. Several layer-building strategies are detailed, and numeric examples validate the full workflow for contact mechanics.

Catalog record

CREATOREugenia Gabriela Loera Villeda, Ivo Steinbrecher, Alexander Popp
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE97212
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.01857

Browse the full library · About the Tower of Babel

Full text

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

An Embedded Mesh Approach for Isogeometric Boundary Layers in Contact Mechanics Eugenia Gabriela Loera Villeda, Ivo Steinbrecher, Alexander Popp Abstract This paper proposes a novel discretization workflow for contact problems in which the discretization of the contact interface is decoupled from that of the bulk domain. This separation enables independently tailored meshes for the contact interface and the bulk volume, allowing local requirements—such as element type and mesh resolution—to be addressed efficiently. Exploiting the boundary representation of CAD models, the contact interface of each body is discretized using a NURBS-based boundary layer mesh. This provides a smooth geometric description of the contact surface and enhanced inter-element continuity. The bulk domain is discretized using a structured Cartesian grid. To couple the resulting non-matching discretizations, an embedded mesh approach based on a mortar-type constraint formulation is employed. The paper describes in detail the proposed discretization workflow for generating both the isogeometric boundary layer and the structured Cartesian grid, and presents several strategies for constructing NURBS-based boundary layer meshes. Finally, a set of numerical examples is provided to validate the proposed approach. Keywords: Embedded mesh method, Contact mechanics, Isogeometric Analysis, Mortar methods 1 Introduction Contact problems are of fundamental importance, as they arise in a wide range of physical and biological systems. They are characterized by complex interface phenomena, including friction, wear, lubrication, and thermomechanical effects. Over the past decades, advances in numerical contact formulations, together with increasing computational resources, have enabled significant technological progress. Notable examples include improvements in vehicle safety through crash simulations and the assessment of wear-induced damage in machine components [ 17 ] . In the medical field, sophisticated contact models have been developed for the planning of maxillofacial surgery [ 51 ] and, more recently, for the development of endovascular treatment strategies for cerebral aneurysms [ 23 ] , with an emphasis on incorporating patient-specific geometries. In the following, we focus on contact formulations based on the finite element method (FEM), for which substantial research efforts have been devoted to improving accuracy and robustness. A central line of research concerns the development of contact discretizations for non-matching meshes, including node-to-segment (NTS) [ 84 , 53 , 90 ] and segment-to-segment (STS) [ 89 , 22 , 61 ] techniques. A particular class of STS approaches is formed by mortar formulations [ 6 , 81 , 56 ] , which have gained increasing relevance in recent years, as they are especially well suited for finite deformation contact problems and finite sliding. Mortar methods enforce contact constraints in a weak sense and enable a variationally consistent treatment of non-penetration and frictional sliding conditions, thereby guaranteeing optimal convergence rates [ 15 ] . In [ 80 ] , it was shown that, for unilateral contact, the maximum expected convergence order under uniform mesh refinement is O(h^{3/2}) for shape functions of polynomial order p\geqslant 2 . This result makes the use of higher-order interpolation questionable for contact problems unless combined with adaptive mesh refinement techniques. For a comprehensive overview of computational contact mechanics based on FEM, the reader is referred to the monograph by [ 15 ] . Classical finite element discretizations based on Lagrange polynomials exhibit significant drawbacks, as their C^{0} interelement continuity leads to a discontinuous normal vector field along the contact interface. This is particularly critical because, regardless of the chosen contact discretization technique, normal projections between the contacting bodies are required throughout the contact evaluation. A commonly adopted strategy to alleviate this issue is the construction of the normal field by averaging the normals of the surface elements surrounding a node. Nevertheless, numerical instabilities may arise, especially in problems involving large sliding, due to the non-smooth variation of contact variables as the slave node slides over the master interface [ 14 , 74 ] . To mitigate these oscillations, geometric smoothing techniques based on Hermite, Bezier, spline, and NURBS basis functions have been proposed. The underlying idea is to employ higher-order interpolations on the contact interface to improve the accuracy of the contact tractions, while keeping the bulk discretization unchanged [ 52 , 83 , 3 , 76 , 47 ] . Although such approaches reduce these oscillations, they do not necessarily eliminate them entirely [ 74 ] . To address these drawbacks, contact formulations based on isogeometric analysis (IGA) have been developed, with many established FEM-based contact formulations adapted to the IGA framework. A comprehensive overview of computational contact formulations within IGA is provided in [ 14 ] . The central concept of IGA is the integration of computer-aided design (CAD) and engineering analysis by employing the same basis functions used in CAD for numerical simulations [ 34 , 11 ] . In [ 20 , 27 ] , it was shown that IGA can deliver increased robustness and accuracy per degree of freedom compared to FEM, as it relies on B-spline-based basis functions such as non-uniform rational B-splines (NURBS) or T-splines. These basis functions possess particularly attractive properties for contact problems, as already emphasized in the first publication on IGA [ 34 ] : they enable an accurate and smooth representation of the contact interface and guarantee an interelement continuity of at least C^{1} , thereby yielding a continuous normal vector field. Nevertheless, while the higher-order continuity of NURBS offers clear advantages for contact algorithms, it may not be required throughout the bulk of the domain, as it does not necessarily improve spatial convergence rates and may instead lead to increased computational cost. In comparison to finite element mesh generation, mesh generation in IGA is not as well established and can be particularly challenging, especially for complex geometries in three dimensions [ 38 , 79 ] . Although the isogeometric boundary representation is provided by a CAD model, it lacks a true volumetric description and is therefore essentially “hollow”. To address this limitation, a volume parametrization conformal to the given boundary representation must be constructed, which can be a demanding task [ 42 , 2 , 91 ] . Building on the previous observations, we propose a novel discretization approach for contact problems. The approach is illustrated in Figure 1 f or a two-dimensional contact configuration; however, it is applicable to both two- and three-dimensional contact problems. The procedure starts from the boundary representation (also referred to as B-rep) of the bodies involved in contact. For the two-dimensional setup depicted in Figure 1(a) , the B-rep is constructed using NURBS curves, whereas for three-dimensional configurations it is given by NURBS surfaces. In the proposed approach, the discretization of the contact interface is decoupled from that of the bulk domain. Starting from the isogeometric B-rep, a boundary layer mesh is generated by offsetting the NURBS B-rep inward into the body, exploiting the tensor-product structure of NURBS. The remaining bulk domain is subsequently discretized using a simple and easy-to-generate Cartesian mesh. The resulting discretization is shown in Figure 1(b) . In this configuration, a subset of the Cartesian mesh elements is intersected by the boundary layer mesh; these elements are referred to as cut elements . The simplified mesh generation strategy underlying the proposed approach follows the idea of immersed boundary methods, in which a B-rep is embedded into a simple background mesh, thereby eliminating the need for

Read the complete text