surface contact approximation

This work addresses the challenge of mesh compatibility arising from the coupled discretization of interfaces and bulk domains in contact mechanics. The authors propose a decoupled discretization approach that employs a NURBS-based boundary layer mesh, constructed directly from CAD boundary representations, to accurately capture the contact interface, while the bulk domain is discretized using a structured Cartesian grid. Non-matching meshes are coupled across scales via mortar-type embedded constraints. This method represents the first formulation that decouples isogeometric boundary layers from structured volume meshes, enabling independent optimization of element type and resolution for both interface and bulk domains. Consequently, it preserves high-order smoothness along the contact surface while significantly enhancing geometric accuracy, computational efficiency, and modeling flexibility.

This study addresses the high computational cost of conventional finite element (FE) dynamic simulations in modeling energy dissipation in frictional joints, which is critically dependent on the true multiscale topography of contact interfaces. To overcome this challenge, the authors propose a novel multiscale framework that couples coarse-grained FE models with fine-mesh boundary element (BE) methods based on half-space theory. For the first time, this FE-BE approach is extended from quasi-static to dynamic contact analysis, enabling long-duration transient simulations with large time steps, without artificial numerical damping and with mesh independence. The method seamlessly integrates both time-domain integration and harmonic balance techniques. Validation on an S4 beam benchmark demonstrates excellent agreement with full FE explicit and implicit solutions, confirming its robustness, while also uncovering load-history-dependent physical response differences under partial slip conditions.

Addressing the challenge of simultaneously achieving geometric adaptability and conservation fidelity in multiphysics interfacial systems, this paper proposes a three-dimensional element-based finite volume method (EbFVM). EbFVM integrates the geometric flexibility of finite element methods—supporting high-order parametric shape functions and body-fitted curvilinear coordinate mappings—with the strict conservation properties of finite volume methods—enforced via flux reconstruction and conservative integral formulations—specifically tailored for heterogeneous interface modeling on unstructured grids. Unlike conventional spatiotemporal discretization schemes, EbFVM concurrently ensures high numerical accuracy and exact satisfaction of physical conservation laws in complex geometries. Validation across multiple lubrication-driven benchmark problems demonstrates its capability to accurately resolve strong field coupling, multiscale transport phenomena, and dynamic interface evolution. The method exhibits superior robustness, accuracy, and broad applicability across diverse interfacial multiphysics scenarios.

This work addresses the challenges in modeling interfacial adhesion and debonding under large-deformation chemo-mechanical-thermal coupled contact by proposing a unified multiphysics framework. Building upon Sauer’s contact theory, the formulation introduces a six-field finite element model that simultaneously accounts for displacements of two bodies, temperature, an interfacial adhesion field, and interfacial temperature. Adhesion evolution is governed by a quadratic contact potential, and strong coupling among all fields is achieved through a monolithic solution strategy. The model accommodates complex phenomena such as pressure- or gap-dependent adhesion, exothermic reactions, thermal hardening, thermal expansion, and concurrent adhesion–debonding processes. It is implemented with either classical or isogeometric shape functions, implicit time integration, and a fully linearized Newton–Raphson algorithm, allowing optional local condensation of adhesion variables at material points when needed. Numerical examples demonstrate the method’s robustness and accuracy in handling highly nonlinear multiphysics interfacial problems.

This work addresses the challenge of efficiently and accurately simulating the electromagnetic–thermal coupled behavior of inter-turn contact layers in no-insulation high-temperature superconducting (HTS) coils. To overcome the limitations of conventional thin-shell models—which require explicit specification of layer thickness, electrical resistivity, and thermal conductivity—the authors propose a surface contact approximation (SCA) method. This approach replaces the volumetric contact layer with an interface-based weak formulation governed by electrical contact resistance and thermal contact conductance. The SCA method is integrated into a magneto-thermal coupled finite element framework featuring the Pancake3D module and an open-source quench simulator. Transient simulations demonstrate that the method accurately captures both electromagnetic and thermal responses while substantially reducing computational complexity, thereby enhancing model usability and numerical robustness.

This study addresses the computational bottleneck in solving cohesive finite element contact systems by proposing the TOGEARI method. The approach constructs a Woodbury right preconditioner via factor space compression, integrating interaction space preconditioning with an optimal truncation strategy to effectively optimize eigenspace approximation. Experimental results on a system with 320,000 degrees of freedom demonstrate that the number of Arnoldi basis vectors is reduced by half, while warm-start solution time decreases from 2.85 to 1.63 seconds. These findings confirm that TOGEARI significantly enhances computational efficiency for large-scale contact problems, validating both the superiority and practical utility of the proposed preconditioning technique.

该文探讨了弱施加的Dirichlet边界条件如何通过允许受控滑移减少计算成本，提高流动分析精度，适用于复杂几何体直接分析。

本文提出一种非侵入式方法，在不匹配网格中施加强Dirichlet边界条件，适用于用户无法访问求解器源代码的情况。

This work addresses the challenge in laser metal processing where conventional diffuse-interface methods fail to accurately resolve the steep thermal gradients across the gas–liquid interface, leading to inaccurate predictions of vapor recoil pressure and surface tension. To overcome this limitation, the authors propose a hybrid interface modeling strategy that combines a sharp-interface CutFEM approach for high-fidelity heat conduction with a level-set-based diffuse-interface single-fluid finite element method for multiphase flow simulation. Coupling between the two solvers is achieved through a narrow-band temperature field extension. The resulting framework preserves robustness in capturing complex interfacial dynamics while significantly enhancing interfacial temperature accuracy, achieving second-order spatial convergence in the thermal model. Compared to purely diffuse-interface approaches, the method permits mesh sizes two orders of magnitude larger at equivalent accuracy, yielding an overall improvement of one order of magnitude in solution accuracy for representative test cases.

提出了一种混合迭代深度Ritz方法（H-IDRM）来解决椭圆型界面问题，利用水平集神经网络架构处理分段光滑解和复杂几何界面。