A matrix-free, GPU-compatible PyTorch implementation of phase-field fracture with explicit dynamics, custom differentiable implicit damage solve, benchmarks on dynamic and quasi-static cases, and inverse recovery of fracture energy G_c via L-BFGS.
Mixed citations
Title resolution pending
Mixed citation behavior. Most common role is background (62%).
21
Pith papers citing it
6,368
external citations · Crossref
Background
62% of classified citations
citation-role summary
background 5
method 3
citation-polarity summary
representative citing papers
Derives three-term singularity asymptotics for surface-sourced Neumann Green's function and implements high-order BIE with Duffy patches to compute the regular part on arbitrary smooth surfaces, validated on spheres and spheroids.
The EXTRA method enables accurate 3D magneto-thermal FE simulations of large-scale NI HTS magnets by explicitly resolving critical turns while homogenizing others, reducing computation time by up to 13x on benchmarks.
SPEC CPU 2026 presents a new benchmark suite using open-source apps, expanded multithreading, and Rolling-Round-Robin Rate to address gaps in evaluating heterogeneous multiprogrammed CPU performance.
New coercive diffuse domain methods for Dirichlet conditions derived from mixed formulations and Nitsche's approach, with coercivity proofs and numerical tests showing improved accuracy on Navier-Stokes benchmarks.
A constrained LLM front-end for FEniCS multi-physics simulations dispatches to human-written templates and achieves 100% valid parses plus 90-100% geometry success on benchmarks while avoiding LLM-generated solver code.
PDE-Agents shows a LangGraph-orchestrated multi-agent LLM framework with GraphRAG that reaches 100% task success and perfect material fidelity on novel materials in ablation tests, with 97.8% success across 1369 production runs.
Presents and verifies a surface contact approximation derived from the thin shell approximation for efficient magneto-thermal FEA of no-insulation HTS coils, showing robustness advantages over thickness-dependent TSA in pancake coil simulations.
EquiNO with Q-DEIM creates reduced-order physics-informed surrogates for 3D hyperelastic RVEs that enforce equilibrium and periodicity by construction, achieve 10^3 speedups, and accurately interpolate and extrapolate stresses from few snapshots.
A continuous 50-nm Permalloy film on 400-nm-period nanopyramid templates forms a 2D magnonic crystal exhibiting a complete tunable in-plane band gap and strongly localized flat-band modes due to curvature-induced demagnetizing fields.
DeepONet learns the operator from signed distance functions of arbitrary 2D scatterer geometries to the resulting scattered fields for the Helmholtz equation, generalizing to unseen shapes as a surrogate for FEM.
EMSL groups material points into clusters, samples a reference strain per cluster once per increment, and computes a linearised stress estimate from the reference tangent and POD strain modes, yielding an affine reduced system that requires no iterations online and Pareto-dominates prior strain-cubc
Data-driven approximation methods are derived for the unitary Koopman-von Neumann operator, its eigenvalues and eigenfunctions, with explicit quantum-circuit representations for finite-dimensional projections.
mLaSDI uses multi-stage residual decoder training with periodic activations to recover high-frequency details in latent space dynamics identification, yielding lower reconstruction and prediction errors than standard LaSDI for PDEs.
Develops an evolving finite element method for parabolic PDEs with evolving interfaces, derives a suitable weak formulation, proves optimal error bounds for isoparametric elements of arbitrary order, and verifies convergence numerically.
Simulations using a Swelling & Random Migration algorithm and finite element homogenization show clustered fiber distributions increase transverse stiffness by up to 20% but reduce tensile strength compared to equilibrium distributions, with mean nearest neighbor distance linearly predicting both pr
A multi-agent LLM framework autonomously completes the full computational mechanics pipeline from a photograph to a code-compliant engineering report on a steel L-bracket example.
Shape optimization of Maxwell eigenvalues via adjoint sensitivities on a reference domain, solved with a damped inverse BFGS method and mixed finite elements.
jNO introduces a unified JAX tracing system for data-driven and physics-informed neural operator training that compiles domains, residuals, losses, and diagnostics into one pipeline.
Smoothing iterations on finite element solutions in an enriched space produce superconvergent approximations for symmetric positive definite problems.
The paper reviews recent developments and unresolved challenges in cardiac mechanics modeling, arguing that identifying essential complexities versus safe simplifications is key to clinical translation.
citing papers explorer
-
Physics-Informed Reduced-Order Operator Learning for Hyperelasticity in Continuum Micromechanics
EquiNO with Q-DEIM creates reduced-order physics-informed surrogates for 3D hyperelastic RVEs that enforce equilibrium and periodicity by construction, achieve 10^3 speedups, and accurately interpolate and extrapolate stresses from few snapshots.
-
Empirical Material Sampling and Linearisation -- A Simple and Efficient Strain-Space Model Order Reduction Approach for Computational Homogenisation in Large-Deformation Hyperelasticity
EMSL groups material points into clusters, samples a reference strain per cluster once per increment, and computes a linearised stress estimate from the reference tangent and POD strain modes, yielding an affine reduced system that requires no iterations online and Pareto-dominates prior strain-cubc
-
Numerical approximation of the Koopman-von Neumann equation: Operator learning and quantum computing
Data-driven approximation methods are derived for the unitary Koopman-von Neumann operator, its eigenvalues and eigenfunctions, with explicit quantum-circuit representations for finite-dimensional projections.