{"title":"Stabilized explicit material point method for fluid flow and fluid-structure interaction simulations using dual high-order B-spline volume averaging","authors":"Zhang Cheng, Shiwei Zhao, Hao Chen, Jidong Zhao","doi":"10.1016/j.cma.2025.118428","DOIUrl":"10.1016/j.cma.2025.118428","url":null,"abstract":"<div><div>Traditional explicit Material Point Methods (MPM) for weakly compressible fluids often suffer from volumetric locking, cell-crossing instability, and excessive energy dissipation, particularly in fluid-structure interaction (FSI) scenarios. This study presents a stabilized explicit MPM framework that integrates dual high-order B-spline volume averaging to address these challenges. The proposed dual averaging technique simultaneously smooths deformation gradients and pressure fields using cubic B-spline basis functions to eliminate cell-crossing errors and reduce volumetric locking. A blended APIC/FLIP mapping scheme is developed to enhance energy conservation and stability at coarse grid resolutions. The framework is further enhanced by seamlessly integrating various complementary techniques such as <span><math><mi>δ</mi></math></span>-correction, pressure smoothing, and specialized boundary handling for more robust and effective modeling of free-surface and FSI problems. The framework is rigorously validated through benchmark cases, including 1D elastic wave propagation, Poiseuille flow, lid-driven cavity flow, water sloshing, dam break, and water impact on elastic obstacles. The simulation results demonstrate a remarkable reduction in pressure oscillations and improved particle distribution uniformity compared to prior MPM approaches. The proposed method establishes a robust and efficient tool for large-deformation FSI problems and bridges gaps in accuracy and stability for industrial-scale applications.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"448 ","pages":"Article 118428"},"PeriodicalIF":7.3,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145222637","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Linearized localized orthogonal decomposition for quasilinear nonmonotone elliptic PDE","authors":"Maher Khrais, Barbara Verfürth","doi":"10.1016/j.cma.2025.118426","DOIUrl":"10.1016/j.cma.2025.118426","url":null,"abstract":"<div><div>In this paper, we propose and analyze a multiscale method for a class of quasilinear elliptic problems of nonmonotone type with spatially multiscale coefficient. The numerical approach is inspired by the Localized Orthogonal Decomposition (LOD), so that we do not require structural assumptions such as periodicity or scale separation and only need minimal regularity assumptions on the coefficient. To construct the multiscale space, we solve linear fine-scale problems on small local subdomains, for which we consider two different linearization techniques. For both, we present a rigorous well-posedness analysis and convergence estimates in the <span><math><msup><mi>H</mi><mn>1</mn></msup></math></span>-semi norm. We compare and discuss theoretically and numerically the performance of our strategies for different linearization points. Both numerical experiments and theoretical analysis demonstrate the validity and applicability of the method.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"448 ","pages":"Article 118426"},"PeriodicalIF":7.3,"publicationDate":"2025-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145222634","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"The isogeometric MITC shell in geometric nonlinear analysis","authors":"Yongzhen Mi","doi":"10.1016/j.cma.2025.118425","DOIUrl":"10.1016/j.cma.2025.118425","url":null,"abstract":"<div><div>This paper extends the isogeometric MITC shell formulation proposed by Mi and Yu (2021) for linear analysis to address geometric nonlinear shell problems. Built on the Reissner-Mindlin shell theory, the original linear formulation employs the Mixed Interpolation of Tensorial Components (MITC) technique to alleviate shear and membrane locking. The present nonlinear extension retains the MITC framework while incorporating a mixed Non-Uniform Rational B-Spline (NURBS)-Lagrange interpolation strategy to address the additional complexities induced by geometric nonlinearity. The interpolatory nature of the Lagrange basis functions is leveraged to simplify the construction of director vectors and assumed strain fields. The nonlinear problem is formulated in a total Lagrangian setting and solved using Newton-Raphson iterations. The effectiveness of the proposed method is demonstrated through a comprehensive set of numerical examples, including both standard benchmarks and a collection of geometric nonlinear shell problems, which features challenging behaviors such as large rotations and the development of local creases. Through Bézier extraction, the method is further evaluated using T-spline and U-spline basis functions. The numerical results confirm that the MITC technique effectively suppresses shear and membrane locking, and the proposed shell formulation exhibits high accuracy and robust convergence, even for coarse and severely distorted meshes. However, it is also observed that the high inter-element continuity inherent in splines-based discretization can inhibit large deformations, introducing a new form of locking. This issue is successfully mitigated using Bézier extraction to reduce the inter-element continuity. Overall, the proposed formulation offers a <em>general</em> and <em>reliable</em> isogeometric framework for geometrically nonlinear analysis, applicable across a wide range of shell geometries, loading conditions, and boundary constraints.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"448 ","pages":"Article 118425"},"PeriodicalIF":7.3,"publicationDate":"2025-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145223343","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A new pre-stressing algorithm for patient-specific simulations of growth and remodeling using the homogenized constrained mixture theory","authors":"Ali Akbar Karkhaneh Yousefi, Stéphane Avril","doi":"10.1016/j.cma.2025.118424","DOIUrl":"10.1016/j.cma.2025.118424","url":null,"abstract":"<div><div>The homogenized constrained mixture theory has been implemented in different Finite-Element software packages for almost ten years to simulate growth and remodeling in soft biological tissues. In these models, it is essential to determine the pre-stresses of each constituent of the mixture in the reference configuration. However, no efficient numerical solution has been proposed so far to solve this problem. We propose to address this lack with a new algorithm based on the concept of anisotropic thermal contraction. In this algorithm, the pre-stretch tensor is incrementally updated by applying a series of anisotropic thermal contractions to each representative volume element of the model to restore its reference configuration. These contractions are proportional to the inverse of the local right stretch tensor, obtained through the polar decomposition of the deformation gradient. We implemented the algorithm in the Abaqus Finite-Element package through a UMAT subroutine and verified it on examples including growth and remodeling of a patient-specific aorta. We demonstrate that the model captures the residual stresses in good agreement with experimental results.</div><div>To highlight the potential clinical relevance of the proposed algorithm, we expanded our model predictions to investigate the influence of the aortic axial pre-stretch on morphological and microstructural evolutions of the aorta under hypertensive conditions. Our simulations show that loss of axial tension, induced by hypertension, increases aortic length and may lead to pathological deformations such as aortic tortuosity. These findings highlight the importance of efficiently determining pre-stresses for simulating long-term vascular growth and remodeling.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"448 ","pages":"Article 118424"},"PeriodicalIF":7.3,"publicationDate":"2025-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145222632","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Anna Corti , Giuseppe De Nisco , Jolanda J. Wentzel , Francesco Migliavacca , Umberto Morbiducci , Claudio Chiastra
{"title":"Personalized multiscale modeling of coronary plaque progression: the interaction between low-density-lipoprotein transport and cellular dynamics","authors":"Anna Corti , Giuseppe De Nisco , Jolanda J. Wentzel , Francesco Migliavacca , Umberto Morbiducci , Claudio Chiastra","doi":"10.1016/j.cma.2025.118427","DOIUrl":"10.1016/j.cma.2025.118427","url":null,"abstract":"<div><div>Multiscale agent-based modeling has shown promise in elucidating the mechanobiological mechanisms underlying atherosclerotic plaque formation and progression. However, the integration of advanced models of low-density lipoprotein (LDL) transport in the lumen and across the endothelium with agent-based models (ABMs) of plaque growth remains underexplored. Furthermore, patient-specific applications are lacking.</div><div>This study introduces a novel agent-based modeling framework for atherosclerosis, integrating hemodynamics and LDL transport in the lumen through computational fluid dynamics simulations, a three-pore model of trans-endothelial LDL migration, and an ABM of lipid and cellular dynamics. For the first time, the framework was applied to a patient-specific coronary artery and validated against 1-year follow-up data. Furthermore, it was used to explore potential plaque evolution over 5 years and under elevated LDL concentration.</div><div>The calibrated model predicted the 1-year variation in wall area in two patient-specific coronary cross-sections with an error of less than 10%. Simulated scenarios indicated that variations in blood LDL concentrations can result in distinct plaque morphologies, from localized to diffuse patterns.</div><div>This study provided an innovative, advanced multiscale model of atherosclerotic plaque formation and progression. As the first patient-specific application of a multiscale agent-based modeling framework for atherosclerosis with initial validation, this study underscored the potential of the approach for deciphering the mechanobiological pathways driving coronary plaque progression. The developed model provided valuable insights into how the interplay between LDL transport and hemodynamics influences arterial wall cellular dynamics in a patient-specific context.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"448 ","pages":"Article 118427"},"PeriodicalIF":7.3,"publicationDate":"2025-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145222633","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Weighted eigenseparation-based residual approach for model reduction of interface failure in heterogeneous materials","authors":"Jacob Fish, Junhe Cui","doi":"10.1016/j.cma.2025.118352","DOIUrl":"10.1016/j.cma.2025.118352","url":null,"abstract":"<div><div>Interface failure plays a critical role in the degradation of heterogeneous materials, often governing structural integrity across a range of applications from fiber-reinforced composites to polycrystalline rocks. This paper introduces a novel model reduction framework—the Weighted Eigenseparation-based Residual (WER) approach—for efficiently simulating interface failure using cohesive zone models. Two variants of the WER are developed: a <em>force-based</em> formulation that weakly enforces equilibrium equations at the interface at the modal level and a <em>separation-based</em> formulation that weakly enforces contact conditions at the modal level. These formulations are supported by precomputed influence functions within representative volume elements (RVEs), significantly reducing computational cost while preserving accuracy. The separation-based variant, in particular, demonstrates broad applicability across microstructures with interface junctions. Numerical examples in two and three dimensions—including fiber composites and geological microstructures—demonstrate the effectiveness, convergence behavior, and computational advantages of the WER over direct numerical simulations. The results show that high-fidelity predictions can be obtained even with coarsely discretized interface modes, confirming the robustness and versatility of the WER methodology.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"448 ","pages":"Article 118352"},"PeriodicalIF":7.3,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145223341","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Bing-Bing Xu , Lourenco Beirao da Veiga , Yongjie Jessica Zhang , Peter Wriggers
{"title":"Second order three-dimensional serendipity virtual elements for hyperelasticity: Static and dynamic analysis","authors":"Bing-Bing Xu , Lourenco Beirao da Veiga , Yongjie Jessica Zhang , Peter Wriggers","doi":"10.1016/j.cma.2025.118432","DOIUrl":"10.1016/j.cma.2025.118432","url":null,"abstract":"<div><div>In this work, a three-dimensional (3D) second-order serendipity virtual element method (S-VEM) is developed for the static and dynamic analysis of hyperelastic materials. The VEM framework is based on the projection of unknown basis functions onto polynomial spaces, allowing for flexible discretization with arbitrary polyhedral meshes. While most existing VEM formulations for 3D mechanical problems are discretized using first-order formulations, higher-order schemes offer improved precision, especially for nonlinear problems. However, conventional second-order VEM formulations introduce additional degrees of freedom (DOFs), such as body and surface moments, which complicate the implementation and reduce computation efficiency. To address this challenge, we propose a novel 3D second-order serendipity VEM that avoids any extra moment-related DOFs. This is the first application of a serendipity VEM to 3D static and dynamic problems in hyperelasticity. Furthermore, by integrating advanced mesh generation techniques, the proposed method enables hybrid simulations that combine second-order serendipity VEM and FEM to efficiently handle complex geometries.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"448 ","pages":"Article 118432"},"PeriodicalIF":7.3,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145223342","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Physics-guided hybrid network for predicting nonlinear dynamic response of structures under bi-directional ground motions","authors":"Zheyi Guo , Jun Xu","doi":"10.1016/j.cma.2025.118422","DOIUrl":"10.1016/j.cma.2025.118422","url":null,"abstract":"<div><div>Seismic structural response is a critical indicator for assessing collapse resistance and ensuring post-earthquake functionality. Therefore, accurate and rapid prediction of these responses is essential, with Deep Learning (DL) models offering a robust alternative to finite element methods due to their computational efficiency and capability to model nonlinear dynamic responses of structures. However, prevailing DL approaches predominantly focus on unidirectional seismic excitations, often neglecting the complex effects of bidirectional seismic excitations on real-world structures. Furthermore, while the development of the physics-informed neural network effectively enhances interpretability under limited data, most existing approaches incorporate physical constraints solely into the loss function, potentially leading to optimization conflicts and slow convergence rates. To bridge these gaps, this study introduces a novel physics-guided hybrid deep learning framework, the Physical Residual Long Short-term Memory network-Kolmogorov–Arnold network model (Phy-RLK), for real-time bidirectional seismic structural response prediction. The proposed DL architecture embeds the Newmark-<span><math><mi>β</mi></math></span> numerical integration scheme as a physical residual within LSTM cells. Simultaneously, the adaptive basis functions of the KAN enhance the feature extraction capability from bidirectional seismic inputs, enabling real-time correction of acceleration, velocity, and displacement predictions at each time step, thereby ensuring improved accuracy and physical consistency. The effectiveness of the proposed model is initially verified on a six-story Reinforced Concrete (RC) frame structure, determining optimal hyperparameters in the process. An ablation study further confirms the necessity of both the physics-residual LSTM cell and KAN layer. The model is subsequently implemented on a five-story RC frame structure to verify accuracy and applicability. The experimental results demonstrate that the Phy-RLK provides superior predictive accuracy and robustness, and significantly higher computational efficiency compared to finite element simulations.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"448 ","pages":"Article 118422"},"PeriodicalIF":7.3,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145160254","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A cohesive zone treatment for the material point method involving problems of large deformation and damage","authors":"Cameron M. Crook , Michael A. Homel","doi":"10.1016/j.cma.2025.118399","DOIUrl":"10.1016/j.cma.2025.118399","url":null,"abstract":"<div><div>A new algorithm is described that permits the use of cohesive zones in the material point method for problems involving large deformation and fracture. In contrast to previous cohesive zone implementations, this method does not utilize massless surface-element particles. Instead, cohesive tractions are computed using the shape function mappings from a reference grid configuration in combination with explicitly defined particle surface normals and surface positions. These normals and relative surface positions are updated each time step according to particle deformation. The tractions are converted to cohesive forces using the nodal areas and mapped back to particles using the same reference shape function mappings. These forces are then remapped by conventional particle-to-grid interpolation as external forces using the current-configuration shape-function mappings. This allows highly compliant cohesive zones to function over jump displacements larger than a grid cell. Upon damage, these interfaces can revert to conventional multi-field contact surfaces. This approach is general and readily applies to two and three dimensions as well as being compatible with damage-field gradient partitioning offering exceptional computational flexibility. The framework for this method enables other capabilities, such as improved contact precision using explicitly defined surface normals and positions, and a method to mitigate spurious material damage at weak discontinuities between stiff brittle materials and soft or compliant materials.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"448 ","pages":"Article 118399"},"PeriodicalIF":7.3,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145181257","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Victor Matray , David Néron , Frédéric Feyel , Faisal Amlani
{"title":"Geometry-agnostic model reduction with GNN-generated reduced POD bases and boosted PGD enrichment for (non)linear structural elastodynamics","authors":"Victor Matray , David Néron , Frédéric Feyel , Faisal Amlani","doi":"10.1016/j.cma.2025.118357","DOIUrl":"10.1016/j.cma.2025.118357","url":null,"abstract":"<div><div>This contribution proposes a new and significantly enhanced extension of a recently-introduced hybrid Graph Neural Network (GNN)-based reduced-order modeling approach for the numerical solution of time-dependent partial differential equations on non-parametric finite element meshes. Building upon previous proof-of-concept work, this more generalized framework presents a number of key novelties: tight integration of graph-based learning with physical information via direct imposition of finite element operators as node and edge level features; introduction of a Grassmannian subspace distance measure as a dedicated training objective; incorporation of a Gated Recurrent Unit (GRU) for a more efficient and lightweight architecture; hybridization with other Galerkin-based reduced-order methods such as the Proper Orthogonal Decomposition (POD); and a first treatment of nonlinear problems. A novel, on-the-fly enrichment mechanism, modified from a classical Proper General Decomposition (PGD) and dubbed ”Boosted PGD”, is additionally introduced to improve prediction accuracy at low computational cost via additional greedy corrective modes. The efficacy of the overall methodology is assessed on two challenging datasets featuring significant geometric and topological variations that include highly heterogeneous spatial discretizations. A variety of performance studies demonstrate very competitive accuracy and computational cost in simulating highly-dynamic behavior when compared to conventional full-order finite element models, including a remarkable capacity to generalize to configurations well outside of the topological scope of the original training and validation sets. Results imply that solvers constructed from such an approach may enable more scalable and robust mechanical simulations for complex, real-world engineering applications related to iterative design.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"448 ","pages":"Article 118357"},"PeriodicalIF":7.3,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145160255","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}