{"title":"Stochastic deep material networks as efficient surrogates for stochastic homogenisation of non-linear heterogeneous materials","authors":"Ling Wu, Ludovic Noels","doi":"10.1016/j.cma.2025.117994","DOIUrl":"10.1016/j.cma.2025.117994","url":null,"abstract":"<div><div>The Interaction-Based Deep Material Network (IB-DMN) is reformulated to decouple the phase volume fraction from the topological parameters of the IB-DMN. Since the phase volume fraction is no longer influenced by the topological parameters, on the one hand the stochastic IB-DMN can predict the response of arbitrary phase volume fraction, and on the other hand the stochastic IB-DMN can be constructed by introducing uncertainties to the topological parameters of a reference IB-DMN, which is trained using data obtained from full-field linear elastic homogenisation, allowing to capture the variability resulting from the micro-structure organisation such as a phase clustering.</div><div>The non-linear predictions of the proposed stochastic IB-DMN are compared to those from Direct Numerical Simulation (DNS) on 2D Stochastic Volume Elements (SVEs) of unidirectional fibre-reinforced matrix composites in a finite-strain setting. The results from in-plane uni-axial stress and shear tests show that the proposed stochastic IB-DMN is capable of reproducing random non-linear responses with the same stochastic characteristics as the predictions of the DNS conducted on SVE realisations.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"441 ","pages":"Article 117994"},"PeriodicalIF":6.9,"publicationDate":"2025-04-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143844965","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}
Jana Wedel , Matjaž Hriberšek , Jure Ravnik , Paul Steinmann
{"title":"Ellipsoidal soft micro-particles suspended in dilute viscous flow","authors":"Jana Wedel , Matjaž Hriberšek , Jure Ravnik , Paul Steinmann","doi":"10.1016/j.cma.2025.117973","DOIUrl":"10.1016/j.cma.2025.117973","url":null,"abstract":"<div><div>Soft particles in viscous flows are prevalent both in nature and in various industrial applications. Notable examples include biological cells such as blood cells and bacteria as well as hydrogels and vesicles. To model these intriguing particles, we present an extension of our recent, efficient, and versatile pseudo-rigid body approach, originally developed for initially spherical soft particles suspended in arbitrary macroscale viscous flows. The novel extension allows modeling the barycenter and shape dynamics of soft initially non-spherical, i.e. ellipsoidal particles by introducing a novel shape and orientation tensor. We consider soft, micrometer-sized, ellipsoidal particles deforming affinely. To this end, we combine affine deformations (as inherent to a pseudo-rigid body) and the Jeffery-Roscoe model to analytically determine the traction exerted on a soft ellipsoidal particle suspended locally in a creeping flow at the particle scale. Without loss of generality, we assume nonlinear hyperelastic material behavior for the particles considered. The novel extension of our recent numerical approach for soft particles demonstrates that the deformation and motion of the particles can be accurately reproduced also for ellipsoidal particles and captures results from the literature, however, at drastically reduced computational costs. Furthermore, we identify both the tumbling and trembling dynamic regime for soft ellipsoidal particles suspended in simple shear flow again capturing results from the literature. Our extended approach is first validated using experimental and numerical studies from the literature for quasi-rigid as well as soft particles, followed by a comparison of the effects of particle deformability for some well-known fluid flow cases, such as laminar pipe flow, lid-driven cavity flow, and a simplified bifurcation. We find that taking particle deformability into account leads to notable deviations in the particle trajectory compared to rigid particles, with increased deviations for higher initial particle aspect ratio. Furthermore, we demonstrate that our approach can track a statistically relevant number of soft particles in complex flow situations.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"441 ","pages":"Article 117973"},"PeriodicalIF":6.9,"publicationDate":"2025-04-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143844967","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":"Mechanical state estimation with a Polynomial-Chaos-Based Statistical Finite Element Method","authors":"Vahab Narouie , Henning Wessels , Fehmi Cirak , Ulrich Römer","doi":"10.1016/j.cma.2025.117970","DOIUrl":"10.1016/j.cma.2025.117970","url":null,"abstract":"<div><div>The Statistical Finite Element Method (statFEM) offers a Bayesian framework for integrating computational models with observational data, thus providing improved predictions for structural health monitoring and digital twinning. This paper presents a sampling-free statFEM tailored for non-conjugate, non-Gaussian prior probability densities. We assume that constitutive parameters, modeled as weakly stationary random fields, are the primary source of uncertainty and approximate them using the Karhunen–Loève (KL) expansion. The resulting stochastic solution field, i.e., the displacement field, is a non-stationary, non-Gaussian random field, which we approximate via the Polynomial Chaos (PC) expansion. The PC coefficients are determined through projection using Smolyak sparse grids. Additionally, we model the measurement noise as a stationary Gaussian random field and the model misspecification as a mean-free, non-stationary Gaussian random field, which is also approximated using the KL expansion and where the coefficients are treated as hyperparameters. The PC coefficients of the stochastic posterior displacement field are computed using the Gauss–Markov–Kálmán filter, while the hyperparameters are determined by maximizing the marginal likelihood. We demonstrate the efficiency and convergence of the proposed method through one- and two-dimensional elastostatic problems.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"441 ","pages":"Article 117970"},"PeriodicalIF":6.9,"publicationDate":"2025-04-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143838578","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}
B. van der Heijden , X. Li , G. Lubineau , E. Florentin
{"title":"Enforcing physics onto PINNs for more accurate inhomogeneous material identification","authors":"B. van der Heijden , X. Li , G. Lubineau , E. Florentin","doi":"10.1016/j.cma.2025.117993","DOIUrl":"10.1016/j.cma.2025.117993","url":null,"abstract":"<div><div>Physics-Informed Neural Networks (PINNs) are computationally efficient tools for addressing inverse problems in solid mechanics, but often face accuracy limitations when compared to traditional methods. We introduce a refined PINN approach that rigorously enforces certain physics constraints, improving accuracy while retaining the computational benefits of PINNs. Unlike conventional PINNs, which are trained to approximate (differential) equations, this method incorporates classical techniques, such as stress potentials, to satisfy certain physical laws. The result is a physics-enforced PINN that combines the precision of the Constitutive Equation Gap Method (CEGM) with the automatic differentiation and optimization frameworks characteristic of PINNs. Numerical comparisons reveal that the enforced PINN approach indeed achieves near-CEGM accuracy while preserving the efficiency advantages of PINNs. Validation through real experimental data demonstrates the ability of the method to accurately identify material properties and inclusion geometries in inhomogeneous samples.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"441 ","pages":"Article 117993"},"PeriodicalIF":6.9,"publicationDate":"2025-04-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143844964","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}
Xi-Wen ZHOU , Yin-Fu JIN , Zhen-Yu YIN , Feng-Tao LIU , Xiangsheng CHEN
{"title":"A novel improved edge-based smoothed particle finite element method for elastoplastic contact analysis using second order cone programming","authors":"Xi-Wen ZHOU , Yin-Fu JIN , Zhen-Yu YIN , Feng-Tao LIU , Xiangsheng CHEN","doi":"10.1016/j.cma.2025.118016","DOIUrl":"10.1016/j.cma.2025.118016","url":null,"abstract":"<div><div>Contact problems are of paramount importance in engineering but present significant challenges for numerical solutions due to their highly nonlinear nature. Recognizing that contact problems can be formulated as optimization problems with inequality constraints has paved the way for advanced techniques such as the Interior Point (IP) method. This study presents an Improved Edge-based Smoothed Particle Finite Element Method (IES-PFEM) with novel contact scheme for elastoplastic analysis involving large deformation using Second-Order Cone Programming (SOCP). Within the proposed framework, classical node-to-surface (NTS) and surface-to-surface (STS) contact discretization schemes in SOCP form are rigorously achieved. The governing equations of elastoplastic boundary value problems are formulated as a min-max problem via the mixed variation principle, and by applying the primal-dual theory of convex optimization, the problem is transformed into a dual formulation with stresses as optimization variables. The Mohr-Coulomb plastic yield criterion and the Coulomb friction law are naturally expressed as second-order cone constraints. A fixed-point iteration scheme is developed to address unphysical normal expansion arising from the natural derivation of an associated friction model within the SOCP formulation. Furthermore, the volumetric locking problem in nearly incompressible materials is alleviated by IES-PFEM formulation without requiring additional stabilization techniques. The proposed method is validated through a series of benchmark examples involving contact and elastoplastic deformations. Numerical results confirm the capability of the proposed approach to handle both contact and elastoplastic nonlinearities effectively, without the need for convergence control, highlighting the superiority of the proposed method.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"441 ","pages":"Article 118016"},"PeriodicalIF":6.9,"publicationDate":"2025-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143834454","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":"Continuum and computational modeling of surface effects in flexoelectric materials","authors":"Mònica Dingle , Irene Arias , David Codony","doi":"10.1016/j.cma.2025.117971","DOIUrl":"10.1016/j.cma.2025.117971","url":null,"abstract":"<div><div>In recent times, with the rise of nanoscale technologies, miniaturization of devices has prompted the need to study electromechanical phenomena at small scales. Most studies focus on the phenomena occurring at the bulk portion of the material, such as flexoelectricity, but neglect the effects that arise from the surfaces of the samples. Given the fact that, at such scales, surface-to-volume ratio is inherently large, surface effects cannot be ignored if the full and accurate description of the material’s response wants to be provided. In this work, we present a model that successfully integrates flexoelectricity and the effects of surfaces, and we properly derive the governing equations and boundary conditions for the boundary value problem. We also present a numerical approach in order to computationally solve it, converging at high-order optimal rates. In addition, we present an analytical 1D Euler–Bernoulli electromechanical beam model. Numerically, we find the presence of boundary layers in the transversal electric field across the beam thickness, which are not accounted for in the analytical 1D model. Finally, we find numerical solutions for geometrically-polarized flexoelectric lattice metamaterials, which have large area-to-volume ratios, giving rise to very relevant surface effects. This work emphasizes the importance of accounting for surface effects in modeling and design of flexoelectric devices, including geometrically-polarized metamaterials.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"441 ","pages":"Article 117971"},"PeriodicalIF":6.9,"publicationDate":"2025-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143838562","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 staggered grid shell particle method for shell structural damage subjected to underwater explosion","authors":"Jiasheng Li , Yong Liang , Zhixin Zeng , Xiong Zhang","doi":"10.1016/j.cma.2025.117996","DOIUrl":"10.1016/j.cma.2025.117996","url":null,"abstract":"<div><div>A novel staggered grid shell particle method (SGSPM) is proposed in this paper to model the shell structural damage subjected to underwater explosion. The material point method (MPM) is used to model the fluid in underwater explosion, and the solid shell material point method (SSMPM) is adopted to model the shell structures. A staggered grid scheme is employed to eliminate the cell crossing noise and improve the accuracy of fluid simulation, and a conversion algorithm is proposed to handle the dynamic fracture of shell structures. In addition, a local multi-mesh contact method is introduced into the staggered grid scheme for modeling the fluid–structure interaction. Several numerical examples, including full hemispherical shell, penetration of a thin plate, large deformation of a plate subjected to underwater explosion, fragmentation of a plate and structural damage of a ship hull subjected to contact underwater explosion, are studied by the proposed SGSPM, and the numerical results agree well with the data in the literature and experiments.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"441 ","pages":"Article 117996"},"PeriodicalIF":6.9,"publicationDate":"2025-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143834451","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}
Juan C. Alzate Cobo , Xiang-Long Peng , Bai-Xiang Xu , Oliver Weeger
{"title":"A finite swelling 3D beam model with axial and radial diffusion","authors":"Juan C. Alzate Cobo , Xiang-Long Peng , Bai-Xiang Xu , Oliver Weeger","doi":"10.1016/j.cma.2025.117983","DOIUrl":"10.1016/j.cma.2025.117983","url":null,"abstract":"<div><div>We present a geometrically exact 3D beam model that incorporates axial and radial swelling strains, both small and large, resulting from a rotationally symmetric, thermal or chemical diffusion. Isogeometric collocation is employed to discretize both the mechanical momentum balances and the axis-symmetric, steady-state 2D diffusion equation along the beam. The resulting coupled nonlinear problem for displacements, rotations, and temperatures or concentrations is solved using a staggered scheme. The approach is further extended to include beam-to-beam interfaces and is therefore well suited for the simulation of lattice structures. The model and its discretization are validated against 3D continuum models in various numerical examples and prove to be both accurate and numerically efficient. The novelty of the presented method is twofold. First, it relates beam theory, and consequently small elastic strains, with large swelling deformation stemming from anisotropic diffusion phenomena. Second, it also provides insight into the implementation of isogeometric collocation for solving diffusion equations subject to large deformations. Ultimately, this novel finite swelling beam model can present the starting point for the efficient modeling of lattice structures under diffusion conditions, such as microstructured Li-ion electrodes or thermoelectric semiconductors.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"441 ","pages":"Article 117983"},"PeriodicalIF":6.9,"publicationDate":"2025-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143834450","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":"Predicting change, not states: An alternate framework for neural PDE surrogates","authors":"Anthony Zhou , Amir Barati Farimani","doi":"10.1016/j.cma.2025.117990","DOIUrl":"10.1016/j.cma.2025.117990","url":null,"abstract":"<div><div>Neural surrogates for partial differential equations (PDEs) have become popular due to their potential to quickly simulate physics. With a few exceptions, neural surrogates generally treat the forward evolution of time-dependent PDEs as a black box by directly predicting the next state. While this is a natural and easy framework for applying neural surrogates, it can be an over-simplified and rigid framework for predicting physics. In this work, we evaluate an alternate framework in which neural solvers predict the temporal derivative and an ODE integrator forwards the solution in time, which has little overhead and is broadly applicable across model architectures and PDEs. We find that by simply changing the training target and introducing numerical integration during inference, neural surrogates can gain accuracy and stability in finely-discretized regimes. Predicting temporal derivatives also allows models to not be constrained to a specific temporal discretization, allowing for flexible time-stepping during inference or training on higher-resolution PDE data. Lastly, we investigate why this framework can be beneficial and in what situations does it work well.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"441 ","pages":"Article 117990"},"PeriodicalIF":6.9,"publicationDate":"2025-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143838577","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}
Sichen Dong, Lei Li, Tianyu Yuan, Xiaotan Yu, Pan Wang, Fusen Jia
{"title":"A random interval coupling-based active learning Kriging with meta-model importance sampling method for hybrid reliability analysis under small failure probability","authors":"Sichen Dong, Lei Li, Tianyu Yuan, Xiaotan Yu, Pan Wang, Fusen Jia","doi":"10.1016/j.cma.2025.117992","DOIUrl":"10.1016/j.cma.2025.117992","url":null,"abstract":"<div><div>In this study, a novel active learning method is proposed and combined with Meta-IS-AK for hybrid reliability analysis under small failure probability. Considering the proportion of responses falling into the failure domain, the interval failure degree is introduced to describe the probability of misjudging the state for random samples. The novel active learning method (IAD) is proposed to select valuable samples for updating Kriging model, considering the interval failure degree and the sample clustering. Additionally, a corresponding convergence criterion based on the similarity of the indicator functions in importance sampling samples is proposed to further enhance efficiency. The accuracy and superiority of the proposed method are validated through seven illustrative examples, accompanied by detailed explanations.</div></div>","PeriodicalId":55222,"journal":{"name":"Computer Methods in Applied Mechanics and Engineering","volume":"441 ","pages":"Article 117992"},"PeriodicalIF":6.9,"publicationDate":"2025-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143834449","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}