
今日更新:Journal of the Mechanics and Physics of Solids 1 篇,Mechanics of Materials 2 篇,Thin-Walled Structures 3 篇
Infimal Representation of Landau Potentials
Christophe Denoual, Nicolas Bruzy
doi:10.1016/j.jmps.2026.106653
朗道势的有限表示
A free energy potential for an arbitrary set of stable or metastable states is defined in a large strain formulation to model displacive transitions such as martensitic transformation or twinning. This potential is based on the use of a graph of stable states (vertices of the graph) and continuous transformations (segments of the graph). This graph is built knowing, for each stable or metastable state, the point group rotations of the lattice, the elasticity tensor, and the transformation (represented as a stretch tensor) describing the continuous evolution between stable states. The potential is based on an infimal convolution of a quadratic potential representing the elastic response and a transformational potential containing the information about the possible succession of states and their energies. Thanks to its expression as an infimal operation, this potential is a function of the total and plastic strain only, and therefore defines the displacive transformation as a non-dissipative process. A particular attention is paid to the representation of elasticity for stable states (including stiffening under compression), and the definition of a non-convex potential for the transition states. Despite the inherent complexity and adaptability of the method, a closed-form expression for stress, characterized as the derivative of free energy, is being proposed. We detail through two illustrative examples the construction of potentials and examine the influence of plasticity and the convexity of the potential on the hysteresis observed in the body-centered cubic (bcc) to hexagonal close-packed (hcp) martensitic transition in iron subjected to pressure.
在大应变公式中定义了任意一组稳定或亚稳态的自由能势,以模拟马氏体相变或孪晶等位移转变。这种潜力是基于稳定状态(图的顶点)和连续转换(图的段)的图的使用。对于每一个稳定或亚稳定状态,我们都知道晶格的点群旋转、弹性张量和描述稳定状态之间连续演化的变换(用拉伸张量表示)。势是基于表示弹性响应的二次势和包含可能的状态序列及其能量信息的转换势的无限卷积。由于它的表达式是一个无穷大的操作,这个势是一个函数的总应变和塑性应变,因此定义位移转换为一个非耗散过程。特别注意稳定状态(包括压缩下的硬化)的弹性表示,以及过渡状态的非凸势的定义。尽管该方法具有固有的复杂性和适应性,但仍提出了一种以自由能导数为特征的应力封闭表达式。我们通过两个例子详细说明了势的构造,并研究了在压力作用下铁的体心立方(bcc)到六方密堆积(hcp)马氏体转变过程中所观察到的磁滞现象的塑性和势的凹凸性对磁滞的影响。
Delaying matrix damage in composite laminates by fibre pre-stressing: A micromechanical numerical study
Pere Maimí, Lucas de Sousa Rodrigues, Laura Carreras, Daniel Sánchez-Rodríguez, Josep Costa
doi:10.1016/j.mechmat.2026.105712
纤维预应力延缓复合材料层合板基体损伤的细观力学数值研究
Composite laminates offer high specific stiffness and strength, but their performance is often limited by matrix cracking, which acts as a precursor to delamination and provides pathways for moisture ingress and environmental degradation. Fibre prestressing can generate beneficial residual stresses in the matrix, delaying the onset of such cracks. However, the magnitude of this effect and the mechanis ms by which prestressing alters the local matrix stress state remain difficult to quantify using experimental observations alone. This work presents a numerical methodology to quantify the influence of prestressing on the matrix stress field in unidirectional plies and multidirectional laminates. A combination of representative volume elements and laminate theory is used to characterise hydrostatic pressure, von Mises stress and stress triaxiality, which are key indicators of cavitation-driven matrix failure. Illustrative results obtained for representative cross-ply laminates show that prestress levels well below the fibre strength can increase the longitudinal load required to reach comparable hydrostatic pressure levels by one order of magnitude.
复合材料层压板具有很高的比刚度和强度,但其性能通常受到基体开裂的限制,这是分层的前兆,并为水分进入和环境退化提供了途径。纤维预应力可以在基体中产生有益的残余应力,延缓此类裂缝的发生。然而,这种影响的大小和预应力改变局部基质应力状态的机制仍然难以单独使用实验观察来量化。本文提出了一种量化预应力对单向层合板和多向层合板基体应力场影响的数值方法。结合代表性体积单元和层压理论来表征静水压力、冯米塞斯应力和应力三轴性,这是空化驱动的基质破坏的关键指标。对代表性的交叉层压板获得的说明性结果表明,远低于纤维强度的预应力水平可以使达到可比静水压力水平所需的纵向载荷增加一个数量级。
Hybrid ANN/FEM-Based inverse identification procedure using the Levenberg–Marquardt method: Application to X100CrMoV5 steel alloy
Najem A., Altmeyer G., Coudray Z., Duchosal A.
doi:10.1016/j.mechmat.2026.105702
基于Levenberg-Marquardt方法的混合ANN/ fem逆识别程序:在X100CrMoV5钢合金中的应用
This study investigates the mechanical response of the X100CrMoV5 steel alloy under severe thermochemical conditions and aims to identify its constitutive parameters through an inverse procedure. This approach integrates experimental force–displacement curves, finite element numerical simulations, and artificial neural network predictions to enhance both accuracy and computational efficiency. The experimental tests are conducted using a thermo-mechanical Gleeble machine under dynamic conditions at ambient and elevated temperature. A finite element model is developed to replicate the experimental test, enabling a detailed comparison with the experimental data. Additionally, an artificial neural network, trained on numerical simulation results, is employed to predict the outcomes of the material tests, significantly accelerating the identification process. The inverse identification procedure is formulated as an optimization problem and solved using the deterministic Levenberg–Marquardt method, minimizing the cost between experimental and numerical results. The proposed hybrid identification strategy demonstrates the effectiveness of coupling numerical simulations with artificial neural networks, particularly Long Short-Term Memory (LSTM) networks, to enhance accuracy and significantly reduce computational costs. The findings of this study highlight the potential of AI-driven approaches in improving material characterization and identification.
本研究研究了X100CrMoV5钢合金在严酷热化学条件下的力学响应,旨在通过反求程序确定其本构参数。该方法将实验力-位移曲线、有限元数值模拟和人工神经网络预测相结合,提高了精度和计算效率。利用热机械式Gleeble试验机在常温和高温动态条件下进行了试验。建立了一个有限元模型来复 制实验测试,以便与实验数据进行详细的比较。此外,利用数值模拟结果训练的人工神经网络对材料试验结果进行预测,大大加快了识别过程。将反辨识过程表述为优化问题,并采用确定性Levenberg-Marquardt方法求解,使实验结果与数值结果之间的代价最小。所提出的混合识别策略证明了将数值模拟与人工神经网络,特别是长短期记忆(LSTM)网络相结合的有效性,可以提高准确性并显着降低计算成本。这项研究的发现突出了人工智能驱动的方法在改善材料表征和鉴定方面的潜力。
Coupled dynamics and scaling relations of thin flexible wings with multiple distributed feathers
Guojun Li, Lanlan Wang, Weitao Jiang, Hongzhong Liu
doi:10.1016/j.tws.2026.115055
多分布羽薄柔性翼的耦合动力学及尺度关系
This study investigates the coupled aeroelastic dynamics and performance enhancement of thin flexible wings equipped with multiple distributed feathers, motivated by bio-inspired design of agile micro-air and unmanned aerial vehicles. A multi-body aeroelastic framework is developed by coupling an incompressible Navier–Stokes solver with large-eddy simulation in an arbitrary Lagrangian–Eulerian frame and a flexible multi-body structural model. The aeroelastic solver is applied to examine the aerodynamic performance of thin flexible wings without feathers, with a single leading-edge feather and with multiple overlapping feathers over a range of angles of attack, feather preset angles and feather lengths at R e = 2500. The results show that multi-feather configurations reorganize leading-edge vortex dynamics, yielding higher lift and lift-to-drag ratios than membrane-only wings. This is particularly evident in pre-stall and transitional regimes, where the design also keeps drag penalties moderate and reduces force fluctuations. A shallow positive preset angle combined with an intermediate feather length establishes a soft chordwise slot that sustains a gently convecting leading-edge vortex, accelerates pressure recovery and improves aerodynamic efficiency. When feathers become too long or too steep, blockage effects re-inflate drag and degrade performance. A new empirical scaling relation incorporating bounded feather coverage and preset-angle modifiers is proposed to predict the maximum membrane deformation under a Weber-number backbone. These findings provide quantitative guidelines for designing feathered morphing membrane wings with tunable lift, efficiency and stability across a wide angle-of-attack envelope.
基于仿生设计的敏捷微型飞行器和无人机,研究了多分布羽毛薄柔性机翼的耦合气动弹性动力学和性能增强问题。将不可压缩Navier-Stokes求解器与任意拉格朗日-欧拉框架的大涡模拟和柔性多体结构模型相结合,建立了多体气动弹性框架。应用气动弹性求解器对无羽毛、单前缘羽毛和多前缘羽毛在不同迎角、羽毛预设角和羽毛长度范围内的气动性能进行了研究。结果表明,多羽毛结构重组了前缘涡动力学,产生了比单膜机翼更高的升力和升阻比。这一点在失速前和过渡阶段尤为明显,在这些阶段,设计还能保持适度的阻力损失,减少力波动。一个较浅的正预设角与中间羽毛长度相结合,建立了一个柔软的弦向槽,维持一个温和的前缘对流涡流,加速压力恢复,提高气动效率。当羽毛变得太长或太陡时,堵塞效应会重新增加阻力并降低性能。提出了一种结合有界羽毛覆盖度和预设角度修正因子的经验标度关系式,用于预测韦伯数骨架下膜的最大变形。这些发现为设计具有可调升力、效率和稳定性的大迎角羽毛变形膜翼提供了定量指导。
LamiNetDef: A novel physics-informed U-Net++ framework for predicting deflection shapes of laminated composite plates
Quang-Viet Vu, Dai-Nhan Le, Nguyen Pham Dinh, George Papazafeiropoulos, Sawekchai Tangaramvong
doi:10.1016/j.tws.2026.115041
LamiNetDef:一种新的基于物理的U-Net++框架,用于预测层压复合材料板的挠度形状
This study introduces LamiNetDef, a novel physics-informed deep learning framework developed to accurately predict the mid-surface deflection patterns of laminated composite plates subjected to uniformly distributed loads. A comprehensive dataset is generated through an automated finite element modeling pipeline, integrating Abaqus and MATLAB via the Abaqus2Matlab toolbox. The model parameters encompass geometric dimensions, stacking sequence, ply orientations, and boundary conditions, sampled using a Latin Hypercube Sampling scheme to efficiently explore the design variable space. The finite element nodal displacements are subsequently normalized and processed to form structured data for training. The proposed LamiNetDef architecture, built upon a modified U-Net++ network, incorporates a Boundary Condition Deflection Form that encodes physical boundary effects as a two-dimensional spatial input to better inform the encoder in learning physically consistent deformation modes. In parallel, design parameters including geometric and material descriptors are processed by Kolmogorov–Arnold Networks to derive adaptive conditioning parameters for the Feature-wise Linear Modulation layers. This mechanis m enables dynamic coupling between physics and features, enriching the model’s representational capacity. The network is trained under a physics-informed loss function that integrates both magnitude and gradient terms, ensuring s moothness and physical consistency in the predicted deformation fields. The LamiNetDef framework achieves outstanding predictive performance, with R 2 = 0.951 and MSSIM = 0.987 on independent test data, demonstrating excellent agreement with finite element results. Furthermore, the framework is deployed as a cloud-based interactive platform via Hugging Face, enabling rapid and reliable deflection prediction directly from user-defined design parameters.
本研究引入了LamiNetDef,这是一种新颖的物理信息深度学习框架,用于准确预测受均匀分布载荷作用的层压复合材料板的中表面挠度模式。通过自动化有限元建模管道生成综合数据集,通过abaquus2matlab工具箱集成Abaqus和MATLAB。模型参数包括几何尺寸、堆叠顺序、层向和边界条件,使用拉丁超立方采样方案进行采样,以有效地探索设计变量空间。随后将有限元节点位移归一化处理,形成结构化数据进行训练。LamiNetDef架构基于改进的U-Net++网络,包含一个边界条件偏转形式,该形式将物理边界效应编码为二维空间输入,以便更好地告知编码器学习物理一致的变形模式。同时,包括几何和材料描述符在内的设计参数由Kolmogorov-Arnold网络处理,以获得面向特征的线性调制层的自适应调节参数。这种机制使物理和特征之间的动态耦合成为可能,从而丰富了模型的表示能力。该网络在一个物理信息损失函数下进行训练,该函数集成了幅度和梯度项,确保了预测变形场的平滑性和物理一致性。LamiNetDef框架具有出色的预测性能,独立试验数据的R 2 = 0.951, MSSIM = 0.987,与有限元结果吻合良好。此外,该框架通过hug Face作为基于云的交互平台部署,可以直接从用户定义的设计参数中快速可靠地预测挠度。
Compatibility an alysis between two-phase local/nonlocal and classical theories for nanobeams in s mooth and non-s mooth fields
Z.W. Song, S.K. Lai, Y.T. Zhang, C.W. Lim
doi:10.1016/j.tws.2026.115031
纳米梁在光滑场和非光滑场中两相局部/非局部与经典理论的相容性分析
In general, it is assumed that both Eringen’s two-phase local/nonlocal theory and the stress-driven two-phase local/nonlocal theory reduce to the classical local theory as the nonlocal length-scale parameter l 0 approaches zero, despite the lack of a rigorous theoretical proof for this fundamental issue. In addition, the relationship between the local phase volume fraction β and l 0 remains unclear. This work demonstrates that the constitutive boundary and interface conditions present in two-phase local/nonlocal theories do not automatically vanish or become satisfied in the classical local theory as l 0 tends to zero. Achieving full compatibility between nonlocal and local theories requires that these conditions are also satisfied within the classical framework. Furthermore, it is concluded that β is associated with l 0 , and the stress-driven nonlocal theory is not fully compatible with the classical local theory. These findings are validated through illustrative examples to highlight the compatibility conditions between nonlocal and local theories and to elucidate the relationship between β and l 0 .
一般认为,当非局部长度尺度参数l0趋近于零时,Eringen的两相局部/非局部理论和应力驱动的两相局部/非局部理论均可归结为经典局部理论,尽管这一基本问题缺乏严格的理论证明。此外,局部相体积分数β与l0之间的关系尚不清楚。本研究表明,两相局域/非局域理论中存在的本构边界和界面条件在经典局域理论中不会随着l0趋于零而自动消失或满足。要实现非局部理论和局部理论之间的完全兼容,需要在经典框架内也满足这些条件。进一步得出β与l0相关的结论,应力驱动的非局部理论与经典局部理论不完全相容。通过举例验证了这些发现,突出了非局部理论和局部理论之间的相容条件,并阐明了β和l0之间的关系。