【新文速递】2026年4月3日复合材料SCI期刊最新文章_ACT_Mechanical_Maxwell

【新文速递】2026年1月2日固体力学SCI期刊最新文章

今日更新：Mechanics of Materials 1 篇，International Journal of Plasticity 1 篇，Thin-Walled Structures 8 篇Mechanics of MaterialsInvestigation to Grain-size Dependent Plasticity of Mg Alloys Based on Phase Field ApproachJiachen Hu, Bo Xu, Junyuan Xiong, Chao Yu, Guozheng Kangdoi:10.1016/j.mechmat.2025.105596 基于相场法的镁合金晶粒尺寸塑性研究A grain size-dependent crystal plasticity-twinning phase field model is proposed by integrating a Hall-Petch-type slipping resistance and a grain misorientation-dependent barrier energy. Monotonic and cyclic deformation simulations are performed on the texture-free and basal-textured polycrystalline Mg alloys to investigate their plastic deformation and underlying mechanis ms governed by both the grain size and texture. The results demonstrate that the grain refinement suppresses twinning and promotes non-basal slipping, shifting the plastic deformation in the texture-free alloys from a twinning-dislocation slipping co-dominance to a dislocation slipping dominance. However, the sustained dominance of twinning and basal slipping in the basal-textured systems indicates a lower critical grain size of twinning compared to the texture-free ones. Under a cyclic loading, the grain refinement mitigates the detwinning-induced inelastic unloading in the texture-free systems and reduces the yield asymmetry in the basal-textured ones. The grain refinement also reduces the number and kind of twins by limiting their available space and nucleation sites, and weakening the twin-twin interactions. The increased deformation homogeneity shifts the twin evolution at peak tensile strain from a thickening- to nucleation- and propagation-dominated modes and makes the re-twinning be more strongly suppressed than primary twinning. Furthermore, the texture enhances twin-twin interactions, intensifying the localization of dislocation slipping, while the grain refinement promotes a more homogeneous distribution of dislocation slipping. The inhomogeneous stress fields generated by accumulated dislocations reversely regulate the twin activity, revealing a coupling mechanis m between them. These findings provide mechanistic insights into strengthening and toughening Mg alloys through a texture-grain size synergistic design.通过集成霍尔-佩奇型滑移阻力和与晶粒取向偏差相关的势垒能，提出了晶粒尺寸相关的晶体塑性-孪晶相场模型。对无织构和基本织构的多晶镁合金进行了单调和循环变形模拟，研究了晶粒尺寸和织构对多晶镁合金塑性变形的影响及其机制。结果表明：晶粒细化抑制了孪晶，促进了非基滑移，使无织构合金的塑性变形从孪晶-位错滑移共主导转变为位错滑移主导；然而，在基底织构体系中，孪晶和基底滑移的持续优势表明，与无织构体系相比，孪晶的临界晶粒尺寸更低。在循环加载下，晶粒细化可以减轻无织构体系中去孪晶引起的非弹性卸载，降低基织构体系中的屈服不对称性。晶粒细化通过限制孪晶的可用空间和成核位置，减弱孪晶间的相互作用，减少孪晶的数量和种类。形变均匀性的增加使孪晶在峰值拉伸应变下的演化从增厚模式转变为以形核和扩展为主的模式，并使再孪晶比初次孪晶受到更强的抑制。织构增强了孪晶间的相互作用，强化了位错滑移的局部化，而晶粒的细化使位错滑移的分布更加均匀。累积位错产生的非均匀应力场反向调节孪晶活动，揭示了它们之间的耦合机制。这些发现为通过织构-晶粒尺寸协同设计强化和增韧镁合金提供了机理见解。International Journal of PlasticityIndentation-informed convolutional neural network for simultaneous prediction of non-equibiaxial residual stress and plastic flowMinwoo Park, Deunbom Chung, Wanchuck Woo, Seungcheol Oh, Kyeongjae Jeong, Heung Nam Handoi:10.1016/j.ijplas.2025.104606 同时预测非等双轴残余应力和塑性流动的凹痕卷积神经网络This study proposes an integrated finite element (FE) simulation and convolutional neural network (CNN) model designed for the simultaneous prediction of plastic properties and surface in-plane non-equibiaxial residual stress from spherical indentation responses. By eliminating the need for stress-free reference specimens, the proposed framework enables non-destructive prediction. The framework leverages indentation load-depth curves and directional deformation profiles derived from validated FE simulations. Sensitivity an alyses identify the indenter radius and penetration ratio as critical factors for improving prediction accuracy and maximizing the sensitivity of indentation responses to variations in residual stress. The influence of non-equibiaxial residual stress states on indentation behavior is further elucidated through a mechanistic investigation, which reveals a close association with cumulative volumetric changes in equivalent plastic strain near the indentation zone. The CNN training performance supports the sensitivity-based determination of optimal indentation settings. The model is shown to achieve a mean absolute error corresponding to below 5% on average for residual stresses, while the plasticity parameters are also well captured. Experimental assess ment on copper specimens with homogeneous residual stress fields verifies the accuracy and adaptability of the developed FE–CNN model. Further validation using additively manufactured stainless steel, exhibiting complex heterogeneous residual stresses, shows strong consistency with neutron diffraction measurements. This FE–CNN framework presents a robust and scalable approach for comprehensive mechanical characterization, offering substantial benefits for assessing structural integrity and reliability across diverse industrial applications without recourse to destructive testing.本研究提出了一种集成有限元（FE）模拟和卷积神经网络（CNN）模型，用于同时预测球面压痕响应的塑性性能和表面面内非等双轴残余应力。通过消除对无应力参考样本的需求，所提出的框架能够进行非破坏性预测。该框架利用了压痕载荷-深度曲线和从经过验证的有限元模拟得出的定向变形曲线。灵敏度分析表明，压痕半径和侵彻比是提高预测精度和最大化压痕响应对残余应力变化敏感性的关键因素。通过力学研究进一步阐明了非等双轴残余应力状态对压痕行为的影响，揭示了压痕区附近等效塑性应变的累积体积变化与压痕区附近等效塑性应变的累积体积变化密切相关。CNN训练性能支持基于灵敏度的最佳缩进设置的确定。结果表明，该模型对残余应力的平均绝对误差小于5%，同时塑性参数也得到了很好的捕捉。对具有均匀残余应力场的铜试样进行了实验评估，验证了所建立的FE-CNN模型的准确性和适应性。用增材制造的不锈钢进一步验证，显示出复杂的非均匀残余应力，与中子衍射测量结果有很强的一致性。这种FE-CNN框架提供了一种强大的、可扩展的方法来进行全面的力学表征，为评估不同工业应用中的结构完整性和可靠性提供了实质性的好处，而无需诉诸破坏性测试。Thin-Walled StructuresDesign and Mechanical An alysis of a Noval Space Orthogonal Feather-like Deployment Structure with High StabilityHeng Li, Ning Kong, Shuo Yang, Jie Zhang, Haowei Wang, Xiaoyu Wang, Yuan Zhuangdoi:10.1016/j.tws.2025.114466一种新型高稳定性空间正交羽状展开结构设计与力学分析The space deployment mechanis m is an important part of the spacecraft. The mechanis m is mainly used to deploy components such as solar wings and satellite antennas, which can reduce the volume of the spacecraft during launch. It can improve the stability of the launch phase, and save launch costs. In this work, a completely elastic deployment structure named the orthogonal feather-like elastic deployment structure has been proposed, which obtains relatively stable stretching force during the unfolding process with high stiffness in the unfolded state. A numerical model has been established by studying the mechanical an alysis and the variation law of stretching force, which can calculate the orthogonal structure’s stretching force with different size parameters in the whole deformation process. The accuracy of the numerical model has been verified by finite element an alysis and expansion test of the 3D printed model. The influence of the substructures of the orthogonal structure on the retraction force has been exhibited by an alyzing the numerical model. This work provides a theoretical basis for the design of the orthogonal feather-like deployment structure.空间展开机构是航天器的重要组成部分。该机构主要用于部署太阳能翼和卫星天线等部件，在发射过程中可以减小航天器的体积。它可以提高发射阶段的稳定性，节省发射成本。本文提出了一种完全弹性展开结构——正交羽状弹性展开结构，该结构在展开过程中获得相对稳定的拉伸力，且在展开状态下具有较高的刚度。通过对拉伸力的力学分析和变化规律的研究，建立了正交结构在整个变形过程中不同尺寸参数下的拉伸力计算模型。通过三维打印模型的有限元分析和展开试验，验证了数值模型的准确性。通过数值模型分析，揭示了正交结构的子结构对回撤力的影响。该工作为正交羽状展开结构的设计提供了理论依据。`Dynamic a nalysis of composite laminated conical-cylindrical shells with random-interval uncertaintiesYugeng Chen, Rui Zhong, Qingshan Wang, Liming Chen, Bin Qindoi:10.1016/j.tws.2025.114467具有随机区间不确定性的复合材料锥形-圆柱形层合壳动力分析The an alysis of dynamic characteristics in composite structures has long been a critical issue in mechanical engineering. However, existing studies have predominantly focused on deterministic parameters, neglecting the influence of uncertainties on the structural dynamic characteristics. This study investigates composite laminated conical-cylindrical shells, systematically incorporating the influence of random-interval hybrid uncertainties originating from both structural parameters and service environments, to develop an advanced modeling framework for uncertain dynamic characteristics and an efficient uncertainty-propagation an alysis method. This study establishes a numerical an alysis model for characterizing uncertain dynamic characteristics through integration of the First-order Shear Deformation Theory, Rayleigh-Ritz method, and P-box method. A new random-interval hybrid uncertainty-propagation an alysis method is subsequently proposed by synergistically combining Kriging, adaptive learning, and interval number ranking methods. Furthermore, the effectiveness of the model and method is rigorously validated through numerical cases and experiments. Based on this, the influence mechanis m of different uncertainties on the uncertain dynamic characteristics of the shell is investigated. Numerical results demonstrates that different uncertainties have different influence on the uncertain dynamic characteristics of the shell. In the precision design of complex mechanical structures, the influence of uncertainties on the dynamic characteristics of structures must be thoroughly considered.复合材料结构的动力特性分析一直是机械工程中的一个重要问题。然而，现有的研究主要集中在确定性参数上，忽略了不确定性对结构动力特性的影响。本文以锥形圆柱复合材料层合壳为研究对象，系统地考虑了结构参数和使用环境的随机区间混合不确定性的影响，建立了一种先进的不确定动态特性建模框架和有效的不确定性传播分析方法。本研究将一阶剪切变形理论、瑞利-里兹法和P-box法相结合，建立了表征不确定动力特性的数值分析模型。将克里格法、自适应学习法和区间数排序法协同结合，提出了一种新的随机-区间混合不确定性传播分析方法。通过数值算例和实验，验证了模型和方法的有效性。在此基础上，研究了不同不确定度对壳不确定动力特性的影响机理。数值结果表明，不同的不确定性对壳的不确定动力特性有不同的影响。在复杂机械结构的精密设计中，必须充分考虑不确定性对结构动力特性的影响。Dynamics and Control of Thin-Walled Liquid-Filled ContainersJiawei Han, Jie Huangdoi:10.1016/j.tws.2025.114471薄壁充液容器动力学与控制The liquid transported in thin-walled containers is widely used in many engineering applications. The elastic sidewall interacting with the liquid sloshing degrade the operation safety. Dynamic equations of the liquid-filled containers with flexible sidewalls undergoing planar motions were derived. The equations depict the dynamic interaction between the liquid sloshing and elastic sidewall. Furthermore, the frequency-estimation equation of the interaction is also described. Then, a new technique for controlling the coupled oscillations caused by the interaction is introduced. The computational simulations were performed to an alyze the complicated dynamic interaction and verify the effectiveness of the control method. The theoretical findings may also benefit other types of thin-walled liquid-filled structures.在许多工程应用中，液体在薄壁容器中运输有着广泛的应用。弹性侧壁与液体晃动相互作用，降低了操作安全性。推导了具有柔性侧壁的充液容器进行平面运动的动力学方程。该方程描述了液体晃动与弹性侧壁之间的动力学相互作用。此外，还描述了相互作用的频率估计方程。然后，介绍了一种控制相互作用引起的耦合振荡的新技术。通过计算仿真分析了复杂的动力学交互作用，验证了控制方法的有效性。这些理论发现也可能对其他类型的薄壁液体填充结构有益。Dynamic response of fully metallic laser-welded trapezoidal corrugated sandwich panels under near-field blast loadingKe Li, Yue Lu, Kun Liu, Yu Gao, Zili Wang, Sanfeng Liudoi:10.1016/j.tws.2025.114472近场爆炸载荷下全金属激光焊接梯形波纹夹层板的动力响应In view of the significant application potential of laser-welded sandwich panels in anti-blast performance and large-scale engineering applications, both experimental and numerical investigations were carried out on fully metallic laser-welded sandwich panels with trapezoidal corrugated cores. Specimens with two different core heights were fabricated to investigate the damage modes of the structures under near-field blast loading. Furthermore, the three-dimensional finite element model was developed and validated to further understand the dynamic response behavior of the sandwich structures subjected to blast loads. The experimental and numerical results demonstrated that the core cell exhibited unidirectional compression on one side, and this behavior became more pronounced with increasing core height. The energy absorption performance of each component of the sandwich panel can be characterized by the critical scaled standoff distance, with the front face sheet and the core cell serving as the primary contributors. Based on the variations in scaled standoff distance and the specific energy absorption of the sandwich panel, the blast loading regime can be divided into three distinct areas: dangerous zone with damage area exceeding 4.51 × 104 mm2, dynamic zone, and safe zone with damage area less than 1.35 × 104 mm2. Furthermore, increasing the core height to transform the sandwich panel into a soft-spring type configuration can also effectively enhance the anti-blast performance under near-field explosive loading. Arranging outer stiffeners perpendicular to the hypotenuses of the core cells in the central blast-affected region can significantly enhance the local structural strength.鉴于激光焊接夹芯板在抗爆破性能和大规模工程应用方面具有巨大的应用潜力，对梯形波纹芯全金属激光焊接夹芯板进行了实验和数值研究。为研究近场爆炸荷载作用下结构的损伤模式，制作了两种不同岩心高度的试件。为进一步了解夹层结构在爆炸荷载作用下的动力响应特性，建立了三维有限元模型并进行了验证。实验和数值结果表明，随着岩心高度的增加，岩心单侧呈现出单向压缩的特征，且这种特征更加明显。夹层板各部件的吸能性能可以用临界尺度距离来表征，其中前面板和芯芯是主要的吸能因子。根据夹芯板尺度距离和比能吸收的变化，将爆炸加载区划分为损伤面积大于4.51 × 104 mm2的危险区、动力区和损伤面积小于1.35 × 104 mm2的安全区。此外，增加芯板高度，将夹层板转变为软弹簧型结构，也能有效提高近场爆炸载荷下的抗爆炸性能。在中心爆炸影响区垂直于核心单元斜边布置外加强筋，可显著提高局部结构强度。Experimental Study on Seis mic Performance of Aluminum Alloy Beam-to-column Joints Using Stainless Steel ConnectionsYing Zhang, Yuanqing Wang, Mingze Wu, Huan Lu, Beibei Li, Shuai Mo, Yuanwen Ouyangdoi:10.1016/j.tws.2025.114473不锈钢连接铝合金梁柱节点抗震性能试验研究A total of six full-scale aluminum alloy cruciform beam-to-column joints using stainless steel connections with different forms were designed and tested under cyclic loads. The experimental program investigated joints with three types of flange connectors (top-and-bottom T-stubs, top-and-seat stiffened T-stubs, and top-and-seat angle cleats), two types of beam web connections (with or without web angle cleats), and two panel zone conditions (with or without stiffeners). The study evaluated the seis mic performance of the joints, including their failure modes, hysteretic response, ductility, and energy dissipation capacity. The key findings are as follows: (1) The high-strength stainless steel swage-locking pins exhibited excellent compatibility with the aluminum alloy structure, failing in a ductile manner through progressive sleeve pull-out under cyclic loading, which enhanced the energy dissipation capacity of the joints. (2) All specimens, except the one with top-and-seat and web connectors, met the seis mic requirements for Special Moment Frames (S MF) and High Ductility Class (DCH) structures, and exhibit the spread of plasticity on the flange connector. (3) A finite element model was developed and validated, providing a fundamental numerical tool for the subsequent formulation of design methods.设计了6个采用不同形式不锈钢连接的全尺寸铝合金十字形梁柱节点，并对其进行了循环荷载试验。实验程序研究了三种类型的法兰连接件（顶部和底部t型存根，顶部和座位加劲t型存根，顶部和座位角夹板），两种类型的梁腹板连接（有或没有腹板角夹板），以及两种面板区域条件（有或没有加强筋）。该研究评估了节点的抗震性能，包括其破坏模式、滞后响应、延性和耗能能力。研究结果表明：(1)高强度不锈钢挤压锁紧销与铝合金结构具有良好的相容性，在循环荷载作用下，通过套管的渐进拉拔，以延性方式失效，增强了节点的耗能能力；(2)除顶座连接件和腹板连接件外，所有试件均满足特殊弯矩框架（S MF）和高延性等级（DCH）结构的抗震要求，且塑性在法兰连接件上表现出扩散。(3)建立并验证了有限元模型，为后续设计方法的制定提供了基本的数值工具。Intelligent optimization method of non-uniformly distributed stiffened cylindrical shells under localized axial load based on CNN model and beluga whale optimization algorithmHuangyang Xu, Peng Jiao, Qinghua Zhou, Mingheng Zhang, Li Yu, Zhiping Chendoi:10.1016/j.tws.2025.114462基于CNN模型和白鲸优化算法的局部轴向载荷下非均匀分布加劲圆柱壳智能优化方法Stiffened cylindrical shells are widely used in engineering due to their excellent load-bearing capacity. However, their buckling behavior under localized axial compression depends on the stiffener layout, and traditional uniform distribution designs often fail to maximize the structural load-bearing potential. This study proposes a method for optimizing the layout of non-uniform stiffeners based on deep learning and intelligent optimization algorithms. The goal is to improve the critical buckling performance of thin-walled cylindrical shells under localized axial compression. First, non-uniform stiffener paths were designed. Grayscale images were used to characterize these paths, and a corresponding finite element an alysis model was created. The an alysis method's validity was verified through numerical examples. Next, a convolutional neural network (CNN) model was introduced as a surrogate model based on the generated diverse dataset of non-uniform stiffener images to enable rapid prediction of structural responses. The CNN model was then combined with the Beluga Whale Optimization (BWO) algorithm to perform a global optimization search for the non-uniform stiffener layout. The optimization results demonstrate that, compared to traditional uniform distribution designs, the proposed method increases the critical buckling load by 9.74%, 20.75%, and 11.19%, respectively, thereby enhancing the structural load-bearing efficiency. This study introduces a novel, image-driven optimization paradigm for non-uniform stiffeners that integrates deep learning with evolutionary algorithms to intelligently design complex stiffened structures. This approach provides a new perspective for reducing the weight of thin-walled structures under local axial compression, offering practical value in improving design accuracy and efficiency.加劲圆柱壳具有良好的承载性能，在工程中得到了广泛的应用。然而，它们在局部轴压下的屈曲行为取决于加劲肋的布置，传统的均匀分布设计往往不能最大限度地发挥结构的承载潜力。提出了一种基于深度学习和智能优化算法的非均匀加强筋布局优化方法。目的是提高薄壁圆柱壳在局部轴压作用下的临界屈曲性能。首先，设计了非均匀加劲路径。利用灰度图像对这些路径进行表征，并建立相应的有限元分析模型。通过数值算例验证了分析方法的有效性。接下来，基于生成的不同非均匀加劲图像数据集，引入卷积神经网络（CNN）模型作为替代模型，实现结构响应的快速预测。然后，将CNN模型与白鲸优化（Beluga Whale Optimization， BWO）算法相结合，对非均匀加劲肋布局进行全局优化搜索。优化结果表明，与传统的均匀分布设计相比，该方法可将临界屈曲载荷分别提高9.74%、20.75%和11.19%，从而提高了结构的承载效率。本研究引入了一种新的、图像驱动的非均匀加筋优化范式，该范式将深度学习与进化算法相结合，以智能地设计复杂的加筋结构。该方法为减轻局部轴压薄壁结构的自重提供了新的视角，对提高设计精度和效率具有实用价值。Influences of internal and external factors on drop impact load characteristics in load generators based on thin conical shellKe Lin, Shuai Wang, Jicheng Li, Ming Qu, Yixia Yan, Gang Chen, Qingping Zhang, Liang Xiangdoi:10.1016/j.tws.2025.114469内外因素对薄锥壳负载发生器跌落冲击载荷特性的影响To study the deformation and failure modes of thin conical shells during drop impact and the evolution of acceleration of key component mounted on its upper end, the finite element simulations on the drop impact of load generator based on thin conical shell are systematically conducted in the present manuscript, in which various thin conical shells with different structural parameters and various drop impact attitudes are all considered, to investigate the deformation and failure morphologies of thin conical shells as well as the resulting changes in the acceleration of key component. The effects of internal and external factors, including shell thickness, half angle, drop impact attitude, and the mechanical properties of shell material, on the deformation and failure modes of thin conical shell and the acceleration history of key component are discussed in detail. Related results indicate that under the drop impact, increasing the thickness of conical shell enhances its bending rigidity, thereby induces increase in the acceleration amplitude of key component while decrease in the acceleration pulse. The half angle of a thin conical shell mainly dominates the deformation and failure morphologies at the bottom end of conical shell; a larger half angle will increase the number of cracks at the shell tip, causing the gradual transition of bottom failure from a triangular failure hole to an approximate circular hole. An oblique drop will reduce the degree of elastic deformation of support frame in the vertical direction, which suppresses the secondary rise in the key component acceleration, and can induce certain decrease in its acceleration vibrations, thereby increasing the acceleration pulse. Under the same drop condition, increasing the strength of thin conical shell material significantly increases the acceleration amplitude and shortens the pulse duration, while changing the material of components mounted on its upper end has only a little effect on the acceleration history of key component. In addition, it is practicable to increase the total mass of load generator by adding the extra mass blocks. This work provides theoretical guidance for designing the acceleration environments of key components through the drop impact of load generators based on thin conical shells.为研究薄锥壳在跌落冲击过程中的变形破坏模式及其上端关键部件加速度演化规律，本文系统地对基于薄锥壳的载荷发生器进行了跌落冲击有限元模拟，同时考虑了不同结构参数的各种薄锥壳以及不同的跌落冲击姿态。研究锥形薄壳的变形和破坏形态，以及由此导致的关键部件加速度变化。详细讨论了壳厚、半角、落点冲击姿态、壳材料力学性能等内外因素对薄锥形壳变形破坏模式及关键部件加速度历史的影响。相关结果表明，在跌落冲击下，增加锥形壳体的厚度可以提高其抗弯刚度，从而导致关键部件的加速度幅值增大，而加速度脉冲减小。薄锥形壳的半角主要支配着锥形壳底端的变形破坏形态；更大的半角会增加壳体尖端的裂纹数量，使底部破坏从三角形破坏孔逐渐过渡到近似圆形破坏孔。斜降会减小支撑架在垂直方向上的弹性变形程度，从而抑制关键部件加速度的二次上升，使其加速度振动产生一定的减小，从而增大加速度脉冲。在相同下降条件下，增加薄锥壳材料的强度可以显著提高加速度幅值，缩短脉冲持续时间，而改变其上端安装部件的材料对关键部件的加速度历史影响很小。此外，通过增加附加质量块来增加负荷发电机的总质量是可行的。该工作为利用薄锥壳负载发生器的跌落冲击设计关键部件的加速度环境提供了理论指导。Prediction of structural behavior of cold-formed steel C-section columns via repetitive backpropagation neural networkXi Zhao, Puxin Zhang, Xuefeng Wang, Meini Su, Xinxia Lidoi:10.1016/j.tws.2025.114470基于重复反向传播神经网络的冷弯型钢c型钢柱结构性能预测Compared to conventional numerical simulations, machine-learning-based surrogate models have demonstrated strong capability in capturing nonlinear behaviours and predicting capacities with high efficiency and accuracy. In this study, an enhanced supervised backpropagation neural network methodology is developed to construct a surrogate model for predicting the load-displacement curves of cold-formed steel (CFS) C-section columns. The machine learning framework integrates physics-based function with a data-driven model, which enables accurate prediction of the implicit curve equations with use of s mall-scale datasets and compact model architectures. First, a parametric representation method, based on an explicit-implicit hybrid function structure, is introduced to characterise the multi-stage behavior of load-displacement curves. Subsequently, a novel repetitive backpropagation neural network (ReBpNN) is developed to establish a surrogate model. Finally, the surrogate model is trained and validated through dataset generated from geometries of CFS members. Comparative an alyses are conducted among ReBpNN and other machine learning methods. Results show that the proposed ReBpNN is feasible for load-displacement curve prediction of the CFS members considering both accuracy and efficiency. The ReBpNN significantly outperforms traditional simulations in computational efficiency, demonstrating strong potential for industrial applications.与传统的数值模拟相比，基于机器学习的代理模型在捕获非线性行为和高效准确的预测能力方面表现出了强大的能力。在这项研究中，开发了一种增强的监督反向传播神经网络方法来构建预测冷弯型钢（CFS） c截面柱的荷载-位移曲线的代理模型。机器学习框架将基于物理的功能与数据驱动的模型集成在一起，可以使用小规模数据集和紧凑的模型架构准确预测隐式曲线方程。首先，引入了一种基于显式-隐式混合函数结构的参数化表示方法来表征荷载-位移曲线的多阶段特性。随后，建立了一种新的重复反向传播神经网络（ReBpNN）来建立代理模型。最后，通过CFS构件几何图形生成的数据集对代理模型进行训练和验证。对ReBpNN和其他机器学习方法进行了对比分析。结果表明，从精度和效率两方面考虑，所提出的ReBpNN对CFS构件的荷载-位移曲线预测是可行的。ReBpNN在计算效率上明显优于传统模拟，显示出强大的工业应用潜力。来源：复合材料力学仿真Composites FEM

【新文速递】2026年4月3日复合材料SCI期刊最新文章

Composite Structures

Composites Part A: Applied Science and Manufacturing

Composites Part B: Engineering

Composites Science and Technology

【新文速递】2026年1月2日固体力学SCI期刊最新文章