Article
Engineering, Civil
Cheng Xiang, Airong Chen, Dalei Wang
Summary: In this paper, an iteration-free method based on convolutional neural network (CNN) is proposed to increase the computational efficiency of stress-based topology optimization (SBTO). With the adoption of p-norm stress aggregation scheme and the method of moving asymptotes (MMA), a dataset is generated to train a deep CNN based on U-Net architecture to find the optimal input mode for SBTO problem. The results show that the proposed method can realize near optimal 3D SBTO prediction using negligible calculation cost.
THIN-WALLED STRUCTURES
(2022)
Article
Computer Science, Interdisciplinary Applications
Shangjun Shi, Pingzhang Zhou, Zhenhua Lu
Summary: This article presents a novel topological design approach inspired by the density method and parametric level set method to control structural complexity and improve computational efficiency. By combining radial basis function and the SIMP formula, the distribution of fictitious density field in the design domain is described to control structural complexity. The proposed method can naturally avoid checkerboard design and reduce the number of design variables by redefining support points.
STRUCTURAL AND MULTIDISCIPLINARY OPTIMIZATION
(2021)
Article
Computer Science, Artificial Intelligence
Bin Wang, Bing Xue, Mengjie Zhang
Summary: A new effective and efficient surrogate-assisted particle swarm optimization (PSO) algorithm is proposed to automatically evolve CNNs by integrating a novel surrogate model, a new method of creating a surrogate data set, and a new encoding strategy. The proposed method achieves competitive error rates on CIFAR-10, CIFAR-100, and SVHN datasets, efficiently learns CNN blocks, and demonstrates transferability of the evolved blocks to other datasets such as CIFAR-100, SVHN, and ImageNet.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Computer Science, Interdisciplinary Applications
Jun Yan, Dongling Geng, Qi Xu, Haijiang Li
Summary: This paper proposes a step-to-step training method to improve the prediction accuracy of a deep learning model for real-time structural topology optimization. By increasing the utilization of optimization history information, the method improves the efficiency of model utilization without increasing the sample set size.
ENGINEERING WITH COMPUTERS
(2023)
Article
Computer Science, Artificial Intelligence
Sergey N. Pozdnyakov, Michele Ceriotti
Summary: Graph neural networks (GNN) are popular in machine learning and have been successful in predicting properties of molecules and materials. However, first-order GNNs are known to be incomplete, leading to the design of more complex schemes. The construction of graph representations for molecules adds a geometric dimension, with the most common approach being to consider atoms as vertices and connect them with bonds. This approach, known as distance graph NNs (dGNN), has shown excellent resolving power in chemical ML. However, the authors present a counterexample that proves dGNNs are not complete even for fully-connected graphs induced by 3D atom clouds.
MACHINE LEARNING-SCIENCE AND TECHNOLOGY
(2022)
Article
Computer Science, Artificial Intelligence
Philipp Christian Petersen, Anna Sepliarskaia
Summary: This study investigates the generalization capacity of group convolutional neural networks and provides precise estimates of VC dimensions for certain simple sets of such networks. It is discovered that even with infinite groups and suitable convolutional kernels, two-parameter families of convolutional neural networks can have infinite VC dimensions, while remaining invariant to the action of an infinite group.
Article
Computer Science, Interdisciplinary Applications
Chungang Zhuang, Zhenhua Xiong, Han Ding
Summary: This paper presents a computationally efficient MATLAB implementation of 2D/3D topology optimization in non-uniform rational basis spline (NURBS) framework. By utilizing page-wise matrix operations and graphics processing unit (GPU), the computational efficiency is greatly improved. The advantages of page-wise matrix operations are demonstrated through performance evaluation of benchmark cases.
STRUCTURAL AND MULTIDISCIPLINARY OPTIMIZATION
(2023)
Article
Computer Science, Interdisciplinary Applications
Rubens Bohrer, Il Yong Kim
Summary: This paper introduces an improved multi-material topology optimization method that avoids the element stacking process and considers a mixture of isotropic and anisotropic materials in commercial finite element engines. By enhancing the element duplication method, it improves numerical efficiency and serves as an alternative for computing sensitivities in discrete material optimization schemes.
STRUCTURAL AND MULTIDISCIPLINARY OPTIMIZATION
(2021)
Article
Environmental Sciences
Xi Wang, Tingfa Xu, Yuhan Zhang, Axin Fan, Chang Xu, Jianan Li
Summary: This paper proposes a method to reconstruct compressed hyperspectral (HS) data using convolutional neural networks (CNNs). By dividing the imaging process into multiple steps and building a subnetwork for each step, this method can reconstruct compressed HS data quickly and accurately, while having superior resistance to noise.
Article
Engineering, Multidisciplinary
Oded Amir
Summary: An efficient computational approach for stress-constrained topology optimization is proposed, which significantly reduces the number of Krylov iterations by enforcing early termination of iterative solves. Results show a savings of 80% in Krylov iterations and very good agreement with accurate direct solver results.
INTERNATIONAL JOURNAL FOR NUMERICAL METHODS IN ENGINEERING
(2021)
Article
Engineering, Multidisciplinary
Martin Ohrt Elingaard, Niels Aage, Jakob Andreas Baerentzen, Ole Sigmund
Summary: This paper presents a deep learning-based de-homogenization method for structural compliance minimization, showing excellent generalization properties and performance within 7-25% of homogenization-based solutions at a fraction of the computational cost, while being robust and insensitive to domain size, boundary conditions, and loading.
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING
(2022)
Article
Engineering, Electrical & Electronic
Taiga Aoyagi, Yoshitsugu Otomo, Hajime Igarashi, Hidenori Sasaki, Yuki Hidaka, Hideaki Arita
Summary: In this study, a fast topology optimization method based on a deep neural network is proposed to predict the current-dependent motor torque characteristics using the cross-sectional image. The trained DNN can provide the current condition that maximizes the torque under the assumed motor control method. The proposed method reduces the number of field computations while maintaining a high search capability.
IEEE TRANSACTIONS ON MAGNETICS
(2022)
Article
Engineering, Civil
Dongling Geng, Jun Yan, Qi Xu, Qi Zhang, Mengfang Zhou, Zhirui Fan, Haijiang Li
Summary: This paper presents a real-time topology optimization algorithm based on the Moving Morphable Component (MMC) method using a Convolutional Neural Network (CNN). The algorithm uses a new data pre-processing method to preserve the numerical characteristics and smoothness of the structure boundary, facilitating the CNN to capture data features with a limited sample set. The effectiveness of the algorithm has been verified with several examples.
ENGINEERING STRUCTURES
(2023)
Article
Environmental Sciences
Jinxiang Liu, Tiejun Wang, Andrew Skidmore, Yaqin Sun, Peng Jia, Kefei Zhang
Summary: In this study, a novel CNN framework that integrates 1D, 2D, and 3D CNNs was proposed to improve the land cover classification accuracy of hyperspectral images. By combining spatial and spectral features, the proposed method achieved high classification accuracy in two datasets while significantly reducing training time.
Article
Mathematics
Benjamin Vial, Yang Hao
Summary: Technological advances in nanofabrication have opened up new applications in nanophotonics, but rigorous and efficient numerical methods are needed. Tremendous advances in algorithmic differentiation have made large-scale optimization of devices possible with minor modifications to the underlying code. Researchers have developed three software libraries for solving Maxwell's equations in different contexts and demonstrate inverse design examples.
Article
Computer Science, Software Engineering
Shuai Zheng, Jun Hong, Kang Zhang, Baotong Li, Xin Li
COMPUTER-AIDED DESIGN
(2016)
Article
Automation & Control Systems
Kang Jia, Jun Hong, Shuai Zheng, Yinhang Zhang
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2017)
Article
Computer Science, Interdisciplinary Applications
Baotong Li, Honglei Liu, Shuai Zheng
STRUCTURAL AND MULTIDISCIPLINARY OPTIMIZATION
(2018)
Article
Automation & Control Systems
Kang Jia, Shuai Zheng, Junkang Guo, Jun Hong
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2019)
Article
Computer Science, Software Engineering
Baotong Li, Congjia Huang, Xin Li, Shuai Zheng, Jun Hong
COMPUTER-AIDED DESIGN
(2019)
Article
Engineering, Multidisciplinary
Kang Jia, Junkang Guo, Shuai Zheng, Jun Hong
APPLIED MATHEMATICAL MODELLING
(2019)
Article
Engineering, Civil
Shuai Zheng, Wenhao Tang, Baotong Li
THIN-WALLED STRUCTURES
(2020)
Article
Engineering, Biomedical
Zhang Chen, Zhiqiang Tian, Yaoyue Zheng, Xiangyu Si, Xulei Qin, Zhong Shi, Shuai Zheng
Summary: This study proposes a weakly supervised method based on prior knowledge for human organ segmentation, which outperforms several state-of-the-art methods in experimental results.
PHYSICS IN MEDICINE AND BIOLOGY
(2021)
Article
Computer Science, Software Engineering
Shuai Zheng, Yabin Wang, Baotong Li, Xin Li
Summary: This paper introduces a hardware-adaptive feature modeling framework for automatically generating and optimizing deep neural networks to support real-time feature extraction and matching. Experimental results demonstrate the effectiveness of the proposed pipelines on hardware platforms with different performance.
COMPUTER-AIDED DESIGN
(2021)
Article
Engineering, Civil
Honglei Liu, Ziyu Zhang, Baotong Li, Miaoxia Xie, Jun Hong, Shuai Zheng
Summary: The study establishes a topology optimization framework based on the energy finite element method, focusing on the topological flexibility of structures. Through basic research, an original EFEM-based topology optimization framework for thin-walled structures is established for the first time.
THIN-WALLED STRUCTURES
(2021)
Article
Engineering, Multidisciplinary
Shuai Zheng, Haojie Fan, Ziyu Zhang, Zhiqiang Tian, Kang Jia
Summary: This study introduces a real-time structural topology optimization method based on a convolutional neural network, replacing traditional iterative calculations with residual learning and attention mechanisms, significantly improving accuracy.
APPLIED MATHEMATICAL MODELLING
(2021)
Article
Engineering, Civil
Shuai Zheng, Fan Gao, Ziyu Zhang, Honglei Liu, Baotong Li
Summary: The paper introduces a topology optimization method for aircraft fuel tank structural design to reduce fuel sloshing effect, using a hybrid fluid-solid particle based simulator and an effective optimizer. The method was validated successfully on a real aircraft wing tank, showing improvements in fuel sloshing time, fuel gauge error, and center of gravity shift.
THIN-WALLED STRUCTURES
(2021)
Proceedings Paper
Computer Science, Theory & Methods
Kang Zhang, Shuai Zheng, Wuyi Yu, Xin Li
10TH INTERNATIONAL CONFERENCE ON COMPUTER SCIENCE & EDUCATION (ICCSE 2015)
(2015)
Article
Engineering, Civil
Jian Xue, Weiwei Zhang, Jing Wu, Chao Wang, Hongwei Ma
Summary: This study integrates a plate-type local resonator with varying free boundaries within a plate to convert the initial low-order global vibration modes into localized vibration modes. A novel semi-analytical method is proposed to analyze the free vibration of the plate with thickness and displacement discontinuities. The results show that by applying free boundary conditions, the low-order localized vibration frequencies can be significantly reduced without affecting the low-order global frequencies.
THIN-WALLED STRUCTURES
(2024)
Article
Engineering, Civil
Merve Tunay
Summary: In recent years, there has been an increasing number of studies on the mechanical properties of sandwich structures manufactured with the Fused Deposition Modeling (FDM) method. However, there is still a lack of experimental data on the mechanical characteristics of FDM-manufactured sandwich structures under different thermal aging durations. In this experiment, the energy absorption capabilities of sandwich structures with different core geometries were investigated under various thermal aging durations. The results showed that the core topology significantly influenced the energy absorption abilities of the sandwich structures.
THIN-WALLED STRUCTURES
(2024)
Article
Engineering, Civil
Zi-qin Jiang, Zi-yao Niu, Ai-Lin Zhang, Xue-chun Liu
Summary: This paper proposes a crosssection corrugated plate steel special-shaped column (CCSC) that improves the bearing capacity and overall stability of structural columns by using smaller material input. Through theoretical analysis and numerical simulation, the overall stability of the CCSC under axial compression is analyzed. The design method and suggestions for the stability of CCSC are put forward. Compared with conventional square steel tube columns, the CCSC has obvious advantages in overall stability and steel consumption.
THIN-WALLED STRUCTURES
(2024)
Article
Engineering, Civil
Yong Zhang, Yangang Chen, Jixiang Li, Jiacheng Wu, Liang Qian, Yuanqiang Tan, Kunyuan Li, Guoyao Zeng
Summary: A hybrid TPMS method was proposed to develop a new TPMS structure, and the mechanical properties of different TPMS structures were studied experimentally and numerically. Results showed that the hybrid TPMS structure had higher energy absorption and lower load-carrying capacity fluctuation. Further investigations revealed that the topological shape and material distribution had significant influence on mechanical properties, and the hybrid additive TPMS structure exhibited significant crashworthiness advantage in in-plane crushing condition.
THIN-WALLED STRUCTURES
(2024)
Article
Engineering, Civil
Tongfei Sun, Ye Liu, Kaoshan Dai, Alfredo Camara, Yujie Lu, Lijie Wang
Summary: This paper presents a series of experimental and numerical studies on the performance of a novel double-stage coupling damper (DSCD). The effects of damper configuration, friction-yield ratio (Rfy), and loading protocol on the hysteresis performance of the DSCD are investigated. The test results demonstrate that the arrangement of ribs in the DSCD increases its energy dissipation capacity. Numerical analysis reveals that the length of the friction mechanism and the clearance between the yield segment and the restraining system affect the energy dissipation and stability of the damper.
THIN-WALLED STRUCTURES
(2024)
Article
Engineering, Civil
Jeonghwa Lee, Young Jong Kang
Summary: This study investigates the local buckling behavior and strength of I-shape structural sections by considering flange-web interactions through three-dimensional finite element analysis. The study provides a more reasonable estimation of local buckling strength by considering the ratio of flange-web slenderness and height-to-width ratio, and presents design equations for flange local and web-bend buckling coefficients.
THIN-WALLED STRUCTURES
(2024)
Article
Engineering, Civil
Yizhe Chen, Wenfeng Xiang, Qingsong Zhang, Hui Wang, Lin Hua
Summary: This study investigates the surface modification of a nickel plate to improve the bonding strength with carbon fiber-reinforced plastics (CFRP). The results show that different surface modification methods, including sandblasting, coupling agent treatment, and compound coupling agent treatment, significantly enhance the bonding strength of CFRP/Ni joints. The research provides insights into improving the connection between nickel and CFRP, as well as other heterogeneous materials.
THIN-WALLED STRUCTURES
(2024)
Article
Engineering, Civil
Agha Intizar Mehdi, Fengping Zhang, Moon-Young Kim
Summary: A spatial stability theory of mono-symmetric thin-walled steel beams pre-stressed by spatially inclined cables is derived and its validity is demonstrated through numerical examples. The effects of initial tension, deviator numbers, inclined cable profiles, and bonded/un-bonded conditions on lateral-torsional buckling of the pre-stressed beams are investigated, with a specific emphasis on the effects of increasing initial tension.
THIN-WALLED STRUCTURES
(2024)
Article
Engineering, Civil
Teng Ma, Jinxiang Wang, Liangtao Liu, Heng Li, Kui Tang, Yangchen Gu, Yifan Zhang
Summary: The structural response of water-back plate under the combined action of shock wave and bubble loads at water depths of 1-300 m was numerically investigated using an arbitrary Lagrange-Euler method. The accuracy of the numerical model was validated by comparing with experimental and theoretical results. The influences of water depth and length-to-diameter ratio of the charge on the combined damage effect were analyzed. The results show that as water depth increases, the plastic deformation energy of the water-back plate decreases, and the permanent deformation mode changes from convex to concave. When the charge has a large length-to-diameter ratio, the plastic deformation energy of the radial plate is higher than that of the axial plate, and the difference decreases with increasing water depth. Increasing the length-to-diameter ratio enhances the combined damage effect in the radial direction in deep-water environments.
THIN-WALLED STRUCTURES
(2024)
Article
Engineering, Civil
Qiu-Yun Li, Ben Young
Summary: This paper investigates the flexural performance of CFS zed section members bent about the neutral axis parallel to the flanges through experimental and numerical analysis. The results show that the current direct strength method generally provides conservative predictions for the flexural strength of unstiffened zed section members, but slightly unconservative design for edge-stiffened zed section beams. The nominal flexural strengths of zed section members with edge stiffeners were found to be underestimated by 17% to 21% on average. Modified DSM formulae are recommended for the design of CFS zed section beams.
THIN-WALLED STRUCTURES
(2024)
Article
Engineering, Civil
Weinan Gao, Bo Song, Xueyan Chen, Guochang Lin, Huifeng Tan
Summary: This paper presents a precise method for predicting deformation in large-scale inflatable structures, utilizing finite element modeling and laser scanning technique. The study shows a good agreement between the predictive model and non-contact measurement results.
THIN-WALLED STRUCTURES
(2024)
Article
Engineering, Civil
Fei Gao, Zongyi Wang, Rui Zhu, Zhenming Chen, Quanxi Ye, Yaqi Duan, Yunlong Jia, Qin Zhang
Summary: This research investigates the mechanical properties of high-strength ring groove rivet assemblies and the load resistances of riveted T-stubs. Experimental tests reveal that Grade 10.9 rivets have higher yield strength and strain, and lower ultimate strain, making them suitable for high-strength ring groove rivet connections. Increasing the rivet diameter benefits the T-stubs, while increasing the flange thickness is not always advantageous. The Eurocode 3 method is not suitable for T-stubs connected through ring groove rivets, while the Demonceau method is conservative.
THIN-WALLED STRUCTURES
(2024)
Article
Engineering, Civil
Shangchun Jiang, Liangfeng Sun, Haifei Zhan, Zhuoqun Zheng, Xijian Peng, Chaofeng Lue
Summary: This study investigates the bending behavior of two-dimensional nanomaterials, diamane and its analogous structure TBGIB, through atomistic simulations. It reveals that diamane experiences structural failure under bending, while TBGIB bends elastically before undergoing structural failure. The study provides valuable insights for the application of these materials in flexible electronics.
THIN-WALLED STRUCTURES
(2024)
Article
Engineering, Civil
Qiang Zhang, Jianian Wen, Qiang Han, Hanqing Zhuge, Yulong Zhou
Summary: In this study, the mechanical properties of Q690 steel H-section columns under bi-directional cyclic loads are investigated, considering the time-varying characteristics of corrosion. A refined finite element (FE) model is built to analyze the degradation of mechanical property and failure mechanisms of steel columns with different design parameters during the whole life-cycle. The study proposes a quantitative calculation method for the ultimate resistance and damage index of steel columns, taking into account the ageing effects. The findings emphasize the importance of considering the ageing effects of steel columns in seismic design.
THIN-WALLED STRUCTURES
(2024)
Article
Engineering, Civil
Yuda Hu, Qi Zhou, Tao Yang
Summary: The magneto-thermo-elastic coupled free vibration of functionally graded materials cylindrical shell is investigated in this study. The vibration equation in multi-physical field is established and solved using the Hamilton principle and the multi-scale method. The numerical results show that the natural frequency is influenced by various factors such as volume fraction index, initial amplitude, temperature, and magnetic induction intensity.
THIN-WALLED STRUCTURES
(2024)