Yexin Zhou

Prof. Dr. Yexin Zhou


Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
13 November 2020 in Journal of Mechanisms and Robotics
Verified
Journal Article
Closed access
PDF file
YHYucai Hu
Yucai Hu
Yexin ZhouYexin Zhou
Yexin Zhou
HLHaiyi Liang
Haiyi Liang
Origami has shown the potential to approximate three-dimensional curved surfaces by folding through designed crease patterns on flat materials. The Miura-ori tessellation is a widely used pattern in engineering and tiles the plane when partially folded. Based on constrained optimization, this article presents the construction of generalized Miura-ori patterns that can approximate three-dimensional parametric surfaces of varying curvatures while preserving the inherent properties of the standard Miura-ori, including developability, flat foldability, and rigid foldability. An initial configuration is constructed by tiling the target surface with triangulated Miura-like unit cells and used as the initial guess for the optimization. For approximation of a single target surface, a portion of the vertexes on the one side is attached to the target surface; for fitting of two target surfaces, a portion of vertexes on the other side is also attached to the second target surface. The parametric coordinates are adopted as the unknown variables for the vertexes on the target surfaces, while the Cartesian coordinates are the unknowns for the other vertexes. The constructed generalized Miura-ori tessellations can be rigidly folded from the flat state to the target state with a single degree-of-freedom.

30 April 2024 in Frontiers in Energy
Verified
Journal Article
Closed access
PDF file
SKShimeng Kang
Shimeng Kang
ZPZehua Pan
Zehua Pan
JGJinjie Guo
Jinjie Guo
Yexin ZhouYexin Zhou
Yexin Zhou
JWJingyi Wang
Jingyi Wang
ZZZheng Zhong
Zheng Zhong

1 January 2024 in Frontiers in Energy
Verified
Journal Article
Closed access
PDF file
CYChengrong Yu
Chengrong Yu
ZPZehua Pan
Zehua Pan
HZHongying Zhang
Hongying Zhang
BCBin Chen
Bin Chen
WGWanbing Guan
Wanbing Guan
Yexin ZhouYexin Zhou
Yexin Zhou

9 August 2023 in Engineering Computations
Verified
Journal Article
Closed access
ZGZiyan Guo
Ziyan Guo
XLXuhao Liu
Xuhao Liu
ZPZehua Pan
Zehua Pan
Yexin ZhouYexin Zhou
Yexin Zhou
ZZZheng Zhong
Zheng Zhong
Zilin YanZilin Yan
Zilin Yan
Purpose: In recent years, the convolutional neural network (CNN) based deep learning approach has succeeded in data-mining the relationship between microstructures and macroscopic properties of materials. However, such CNN models usually rely heavily on a large set of labeled images to ensure the accuracy and generalization ability of the predictive models. Unfortunately, in many fields, acquiring image data is expensive and inconvenient. This study aims to propose a data augmentation technique to enhance the performance of the CNN models for linking microstructural images to the macroscopic properties of composites.Design/methodology/approach: Microstructures of composites are synthesized using discrete element simulations and Potts kinetic Monte Carlo simulations. Macroscopic properties such as the elastic modulus, Poisson's ratio, shear modulus, coefficient of thermal expansion, and triple-phase boundary length density are extracted on representative volume elements. The CNN model is trained using the 3D microstructural images as inputs and corresponding macroscopic properties as the labels. The comparison of the predictive performance of the CNN models with and without data augmentation treatment are compared.Findings: The comparison between the prediction performance of CNN models with and without data augmentation showed that the former reduced the weighted mean absolute percentage error (WMAPE) for the prediction from 5.1627% to 1.7014%. This significant reduction signifies that the proposed data augmentation method can effectively enhance the generalization ability and robustness of CNN models.Originality/value: This study demonstrates that data augmentation is beneficial for solving the problems of model overfitting, data scarcity, and sample imbalance for CNN-based deep learning tasks at a low cost. By developing more and advanced data augmentation techniques, deep learning accelerated homogenization will boost the multi-scale computational mechanics and materials.

8 October 2025 in Processes
Verified
Journal Article
Open access
PDF file
CCChaolong Cheng
Chaolong Cheng
WDWen Ding
Wen Ding
JSJunfeng Shen
Junfeng Shen
PLPenghui Liao
Penghui Liao
CYChengrong Yu
Chengrong Yu
ZZZheng Zhong
Zheng Zhong
Solid oxide electrolysis cells (SOECs) have emerged as a promising technology for efficient energy storage and CO2 utilization via H2O–CO2 co-electrolysis. While most previous studies focused on planar or tubular configurations, this work investigated a novel flat, tubular SOEC design using a comprehensive 3D multi-physics model developed in COMSOL Multiphysics 5.6. This model integrates charge transfer, gas flow, heat transfer, chemical/electrochemical reactions, and structural mechanics to analyze operational behavior and thermo-mechanical stress under different voltages and pressures. Simulation results indicate that increasing operating voltage leads to significant temperature and current density inhomogeneity. Furthermore, elevated pressure improves electrochemical performance, possibly due to increased reactant concentrations and reduced mass transfer limitations; however, it also increases temperature gradients and the maximum first principal stress. These findings underscore that the design and optimization of flat tubular SOECs in H2O–CO2 co-electrolysis should take the trade-off between performance and durability into consideration.

23 July 2021 in Engineering Computations
Verified
Journal Article
Closed access
BQBao Qin
Bao Qin
Yexin ZhouYexin Zhou
Yexin Zhou
ZZZheng Zhong
Zheng Zhong
Purpose: A diffusion-reaction-deformation coupled model is employed and implemented as a user-defined element (UEL) subroutine in the commercial finite element software package ABAQUS.Design/methodology/approach: Chemical reaction and diffusion are treated as two distinct processes by introducing the extent of reaction and the diffusion concentration as two kinds of independent variables, for which the independent governing equations for chemical reaction and diffusion processes are obtained. Furthermore, an exponential form of chemical kinetics, instead of the linearly phenomenological relation, between the reaction rate and the chemical affinity is used to describe reaction process. As a result, complex chemical reaction can be simulated, no matter it is around or away from equilibrium.Findings: Two numerical examples are presented, one for validation of the model and another for the modeling of the deflection of a plane caused by a chemical reaction.Originality/value: 1. Independent governing equations for diffusion and reaction processes are given. 2. An exponential relation between the reaction rate and its driving force is employed. 3. The UEL subroutine is used to implement the finite element procedure.