Data-Driven Topology Optimization With Multiclass Microstructures Using Latent Variable Gaussian Process

被引:55
作者
Wang, Liwei [1 ,2 ]
Tao, Siyu [2 ]
Zhu, Ping [1 ]
Chen, Wei [2 ]
机构
[1] Shanghai Jiao Tong Univ, State Key Lab Mech Syst & Vibrat, Shanghai Key Lab Digital Manufacture Thin Walled, Sch Mech Engn, 800 Dongchuan Rd, Shanghai 200240, Peoples R China
[2] Northwestern Univ, Dept Mech Engn, 2145 Sheridan Rd, Evanston, IL 60208 USA
基金
美国国家科学基金会;
关键词
multiscale topology optimization; Gaussian process; mixed variables; multiclass; data-driven design; CELLULAR STRUCTURES; INVERSE REGRESSION; DESIGN; COMPOSITES; MODEL;
D O I
10.1115/1.4048628
中图分类号
TH [机械、仪表工业];
学科分类号
0802 ;
摘要
The data-driven approach is emerging as a promising method for the topological design of multiscale structures with greater efficiency. However, existing data-driven methods mostly focus on a single class of microstructures without considering multiple classes to accommodate spatially varying desired properties. The key challenge is the lack of an inherent ordering or "distance" measure between different classes of microstructures in meeting a range of properties. To overcome this hurdle, we extend the newly developed latent-variable Gaussian process (LVGP) models to create multi-response LVGP (MR-LVGP) models for the microstructure libraries of metamaterials, taking both qualitative microstructure concepts and quantitative microstructure design variables as mixed-variable inputs. The MR-LVGP model embeds the mixed variables into a continuous design space based on their collective effects on the responses, providing substantial insights into the interplay between different geometrical classes and material parameters of microstructures. With this model, we can easily obtain a continuous and differentiable transition between different microstructure concepts that can render gradient information for multiscale topology optimization. We demonstrate its benefits through multiscale topology optimization with aperiodic microstructures. Design examples reveal that considering multiclass microstructures can lead to improved performance due to the consistent load-transfer paths for micro- and macro-structures.
引用
收藏
页数:13
相关论文
共 54 条
  • [1] GrandPrix: scaling up the Bayesian GPLVM for single-cell data
    Ahmed, Sumon
    Rattray, Magnus
    Boukouvalas, Alexis
    [J]. BIOINFORMATICS, 2019, 35 (01) : 47 - 54
  • [2] Constrained mixed-integer Gaussian mixture Bayesian optimization and its applications in designing fractal and auxetic metamaterials
    Anh Tran
    Minh Tran
    Wang, Yan
    [J]. STRUCTURAL AND MULTIDISCIPLINARY OPTIMIZATION, 2019, 59 (06) : 2131 - 2154
  • [3] Ba J., 2014, INT C LEARN REPR 201
  • [4] Barber D., 2012, BAYESIAN REASONING M, P253
  • [5] Model-based methods for continuous and discrete global optimization
    Bartz-Beielstein, Thomas
    Zaefferer, Martin
    [J]. APPLIED SOFT COMPUTING, 2017, 55 : 154 - 167
  • [6] Globally Approximate Gaussian Processes for Big Data With Application to Data-Driven Metamaterials Design
    Bostanabad, Ramin
    Chan, Yu-Chin
    Wang, Liwei
    Zhu, Ping
    Chen, Wei
    [J]. JOURNAL OF MECHANICAL DESIGN, 2019, 141 (11)
  • [7] Concurrent topology design of structure and material using a two-scale topology optimization
    Chen, Wenjiong
    Tong, Liyong
    Liu, Shutian
    [J]. COMPUTERS & STRUCTURES, 2017, 178 : 119 - 128
  • [8] Functionally graded lattice structure topology optimization for the design of additive manufactured components with stress constraints
    Cheng, Lin
    Bai, Jiaxi
    To, Albert C.
    [J]. COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING, 2019, 344 : 334 - 359
  • [9] Gaussian process emulation of dynamic computer codes
    Conti, S.
    Gosling, J. P.
    Oakley, J. E.
    O'Hagan, A.
    [J]. BIOMETRIKA, 2009, 96 (03) : 663 - 676
  • [10] Sufficient dimension reduction via inverse regression: A minimum discrepancy approach
    Cook, RD
    Ni, LQ
    [J]. JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 2005, 100 (470) : 410 - 428