TY - GEN
T1 - Data-driven multiscale topology optimization using multi-response latent variable gaussian process
AU - Wang, Liwei
AU - Tao, Siyu
AU - Zhu, Ping
AU - Chen, Wei
N1 - Funding Information:
The authors are grateful to Prof. K. Svanberg, from the Royal Institute of Technology, Sweden, for providing a copy of the MMA code for metamaterial designs. Support from the National Science Foundation (NSF) (Grant No. OAC 1835782) is greatly appreciated. Mr. Liwei Wang would like to acknowledge the support from the Zhiyuan Honors Program for Graduate Students of Shanghai Jiao Tong University.
Publisher Copyright:
© 2020 American Society of Mechanical Engineers (ASME). All rights reserved.
PY - 2020
Y1 - 2020
N2 - The data-driven approach is emerging as a promising method for the topological design of the multiscale structure with greater efficiency. However, existing data-driven methods mostly focus on a single class of unit cells without considering multiple classes to accommodate spatially varying desired properties. The key challenge is the lack of inherent ordering or "distance" measure between different classes of unit cells in meeting a range of properties. To overcome this hurdle, we extend the newly developed latent-variable Gaussian process (LVGP) to creating multi-response LVGP (MRLVGP) for the unit cell libraries of metamaterials, taking both qualitative unit cell concepts and quantitative unit cell design variables as mixed-variable inputs. The MRLVGP embeds the mixed variables into a continuous design space based on their collective effect on the responses, providing substantial insights into the interplay between different geometrical classes and unit cell materials. With this model, we can easily obtain a continuous and differentiable transition between different unit cell concepts that can render gradient information for multiscale topology optimization. While the proposed approach has a broader impact on the concurrent topological and material design of engineered systems, we demonstrate its benefits through multiscale topology optimization with aperiodic unit cells. Design examples reveal that considering multiple unit cell types can lead to improved performance due to the consistent load-transferred paths for micro- and macrostructures.
AB - The data-driven approach is emerging as a promising method for the topological design of the multiscale structure with greater efficiency. However, existing data-driven methods mostly focus on a single class of unit cells without considering multiple classes to accommodate spatially varying desired properties. The key challenge is the lack of inherent ordering or "distance" measure between different classes of unit cells in meeting a range of properties. To overcome this hurdle, we extend the newly developed latent-variable Gaussian process (LVGP) to creating multi-response LVGP (MRLVGP) for the unit cell libraries of metamaterials, taking both qualitative unit cell concepts and quantitative unit cell design variables as mixed-variable inputs. The MRLVGP embeds the mixed variables into a continuous design space based on their collective effect on the responses, providing substantial insights into the interplay between different geometrical classes and unit cell materials. With this model, we can easily obtain a continuous and differentiable transition between different unit cell concepts that can render gradient information for multiscale topology optimization. While the proposed approach has a broader impact on the concurrent topological and material design of engineered systems, we demonstrate its benefits through multiscale topology optimization with aperiodic unit cells. Design examples reveal that considering multiple unit cell types can lead to improved performance due to the consistent load-transferred paths for micro- and macrostructures.
KW - Data-driven design
KW - Gaussian process
KW - Mixed variables
KW - Multi-class
KW - Multiscale topology optimization
UR - http://www.scopus.com/inward/record.url?scp=85096331792&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85096331792&partnerID=8YFLogxK
U2 - 10.1115/DETC2020-22595
DO - 10.1115/DETC2020-22595
M3 - Conference contribution
AN - SCOPUS:85096331792
T3 - Proceedings of the ASME Design Engineering Technical Conference
BT - 46th Design Automation Conference (DAC)
PB - American Society of Mechanical Engineers (ASME)
T2 - ASME 2020 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, IDETC-CIE 2020
Y2 - 17 August 2020 through 19 August 2020
ER -