@Article{icces.2021.08311,
AUTHOR = {Yuan Zhu, Xin Ning, Yao Zhang, Yuwan Yin},
TITLE = {Multiscale Topology Optimization using Subspace-based Model  Reduction Method},
JOURNAL = {The International Conference on Computational \& Experimental Engineering and Sciences},
VOLUME = {23},
YEAR = {2021},
NUMBER = {1},
PAGES = {11--12},
URL = {http://www.techscience.com/icces/v23n1/42026},
ISSN = {1933-2815},
ABSTRACT = {High performance of the spacecraft structure is required in 
the special environment, it includes mechanical performance and 
operational performance, etc. When performing tasks, the 
spaceborne equipment requires high precision. Therefore, the design 
of lightweight, high stability and high reliability structure is essential 
for spacecraft. Topology optimization is widely used in structural 
design. However, there are some problems in the structure after 
macro topology optimization, such as checkerboard, local optimal 
solution and other phenomena. Despite a long calculation period, the 
obtained structure is often not smooth enough and hard to 
manufacture. Aiming to this issue, this paper proposes a combined 
method of multiscale topology optimization method and multisubstructure multi-frequency quasi-static Ritz vector (MMQSRV) 
method. Firstly, a shape interpolation technology is used to generate 
microstructures. Those microstructures are predicted the effective 
characteristics by Kriging metamodel, and then they are used to build 
the macrostructure. Furthermore, variable thickness sheet (VTS) 
method is used to get the structure of free distribution of materials. 
Due to the similar topological characteristics, the interfaces of 
microstructures are well connected. In addition, on the macro scale, 
based on the Ritz vector method, the MMQSRV method simplifies 
the model matrix and effectively guarantees the computational 
efficiency and accuracy by Krylov Subspace Method. This method 
optimizes the microstructure and macrostructure, and curtail the the 
iteration period of topology optimization.},
DOI = {10.32604/icces.2021.08311}
}