DocumentCode :
441921
Title :
Wavelet-based interpolation algorithm for topology extraction of compliant mechanisms
Author :
Fu, Yong-Qing ; Zhang, Xian-Min
Author_Institution :
Coll. of Mech. Eng., South China Univ. of Technol., Guangzhou, China
Volume :
5
fYear :
2005
fDate :
18-21 Aug. 2005
Firstpage :
2802
Abstract :
An important issue is to extract a practicable topology from the solution to the topology optimization design of a compliant mechanism. However, there exist checkerboard patterns in most topology results, which makes topology extraction very difficult. In this investigation, a wavelet-based interpolation algorithm, which directly works with the optimization result to extract a checkerboard-free topology, is proposed. Firstly, auxiliary variables, which are mapped from topology variables, are decomposed by an orthogonal wavelet into approximation and detail coefficients. Secondly, the approximation coefficients are served as reference data for a cubic spline interpolation. The approximation coefficients are reconstructed after the interpolation. In order to evaluate the goodness of the algorithm proposed and determine an optimal resolution level selected in the wavelet decomposition, a method based on a volume constraint of the topology optimization is considered. Numerical simulation is carried out on a compliant force inverting mechanism. Simulation results show that the algorithm proposed is an effective remedy for checkerboard patterns in topology solutions and as a result a desired topology is extracted.
Keywords :
interpolation; mechanical engineering computing; optimisation; splines (mathematics); topology; wavelet transforms; checkerboard pattern; compliant force inverting mechanism; compliant mechanism; cubic spline interpolation; numerical simulation; orthogonal wavelet; topology extraction; topology optimization design; volume constraint; wavelet decomposition; wavelet-based interpolation; Constraint optimization; Data mining; Design optimization; Educational institutions; Filters; Interpolation; Machine learning algorithms; Numerical simulation; Spline; Topology; Compliant mechanisms; cubic spline interpolation; topology extraction; wavelet;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Machine Learning and Cybernetics, 2005. Proceedings of 2005 International Conference on
Conference_Location :
Guangzhou, China
Print_ISBN :
0-7803-9091-1
Type :
conf
DOI :
10.1109/ICMLC.2005.1527419
Filename :
1527419
Link To Document :
بازگشت