and multi objective optimization

 

 

ZHOU Ding zhi, WANG Xiao, LIU Hui xia, SU Zhan long, GAO Chuan yu

 

(School of Mechanical Engineering, Jiangsu University, Zhenjiang 212013, China)

 

 

Abstract: Aiming at the problem that research on lightweight of the roof structure at home and abroad are mostly single objective optimization and do not consider the influence of uncertainty factors, a method of roof structure lightweight design based on multi objective optimization and 6σ robustness design was presented. The sensitivity of design variable was analyzed through parameter study and the thickness of structural parts which contribute more to the response of the roof structure were selected as a design variable. Optimal Latin hypercube design and response surface methodology were applied to establish approximate models of each response of the roof. Total mass of design variables, strength of roof snow and dent resistance as the optimization objectives, the first three modal frequencies as constraint, based on approximate models, deterministic optimal solution was obtained by using the Non dominated sorting genetic algorithm Ⅱ (NSGAⅡ). With the using of monte carlo simulation techniques, 6σ quality of deterministic multi objective optimization results of roof structure was analyzed and 6σ robustness multi objective optimization of roof structure was conducted. The optimization results indicate that under the situation of meeting requirements of all performance, the mass of the total mass of design variables was reduced by 3.77kg and the strength of roof snow is increased. At the same time, the robustness of each working condition performance of the roof structure also was improved, reaching six sigma quality level.

 

Key words: roof structure; lightweight; 6σ robustness multi objective optimization; DOE; RSM