关键词: Kriging插值, 多项式回归响应曲面法, 拉丁方试验设计, 冲压成形, 优化设计

Abstract:

A sample data design was carried out by adopting Latin square experimental design method, moreover, in order to improve computing efficiency, Kriging surrogate model and polynomial regression response surface methodology were introduced into the complex system of optimization design of automotive panel sheet forming, and was employed for producing coarsen response surface. And then, the Kriging interpolation method was used for constructing the better robust surface to decide the better initial values of design variables and reduce constraints in considerable range. Finally, genetic algorithm was applied for finding the best optimization of the initial values of the design variables and the range of constraints updated. The optimization results show that is able to quantify the warpage of sheet metal forming and crack in the trends and characteristics accurately, the comparison between sequential linear programming and Kriging-based optimization method is done, and Kriging model has better predict ability than the polynomial regression model based on the classic fractional factor design, with the accuracy of the interest region and the points with larger predicted uncertainty for the accuracy and efficiency of the global approximation.

Key words: Kriging interpolation, polynomial regression response surface methodology, Latin square experimental design, sheet forming, optimization design