Development of hybrid fuzzy regression-based metamodeling technique for optimization of sheet metal forming problems

In this work, a hybrid fuzzy regression-based metamodeling technique is proposed to optimize the sheet metal forming design. Compared with other metamodeling techniques, the distinctive character of the proposed approach is to consider uncertainties and certainties in one metamodel. In order to further improve the efficiency of optimization, an intelligent scheme is incorporated to generate samples at the stage of design of experiment. Additionally, one step and incremental FE forming methods are implemented to calculate values of response functions. Finally, the proposed metamodeling technique and sampling schemes are integrated and applied to optimize sheet forming design; the corresponding optimized results demonstrate that the proposed hybrid metamodeling technique is able to produce good approximation model of sheet metal forming system. (C) 2009 Elsevier Ltd. All rights reserved.