A new Computational Design of Experiment for Process Optimization Application for Bending Process

Abstract

The aim of this paper is to propose an original optimization approach using the concept of computational design of experiments and response surface methodology to check the curvature in numerical experiments. Generally, for the study of the strongly nonlinear phenomena, it is recommended to use a second-order model. This paper proposes a new technique to estimate a quadratic effect for a model based on finite element simulations and the Design of Experiments. Therefore, it is very important that the proposed technique (adding center points) controls other properties of the design matrix, especially orthogo¬nality for the response surface methodology and brings more information for the quality of the proposed model. The computational design of experiments is considered an advantage in comparison with design of physical experiments because generally computer solutions cost less than physical testing. The bending process optimization is used to improve the performances of this approach.

Keywords

• -