یک مدل پیش بینی کننده فرآیند تشکیل رول انعطاف پذیر و قابل تنظیم با استفاده از تحلیل رگرسیون

Flexibly reconfigurable roll forming (FRRF) is a new and flexible forming process that can be used to produce a multi-curvature surface. The basic concept of FRRF is the formation of a curvature in the longitudinal direction by controlling the strain distribution. By the arrangement of curvature adjustment rods, various curvatures can be produced by means of rollers. The shape of the formed surface is determined by the curvature of the reconfigurable rollers and the gaps between the rollers. However, it is difficult to intuitively predict the longitudinal curvature of the formed sheet because the FRRF process produces a three-dimensional curved surface from a two-dimensional curve. To fabricate the objective surface, it is necessary to determine the appropriate input parameters. This can be rapidly done by developing a predictive model using regression analysis. The input parameters selected for the regression analysis in this study are the compression ratio of the formed sheet and curvature radius in the transverse direction. The dependent variable is the longitudinal curvature of the formed sheet. The curvatures are obtained by experiments using FRRF apparatus, and the predictions are used for the regression analysis. To verify the predictive regression model, r-squared values and root mean square errors are calculated. Through the employed procedure, it is confirmed that a statistical formula for predictive model is reasonable.