A MONTE CARLO ANALYSIS FOR THE INFLUENCE OF RANDOM LOCAL CURVATURE OF BLANK ON ROLL BENDING PRODUCT

The influence of local random curvatures of the work-piece on the roll bending result is studied by Monte Carlo finite element simulations. To improve the simulation efficiency, a finite element scheme based on Eulerian grid and classical beam element is proposed, whose correctness is verified against the conventional finite element model and the theoretical solution. Based on the proposed scheme, the roll bending process of an ultra-long work-piece, whose initial curvature is normally distributed with zero mean, is simulated. The results show that the distribution of output curvature radius approximately follows the normal distribution. With the increase of the variance of the initial curvature, the mean value of the output radius decreases, while the standard deviation increases. As a result, the macro radius of the product decreases. The influence of initial curvature is even greater if the goal curvature radius or the representative elementary length gets larger. However, the roller positions do not significantly influence the output distribution for a given goal curvature.