Development of a geometric modelling strategy for roll pass optimal design

Abstract

Roll pass design is one of the most important tasks in shape rolling operations that are employed to provide raw materials with appropriate cross-section profiles for various industrial applications. Currently, many approaches, such as experience-based trial-and-error strategies, finite element methods, and expert systems, are applied to improve both quality and efficiency of roll pass design. However, due to lack of a flexible geometrical modelling strategy, the application of extant approaches is largely limited. This study attempts to develop a novel approach for generic geometrical modelling to support optimal design of roll passes. Features of the proposed model are analysed to support its application. Furthermore, a parameters estimation approach based on genetic algorithm is also developed to facilitate the transformation between the generic model and other geometrical models, as well as to improve its flexibility and applicability. The results from the case study presented in the paper indicate that the new model is more flexible and efficient, and that the parameters estimation approach also can achieve high transformation accuracy and efficiency.