A novel multi-pass machining accuracy prediction method for thin-walled parts

Abstract

Thin-Walled Parts (TWP) are widely used in aerospace, whose service performance is significantly affected by the machining accuracy. Considering the poor stiffness of TWP, multi-pass machining is used to machine TWP for better surface accuracy. However, it is difficult to accurately predict the final machining accuracy due to the surface topography error's propagation and accumulation in multi-pass machining. Therefore, this paper proposes a multi-pass machining accuracy prediction method for TWP based on dynamic factors (cutting force and stiffness). Firstly, a flexible cutting force prediction model, which considers the axial errors determined by the initial surface topography and part deflection, is proposed. Secondly, a Position-Pass-Dependent Stiffness (PPDS) model is established considering position-dependent of stiffness and multi-pass machining material removal. Finally, combining the two models above, a multi-pass machining accuracy prediction method based on Genetic Algorithm -Back Propagation (GA-BP) neural network is proposed. The experiments under various conditions have been carried out to validate the proposed method. The machining accuracy (flatness as an example) is as high as 90.8% using the method in this paper, while it is only 73.9% when the accumulative error is neglected. The proposed method can significantly improve the performance of machining accuracy prediction by analyzing the error propagation mechanism and the effect of dynamic factors between multi-pass machining. Furthermore, this also provides a theoretical basis for process parameters optimization and machining accuracy improvement in TWP machining.