Kinematics and improved surface roughness model in milling

Surface roughness has a significant influence on the mechanical properties and service life of a component. During face milling, surface roughness greatly varies in the tool step direction and can be controlled by using a surface roughness prediction model. However, the issues of accuracy and efficiency of surface roughness prediction models have not been adequately addressed. This study aims to address these research constraints. An improved surface roughness prediction model is proposed, taking into consideration the influences of insert back cutting and stepover ratio. First, the profile-forming mechanism is analyzed based on geometry and kinematics. Subsequently, an improved surface roughness prediction model is established. Thereafter, the influence of feed per tooth, stepover ratio, corner radius, and minor cutting edge angle on surface roughness are analyzed through numerical simulation. Finally, the experiment of face milling aerospace aluminum alloy 7075 is suggested to verify the improved model, and the Z-Map model is introduced for comparison. Results show that the surface roughness is nonlinear with a feed per tooth and stepover ratio, a monotonic variation with corner radius, and a minor cutting edge angle. The predicted values of the improved model and the Z-Map model for the Rsm are equal to the experimental values. However, the improved model reduces the prediction error of R-a from 11.2 to 4.2% in the non-overlapping compared with the Z-Map model and from 62.58 to 13.34% in the overlapping. In addition, the improved model performs better than the Z-Map model in predicting the shape parameters. This work serves as a significant reference for selecting and optimizing the milling parameters to enable machining quality control.

Automated summary

This study proposes an improved surface roughness prediction model for face milling processes by considering the influences of insert back cutting and stepover ratio. Through numerical simulation and experimental verification, the results show that the improved model has higher prediction accuracy compared to the Z-Map model.