Optimization of geometry parameters and cutting parameters of threading tool based on neural network prediction model for chip flow angle of single-edged oblique cutting

A method for optimizing the threading tool’s geometry and cutting parameters based on a neural network prediction model for chip flow angle (CFA) of single-edged oblique cutting is proposed. Firstly, four control variables (i.e., inclination angle \({\lambda }_{s}\) ${\lambda }_s$ , normal rake angle \({\gamma }_{n}\) ${\gamma }_n$ , feed \(f\) $f$ , and the principal cutting edge angle \({{K}}_{r}\) $K_r$ ) that would have a significant effect on the CFA ( \({\psi }_{\lambda }\) ${\psi }_{\lambda }$ ) are analyzed and determined based on the simulation result of oblique cutting AISI 304 workpiece. Then, the 4-factor 8-level orthogonal simulation test of oblique cutting AISI 304 stainless steel workpieces is designed and completed, and the neural network prediction model for CFA is established according to the simulation data. Next, a mathematical model is established to calculate the deviation between the natural chip flow angle (NCFA) of the threading tool’s two straight cutting edges and the whole tool’s actual chip flow angle (ACFA) based on the neural network prediction model. Finally, based on the mathematical model, a method of minimizing the deviation between the NCFA and the ACFA by optimizing the threading tool’s geometry parameters and cutting parameters is proposed to reduce the degree of chip-ejection interference and cutting force in threading process. Moreover, a set of optimal values ( \({\lambda }_{s}\) ${\lambda }_s$ ,  \({\gamma }_{n}\) ${\gamma }_n$ , and \(f\) $f$ ) are obtained. The results of the thread turning comparison experiment show that the optimized triangle thread tool has the function of chip splitting, which results in a smaller degree of chip-ejection interference and cutting force (23.96% reduction compared to standard threading tool), showing more excellent cutting performance.