Prediction of Waviness Values in Skew Rolling Using Machine Learning Methods

Abstract

Skew rolling with three rolls is used for producing axisymmetric parts. In this method, the tools are spaced every 120° on the circumference of the workpiece. They are also set askew relative to the rolling axis. Cross sectional reduction is made effective by moving the tapered rolls closer to or away from the center line of the workpiece. Experiments were conducted with variable initial conditions of the rolling process to examine surface topography of rolled parts. Obtained experimental results were then analyzed using machine learning methods in order to determine the most effective regression model with the highest coefficient of determination R2 for waviness prediction.

Keywords

• machine learning
• skew rolling
• cnc rolling mill
• stepped axles and shafts