Research on Ploughing to Shearing Transition in Micro Milling of Titanium Alloy Using Gramian Angular Field and Machine Learning Methods

Micro milling tools play a key role in the fabrication of miniaturized components. However, their efficacy is hindered by inherent limitations such as reduced rigidity stemming from small dimensional features. Notably, the material removal dynamics in micro milling diverge from those observed in macro milling operations. Among the prominent obstacles encountered in micro milling is the occurrence of high amount of ploughing. The accurate identification of the transition from ploughing to shearing is essential in optimizing machining performance. Various methodologies have been proposed to define this critical transition zone. This study aims to forecast machining conditions in the context of micro milling of Ti6Al4V, with a specific focus on addressing ploughing phenomena. Machining forces measured during the micro milling of Ti6Al4V were integrated into an AI model using a novel approach utilizing Gramian Angular Field (GAF) transformation to convert force curves into images for classification training. The AI model, validated through K-fold cross-validation, effectively distinguished between critical and safe machining conditions, particularly in relation to the ploughing effect (with 100% accuracy). While the model struggled with predictions near the transition from ploughing to shearing, it showed strong performance overall, especially when appropriate datasets were selected. The study highlights the potential of the GAF and Convolutional Neural Network classification model for detecting critical conditions in micromachining, particularly within the ploughing-to-shearing transition regime.