Chip Bifurcation Machine Learning Forecasting Model Based on Experimental Data and Random Forest Algorithm

The phenomenon of chip bifurcation catastrophe that greatly reduces the cutting force when cutting with a linear double-edged turning tool has important utilization value in energy saving and emission reduction. In order to reveal the occurrence law of chip bifurcation catastrophe, based on the dichotomy idea, a set of cutting experiments were designed and completed for symmetrical cross-turning V-groove of disc work-pieces with linear double-edged turning tools, and 346 sets of experimental data containing the critical conditions for chip bifurcation catastrophe were obtained. Based on experimental data, a machine learning model for predicting the occurrence of chip bifurcation for a given cutting condition was developed using a random forest algorithm, and the model parameters were optimised. Validation experiments found that the prediction accuracy of the proposed model reached 100%, which is significantly higher than the previous empirical model of chip bifurcation critical surface based on surface fitting technology and the mathematical model of chip bifurcation catastrophe based on cutting power calculation function and catastrophe theory.