Mill-cut: a neural network system for the prediction of thermo-mechanical loads induced in end-milling operations

This paper presents the design and implementation issues of a generalized system called mill-cut, developed for the prediction of cutting forces and temperature in end-milling operations. Based on an ANN approach, mill-cut predicts all the three components of cutting forces and average shear plane temperature for a given set of machining parameters broadly categorized into three groups viz. (i) cutting tool geometrical parameters (ii) cutting parameters and (iii) workpiece material properties. In the present work, for representing overall machining condition, 15 machining parameters having major impact on the cutting forces and cutting temperature were chosen. The feed-forward back-propagated ANN architecture has been incorporated, which was initially trained with analytical data before incorporating it as part of an integrated system. Results obtained from the proposed model show good agreement with the experimental/numerical (FEM based) results available in the literature.