Analytical Models for Tool Wear Prediction During AISI 1045 Turning Operations

Abstract

Tool wear is one of the most important topic in cutting field. Its interest is due to the influence of tool wear on surface integrity of the final parts and on tool life, and, consequently, on the substitution policies and production costs. Analytical models, able to forecast the tool wear with a satisfactory accuracy, can give to the companies working in the material removal field a valid instrument to optimize the cutting processes. In the present work a comparison between response surface methodology (RSM) and artificial neural networks (ANNs) fitting techniques for tool wear forecasting was performed. For developing these predictive models, tool life tests, consisting of longitudinal turning operations of AISI 1045 steel bars using uncoated tungsten carbide inserts and variable cutting parameters, were conducted. Both flank (VB) and crater wears (KT) of the tool were monitored. The models were validated comparing the calculated tool wear values with the experimental ones, showing that ANNs model provides better approximation than RSM in the prediction of the amount of the tool wear parameters. So, from an industrial point of view, this model should be implemented into a production management software in order to correctly define the tool substitution policy during batch production.