Neural network based prediction of drill wear from theoretically analysed and experimentally measured values of thrust force and torque

Abstract

This work introduces a new approach of drill wear monitoring by combining model-based and experimental values of thrust force and torque. The drill wear prediction performance of the neural network using the difference of experimental and theoretical values of thrust force and torque is shown to be better than the performance of the same using only experimental values. Experimental data were acquired from strain gauge type dynamometer during drilling on mild steel work piece with high-speed steel (HSS) drill bits. In this work, Williams' orthogonal model has been implemented for the theoretical prediction of thrust force and torque in drilling under different cutting conditions.