An approach to determine condition of boring tool using acoustic & vibration signals

Abstract

The work presented in the paper focuses on an approach for automatic system to determine boring tool condition, using acoustic signatures. This acoustic signal based tool wear prediction system removes the need for manual inspection of the tool cutting edges and minimizes the time used on wear measurement. The tool condition prediction system proposed in this paper for boring operation essentially utilizes acoustic sensor i.e. microphone. In order to understand effects of vibrations generated during boring operation on tool, vibration signals were also recorded and processed to compare it with acoustic signals. The signals were acquired by data acquisition system in LABVIEW environment. The data thus obtained were further processed using Fourier transform and wavelet transform. Using this transforms,12 number of time domain, frequency domain and time frequency domain features were collected then with soft computing techniques which are learning from data strategy i.e.RBF and MLP, prediction of tool condition was done. The experimental results were analyzed with respect to various depth of cuts, feed rates and cutting speed for EN8, EN-9 and EN-31 steel. This approach is well suited for determining condition of boring tool using acoustic and vibration signals.