Tool wear Modelling and Monitoring in Drilling Using Vibration Signals

Abstract

Prediction of drill life and tool replacement strategies greatly depend on newlineconservative estimates of tool life from past tool wear data. The tool may be newlineunder utilized or over utilized. For a fully automated machine, prediction of tool newlinewear is significantly increased in hindering tool failure and is required to change newlineor regrind the tool for avoiding longer machine down time. On the other hand, newlinedrill wear beyond the recommended range affects the quality of the workpiece. newlineVibration based on-line monitoring system has been presented by many of the newlinereasearchers. In this study, the vibration parameters are used for the prediction of newlinedrill flank wear during drilling operation. In the past studies, the sensor was newlinepositioned at a constant distance from the workpiece to the holes requiring newlinedrilling for monitoring tool wear. But in actual practice, all the workpieces do not newlinelend themselves to the arrangement of the sensor in the same place where the newlinedistance from all holes is equal. The position of sensor has a significant effect on newlinethe vibration signal which determines the state of drill flank wear. The present newlineresearch study offers normalization methodologies using Frequency and Time newlinedomain features thereby overcoming the signal attenuation effect of sensor newlineii newlinepositioning and focuses on the development of mathematical model for newlinedetermining drill flank wear, which is a function of independent cutting newlineparameters and vibration amplitude ratio. newlineFree vibration analysis using Finite Element Analysis (FEA) study on the newlineworkpiece and drill bit were carried out. This analysis reveals drill bit frequencies newlinedetermined as corresponding to bending frequency and torsional-axial coupling newlinefrequencies which occur in the drilling process and verified using Experimental newlineModal Analysis (EMA) by the resonance frequency test. Also Fast Fourier newlineTransformation (FFT) analysis was carried out and dominant peaks of bending newlineand torsional frequencies are identified. The proposed strategy uses dominant newlinepeaks of torsional axial first mode