Wear State Monitoring of Drills Based on Mahalanobis Distance of Welch Spectrum Energy of Wavelet Packet Coefficients from Cutting Torque Signals

In this paper, drills wear monitoring method that based on the mahalanobis distance (MD) of spectrum energy features form cutting torque is proposed. The method consists of three steps. First, using the wavelet transform carried out separating cutting torque components from the original signals. Second, signals feature is adopted to Welch spectrum method extracted on wavelet coefficients of differ tool state (i.e., slight wear, normal wear, and serious wear). Welch spectrum energy of wavelet coefficients compose the feature Vector. Finally, wear states monitoring is used MD to identify recognition of these during drilling process of cutting signal features. Practical application results on the proposed method is accurate, that applied to extract the feature and evidence for wear states monitoring of cutting tools in drilling process.