Prediction of drilled hole quality in bidirectional woven carbon fiber reinforced plastic using wavelet packets of force–torque signals

Drilling of fibrous composites like carbon fiber reinforced plastics (CFRPs) is always considered a grim job to perform owing to its abrasiveness and inhomogeneous properties. Compared to the available studies, the present work addresses the parametric investigation on drilling capability of TiAlN- and TiN-coated drills over the uncoated one. The delamination factor, surface roughness, and circularity error in CFRP were considered as the quality features. The acquired thrust and torque were analyzed in the time domain as well as the time–frequency wavelet field, followed by a sensitivity analysis between decomposed original force–torque signals in different phases. Finally, an attempt was made to improve the predictability of drilled hole quality using decomposed wavelets of force and torque extended to hybrid force–torque wavelets of the best features to the developed regression models. The feed rate was more predominant than spindle speed on delamination evidenced by higher axial thrust particularly using coated tools. However, the coated drills generated more uniform torque than the uncoated drills and improved the drilled hole quality as well. The torque wavelet was a better indicator of hole surface integrity than force wavelets in which intermediate frequency band torque and low-/high-frequency band force wavelets were found to be more significant.