Wavelet packet analysis based on power signature for weld quality monitoring in friction stir lap welding of nylon-6

The solid-state friction stirring processes are often recommended for the joining of thermoplastic polymers. This work addresses different tool pin profiles on lap-weld characteristics for semi-crystalline nylon-6 in friction stir welding. The objective was to achieve uniform weld profile over the weld line for improved joint strength and ductility. Various sensor signals like tool thrust, stirring torque, thermal cycles and input electrical power were acquired during welding. The longitudinal weld profile was monitored using time-domain integrated thrust-torque-power signatures, while joint mechanical behaviour was assessed through peak temperature and cooling rate. The hidden time-frequency information in power signal was processed through wavelet packet analysis to characterize process stability. The higher swept-ratio square pin provided the highest strength efficiency (≥ 68%) with smoother torque-power at high tool speed (2400 rpm) while cylindrical pin improved joint ductility despite fluctuated stirring torque. Minimal deviation in peak temperature in advancing to retreating side indicated uniform heat balance with square and cylindrical pin. Power loss was significantly reduced (≤ 7%) at higher rotational speeds using square pin. The low frequency band (1–125 Hz) power wavelet was predominant in dictating process stability. The results confirm that wavelet analysis of power is highly effective for weld quality monitoring.