Weld quality monitoring of dissimilar polycarbonate to nylon-6 in friction stir lap welding using wavelet packets of power signal

The present work addresses the influence of process variables on friction stirred lap weld quality for dissimilar amorphous polycarbonate to semi-crystalline nylon-6 using three different tool pin profiles. The input electrical power along with tool thrust, stirring torque sensors and thermocouples were used to acquire dynamic variation in process stability and monitoring of weld quality using significant sensory responses. The joint strength was significantly higher (31 MPa) when both tool traverse speed and plunge depth were low (15 mm/min and 0.1 mm, respectively) at elevated tool revolving speed (2400 rpm) using a cylindrical tool pin, whereas it was slightly upgraded at low tool spinning speed (1600 rpm) with a high-swept-ratio hexagonal pin. The deviation in axial tool thrust was better indicator of weld strength, while the thermal history along the tool retreating side dictated the bonding mechanism. Moreover, the time-frequency information of power was processed through wavelet packet analysis to describe the process stability. The low‑frequency band power wavelet was found to be better indicator of joint strength than time‑domain features of sensory signals. However, the instability in intermediate to high‑frequency power was also important to characterise the non-uniformity in the weld bead profile.