TWI

Coupons with both triangular and rectangular gaps were prepared from AA 2024 aluminium alloy plate stock by friction stir welding (FSW) and volumetric defect formation was investigated by a neural network-based electronic NDE (e-NDE) technique. The stock material comprised AA 2024-T351 plate (thickness 6.35 mm). Specimens were butt welded at a spindle speed of 400 rpm, a travel speed of 330 mm/min and an axial load of 29.6 kN using an FSW tool incorporating the "Wiper" feature and a threaded twisted-flat probe. The joining surfaces of the plates were machined to have gaps with geometries including full and half diamonds or full and half rectangles. The welded joints were characterised by microstructure observations and tensile testing. The FSW machine was equipped with force sensors and a data acquisition system to collect feedback signals (such as travel and spindle speed feedback and tool position, force and torque) which were analysed by a neural network model. The detection of voids caused by the presence of preexisting gaps along the joint line is discussed.

9th International Symposium, 15-17 May 2012, Session 2B: Inspection, Paper 02