Shery, P., Altenhof, W., Smith, R., Beeh, E. et al., "Experimental Observations on the Mechanical Response of AZ31B Magnesium and AA6061-T6 Aluminum Extrusions Subjected to Compression and Cutting Modes of Deformation," SAE Technical Paper 2017-01-0377, 2017, doi:10.4271/2017-01-0377.
Cylindrical extrusions of magnesium AZ31B were subjected to quasi-static axial compression and cutting modes of deformation to study this alloy’s effectiveness as an energy absorber. For comparison, the tests were repeated using extrusions of AA6061-T6 aluminum of the same geometry. For the axial compression tests, three different end geometries were considered, namely (1) a flat cutoff, (2) a 45 degree chamfer, and (3) a square circumferential notch. AZ31B extrusions with the 45 degree chamfer produced the most repeatable and stable deformation of a progressive fracturing nature, referred to as sharding, with an average SEA of 40 kJ/kg and an average CFE of 45 %, which are nearly equal to the performance of the AA6061-T6. Both the AZ31B specimens with the flat cutoff and the circumferential notch conditions were more prone to tilt mid-test, and lead to an unstable helical fracture, which significantly reduced the SEA. Axial cutting of AA6061-T6 extrusions has been shown to be an effective, ductile mode of energy dissipation, yielding a repeatable, nearly constant load/deflection response with a crush force efficiency (CFE) up to 96%. In the present tests, the quasi-static cutting deformation of AZ31B extrusions achieved a respectable CFE of 80%, but revealed a load/deflection response with sharp, minute, rapid fluctuations, indicating an undesirable fracturing failure. Additionally, the average specific energy absorption (SEA) of AZ31B was 11 kJ/kg, which is less than half that seen for AA6061-T6 extrusions of the same geometry (24 kJ/kg). An analytical model of the cutting deformation of AA6061-T6 extrusions can predict the steady state cutting force to within 10%. However, the model did not agree well with the experimental results of AZ31B, yielding approximately 150% error. This deviation is likely attributed to the brittle deformation nature of AZ31B that is not accounted for in the model.