Based on the tensile and compressive yield strengths reported by Wilson for 2024-T351 aluminum, presumably using the 0.2% offset strain definition; the yield function parameters can be calculated and compared with results from Spitzig and Richmond in Table 1.
The pressure-dependence of 1100 and 2024-T351 is similar, but 2024-T351 exhibits a strength-differential of 5.9%, while 1100 does not exhibit an appreciable strength-differential. While Wilson did not measure volume change, Spitzig and Richmond did, and found there to be no significant dilation; indicating that an associated flow rule will not correctly predict plastic strain. This is also the case for frictional materials, where it is common to employ a nonassociated flow rule.
We have observed strength-differential in laboratory experiments using aged Inconel 718 (a precipitation strengthened nickel-base alloy) (2,3), 6061-T6 aluminum and 6092/SiC/17.5-T6 (a particulate reinforced aluminum alloy) (4). The Mises yield criterion does not apply well to these materials either. Our work on Inconel 718 (3) indicates that a yield function, which we called a threshold function because we were working in the realm of viscoplasticity, along the lines of that proposed by Drucker 5 for an aluminum alloy was most suitable.
Finally, while it is fairly obvious, it is worth pointing out that the Drucker-Prager yield criterion predicts more flow for the same tensile stress than the Mises yield criterion simply due to the presence of the positive term. Thus, the finite element results of Wilson for Mises and Drucker-Prager yield criteria are self-consistent. It would be interesting to know the range of for a particular notch geometry.