Plastic potentials/ Flow rules
I have a fundamental question regarding flow rules of finite plasticity models especially those used in soil mechanics. In most of the papers and books, I have seen the usage of an associated flow rule with the plastic potential similar to the yield surface. However, I am unable to understand the means of obtaining a non-associative flow rule. I am using Abaqus with cap plasticity model (modified Drucker Prager Cap model) to simulate powder compaction process. Abaqus uses a non-associative plastic flow rule definition in the shear yield space but doesn't provide any valid explanation for the equation of the flow potential used. The non-associated flow rule definition results in a lower dilation than that predicted by an associated flow rule.Is the flow rule equation something that is obtained arbitrarily and then fit using experiments or is it driven by plastic strain data obtained by compaction experiments?
If it is the latter case, then I am confused how does on obtained the plastic flow potential surface by plotting incremental plastic strains in the stress space. Wouldn’t it just give a qualitative estimate of the surface and one will have to come up with a proper surface by trial and error.