The Anand’s viscoplasticity model is different from other rate-independent plasticity and Creep models so this behaviour is expected, – Anand’s model is a unified plasticity model where the total inelastic strain cannot be decomposed into plastic strains due to rate independent plasticity theories and those due to creep effects. – The inelastic strain is calculated by an internal variable, s, which is called as deformation resistance which is nothing but resistance to inelastic flow of the material. – The evolution of this internal variable can result in either strain hardening or strain softening depending on the saturation value of deformation resistance. – The strain softening occurs when either the strain rate decreases or the temperature increases. In such a state the inelastic strain in the material increases without developing significant stresses (in other words creep is dominant). But the total inelastic strain cannot be decomposed into plastic and creep strains. – ANAND is very similar to the Generalised Garofalo creep equation with the difference that instead of C2 being a constant (for the creep only representation) it is a variable that encompasses both rate dependent creep and plasticity together. The rate dependent inelastic strains could potentially continue to develop even as stress remains constant or maybe even decreases, which could be a reason why the the two max. values are appearing at different locations. – The disconnect between peak stresses and strains might also be related to the complexity of a non-proportional cyclic loading in 3D stress state. I am not totally sure about that, but perhaps they hang together better in a 1D uni-axial model.