P. Haupt et A. Lion, EXPERIMENTAL IDENTIFICATION AND MATHEMATICAL-MODELING OF VISCOPLASTICMATERIAL BEHAVIOR, Continuum mechanics and thermodynamics, 7(1), 1995, pp. 73-96
Uniaxial torsion and biaxial torsion-tension experiments on thin-walle
d tubes were carried out to investigate the viscoplastic behavior of s
tainless steel XCrNi18.9. A series of monotonic tests under strain and
stress control shows nonlinear rate dependence and suggests the exist
ence of equilibrium states, which are asymptotically approached during
relaxation and creep processes. Strain controlled cyclic experiments
display various hardening and softening phenomena that depend on strai
n amplitude and mean strain. All experiments indicate that the equilib
rium states within the material depend on the history of the input pro
cess, whereas the history-dependence of the relaxation and creep behav
ior appears less significant. From the experiments the design of a con
stitutive model of viscoplasticity is motivated: The basic assumption
is a decomposition of the total stress into an equilibrium stress and
a non-equilibrium overstress: At constant strain, the overstress relax
es to zero, where the relaxation time depends on the overstress in ord
er to account for the nonlinear rate-dependence. The equilibrium stres
s is assumed to be a rate independent functional of the total strain h
istory. Classical plasticity is utilized with a kinematic hardening ru
le of the Armstrong-Frederick type. In order to incorporate the amplit
ude-dependent hardening and softening behavior, a generalized are leng
th representation is applied [14]. The introduction of an additional k
inematic hardening variable facilitates consideration of additional ha
rdening effects resulting from the non-radiality of the input process.
Apart from the common yield and loading criterion of classical plasti
city, the proposed constitutive model does not contain any further dis
tinction of different cases. The experimental data are sufficient to i
dentify the material parameters of the constitutive model. The results
of the identification procedure demonstrate the ability of the model
to represent the observed phenomena with satisfactory approximation.