THE ONSET OF NONLINEAR VISCOELASTICITY IN MULTIAXIAL CREEP OF GLASSY-POLYMERS - A CONSTITUTIVE MODEL AND ITS APPLICATION TO PMMA

Citation
Pj. Dooling et al., THE ONSET OF NONLINEAR VISCOELASTICITY IN MULTIAXIAL CREEP OF GLASSY-POLYMERS - A CONSTITUTIVE MODEL AND ITS APPLICATION TO PMMA, Polymer engineering and science, 38(6), 1998, pp. 892-904
Citations number
31
Categorie Soggetti
Polymer Sciences","Engineering, Chemical
ISSN journal
00323888
Volume
38
Issue
6
Year of publication
1998
Pages
892 - 904
Database
ISI
SICI code
0032-3888(1998)38:6<892:TOONVI>2.0.ZU;2-E
Abstract
A physically based, isostructural, constitutive model is described for simulating the onset of nonlinear viscoelasticity in multiaxial creep of glassy polymers, as needed in stress analyses of load-bearing comp onents. In the linear viscoelastic limit, shear response reduces to th at of a generalized Maxwell model, while hydrostatic response is Hooke an. Nonlinearity enters through Eyring-type rate process kinetics. The equations of the model are solved numerically using a pseudo-linear a pproximation through each time step, leading to an incremental equatio n for stress that would be convenient for use in finite element analys es. The model and its assumptions were tested using tension, shear and combined tension/shear creep experiments on well-aged poly(methyl met hacrylate) at 70 degrees C. Reproducibility tests confirmed the assump tion of constant glass structure for strains up to similar to 1.5 x 10 (-2). Shear and pressure activation volumes were obtained by fitting t he dependence of the shear compliance on stress invariants. The data s howed unequivocally that shear activation volumes vary with log(relaxa tion time), and excellent agreement was obtained for a linear variatio n, consistent with the ''compensation rule'' of polymer thermo-viscoel asticity. The activation volumes are large (many monomer units), indic ating the cooperative nature of the elementary flow process. Interesti ngly, they are of the same order as those applying to yield and plasti c flow. Although the model finds success in simulating creep, it fails to describe so accurately the strain recovery on unloading Possible e xplanations are suggested.