Natural fiber composite: Challenges simulating inelastic response in strain-controlled tensile tests

Abstract

Problems occurring, when nonlinear time-dependent material model with parameters identified in creep tests is applied to simulate high-strain response in strain-controlled tests, are described and analyzed. Reasons for discrepancies with experimental loading curves are revealed. Presented numerical/experimental examples deal with three bio-based composites showing highly nonlinear behavior due to damage, nonlinear viscoelasticity and viscoplasticity. Schapery's approach for viscoelasticity and Zapas' model for viscoplasticity are used. The model is generalized to include microdamage effect. It is shown that the main problem in simulations at high stresses is the reliability of data from creep test for model identification in this region because creep rupture limits the available data region and extrapolation to higher stresses is rather uncertain. Alternative solution is to employ relaxation tests at high strains to obtain the missing information. However, it would work only in absence of viscoplastic strains: viscoelastic relaxation functions cannot be determined by maintaining constant total strain if viscoplastic-strain is developing. Based on sensitivity analysis of composite response to variations of the elastic modulus, damage, viscoelastic and viscoplastic parameters, suggestions are made for improving (further “tuning”) the model in high stress region by using tensile stress–strain curves in quasi-static loading.