A parametric study of the statistical parameters of the creep compliance of a vinyl ester polymer

In this study, due to the large scatter in the viscoelastic response of polymers, probability distributions are used in the formulation of the creep compliance functions of a vinyl ester polymer. Short-term tensile creep experiments are conducted at two stress levels and at four temperatures below the glass transition temperature, with 10 replicates for each test configuration. Creep compliance is calculated from the generalized 3-D viscoelastic constitutive equation with a Prony series representation. Two numerical methods, the maximum likelihood estimation technique and the method of moments, are used for estimating the distribution parameters. The Weibull and the log-normal distributions of creep compliance functions, estimated by both numerical methods, are tested for goodness-of-fit using the Kolmogorov-Smirnov hypothesis test. The results show that the probability density distributions of the creep compliances are highly time-dependent, significantly affected by the choice of the parameter estimation method, and better represented by the method of moments. Additionally, the Weibull distribution provides better representation of the data for short-term creep whereas the lognormal distribution is better suited for long-term creep.