This thesis is concerned with the experiments and modelling for the time-dependent behavior of semicrystalline polymers. Relaxation test had long been used to evaluate the performance of materials. In this study, a novel multi-relaxation-recovery test was proposed based on cyclic stages of stress relaxation and stress recovery. Three nonlinear viscoelastic models, that is, the standard model and two models with two dashpots connected either in parallel or in series, were examined for the analysis of the test results. Each model contains a time-dependent, viscous branch and a time-independent, quasi-static branch. The examination suggests that the standard model can determine the long-term, load-carrying performance of polyethylene and identify a transition point for the onset of plastic deformation in the crystalline phase, but the models with two dashpots connected either in parallel or in series are needed to provide a close simulation of the experimentally measured stress response in both relaxation and recovery stages of the multi-relaxation-recovery test. In this work, the mechanical performance of two types of polyethylene was compared based on multi-relaxation-recovery test results at room temperature. The multi-relaxation-recovery tests were also conducted at elevated temperatures to explore the possibility of quantifying the activation energies for deformation of the dashpots at the relaxation stage. It was found the multi-relaxation-recovery test has the advantage of separating the time-dependent and time-independent components of stiffness of the materials. The study concludes that the multi-relaxation-recovery test can provide data for determining parameters in Eyring’s model in order to characterize the contribution of time-dependent and time-independent components of the stress response to polyethylene’s deformation.

This study also presents an analysis of the stress evolution of high-density polyethylene at loading, relaxation, and recovery stages in a multi-relaxation-recovery test. The analysis is based on a three-branch spring-dashpot model that uses the Eyring’s law to govern the viscous behavior. The spring-dashpot model comprises two viscous branches to represent the short- and long-term time-dependent stress responses to deformation, and a quasi-static branch to represent the time-independent stress response. A fast numerical analysis framework based on genetic algorithms was developed to determine values for the model parameters so that the difference between the simulation and the experimental data could be less than 0.08 MPa. Using this approach, values of the model parameters were determined as functions of deformation and time so that the model can simulate the stress response at loading, relaxation, and recovery stages of the multi-relaxation-recovery test. The simulation also generated ten sets of model parameter values to examine their consistency. The study concludes that the three-branch model can serve as a suitable tool for analyzing the mechanical properties of high-density polyethylene, and values for the model parameters can potentially be used to characterize the difference among different types of polyethylene for their mechanical performance.

It is important to explore the possibility of identifying a unique set of the parameter values so that the parameters can be used to establish the relationship between deformation and microstructural changes. The study developed an approach for this purpose based on stress variation during loading, relaxation, and recovery of polyethylene. One thousand sets of parameter values were determined for fitting data at the relaxation stages with discrepancy within 0.08 MPa. The study found that even with such a small discrepancy, the 1000 sets of parameter values showed a wide range of variation, but one of the model parameters, σ_(v,L) (0), followed two distinct paths rather than showing a random distribution. The study further found that five selected sets of parameter values which showed discrepancy below 0.04 MPa yielded highly consistent values for the model parameters, except for the characteristic relaxation time. Therefore, the study concludes that a unique set of model parameter values can be identified to characterize mechanical behavior of polyethylene. This approach was then applied to four types of polyethylene pipes, to determine their quasi-static stress. The results showed that these polyethylene pipes have very close quasi-static stress despite the clear difference in the measured stress. This indicates that a unique set of model parameter values could be identified for the spring-dashpot model, enabling a further study of using spring-dashpot models to characterize microstructural changes of polyethylene during deformation.