Evaluating Parameter Value Identification Methods for Modeling of Nonlinear Stress Relaxation in Polyethylene.

Viscous properties play a major role in the time-dependent deformation behavior of polymers and have long been characterized using spring-dashpot models. However, such models face a bottleneck of having multiple sets of model parameter values that can all be used to simulate the same deformation behavior. As a result, these model parameters have not been widely used to quantify the viscous properties. In this study, a newly developed multi-relaxation-recovery test was used to obtain the variation in stress response to deformation of polyethylene (PE) and its pipes during relaxation, revealing the complexity of PE's nonlinear viscous stress response to deformation. Using a three-branch spring-dashpot model with two Eyring's dashpots, this study shows the possibility of determining the model parameter values using four different analysis methods, namely, the mode method, peak-point method, highest-frequency method, and best-five-fits method. Model parameter values from these methods are compared and discussed in this paper, to reach the conclusion that the best-five-fits method provides the most reliable and relatively unique set of model parameter values for characterizing the mechanical performance of PE and its pipes. The best-five-fits method is expected to enable the use of the model parameters to quantify PE's viscous properties so that PE's load-carrying performance can be properly characterized, even for long-term applications.