A comparison of limited-stretch models of rubber elasticity

TL;DR

This paper reviews and compares various limited-stretch models of rubber elasticity, introducing a new approximation that balances simplicity and accuracy, and evaluates their performance against experimental data.

Contribution

It provides a comprehensive comparison of existing limited-stretch models and proposes a new, more accurate approximation to the inverse Langevin function.

Findings

The eight-chain model effectively combines neo-Hookean and Gent models.

The new approximation significantly improves accuracy over Cohen's model.

The models show good agreement with experimental data across deformation modes.

Abstract

In this paper we describe various limited-stretch models of nonlinear rubber elasticity, each dependent on only the first invariant of the left Cauchy-Green strain tensor and having only two independent material constants. The models are described as limited-stretch, or restricted elastic, because the strain energy and stress response become infinite at a finite value of the first invariant. These models describe well the limited stretch of the polymer chains of which rubber is composed. We discuss Gent's model which is the simplest limited-stretch model and agrees well with experiment. Various statistical models are then described: the one-chain, three-chain, four-chain and Arruda-Boyce eight-chain models, all of which involve the inverse Langevin function. A numerical comparison between the three-chain and eight-chain models is provided. Next, we compare various models which involve approximations to the inverse Langevin function with the exact inverse Langevin function of the eight-chain model. A new approximate model is proposed that is as simple as Cohen's original model but significantly more accurate. We show that effectively the eight-chain model may be regarded as a linear combination of the neo-Hookean and Gent models. Treloar's model is shown to have about half the percentage error of our new model but it is much more complicated. For completeness a modified Treloar model is introduced but this is only slightly more accurate than Treloar's original model. For the deformations of uniaxial tension, biaxial tension, pure shear and simple shear we compare the accuracy of these models, and that of Puso, with the eight-chain model by means of graphs and a table. Our approximations compare extremely well with models frequently used and described in the literature, having the smallest mean percentage error over most of the range of the argument.