A micro-macro approach to rubber-like materials?Part I: the non-affine micro-sphere model of rubber elasticity
MetadataShow full item record
The contribution presents a new micro-mechanically based network model for the description of the elastic response of rubbery polymers at large strains and considers details of its numerical implementation. The approach models a rubber-like material based on a micro-structure that can be symbolized by a micro-sphere where the surface represents a continuous distribution of chain orientations in space. Core of the model is a new two-dimensional constitutive setting of the micro-mechanical response of a single polymer chain in a constrained environment defined by two micro-kinematic variables: the stretch of the chain and the contraction of the cross section of a micro-tube that contains the chain. The second key feature is a new non-affine micro-to-macro transition that defines the three-dimensional overall response of the polymer network based on a characteristic homogenization procedure of micro-variables defined on the micro-sphere of space orientations. It determines a stretch fluctuation field on the micro-sphere by a principle of minimum averaged free energy and links the two micro-kinematic variables in a non-affine format to the line-stretch and the area-stretch of the macro-continuum. Hence, the new model describes two superimposed contributions resulting from free chain motions and their topological constraints in an attractive dual geometric structure on both the micro- and the macro-level. Averaging operations on the micro-sphere are directly evaluated by an efficient numerical integration scheme. The overall model contains five effective material parameters obtained from the single chain statistics and properties of the network with clearly identifiable relationships to characteristic phenomena observed in stress-strain experiments. The approach advances features of the affine full network and the eight chain models by a substantial improvement of their modeling capacity. The excellent predictive performance is illustrated by comparative studies with previously developed network models and by fitting of various available experimental data of homogeneous and non-homogeneous tests.