Abstract
It is common practice in the application of finite element analysis to model compression set (CS) of elastomers during aging with two different material models according to the two-network theory of Tobolsky. The theory relies on the existence of two networks. The first one represents the original network after vulcanization and is sensitive to chain scission. The second network accounts for the formation of additional crosslinking during aging. Besides the use of user subroutines to describe the two-network model, an element overlay technique is also needed as the full set of both material behaviors did not exist for assignment to a single element. This element overlay technique is valuable for research and developmental purposes but makes extension to industrial usage quite challenging. Our goal is to simulate the CS of elastomers after long-term aging in a commercial finite element software with no need for extra subroutine codes or mesh superposition.
Ethylene propylene diene (EPDM) O-rings were aged in a compressed state at 75 °C, 100 °C, 125 °C and 150 °C for up to 183 days. Investigations of the experimental test results were used to identify material models and their parameters to develop a finite element model to simulate CS. The model was implemented in the finite element software ABAQUS/Standard® with a sequential temperature-displacement coupling.
Regarding the influence of temperature, the Arrhenius equation is adopted for the time-temperature relationship. The activation energy value that is required for the simulation is firstly determined from shifting the experimental CS results with the time-temperature superposition technique and plotting the shift factors in an Arrhenius diagram. The experiments were compared with the simulation results. Afterwards different activation energies were used in the simulation and discussed. A suitable choice of the activation energy value with regard to the reference temperature and the test temperature is presented. With the chosen activation energies, the match between numerical CS values after long-term aging and the experimental results was improved.
