A model has been developed to predict creep and creep-rupture of continuous carbon fiber reinforced epoxy composite (AS4 / 3501-6) with different fiber orientations in a wide range of temperature and stress. The model is based on a three element mechanical system consisting of two elastic springs and a third element representing the anleastic behavior for which a modified thermal activation theory is utilized. Two elastic moduli and two model parameters (activation energy and activation volume) were determined experimentally. The other two parameters were obtained by numerical fitting of the experimental creep curves using the creep model. The agreement between the predicted and experimental results (creep strain or time dependent total compliance versus time) was excellent within 4% error in the test temperature range 295 to 433 K and at stresses ranging from 10% to 80% of the ultimate tensile strength. The creep of the composite was found to be due to the mechanism, causing CC transition in the matrix, with an activation energy of 494 kJmole⁻¹. Reinforcing carbon fibers did not alter the creep mechanism in the epoxy. However, they altered the creep behavior of the epoxy resulting in reduction in creep rate and the energy dissipated by creep. The orientation of the fibers with respect to loading axis did not influence the creep behavior of the epoxy.

The creep-rupture was defined in the present model by equating the stored energy in the two elastic springs with a critical energy. This critical energy was estimated using the constant strain rate tensile test results and was found to be dependent on strain rate and temperature due to change in fracture mode with temperature and strain rate. Very good agreement between the predicted and experimental creep-rupture time at an applied stress was obtained in the temperature range 295 to 433 K for [90]₁₆ composite and in the temperature range 295 to 403 K for [10]₈ composite. The agreement in [90]₁₆ composite was within a factor of 2 in the experimental time window of 14 days. In [10]₈ composite, the agreement was within 30% at 295 K during a test time of 4 months. At higher temperatures and longer time, an overestimation by a factor in the range 8.7 - 10 was obtained. This apparent deviation is interpreted to be due to an additional creep mechanism that becomes significant for the test conditions which produce relatively large strains (~0.87%) before rupture.