| A microstructural model for creep crack growth has been constructed incorporating a number of random features as a means of estimating the scatter in creep crack growth rates. Experimental data have indicated that creep crack growth rates can exhibit large amounts of scatter. The sources of this scatter are thought to arise from random variations in microscopic features and randomness in the damage mechanism.;In this model, crack propagation is assumed to result from the diffusive transport of matter from the crack tip onto the grain boundary, according to the Chuang-Rice (1973) model. The crack is assumed to advance along grain boundaries having random grain boundary and surface diffusivities. Crack growth is enhanced by the presence of a random number of randomly-spaced creep cavities, which form on the grain boundary in front of the advancing crack tip. The stress field due to the advancing crack is assumed to have an arbitrary power-law singular behavior, so that the model is capable of representing both K- and C*-controlled crack growth. In the model, given the near-tip stress field, the transport equations govern the flux of matter from the cavities and the crack tip onto the grain boundary. Due to the complicated nature of the resulting model, Monte Carlo simulation is needed to assess the statistical nature of the crack growth behavior.;Detailed calculations have been carried out using material parameters representative of type 304 stainless steel under C*-controlled crack growth conditions and of Inconel 718 under K-controlled crack growth conditions. The resulting scatter bands for the predicted crack growth rates are compared with available experimental data.;The model is more successful in predicting creep crack growth rates in inert environments, e.g. vacuum, pure helium and pure argon, or under lower crack tip stresses. These conditions result in lower crack rates, where voids have more time to grow and to have significant influence on the fracture process. |
