Modeling the Creep Deformation, Damage, and Rupture of Hastelloy X Using MPC Omega, Theta, and Sin-Hyperbolic Models

Abstract

Combined cycle power plants components such as steam pipe work, pressure vessels, boilers, heat exchangers, and gas turbine disks, etc. are exposed to elevated temperature and pressure operation conditions for longer durations. Components may fail within the elastic limit due to a time dependent deformation and damage mechanism called creep. Creep prediction models are used to determine the state of these components and to schedule optimum inspection, maintenance, and replacement intervals. In this study, the deformation, damage, and life of Hastelloy X is characterized using three recently developed models; the Omega, Theta projection, and Sin-hyperbolic models. An analysis is performed to compare the models in terms of accuracy, assumptions, constant identification techniques, flexibility in use, and limitations. The influence that final creep strain has on Theta and Omega model is discussed. Sixteen tests were performed at four different configurations of stress (2.1–36.5 ksi) and temperature (1200–1800°F). In the experimental data, Hastelloy X does not exhibit the primary stage. In this study, the secondary and tertiary creep stages are modeled. Creep deformation and rupture life data is used to optimize the constants for the three models. Predictions using these models are compared with experimental data. It is found that the novel Sin-hyperbolic model better fits the experimental data, and is easier to apply. The Omega model predicts longer life than the Sinh and the Theta Projection model. The rupture life prediction of the Theta projection model is the worst due to dependence on the critical creep strain rate. It is observed that the Hastelloy X final creep strain depends on stress and temperature; this leads to a less accurate critical creep strain rate prediction resulting in inaccurate rupture life predictions for the Theta projection model. The analytical damage of the Omega model exhibits a linear evolution with time while the Sinh model show a more realistic elliptical creep damage evolution with time. A process to determine the constants of all the models is clearly described. The dependence of the trajectory of the creep curves with respect to the constants is discussed in detail. An analytical derivation of each model is provided. Predictions of these three models show that the Sinh model produces a better creep deformation curve by normalizing the experimental creep strain rate data. It is found that overall the Sinh model offers more flexibility, prediction accuracy, and is easier to apply.