Alexander Staroselsky

Creep, Plasticity and Fatigue of Single Crystal Superalloys: Physics-Based Life Prediction for Turbine Components in Severe Operating Environments

The chapter reports the process and computer methodology for a physics-based prediction of overall deformation and local failure modes in cooled turbine airfoils, blade outer air seals, and other turbomachinery parts operating in severe high temperature and high stress environments. The computational analysis incorporated coupled aero-thermal CFD with non-linear deformation finite element calculations with a crystallographic slip-based constitutive model. The methodology utilized a fully-coupled elastic-viscoplastic model that was based on crystal viscoplasticity, and a semi-empirical lifing model introduced the use of dissipated energy to estimate the remaining part life in terms of cycles to failure. The viscoplastic model used an incremental large strain formulation additively that decomposed the inelastic strain rate into components along the octahedral and cubic slip planes of single crystal nickel-based superalloys. This crystallographic-based viscoplastic constitutive model based on Orowan’s law was developed to represent sigmoidal creep behavior. Inelastic shear rate along each slip system was expressed as a sum of a time dependent creep component and a rate independent plastic component. A new robust and computationally efficient rate-independent crystal plasticity formulation was developed and combined with the creep flow model. The transient variation of each of the inelastic components included a back stress for kinematic hardening and latent hardening parameters to account for the stress evolution with inelastic strain as well as the evolution for dislocation densities. The model was evaluated at real engine characteristic mission times and flight points for part life prediction. The method was effective for use with three-dimensional finite element models of realistic turbine airfoils using commercial finite element applications. The computationally predicted part life was calibrated and verified against test data for deformation and crack growth.