Abstract
A theoretical treatment on the oxide-controlled dwell fatigue crack growth of a y'strengthened nickel-based superalloys is presented.In particular,this study investigates the influence of an externally applied load and variations in the y'dispersion on the grain boundary oxide growth kinetics.A dislocation-based viscoplastic constitutive description for high temperature deformation is used to simulate the stress state evolution in the vicinity of a crack at elevated temperature.The viscoplastic model explicitly accounts for multimodal y'particle size distributions.A multicomponent mass transport formulation is used to simulate the formation/evolution of an oxide wedge ahead of the crack tip,where stress-assisted vacancy diffusion is assumed to operate.The resulting set of constitutive and mass transport equations have been implemented within a finite element scheme.Comparison of predicted compositional fields across the matrix/oxide interface are compared with experiments and shown to be in good agreement.Simulations indicate that the presence of a fine γ'size distribution has a strong influence on the predicted ow stress of the material and consequently on the relaxation in the vicinity of the crack-tip/oxide wedge.It is shown that a unimodal dispersion leads to reduced oxide growth rates(parabolic behavior)when com-pared to a bimodal one.Stability conditions for oxide formation are investigated and is associated with the prediction of compressive stresses within the oxide layer just ahead of the crack tip,which become progressively negative as the oxide wedge develops.However,mechanical equilibrium requirements in-duce tensile stresses at the tip of the oxide wedge,where failure of the oxide is predicted.The time taken to reach this critical stress for oxide failure has been calculated,from which dwell crack growth rates are computationally derived.The predicted rates are shown to be in good agreement with available experimental data.
基金项目
Engineering and Physical Sciences Research Council EPSRC(Rolls-RoyceawardEP/H022309/1)
NETL
NSTL
维普
万方数据