Development of a Constitutive Backstress Model for the Prediction of Creep and Stress Relaxation in Gas Turbine Materials

Abstract

There is a drive towards a broader range of fuels in industrial gas turbines, with higher levels of sulphur and potentially hydrogen. Due to these harsher environments, there is also a drive for corrosion resistant alloys and coatings. A number of key corrosion resistant superalloys, which are being employed to cope with these evolving conditions, exhibit primary creep. It is therefore imperative that fundamental material models, such as those for creep deformation, are developed to ensure they can accurately predict the material response to evolving operating conditions. The requirements for a creep model are complex. The model must be able to: predict forward creep deformation in regions dominated by primary loads (such as pressure); predict stress relaxation in regions dominated by secondary loads (such as differential thermal expansion); predict the effects of different creep hardening mechanisms. It is also clear that there is an interaction between fatigue and creep. With flexible operation, this interaction can be significant and should be catered for in lifing methods. A model that has the potential to account for the effect of plasticity on creep, and creep on plasticity is therefore desirable. In previous work the authors presented the concept for a backstress model to predict creep strain rates in superalloys. This model was fitted to a limited dataset at a single temperature. The approach was validated using simple creep-dwell tests at the same temperature. This paper expands on the previous work in several ways: 1) The creep model has been fitted over a wide range of temperatures. Including the effect of temperature in complex creep models presents a number of difficulties in model fitting and these are explored. 2) The model was fitted to constant load (forward creep) and constant strain (stress relaxation) tests since any creep model should be able to predict both forms of creep deformation. However, these are often considered separately due to the difficulty of fitting models to two different datasets. 3) The creep deformation model was validated on stress change tests to ensure the creep deformation response can cope with changes in response variables. 4) The approach was validated using creep-fatigue tests to show that the creep deformation model, in addition to our established fatigue models, can predict damage in materials under complex loading.