Crystal plasticity and multiscale modelling of superalloy creep

Abstract

A crystal plasticity approach for superalloy creep has been presented which employs a finite element-based representative volume element (RVE) methodology. The γ channels are assumed to undergo crystal slip and the γ′ particles to deform elastically. A range of superalloys has been studied. Thermocalc computations provide the γ′ volume fraction and an automated scheme for generating the resulting RVE has been developed. It has been shown that primary creep response in a wide range of superalloys over high stress, low temperature regimes is represented excellently by the model, by determination of just an activation energy and an alloying element density. It has been hypothesised that the transition from primary to secondary creep results from the development of geometrically necessary dislocations within the γ channels at the γ′ interfaces. Without the need of further material parameters, it has been shown that secondary creep rates over a broad range of stress and temperature can be accurately predicted, hence supporting the hypothesis. An empirical relationship has been established between the alloying element density and the atomic weight percentages of the alloying elements, using a range of superalloy data. It is hypothesised that a role of the alloying elements within the γ channels is to act as inhibitors of ribbon dislocation motion, hence leading to the large range of macro-level primary and secondary creep responses observed in the alloys with variations in constituent alloys. The empirical relationship established, when combined with the crystal RVE methodology, then allows the prediction of superalloy creep rates from knowledge of alloying constituents.

Keywords:

• superalloys
• creep
• plasticity of crystals
• dislocation mechanics
• alloying
• superalloy creep
• crystal plasticity

Acknowledgements

The EPSRC is gratefully acknowledged for financial support for this work and Rolls-Royce Plc. for the provision of nickel alloy CMSX-4 creep data. In particular, Dr Duncan Maclachlan is thanked for his contributions to the manuscript.