Modelling of cyclic and viscous behaviour of materials for gas turbine applications - with a view towards fatigue life predictions

Abstract: Critical jet engine components are exposed to high temperatures. These components are also required to have low weight in order to reduce fuel consumption and allow for heavier payload, higher speed, and longer flight distance of the aircraft. The combined requirements of high temperature resistance and low weight raise the demands on the material strength and durability properties, as well as the demands on a robust and accurate component design methodology including thermal, structural and fatigue analyses. Uncertainties in these analyses entails that higher safety factors and more conservative designs must be utilised. Therefore, it is of great importance to continuously improve the accuracy of these analyses by e.g. taking into account experimentally observed mechanical behaviour of the material. The aim of this thesis is to tailor material models that can capture the observed physical deformation and failure mechanisms in the studied high temperature alloys; thereby providing a basis for an accurate and robust design methodology. The focus is placed on calibrating and evaluating the material models with regard to isothermal strain controlled low-cycle fatigue tests of a Ni-based superalloy, Haynes 282, and a high temperature Ti-alloy, Ti-6242. Different cyclic and viscous material phenomena are studied, such as: cyclic plasticity, cyclic hardening/softening, mean stress relaxation, and stress relaxation. The material models are formulated to accurately capture the stress-strain behaviour; thereby, they may also serve as a basis for crack initiation fatigue analysis. A criterion for predicting the crack initiation phase of the fatigue life, accounting for the evolution of the cyclic material behaviour and viscous effects, is adopted in the thesis and the influence of the material model on the predicted fatigue life is investigated. To be able to account for the evolution of the cyclic material behaviour, many load cycles need to be simulated, which is computationally demanding. Therefore, methods for improving the computational efficiency of such simulations are assessed. Moreover, the complexity of the material parameter identification of the formulated material models is highlighted, and methodologies for improving and simplifying this identification process are discussed. The sensitivity of experimental scatter on the identified material parameter values and the resulting model response is also evaluated. Furthermore, the intergranular fracture of Haynes 282 forgings is investigated using micromechanical models. In these models, the nonuniform distribution of grain size and carbides along the grain boundaries results in an anisotropic macroscopic tensile ductility similar to that observed in experiments.