Experimental study & multiscale modelling of the high temperature deformation of P91 under multiaxial loading

P91 steel, a tempered martensitic steel containing 9% Cr and 1% Mo, is a widely used material in the power generation industry. In order to optimise power plant performance and minimise in-service failures the mechanical behaviour of this material must be understood at multiple length scales and at temperatures representative of in-service conditions. In this thesis, P91 steel is studied using a combination of experimental and computational methods. As-received and ex-service material has been examined, and the effect of service exposure on the tensile strength has been determined. Significant differences between the strength of as-received and service-exposed material have been observed, which has been attributed to block coarsening during service. A crystal plasticity finite-element model developed for P91 has been extended in this work. The material subroutine, initially using 12 of a possible 48 slip systems in the BCC structure, has been modified to include slip activation on an additional 12 {211}<111> slip systems. An orientation analysis, combining of scanning electron microscopy (SEM) and electron backscatter diffraction (EBSD) revealed that the Kurjumow-Sachs (K–S) relation best described the relationship between the orientation of prior austenite grains (PAGs) and the martensite blocks. This discovery allowed for the development of a modified Voronoi tessellation (VT) model providing a systematic method to study the influence of block/packet size and morphology on the mechanical response of the material. A multiscale experimentally-based modelling strategy is presented in the thesis to study the high-temperature deformation of P91 at multiple length scales. This multiscale approach combined a macroscale (specimen level) FE model and a microscale crystal plasticity model of the local microstructure. Damage evolution has been implemented in the multiscale model, making use of an experimentally calibrated void growth model at the macroscale and a linear Lemaitre-type damage model to account for void nucleation and growth at the microscale. The model has been experimentally validated using high-temperature mechanical measurements and EBSD analysis. The experimental studies include both uniaxial tension specimens and multiaxial (notched) specimens. A study of the model response and predicted orientation change during loading found that 12 slip systems of type {110}<111> are sufficient to model to microscale deformation in these materials. Overall, it has been found that the modelling approach can accurately account for crystal orientation changes during deformation, particularly at large deformations. For lower deformations, the orientation changes are not sufficiently large to provide clear validation, but at these deformation levels, good agreement was obtained in measured inelastic strains determined directly from the EBSD maps.