Advancing the synergy between models and experiments to investigate environmentally and mechanically driven crack propagation

Aero-gas turbine running temperatures are rapidly increasing in order to improve their efficiency, and as a consequence components are subjected to more extreme environ- ments. With higher operational temperatures and improved reliability, there is an in- creased chance of both corrosion and mechanical degradation. In addition to operational temperatures, the environment in which an aircraft flies has a significant effect on the material life. Many contaminants are ingested by the engine and deposited on the turbine blades, which often leads to surface degradation. Depending on the ingested contami- nants, temperature, and applied stresses, cracking can be initiated and propagated rapidly. This is particularly evident in the lower-shank regions of single-crystal nickel-based su- peralloy blades, which have recently experienced significant cracking. This study aims to understand the mechanisms behind crack propagation in single- crystal nickel alloys exposed to intermediate temperatures, and when this propagation is either mechanically or chemically driven. This research started by assessing crack inter- action mechanisms that were hypothesised to be both stagnating and accelerating crack growth, depending on specific length scales and crack formations. This was performed by integrating available experimental data to calibrate a phase field model that could predict the extension of cracks for different crack separations and layouts. The modelling results clearly characterised the length scales needed to encourage crack shielding, and which crack formations would see a stress intensification and consequently crack coalescence. These results informed the decision to revisit the experimental setup to optimise which experiments were performed. Using this newly developed methodology, the salt deposi- tion method was amended with the aim of isolating the deposition sites to minimise crack interaction mechanisms. The hypothesis was that significantly longer cracks would be ob- ii served if this could be achieved. This was performed for both the C-ring (at 550°C), and corrosion-fatigue (at 700°C) tests. In the case of CMSX-4, the results were striking, with the C-ring seeing cracks as much as ten times the size of those previously seen. CMSX-10 however, did not show a significant difference, as such, a microstructural characterisation analysis was conducted, whereby the γ/γ′ structure for the two alloys was replicated from microscopy data and further phase field models were run. The results showed that a more regular structure was more resistant to crack propagation owing to the misalignment of γ′ , which caused stress relaxation in the γ channel and at the interface. Finally, this thesis shows how the model, once calibrated for one material and species, can be used to approximate the response expected for another single-crystal nickel alloy or a change in the embrittling species, while accounting for a degree of uncertainty. This is not to say that modelling can or should replace experiments but rather to highlight that preliminary modelling results can be used to build a test matrix that can reduce the number of experiments that are run. It should be noted that this thesis does not focus on the chemical/corrosive aspects in much detail, but rather investigates the importance of stress. This thesis summarises the importance of integrating modelling, microscopy, and experiments to set and answer hypotheses more efficiently.