Overall Framework and Major Contributions

This research is framed as a Bayesian analysis of a reliability problem. As such, techniques are drawn from various disciplines. The main contribution is that insight into the engineering process is obtained through consideration of the process and data from a statistical perspective. Specifically, the following are the main contributions made by this research:

• A hierarchical model is developed to model short crack growth in steel. This constitutes a new application of the methodology in an engineering context;
• A model for rate of coalescence is developed, which is motivated by work on an analysis of damage accumulation in orthopaedic hip replacement fixation;
• A unification of the model for growth and the model for coalescence is carried out, using auxiliary variables for unobserved parameters. Practical conclusions as to what spatial data would be useful and appropriate are drawn, in order to guide future engineering research.

Thus this work develops new statistical models for an engineering problem, and draws practical conclusions in order to focus future research.