Ultrasonic testing is routinely used in the nuclear power generation industry to assess the structural integrity of plant components. The regulatory nature of the industry means that ultrasonic inspection procedures require substantiation through experimental trials and semi-analytical simulation, which can be costly and time-consuming. Furthermore, due to the complex defect and component geometries that are commonly encountered in the industry, the generation of evidence that accurately represents real inspections can be challenging. Numerical modelling techniques offer an effective alternative for substantiating inspections as they can accurately simulate the ultrasonic scattering from complex geometries, yet such techniques lead to increased computational cost. Hybrid techniques, which combine both semi-analytical and numerical methods, offer an approach to rapidly simulate an entire ultrasonic inspection whilst maintaining the ability to simulate the scattering from complex defects. However, there is a lack of experimental validation for extant hybrid techniques, particularly for three-dimensional simulations, and they have not been widely applied to the simulation of surface-breaking defect inspections, which are commonly encountered in both manufacturing and in-service inspections.

This thesis has extended the functionality of a three dimensional hybrid technique to simulate the inspection of surface-breaking defects. Validation evidence demonstrates that the hybrid model accurately predicts the ultrasonic scattering from a well-characterised reflector. An arbitrary transduction system has been incorporated to the model, enabling the model to be more broadly applied to the simulation of ultrasonic inspections of plant components. Furthermore, validation on a real defect provides a high level of confidence that the model can be used to assess and quantify the performance of inspections where qualification becomes challenging. This work has made a clear step forward in the use of simulation for inspection qualification and further work is required to fully optimise the simulation methodology.