The aim of this project is to refine existing crack growth models to be more realistic of the early stages of crack growth, by considering the spatially-resolved crack-front geometry and microstructural interactions. 3D X-ray computed tomography (CT) data of crack propagation will be directly imported into finite element (FE) simulations to obtain the crack tip stress field and a local description of crack-front driving forces. This will provide insight into why some cracks propagate to form long cracks, whilst other cracks arrest whilst they are still small, and why in some cases the growth of arrested cracks restarts.