Applying Machine Learning and Computer Vision to the Study of Fatigue Fracture Surfaces

Images of fracture surfaces are rich in information and have long been examined to determine fracture characteristics such as initiation site and crack growth mode. Computer vision and machine learning techniques, including convolutional neural networks (CNNs), are novel methods that can be applied to microstructural images to make predictions and connections between the microstructure and material responses. For fatigue failure, the goal is to connect the visual appearance of the fatigue fracture surface to fatigue characteristics such as loading values, crack length, and crack growth rate. This allows quantitative information to be obtained from fracture surfaces where specific loading conditions are unknown. Scanning electron microscopy (SEM) has been the primary image modality for studying fracture surfaces, but surface height information acquired from Scanning White Light Interferometry (SWLI) can provide additional useful information. Backscattered electron (BSE) images provide compositional information that can be used to differentiate phases in a material and help determine how they influence the fracture surface. This project collects multimodal data from Ti-6Al-4V fracture surfaces in the form of Secondary Electron and BSE images and SWLI height data and combines them to train a CNN and identify high stress points, crack initiation sites, and predict values such as stress intensity factor, crack length, and crack growth rate. Through data fusion, these image modalities can be combined to improve the recognition of subtle features on the fatigue fracture surface. Using standard computer vision techniques to register images and transfer learning to construct CNNs, the stated quantitative values can be predicted. The images used to develop this model, the image registration technique, the creation of CNNs, identified fatigue properties, and fracture characteristics will be presented.