A novel physical-digital approach for the assessing a large critical asset. This project aims to deliver an artificial intelligence-enabled decision-making tool to maintain and manage the floating covers of vast lagoons that treat raw sewage. The cover harvests the biogas released from the anaerobic digestion of sewage for electric power generation that exceeds the plant’s requirement. The approach involves an innovative thermographic technique and exploits transfer learning to adapt neural netw ....A novel physical-digital approach for the assessing a large critical asset. This project aims to deliver an artificial intelligence-enabled decision-making tool to maintain and manage the floating covers of vast lagoons that treat raw sewage. The cover harvests the biogas released from the anaerobic digestion of sewage for electric power generation that exceeds the plant’s requirement. The approach involves an innovative thermographic technique and exploits transfer learning to adapt neural networks trained on lab-scale and synthetic data to field implementation. The outcome is a machine learning framework to optimise biogas harvesting and renewable energy generation, and to avoid structural failure, that is capable of continuous improvement to take into account improved data and/or modelling capabilities.Read moreRead less
Unlocking the potential of quantitative x-ray micro-tomography. This project aims to build on two new ideas in data acquisition and 3D image reconstruction to bring 3D X-ray microscopy or computed tomography (CT) into advanced research use as well as common industrial applications. In the past 10 years, CT has improved our understanding in areas ranging from the evolution of life and osteoporosis to composite material failure and oil recovery. However, the full potential of CT remains unrealised ....Unlocking the potential of quantitative x-ray micro-tomography. This project aims to build on two new ideas in data acquisition and 3D image reconstruction to bring 3D X-ray microscopy or computed tomography (CT) into advanced research use as well as common industrial applications. In the past 10 years, CT has improved our understanding in areas ranging from the evolution of life and osteoporosis to composite material failure and oil recovery. However, the full potential of CT remains unrealised because crucial features in structure and composition are overlooked by simplistic algorithms. Users cannot directly capture quantities of interest such as key compositional variation or defects, and workflows are poorly adapted for large-scale use in industrial fabrication or phenomics. This project aims to address these shortcomings using advanced mathematics and algorithms.Read moreRead less