Project title: Engineering-informed foundational models for sustainable bridges
PI: Barbara Rossi
Structures such as bridges are crucial to our societies, cities, and living spaces. These complex structures have been designed and built over the past millennia. Today, architects and engineers can construct aesthetic, reliable, and sometimes truly magnificent structures that last for hundreds of years, often relying on time-tested theories, traditional computational methods (e.g. finite element method), and significant time investment. Rarely does this process take advantage of complex, multi-criteria calculations: instead, comparisons are typically used, and AI is almost never employed to assist in the discovery, optimisation and design process.
While engineers use computers daily to assist bridge design, the process of delivering infrastructure has not significantly changed in a hundred years. What has changed are the accelerating threats posed by climate change to our structures and the pressing need for truly carbon-neutral designs.
AI, as a tool that can learn patterns and generate novel designs, offers a promising solution to these challenges. It can help design more sustainable and efficient structures faster, with fewer mistakes, and at a lower cost. Not only can AI analyse vast amounts of data, but it can also "learn from the past" to create new bridge designs that would respond to sustainable objectives.
