Artificial intelligence (AI) has changed our daily life and is also progressively becoming an important tool in the design of functional components and in ensuring their structural integrity under critical loading conditions. Its contribution is particularly significant when addressing complex physical phenomena and multivariate problems, which are quite common when components to be produced with advanced manufacturing processes are designed. One example is the structural assessment of parts produced through Additive Manufacturing (AM) processes, whose design involves a multitude of interacting factors, making the prediction of mechanical performance particularly challenging using conventional methods.
This talk deals with the recent research activities carried out within my research group and focusing on the application of AI for the design of AM components. In particular, AI has been applied to support the design against fatigue failures of metal AM parts, with the developed algorithms providing the fatigue response starting from the manufacturing process parameters. In particular, Physics-Informed Neural Networks (PINNs) have been exploited to incorporate physical constraints, to enhance the consistency of the outcomes. Examples of integration of topology optimization and Machine Learning algorithms, applied for the optimized design of industrial components, are also shown. Another topic addressed in this talk is the use of Machine Learning algorithms, integrated with numerical simulations and a multiscale approach, to enable damage-tolerant design of lattice structures for crash-absorbing applications, as well as for assessing the structural integrity of polymeric AM parts. Finally, current research activities and future research trends are discussed and outlined.