Computational methods pervade all branches of science and engineering, helping us to advance our understanding of Nature and foster the development of new technologies. They are essential for supporting the industry in the process of embracing the ongoing digital transformation, effectively taking advantage of the ever-growing amount of available digital data and computer power.

To advance computational methods for real-world problems, we combine theory, simulations, monitoring, and experiments. Our goals are (i) to devise novel approaches and algorithms, (ii) to employ numerical simulations to understand complex systems and processes, and (iii) to solve practically relevant problems. Our interests are intrinsically multidisciplinary and concern applications to support a healthy and sustainable society, to enhance security (in terms of energy supply and public health), and to design new materials that can contribute to a carbon-negative economy.

Our main focus is on multiscale and data-driven simulations, with the objective to integrate mechanistic modeling, machine learning, and experiments into algorithms that interact with the environment (embodied machine learning). We have four main research areas:

• multiscale porous media for a sustainable economy
• modeling, experiments, and machine learning for fluid dynamics
• network models for energy and society
• data-science and autonomous experimentation