Artificial Intelligence and Civil Engineering
27 November, 2021
Juan Chiachío and Manuel Chiachío
The most advanced countries are seeing the onset of a new industrial revolution brought about by digital technologies such as artificial intelligence, remote sensing and robotics. The civil engineering industry, which comprises the construction and management of large infrastructures (roads, railways, bridges, power plants, dams, ports, etc.), is being part of this revolution, due to the relatively low cost of digitalisation technologies in relation to the huge construction and management costs of the infrastructures.
The amount of real-time data and information coming from monitored infrastructures is expected to increase exponentially in the coming decades. This information has the potential to reduce by billions the national expenditure on infrastructure operation and maintenance. Furthermore, it is speculated that these data will change the way in which 21st century infrastructures are designed, built and operated.
However, although a significant progress is being carried out in this direction, we are not yet in such a state of data abundance that will allow an autonomous and predictive management of infrastructures as ciber-physical entities. The complexity, the enormous size, as well as the variability and uncertainty inherent to the behaviour of infrastructures, make the acquisition of relevant data for accurate prognostics a challenge in itself. In the meantime, state-of-the-art methodologies such as physics-enriched artificial intelligence are emerging as candidates to fill this initial “data gap”.
In particular, over the last century, civil engineering has greatly benefited from the theoretical and scientific advances carried out in the fields of Mechanics, Hydraulics, Electricity, Geotechnics, etc. Emerging tools such as the Physics-Informed Artificial Neural Networks have the potential to exploit all this valuable scientific knowledge to reliably predict the medium and long-term behaviour of our infrastructures, and thereby, their degradation and maintenance needs.
Subsequently, in a future scenario of data abundance, these data will not only be valuable to make long-term predictions that have not been possible so far. Also, these will be useful to improve the available physics-based models and even to discover new scientific knowledge based on them.
In this context, the aim of DaSCI in this research field is to contribute to this revolution by working at the interface between civil engineering and artificial intelligence, and leading at an academic level the development of groundbreaking methodologies for the civil engineering industry of the 21st century.