GB-DT for real-time risk monitoring in the high-hazard energy industry


Digital Twins (DTs) can improve real-time risk monitoring and predictive maintenance of high-hazard industry (process, transport, energy, etc.), therefore enabling the Safe and Sustainable by design (SSbD) approach recently recommended by the EU Commission to guide the industrial innovation process.

Our competences will be offered to address the issues related with the computational feasibility of relying on DTs in such context. DTs typically include Artificial Intelligence (AI) /Machine Learning (ML) Black-Box (BB) models trained on historical data collected during past system operation: the need of large amount of data (often not available in raising industry, like hydrogen) and the lack of physical explanation of the inputoutput relationships in BB models, render DTs of difficult acceptance.

To address this issue, we propose to rely on Grey-Box (GB) DTs that, by coupling the DT win a consequence assessment model enables real-time risk monitoring of high-hazard industry assets (hydrogen, nuclear, etc.), trading-off accuracy, computational burden, and interpretability by exploiting the capabilities of both White-Box (WB) and BB modelling approaches. Uncertainty analysis, and decision-making tools will have to be throughoutly investigated to demonstrate the feasibility of the approach on real industrial cases.

Prof. Francesco Di Maio (POLITECNICO DI MILANO, Italy)
Prof. Enrico Zio (POLITECNICO DI MILANO, Italy, and MINES Paris-PSL, Sophia Antipolis, France)

More details



Country Italy
Contact email

Return to the list of expressions of interest.

Pre-proposal deadline June 12 2024