The SAFE-10-T project will develop a Safety Framework to ensure high safety performance while allowing longer life-cycles for critical infrastructure
across the road, rail and inland waterway modes. Moving from considering critical infrastructure such as bridges, tunnels and earthworks as inert
objects to being intelligent (self-learning objects) the SAFE-10-T project will provide a means of virtually eradicating sudden failures. This will
be achieved by:
- The Safety framework will incorporate remote monitoring data stored in a BIM model that feeds into a decision support framework (DST) that enables
decisions to be made automatically with maintenance prioritised for elements exhibiting stress.
- A major advance that will be achieved in the project is that the algorithms at an object level and at a network level will incorporate machine
learning to train the system to evolve with time using available monitoring data.
- A trans-disciplinary approach with experts in Artificial Intelligence and big data management working with owners, engineers with expertise in
risk and modelling and sociologists to make decisions.
- Our major European infrastructure managers will undertake demonstration projects at critical interchanges and nodes of the TEN-T transport network.
The involvement of the MoTeSy group in SAFE-10-T focuses on supporting the civil engineers in the project with the use of machine learning technology
to improve the prediction accuracy of infrastructure health and usage predictions on the object and network levels. We will furthermore use formal
methods to verify the machine-learned models against the known boundary conditions established by civil engineers. In this way, the confidence in
the learned models will be improved, which is needed to base costly infrastructure construction and maintenance decisions on them.
Funded by the European Commission under the Horizon 2020 program.