PINN-PORT
Automated Physics-Informed Neural Network Modeling and Application Software for Tsunami Risk Assessment in PORTs
Automated Physics-Informed Neural Network Modeling and Application Software for Tsunami Risk Assessment in PORTs
Project PINN-PORT aims to develop an automated GIS framework and two application software for assessing tsunami hazards and their impact on industrial infrastructures in ports. The model will be applied to three ports in Cilegon, northwest Java, Indonesia, which are within 200 km of the identified seismic gap region of the Sunda megathrust. This region is estimated to be capable of generating Mw 8.9 tsunamigenic earthquakes in the future. These locations are of particular interest due to Indonesia's vulnerability to tsunamis, evidenced by the significant loss of life in tsunamis like those in Palu and Anak Krakatau in 2018, which resulted in over 5,000 fatalities. Tsunamis have the potential to inflict damage on structures within port areas, in turn, causing operational disruptions with cascading effects on maritime trade. Hence, understanding the tsunami impact on port structures and moored ships is essential.
The objectives of this project are to (1): Make ports more resilient and safer to tsunamis by developing innovative hazard analysis tools and harnessing the latest modelling technique namely Physics-Informed Neural Network modelling. (2): Automate the process of tsunami risk analysis to ports with the aid of a GIS-based tool and standalone applications that are user-friendly, even for non-experts. (3): Communicate, collaborate and co-create the resilience tools by consulting with port authorities.
PINN-PORT is funded by the European Commission, under the Horizon scheme and the UKRI EPSRC under project code EP/Z001080/1. See more details here: https://gtr.ukri.org/projects?ref=EP%2FZ001080%2F1
PINN-PORT is managed by the CERLAB.