WP leaders: (Health) and (Centre of Maritime Health and Society, CMSS)
Analysing safety risks for energy hubs in relation to maritime operations reinforced by Natural Language Processing
Postdoc Ibsen Chivata Cardenas, (Centre of Maritime Health and Society, CMSS) (May 2023 - )
The work package explores the application of the new developments in Natural Language Processing (NLP) to improve accident analysis completeness and predictability. The findings will be illustrated by analysing ongoing safety challenges in the construction, operation, and maintenance of energy hubs in the North and Baltic Seas. Energinet and its partners (research collaborators involved in the energy hubs development), have suggested the analysis of diving and deck operations during the installation, maintenance, and repair of subsea cables, as well as operations related to ship traffic.
The following key activities have been set out for the work package:
- Accident analysis: Using the Accident Anatomy method (Brunn, Rasmussen, and Taylor, 1979), which is based on cause-consequence diagrams (CCDs) (Nielsen, 1974). This method supports both post-accident and predictive cause-consequence analysis (Svedung and Rasmussen, 2002).
- Data collection: Gathering over 10,000 records of accidents and incidents from sources like DMAIB, EMSA, HSE, MAIB, OSHA, and the United States Coast Guard. The focus is on accidents involving personnel injuries (subsea work, deck, and platforms) and ship interactions. Compiling specifications and procedures for the construction, operation, and maintenance of energy hubs.
- Tools: Implementing automation tools from NLP to rapidly conduct the analysis of the collected data (10,000+ records)