The paper shows how text analysis techniques are used to automatically select text-based incident reports to import in commercial BowTie software. Text is analysed with the TFIDF method to create an ontology that facilitates the creation of search queries to match individual reports to threats in a BowTie. The method is illustrated with an example from the railway industry but it is equally applicable in the chemical industry. The approach saves time and allows for the analysis of huge volumes of incident reports. This work demonstrates big-data techniques can add value to chemical safety and paves the way to digital safety management systems.
|Title of host publication||Hazards 28 Conference Proceedings|
|Number of pages||9|
|Publication status||Published - 1 May 2018|
|Event||Hazards 28 Conference Proceedings: In Association with the Mary Kay O'Connor Process Safety Centre - Edinburgh International Conference Centre , Edinburgh, United Kingdom|
Duration: 15 May 2018 → 17 May 2018
Conference number: 28
|Name||Institution of Chemical Engineers Symposium Series|
|Conference||Hazards 28 Conference Proceedings|
|Period||15/05/18 → 17/05/18|
Van Gulijk, C., Figueres Esteban, M., Hughes, P., & McCullogh, P. (2018). Incorporating incident reports in bow-ties with Big Data techniques. In Hazards 28 Conference Proceedings (163 ed., Vol. 2018-May). (Institution of Chemical Engineers Symposium Series). ImechE.