Loss of containment of toxic and flammable inventories from process plant is associated with a long history of major accidents including fires, explosions and toxic releases. Such accidents affect both workers and the offsite public. These issues are often associated with the onshore process industries which incorporate a very wide range of segments including pharmaceutical manufacture, tank storage, downstream oil & gas, fine and speciality chemical manufacture as well as many others. What may be less well appreciated is that while the Civil Nuclear sector has a key focus on containment of radiological materials, it also maintains significant inventories of flammable and toxic materials, which it terms ‘chemotoxic’ hazards. It follows that a very broad range of industries have a desire to prevent and mitigate the potential for loss of containment events which release chemotoxic materials. Existing sources of loss of containment intelligence include the Health & Safety Executive (HSE) and other databases which can be interrogated to glean process safety insights. Such systems incorporate some limited coding of data, but often feature much greater detail within unstructured free text. Systematic interrogation of such free text fields could yield greater detail within process safety insights as well as a potentially larger number of records with which to draw insight. The Discovering Safety Programme is a multidisciplinary initiative funded by the Lloyd’s Register Foundation. The programme aims to improve plateaued safety performance through better insight via data analysis tools including text mining and natural language processing. This paper describes the early stages of a project within the Discovering Safety Programme to obtain process safety insights from HSE’s regulatory database. This work includes analysis of coded information, proposals to extract intelligence from unstructured free text and also exploration of whether process safety intelligence can be extracted from a subset of occupational safety incidents. The paper describes the findings from industry consultation, including the civil nuclear sector.
