Development of a new chemical process-industry accident database to assist in past accident analysis

Past accident analysis (PAA) is one of the most potent and oft-used exercises for gaining insights into the reasons why accidents occur in chemical process industry (CPI) and the damage they cause. PAA provides invaluable ‘wisdom of hindsight’ with which strategies to prevent accidents or cushion the impact of inevitable accidents can be developed.A number of databases maintain record of past accidents in CPI. The most comprehensive of the existing databases include Major Hazard Incident Data Service (MHIDAS), Major Accident Reporting System (MARS), and Failure and Accidents Technical Information Systems (FACTS). But each of these databases have some limitations. For example MHIDAS can be accessed only after paying a substantial fee. Moreover, as detailed in the paper, it is not infallible and has some inaccuracies. Other databases, besides having similar problems, are seldom confined to accidents in chemical process industries but also cover accidents from other domains such as nuclear power plants, construction industry, and natural disasters. This makes them difficult to use for PAA relating to CPI. Operational injuries not related to loss of containment, are also often included. Moreover, the detailing of events doesn’t follow a consistent pattern or classification; a good deal of relevant information is either missing or is misclassified.The present work is an attempt to develop a comprehensive open-source database to assist PAA. To this end, information on about 8000 accidents, available in different open-source clearing houses has been brought into a new database named by us PUPAD (Pondicherry University Process-industry Accident Database). Multiple and overlapping accident records have been carefully eliminated and a search engine has been developed for retrieval of the records on the basis of appropriate classification. PUPAD doesn’t aim to replace or substitute the well established databases such as MHIDAS and MARS but, rather, aims to compliment them.

► A novel database covering over 8000 records of accidents in the chemical process industry is presented.
► A comparison indicates that the new database can complement the existing ones and help reduce the uncertainties associated with past accident analysis.
► The novelty and widespread utility of the work justifies its publication in the journal.