Logic based methods for dynamic risk assessment

Abstract For the origin of logic based methods to assess process plant risks, we go back to the first symposium on Loss Prevention in the United Kingdom in 1971, where concepts in part developed for nuclear power plant risk assessment and extended to process plant have been presented. From there we follow the developments in the 1980s and the benchmarks in Europe in the 1990s revealing the large differences in outcomes when various teams work out the risks of a same plant. It showed the uncertainties intrinsic to the methodology of that time. The last 2 decades have seen various kinds of improvements but also awareness that other factors, such as human failure and organizational ones are important to include. The last subchapter is highlighting approaches partly based on machine learning and artificial intelligence that will make use of “big data,” even enabling dynamic operational risk assessment.