Publication Date Nov 14, 2013 by Garrick B.J.
The early years in the development of probabilistic risk assessment (PRA).
Closing Plenary Session - Embedded Topical Meeting: Risk Management
November 14, 2013
Many excellent papers have been written on the evolution and history of probabilistic risk assessment. It is the purpose of this paper to put a little different spin on the usual history of PRA by making more visible the contributions of nuclear plant owners and operators, including their consultants and suppliers. It is not the intent to provide a detailed chronological history of the development of PRA; rather it is the desire to highlight selected milestones in the evolution of PRA with emphasis on risk management practices.
The events that led to the development of PRA were primarily related to the inadequacies of the early methods used to assess the safety of nuclear power plants. Early nuclear reactor safety analysis involved the defense in depth concept and a design basis accident approach to assure safety. Both concepts had their roots in the Manhattan Project and both have their merits, although this author has never been a proponent of the design basis accident approach. The primary shortcoming of the methods had to do with the absence of knowledge about the likelihood or frequency of severe accidents. In particular, quantitative methods were lacking for determining the safety margins of engineered barriers and safeguards to protect the plants. In the mid to late 1950s, reactor safety analysts (including this author) recognized the need to embrace the uncertainty sciences to better represent the risks involved in the operation of nuclear power plants. While there was growing concern about the lack of quantitative methods, there wasn’t much action until sometime later.