In stressful situations such as severe accidents in nuclear power plants, operators need support tools to
ease decision making in the selection of accident management measures. Following the Three Mile Island
(TMI) accident in 1979, the first severe accident in a nuclear power plant, Accident Management Support
Tools (AMSTs) were extensively developed and installed in a number of nuclear power plants. Lessons
learned from the Fukushima accident highlighted the importance of accident management in mitigation
severe accidents and suggested the reconsideration of accident management programs, which in turn
created the need for AMSTs adaption and modernization.
This paper provides the first post-Fukushima comprehensive review of AMSTs, covering the particularities of all their elements, unlike other previous review papers which are limited to the study of individual fields, such as fault detection or decision-making support. Applications, advantages, and
disadvantages of various methods which can be used in the design of AMSTs are investigated, categorized
and compared in general well-known categories (Artificial neural networks, fuzzy logic, etc.). Moreover,
human factor related issues in implementation of AMSTs are introduced and discussed. It was concluded
that a modern AMST can provide vital information about the plant states, e.g. timing of critical events and
a quantitative estimation of important parameters, which cannot be provided by typical Severe Accident
Management Guidelines (SAMG). Nonetheless, it is emphasized that AMSTs should only have a supporting role in accident management, not replace SAMG.