Research Programs

Advanced Nuclear Power (ANP) Program

Projects

Publications:

  1. M.J. Delaney, G.E. Apostolakis, “A Probabilistic Analysis of General Design Criterion 35 for a Gas-Cooled Fast Reactor,” Proceedings of ICAPP ‘04, Pittsburgh, PA, Vol. 90, June 13-17, 2004
  2. G. Jourdan, Using Risk-Based Regulations for Licensing Nuclear Power Plants: Case Study of the Gas-Cooled Fast Reactor, SM Thesis, MIT Nucl. Eng. Dept., February 200
  3. M. J. Delaney, Design Guidance of a Generation-IV Gas-Cooled Fast Reactor Emergency Core Cooling System, SM Thesis, MIT, April 2005
  4. C. Matos, Feasibility of Risk-Informed Regulation for Generation-IV Reactors, SM Thesis, May 2005
  5. Delaney, M.J., Apostolakis, G.E., and Driscoll, M.J., “Risk-Informed Design Guidance for Future Reactor Systems,” Nuclear Engineering and Design , 235:1537-1556, 2005

Investigators

  • G.E. Apostolakis
  • M.J. Driscoll
  • M.W. Golay
  • M. Delaney
  • G. Jourdan
  • C. Matos

Methods for Risk-Informing Future Reactor Regulation

New nuclear reactor concepts face many design and licensing challenges. A fundamental challenge when dealing with safety issues is how to handle the relevant uncertainties. Before the advent of Probabilistic Risk Assessment (PRA), these uncertainties were largely unquantified and the design and regulatory philosophy relied on the concept of defense in depth (DID) and large safety margins to ensure that accident frequencies were low. This “structuralist” approach to safety is embodied in the structure of the regulations. For example, in the United States, Title 10, Part 50 of the Code of Federal Regulations establishes the minimum design requirements for water-cooled reactors in Appendix A, “General Design Criteria for Nuclear Power Plants”. The General Design Criteria (GDC) require reactors to be designed with sufficient margin to assure that postulated accident sequences are protected against. The postulated accidents are also known as Design Basis Accidents (DBAs). The unquantified accident frequencies are addressed by protecting against DBAs and by meeting or exceeding the GDC.

PRA, which quantifies accident frequencies, has matured sufficiently since its introduction in 1975 so that the US Nuclear Regulatory Commission (USNRC) has started to use it in regulatory decision-making. These decisions are risk-informed rather than risk-based . This means that risk information is one input to an integrated decision-making process that also utilizes traditional requirements and safety philosophies, i.e., structuralist DID requirements. The quantification of uncertainties has led to the emergence of the “rationalist” model of DID, which asserts that DID is “the aggregate of provisions made to compensate for uncertainty and incompleteness in our knowledge of accident initiation and progression”

The focus of this work is to explore the use of risk information in the design of future reactors and, in particular, how the uncertainties could be handled in the licensing process.