APPLICATION OF BEST ESTIMATE ANALYSIS
AND UNCERTAINTY METHODOLOGY
IN NUCLEAR REACTOR SAFETY ANALYSIS AT AECL
Nikola K. Popov
, Amad Abdul-Razzak
, Romney B. Duffey
, Harve E. Sills
Atomic Energy of Canada Ltd.,
2251 Speakman Drive, Mississauga, Ontario, L5K 1B2, Canada
Abstract – The Canadian nuclear industry has
traditionally used a conservative approach in
performing licensing safety analysis for CANDU
nuclear reactor power plants world-wide.
computer codes used by the Canadian nuclear
industry for licensing safety analysis are mostly
based on best estimate models, whereas the
assumptions are selected as boundary values or
assumptions to yield very conservative results.
Similar to practice used in the US and in many
other countries, an alternative approach has been
promoted for the past several years based on the
best estimate and uncertainty methodology.
AECL developed the so-called Best Estimate
Analysis and Uncertainty (BEAU) methodology
for potential use in licensing applications for
CANDU nuclear reactor power plants in Canada
The BEAU methodology is consistent
with the Code Scaling, Applicability, and
Uncertainty methodology (CSAU) developed in
the early 1990s in the US.
This paper provides a summary of the BEAU
development effort at Atomic Energy of Canada
Ltd. and describes the key elements of the
The paper also presents a summary
of the pilot applications of the BEAU methodology
along with key results.
– nuclear reactor, safety analysis, best-
estimate, uncertainty analysis
Safety analysis requires use of computer programs,
Probabilistic Safety Analysis and Deterministic
These approaches require validated
methods that include specific statements on the
applicability and accuracy for the intended
application for a selected power plant and selected
To perform safety analysis with
acceptable level of credibility for a regulatory review,
it is necessary to consider the uncertainties in the
safety analysis itself, which includes modelling of the
important phenomena, physical representation of the
plant, equipment, and plant-specific data.
To date, the Limit of Operating Envelope (LOE)
approach has been the basis of safety analyses of
submissions, which requires that input assumptions
and parameters used for safety analyses be set
simultaneously to pessimistic or bounding values.
For certain accident scenarios, such analysis
assumptions can lead to predictions of small safety
margins and unrealistic results, which can obscure the
real margins and may limit operational flexibility.