Review of Deterministic Safety Analysis Methodologies for Beyond Design Basis Accidents

Abstract of the technical paper/presentation presented at:
International Conference on Topical Issues in Nuclear Installation Safety
June 29 – July 3, 2026

Prepared by:
M. Shawkat
Canadian Nuclear Safety Commission

Abstract:

Beyond Design Basis Accidents (BDBAs) in Nuclear Power Plants (NPPs) are low- frequency events that may result in more severe consequences than those considered during plant design. Mitigation of BDBAs is essential to minimize their radiological impact on both the population and the environment. All Canadian NPPs have implemented Severe Accident Management Programs (SAMPs), which outline strategies for plant personnel to halt or delay the progression of BDBAs. Analyzing these events is important to ensure the effectiveness of both engineered safety and non-safety systems employed in mitigation strategies.

Canadian Nuclear Safety Commission (CNSC), Regulatory Document REGDOC-2.4.1 (issued in 2014) details the requirements to perform deterministic safety analysis for all accident categories including BDBA. The objective of this paper is to assess REGDOC-2.4.1 requirements to ensure they cover all BDBA safety analysis aspects.

This paper reviews several methodologies for modeling and analyzing BDBAs in NPPs, with the objective of demonstrating the effectiveness of the SAMP mitigation measures and complementary design features. The review encompasses approaches proposed by the International Atomic Energy Agency (IAEA), Canadian Standards Association (CSA), and peer-reviewed scientific literature. It covers techniques for event identification, definition of acceptance criteria and figures-of-merit, input selection and assumptions, and treatment of uncertainties. These methodologies are then compared against the REGDOC-2.4.1 requirements for analyzing BDBA deterministically.

Probabilistic safety assessments with engineering judgment should be used to support identification of the BDBA scenarios. Deterministic BDBA analyses are required to simulate the relevant physical sequences expected to challenge the capabilities of the accident management mitigation strategies as BDBA progresses such as core heat-up, core relocation, hydrogen production, in-vessel and ex-vessel retention, and containment integrity. Qualitative acceptance criteria may include maintaining coolable core geometry for no or limited core damage under design extension conditions or achieving in-vessel retention for severe accidents with widespread core damage or core melt. Quantitative thresholds for demonstrating these criteria will vary depending on reactor design but parameters such as peak clad temperature, minimum coolant inventory, corium mass/geometry bounds, or containment pressure/time limits may be used.

The reviewed methodologies fall into two broad categories: deterministic and probabilistic (stochastic) modelling. Deterministic approaches use best-estimate values for input parameters, often based on operational data or nominal design specifications if there are no statistical data available. Sensitivity cases are usually used to address the uncertainties in the key parameters. Stochastic methods, by contrast, incorporate predefined ranges and distributions for the input key parameters then use propagation techniques to simulate and characterize the spectrum of possible outcomes. While stochastic methods offer broader coverage and explicit quantification of uncertainty, deterministic methods can achieve comparable robustness when accompanied by a structured uncertainty/sensitivity evaluation. Stochastic modelling can also be used to assess the adequacy of a deterministic approach.

Common features across the reviewed BDBA safety analysis methodologies include:

1. Use of best-estimate, physically realistic inputs and assumptions, with recognition that no standardized numeric definition exists
2. Realistic initial and boundary conditions, with conservative assumptions applied when data is uncertain or incomplete
3. Systematic treatment of parameter uncertainties
4. Avoidance of reliance on the single-failure criterion, which is limited to design-basis analysis
5. Identification and mitigation of cliff-edge effects
6. Consideration of credited performance of Systems, Structures, and Components (SSCs) beyond their original design basis, including safety, non-safety, and temporary systems

REGDOC-2.4.1 does not prescribe a specific methodology for BDBAs safety analysis but instead sets high-level requirements that licensee must meet. These requirements align with the common features identified in the reviewed approaches.

To obtain a copy of the abstract’s document, please contact us at info@cnsc-ccsn.gc.ca or call 613-995-5894 or 1-800-668-5284 (in Canada). When contacting us, please provide the title and date of the abstract.