R564.1 Third Party Review of PRAISE-CANDU Probabilistic Fracture Mechanics Code

1.0 	Background

The CANDU industry has asserted that small safety margins associated with the analysis of A Large Break Loss of Coolant Accident (LBLOCA) as a Design Basis Accident (DBA) can be attributed to conservatism in the Traditional Safety Analysis (TSA) evaluation procedure. Industry has presented its case that the expected frequency of large pipe breaks is much lower than the assumed frequency attributed to DBAs. To support its position, industry undertook a COG project to develop a Composite Analytical Approach (CAA) to characterize, demonstrate and reduce conservatism in TSA for LBLOCA scenarios. The CAA is currently under review by CNSC staff.

One element of the CAA involved the development of a Probabilistic Fracture Mechanics (PFM) code, PRAISE-CANDU, to generate estimates of the frequency of primary system pipe failures which would lead to LOCAs. One of the intended functions of this code is to provide estimates of failure frequencies for piping of different sizes to identify which failures are expected to occur at frequencies lower than 1x10-5, which would support their reclassification to Beyond Design Basis Accidents (BDBAs). In theory a double-ended paper cutter failure of large diameter piping, which would stress safety systems and challenge post-accident fuel cooling, is far less probable than the failure of small diameter piping.

The concept of PFM is not new, but its application to nuclear piping systems has been approached with caution. The level of uncertainty associated with modelling parameters, lack of documentation, Quality Assurance control during software development and the variability in results achieved from different PFM codes have resulted in limited acceptance of PFM codes for regulatory decision making. There is an inherent difficulty in acceptance of PFM code given the complexity of some of the piping degradation mechanisms which may lead to a LOCA and the fact that no operational data exists for these low frequency events, thereby, making validation of code results difficult. No regulatory jurisdiction has ever accepted a PFM assessment result in isolation to justify regulatory decision making. There have only been limited examples of the adoption of PFM insights into Risk Informed Decision Making (RIDM) applications.

2.0	Objectives

To obtain an independent third-party evaluation of the PRAISE-CANDU Probabilistic Fracture Mechanics code development process carried out for the Composite Analytical Approach to address Large Break Loss of Coolant Accident safety margins.

3.0	Scope of Work

The successful contractor will carry out a technical evaluation of the CANDU Owner’s Group (COG) Reports prepared during the development of PRAISE-CANDU to provide an independent evaluation of:

•	The suitability of the PFM code degradation models to assess pipe failure frequencies due to plausible CANDU specific degradation mechanisms.
•	A comparison of the PRAISE-CANDU validation and verification (V&V) and benchmarking activities to public domain information describing V&V and benchmarking of other PFM codes, for example (but not necessarily limited to) PRO-LOCA, PASCAL, NURBIT, ProSACC, PRODIGAL and PROST. 
•	The extent to which PFM codes have been accepted in regulatory decision making in the United States, Europe and Japan.

 
The contractor shall address the following elements of PRAISE-CANDU in the project deliverables:

•	Limitations of the probabilistic simulation techniques 
•	Limitations of the uncertainty analysis techniques


4.0	Tasks to be Performed

4.1	Prepare a report, based upon a literature review, which documents the current level of acceptance of PFM codes in nuclear regulatory decision making in the United States, Europe and Japan.

4.2	Review the following COG reports generated during the development of the PRAISE-CANDU code:
•	COG-JP-4367-041-R0, FPBOC-WG Integrated Close-out Report
•	COG-11-2096, LOCA Break-Opening Characteristics for CANDU HTS Piping – Literature Review
•	COG-JP-4367-005, Revision 2, Specification for PRAISE-CANDU 1.0
•	COG-JP-4367-011, Revision 0, Theory Manual for PRAISE-CANDU 1.0
•	COG-JP-4367-046, Revision 0, PRAISE-CANDU Version 1.0 Pilot Study
•	COG-JP-4367-026, Verification and Validation Report for PRAISE-CANDU 1.0

Note: Bidders should make a request to the Contracting Authority for the redacted summaries of the above-referenced reports for the purpose of bid preparation. Un-redacted reports will be provided to the winning bidder once the security requirements have been met. A non-disclosure agreement may also be required at that time.

4.2.1	Assess the applicability of degradation assessment models used in PRAISE-CANDU (fatigue, primary water stress corrosion cracking and flow accelerated corrosion) for the materials, manufacturing process and operating conditions typical for CANDU nuclear piping systems. Identify potential gaps in the degradation models.

4.2.2	Assess whether the theory presented in COG-JP-4367-026 adequately describes the basis for the selection of input parameters to ensure that consistent results would be generated by different users.

4.2.3	Assess the suitability PRAISE-CANDU input parameters by answering the following questions for the degradation mechanisms considered:
1) Do the input parameters accurately describe the loading conditions which could lead to crack initiation and subsequent pipe breaks?
2) Do the input parameters accurately describe the environmental conditions which could lead to crack initiation and subsequent pipe breaks?

4.2.4	Assess the suitability of the methodology used to address uncertainties in input and output parameters.

4.3	Review CNSC staff comments generated during the development of PRAISE-CANDU code along with the developer’s response to those comments. Assess whether the developer has dispositioned all CNSC staff comments.

4.4	Identify potential gaps in the code V&V and benchmarking activities presented in COG-JP-4367-026 compared to similar activities undertaken for the development of other PFM codes including, but not necessarily limited to, PRO-LOCA, PASCAL, NURBIT, ProSACC, PRODIGAL and PROST. Propose an action plan that will address any gaps identified.