Language selection


Research report summaries 2007–2008

Archived information

Archived information is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please contact us to request a format other than those available.

Contractors' reports are only available in the language in which they are submitted to the Canadian Nuclear safety Commission (CNSC).

Top of page

RSP-0223 – Basis for the development of licensing requirements to prepare a site for new nuclear installations in Canada

The site evaluation requirements contained in the regulatory frameworks of the CNSC, the United States Nuclear Regulatory Commission (USNRC), Finland's Radiation and Nuclear Safety Authority (STUK) and the International Atomic Energy Agency's (IAEA) Safety Requirements and Safety Requirements series have been reviewed. Supplied documents from France's Nuclear Safety Authority have also been reviewed for any requirements applied to nuclear power plant Site Evaluation programs.

These reviews identified similarities and differences in site evaluation requirements from these jurisdictions that are the basis of observations and recommendations from the contractor to the CNSC regarding the development of requirements for a Licence to Prepare Site for new nuclear installations in Canada.

Top of page

RSP-0224 – Evaluation of a new approach for the assessment and disposition of pressure tube crevice corrosion flaws

The proposed flaw assessment procedure for crevice corrosion flaws at bearing pad contacts in CANDU pressure tubes has been evaluated with the following aspects in mind: conservatism of the mechanics analysis; statistics of the flaw dimensions; the corrosion mechanism and its propensity for self-arrest or other desirable features.

I find that the mechanics analysis cannot be faulted in terms of its conservatism. I would have liked to see a more explicit statement that since the peak local stress at the worst crevice corrosion flaws is very close to the threshold stress for DHC at a planar surface, considerations of ‘notch stiffening' dictate that we will not approach the local threshold which must be higher even than that.

There do not appear to be any hidden hazards in the flaw size data. However, the possibility of multiple landings of a bearing pad on one site has not been analyzed in a fully convincing way. I believe a more convincing case can be made for the improbability of such a mechanism.

I find that the prevailing industry description of the corrosion mechanism leaves something to be desired. There is too much focus on solid-state transport concepts and not enough on the propensity for Zr to dissolve (albeit temporarily) in alkaline solutions as a zirconate species. I cite literature from the geochemistry community to illustrate that such dissolution is taken for granted amongst a certain school of chemists. Regarding the specific effect of Li cations, I find it odd that the industry has been prepared to deny altogether the role of alkalinity and complexation, or even dissolution, in the corrosion of Zr in LiOH solutions, simply because LiOH is more aggressive to a particular alloy (not necessarily all alloys) than NaOH or KOH. Analogies in the corrosion of aluminum come to mind, yet no-one would deny that aluminum corrodes in alkaline solutions because it is solubilized as aluminate.

I consider that it would be beneficial to the industry to acknowledge that there is a role of dissolution in the crevice corrosion, since this would argue for a kind of diffusion control that prevents ‘sharpening' of the flaws.

We need to explain, though, why LiOH solutions appear to be more aggressive than NaOH or KOH solutions of similar room-temperature pH, at least to Zircaloys – the effect on Zr-Nb is not so clear. Authors have ascribed this to the small size of the Li cation, that supposedly enables it to enter the zirconium oxide lattice. It is not clear how this would produce porosity or otherwise promote the high observed corrosion rates. I mention an alternative speculation, based on recent research on zirconates as CO2 storage media, that might appeal to the differing stability of Li, Na and K solid zirconates. I do not find any evidence that during one residence of a bearing pad the corrosion reaction is self-limiting in any significant way. Old Westinghouse data using model bearing-pad contacts have sometimes been misinterpreted – in fact they show that the corrosion rate becomes nearly constant with time.

Top of page

RSP-0225 – International Common Cause Data Exchange (ICDE) Project: Preparation of data to be submitted to the clearing house

This report describes the activities and methods used for Canada's data submission to the International Common Cause Data Exchange Project (ICDE) Project. The International ICDE Project was established by the Organization for Economic Cooperation and Development (OECD) to encourage multilateral cooperation in the collection and analysis of data relating to Common Cause Failure (CCF) events at nuclear power stations. The CNSC established its own project to collect data on Canadian nuclear power plants and fulfill its obligations under the International ICDE agreement. This is the third CNSC data collection campaign for contribution of data to the ICDE. The report describes the data sources used for collecting data from Canadian nuclear power plants, the methods employed to analyze the data and code it into the ICDE database and some summary statistics on the data.

Top of page

RSP-0226 – Type 2 inspection checklists development project

As part of the CNSC's Compliance Program, the CNSC conducts systematic evaluations of all Canadian nuclear power stations radiation protection programs. The CNSC radiation specialist and the site inspector responsible for the radiation protection need specific and efficient tools to conduct this surveillance.

The Director of Nuclear Substance Regulations (DNSR) has a proven compliance system called LOUIS. This computerize tool provides direction for achieving a consistent and uniform plan. It evaluates, follows trends, identifies and maintains records of many activities such as document review; Type I inspections; and Type II inspections. By the end of 2007, the Radiation Protection Division (RPD) anticipates using the DNSR LOUIS design for its surveillance of the nuclear power station Radiation Protection (RP) Programs in Canada.

The objective of this project is to assist the CNSC in developing compliance tools to be used by radiation specialist/site inspectors to assess nuclear power station radiation protection program.

Expertise Omega Inc. was chosen to develop type II checklists for Point Lepreau and Gentilly, which will be used as part of its compliance program by the CNSC radiation specialists and the site inspectors to monitor closely the activities of the nuclear power station radiation protection program.

The first task was to establish the scope of inspections in the radiation protection domain before developing any type 2 inspection checklists. Having done this, the second challenge was to elaborate a systematic approach to ensure consistency in the development no matter to what nuclear plants it will apply and keep inspection checklists as free as possible from site procedure revisions.

The resulting radiation protection model represents the occupational and public radiation safety in 10 modules that characterize all the areas of radiation protection that must be inspected or verified by the CNSC inspectors. The model has served as a template to develop a set of type 2 inspection procedures for both Gentilly 2 and Point Lepreau sites.

The proposed model and the resulting deliverables will require validation from the CNSC experts especially on the identification of the regulations that apply and the characteristics of the model if ever some aspects of the radiation protection domain have not been entirely covered. This has to be done prior to initiate any further development on the other Canadian utilities.

Top of page

RSP-0228 – OECD Piping Failure Data Exchange (OPDE) Project: Results and insight into the first phase

Structural integrity of piping systems is important to plant safety and operability. In recognition of this, information on degradation and failure of piping components and systems is collected and evaluated by regulatory agencies, international organizations (e.g., OECD/NEA and IAEA) and industry organizations worldwide to establish systematic feedback to reactor regulation and research and development programs associated with non-destructive examination (NDE) technology, in-service inspection (ISI) programs, leak-before-break evaluations, risk-informed ISI, and probabilistic safety assessment (PSA) applications involving passive component reliability.

Several OECD member countries have agreed to establish the OECD-NEA Piping Failure Data Exchange Project (OECD-NEA OPDE) to encourage multilateral co-operation in the collection and analysis of data related to piping failure events in nuclear power plants.

The project was formally launched in May 2002 under the auspices of the OECD/NEA. Organizations producing or regulating more than 80 percent of nuclear energy generation worldwide contribute data to the OECD-NEA OPDE data project. Currently twelve countries have signed the OECD OPDE agreement (Belgium, Canada, Czech Republic, Finland, France, Germany, Korea, Japan, Spain, Sweden, Switzerland and United States). The OECD Pipe Failure Data Exchange project was established to compile a database containing a list of records that pertained to pipe failures in nuclear power plants (NPP). At present, the database contains 3547 records to which twelve participating countries contributed. The database was designed for multiple activities, and although its current goals could be subject to expansion in the future, the database supports the following items:

  1. trend analysis, including aging analysis
  2. statistical analyses to determine pipe failure rates and rupture frequencies for use in risk-informed activities (e.g., loss-of-coolant-accident frequency assessment, internal flooding initiating frequency assessment, high-energy-line-break frequency assessment, RI-ISI and risk assessment, etc.)
  3. source of data parameters for input to probabilistic fracture mechanics codes
  4. degradation mechanism analysis (DMA) in RI-ISI applications
  5. development of defenses against recurring (e.g., systematic) pipe failures
  6. exchange of service data in order to pinpoint potential generic implications of a specific, significant pipe failure

This report describes the current status of the OECD-NEA OPDE database after three years of operation, and gives some insights based on piping events collected into the database. The report contains a number of tables and figures that outline a selection the current content that can be obtained from these records, with selected data ranging from broad (i.e., level of participation in the database from each member country) to very specific (i.e., plant operational state at time of pipe failure discovery for CANDU reactors). In some cases, graphs are also used to compare data from CANDU reactors to the data from all other NPPs. From analysis of the data found in this report, as well as the data not covered in these pages, the specialist can derive applications which can be used anywhere from probabilistic safety assessment to material degradation analysis and to code development for their respective NPP.

Top of page

RSP-0229 – Review of the analysis basis for Pickering B LBLOCA BEAU analysis

In the recent years, the Canadian CANDU industry has been developing the best estimate analysis and uncertainty (BEAU) methodology as an alternate safety analysis tool. This methodology is meant to be used when margins, predicted using the traditional conservative methodologies, are small. Several prototypical applications of BEAU have been undertaken to date: in particular, the first full application of BEAU to LBLOCA was performed for Darlington NPP in 2003. The industry is now collaborating in the application of BEAU to Pickering Units 4 to 8 LBLOCA analysis including the Best Estimate Analysis and Uncertainty (BEAU) methodology in Canada.

The motivation to use best estimate (BE) system analysis codes and to calculate the uncertainty of the final results is compelling. When the calculated values of key parameters, such as maximum fuel sheath temperature, are known with great confidence for limiting transients, plant operators may then fully exploit a large number of techniques to maximize plant operational efficiencies, power output and plant operational cycles. These capabilities, in turn, enable a utility to reduce the cost of operating a plant.

A review is performed for five reports:

  • Guidelines for Application of the Best Estimate Analysis and Uncertainty (BEAU) Methodology to Design Basis Accident (DBA) Licensing Analysis by R.G. Huget and Amad Abdul-Razzak; COG-06-9012 Rev 0, November 2006
  • Pickering B Large Break LOCA BEAU Analysis by Hong Xie and Chick-Bong So; Technical Basis Document G0111/RP/019, November 10, 2006
  • Proposed Code Uncertainty Estimation Methodology and Results of Demonstration Application by B. McLaughlin and J. Pascoe; NSS Report G0209/RP/002 R01, April 2007
  • Initial Phenomena and Key Parameter Identification and Ranking Table by A.C. Dennier, C. Wong and S. Tantirige; NSS Report G0242/RP/002 R01, April 2007
  • Darlington NGS Fuel Channel Integrity Phenomena Identification and Ranking Table for a Critical Break LOCA by H.E. Sills; NSS Report G0198/RP/001 R0), February 2007. The main findings are set out in this report.

Top of page

RSP-0230 – Modeling of molten-fuel-moderator interactions

CANDU reactors are pressurized heavy-water moderated and cooled reactor designs. During commissioning of nuclear power plants a range of possible accidents must be considered to assure the plants' robust design. Consider a complete channel blockage in the CANDU reactor. Such an extreme flow blockage event would result in fuel overheating, pressure tube failure, partial melting of fuel rods and possible molten fuel-moderator interactions (MFMI). The MFMI phenomenon would occur immediately after tube rupture, and would involve a mixture of steam, hydrogen and molten fuel being ejected into the surrounding moderator water in the form of a high-pressure vapor bubble mixture. This bubble mixture would accelerate the surrounding denser water, causing interfacial mixing due to hydrodynamic instabilities at the interface. As a result of these interfacial instabilities, water is entrained into the growing two-phase bubble mixture with the attendant mass and heat transfer – e.g., water vaporization, fuel oxidation.

A comprehensive model is developed to investigate these complex phenomena resulting from a postulated complete flow blockage and complete pressure tube failure. This dynamic model serves as a baseline to characterize the pressure response due to a pressure tube rupture and the associated MFMI phenomena. Theoretical modeling of these interrelated complex phenomena is not known a priori and therefore a semi-empirical approach is adopted. Consequently, experimental work is being proposed as part of the thesis work to verify key hypotheses regarding these interfacial fluid instabilities, such as the entrainment fraction into the rapidly expanding bubble.

Top of page

RSP-0231 – Numerical model of the thermal and mechanical behaviour of a CANDU 37-element bundle

The objective of the present contract is to develop a reliable numerical model, which is capable of predicting macroscopic thermal and mechanical behavior of the fuel bundle assembly, and suitable for simulating fuel bundle to pressure tube interactions under different accident conditions. This objective was accomplished in three stages.

The first stage involved development and verification of a finite element-based numerical model for predicting temperature distributions and thermal-induced barreling deformations of a complete 37-element fuel bundle. For structural analysis, a 3D beam model was developed which was able to take into account axial and radial temperature variations in fuel elements and endplates, and use them to compute thermal-included macroscopic deformations. On the other hand, a 2D planar heat transfer model was developed to calculate the asymmetric temperature distribution over the mid-section of fuel elements due to effects of neutron flux gradient and uneven coolant conditions. The results from this 2D planar heat transfer model were then combined with the previously developed 2D axi-symmetric temperature solution to form a complete 3D approximation of the temperature field in fuel bundle. Both the structural and heat transfer analysis models were verified against available experimental and numerical solutions.

The second stage of the present contract involved application of the developed bundle model to predict the static response of a 37-element fuel bundle to various steady-state power and coolant conditions. Formulation of the heat transfer problem required extraction of asymmetric coolant conditions from ASSERT output files, adjustment of power generation rate to account for neutron flux gradients, and consideration of axial temperature variation using axial scaling functions constructed from the previous axi-symmetric temperature solution. The endplates were assumed to be at the coolant temperature, which was normally variable over the bundle cross-section. The approximate 3D temperature field was then applied to the beam finite element model to predict bundle deformation under two different steady-state thermal conditions. These solutions indicated that all the fuel elements bowed in the anticipated direction and the overall pattern of the deformation was consistent with the temperature distribution. Element-to-element interactions through the spacer pads played a very important role in maintaining the shape of bundle cross-section. If these interactions were not considered, the maximum barrelling deformation of each fuel element occurred at the mid-span and the magnitude of deformation was nearly proportional to the temperature gradient over the fuel element, suggesting that the temperature gradients in fuel rods are the primary driving force for bundle barrelling.

In the third stage, the steady-state solution procedure was extended to handle transient thermal situations, such as LOCA conditions. This required implementation of the nonlinear visco-plastic material model for Zircaloy-4 into the beam finite element and generalization of the axial scaling technique, which combined the 2D planar and axisymmetric heat transfer results to form a full 3D transient temperature field of the entire fuel bundle. An investigation was conducted to identify the most important mechanism for barreling deformation of the fuel bundle and it indicated that although the differential on rod elongations played a role on bundle barreling, the most important driving force was the temperature gradients in the fuel rods.

Numerical studies presented in this report demonstrated that the finite, element-based fuel bundle model is reliable and predicts credible behaviour in response to accident type conditions in a fuel channel. The intended application of this model is to assess pressure tube integrity in large break loss of coolant accidents. An evaluation of the completeness of the models and how the application should be performed is the next step towards this goal.

Top of page

RSP-0232 – Application of the earned quality method to the probabilistic assessment of leak rate through steam generator tubes research project

This report describes how the earned quality method (EQM) is used to measure and to control the quality of the deliverables of the Probabilistic Assessment Model of Leak Rates through Steam Generator Tubes Project. For this project, implementing EQM is done in four steps. In the first step, an overall project vision is obtained using the logical framework approach (LFA). In the second step, a quality breakdown structure (QBS) is developed for each of the project deliverables to identify their quality attributes and criteria, and for each criterion a quality assessment function is proposed. In the third step, a work breakdown structure (WBS) is developed to identify all the tasks required to produce the deliverables according to their cost, time and quality objectives. In the last step, the activities of the WBS are linked to at least one quality criterion of the QBS to measure the earned quality. The earned quality is used to control quality throughout the project. In the first step, the LFA is used to provide a concise and complete overview and understanding of the project. The result of the implementation of the LFA is a matrix called the logframe, which summarizes why the project should be undertaken, what it intends to do, what are the deliverables, how the deliverables will be obtained and what are the risk events that can jeopardize the successful completion of the project. The organizational value of the project is defined as follow:

  1. to be able to assess steam generator tube inspection strategies
  2. to be able to predict all significant leaks

To achieve the above organizational value, the scope of the project was established and the deliverables identified as follow:

  1. experimental data on leak rates from fretted tubes, from pitting flaws and from crack flaws (December 2008)
  2. a probabilistic assessment model (December 2008)
  3. progress reports (each month)
  4. status reports (every 3 months)
  5. a seminar (October 2008)
  6. a final report (December 2008)

For each of the above project deliverables, objectives in terms of cost, time and quality, were determined. In the second step of the implementation of EQM, the QBS is used to identify the quality criteria for each of the project deliverable. The QBS helps the project team members in developing a set of meaningful and measurable quality criteria for each deliverable as well as their evaluation protocol.

The quality attributes and the criteria of the experimental data QBS are defined as follow:

Validity – the degree to which (i) bias are absent from the data, that is, there is no systematic bias due to known or unknown causes that makes the data lean more in one direction than the other (accuracy), (ii) the data are reliable, that is, they have little random or unstable error, or low variance (reliability), and the data cannot be affected by the person gathering them nor the time or location of their gathering (stability)

Relevancy – the degree to which the data are sufficient but not excessive (parsimony) and at the right level of detail (meaningfulness) to be useful for the probabilistic assessment model

Practicality – the degree to which the data can be easily obtained (convenience), easy to understand using Excel spreadsheets, charts, raw data expressed in millivolts and milliamperes (interpretability) and do not cause changes or deteriorations to the system being measured (noninvasiveness)

The two attributes the probabilistic assessment model QBS are described below:

Validity – the extent to which the model assesses what actually have to be assessed (absence of any significant bias). It includes: i) content validity (the model takes into account all known factors significantly influencing failures, leakages, the probability of detection, and flaw growth distributions of the steam generator tubes), ii) consistency (the model can predict successfully actual results), iii) reliability (the model produces reliable results, that is, the results of the models have little random or unstable error, or low variance), iv) stability (the results obtained from the model cannot be affected by the person using it nor the time of its use), v) adaptability (the model is not significantly influenced by changes to the environment of the system or its operation, for example using the reactors over their operating capacity for a long period of time)

Practicality – the degree to which the model is easy to use (convenience), to understand (interpretability) and to update (flexibility).

The following two attributes compose the QBS of the seminar:

Validity – the degree to which the seminar presents all the essential elements of the probabilistic assessment model (content validity), and the degree to which it presents and discusses all the actual results of the model (accuracy)

Usefulness – the degree to which the seminar is well structured and easy to understand (comprehensibility), the degree to which the seminar material covers all the essential elements of the model (completeness), and contains only the essential elements of the model required to understand it (parsimony)

The QBS of the reports contains the following two attributes:

Practicality – the degree to which the report is easy to read and to understand (readability), and in the appropriate Word and Adobe format in accordance with the format and style requirements as specified in the tendering document (convenience)

Relevancy – the report includes all the required information (completeness), contains only the essential information about the project required to understand it (parsimony) and it conveys the information in a precise and accurate manner (content validity).

Then, in the third step, the activities required to produce the deliverables according to their objectives were identified using the WBS. Four main project activities compose the WBS as follows: generate experimental data, develop a probabilistic assessment model, organize a seminar, and produce the reports. The WBS helps the project team members in identifying all the tasks that must be done to achieve the deliverables as well as their duration and the resources required to realize them. Then a project schedule using Gantt charts was developed using Microsoft Project.

Finally, in the last step, each quality criteria of the QBS are linked to the work elements of the WBS using an Excel spreadsheet. The spreadsheet is used to assessed and controlled quality throughout the project life cycle.

Top of page

RSP-0233 – Pickering B Unit 6 probabilistic assessment

This document presents a review of Pickering B Unit 6 Probabilistic Blister Susceptibility Assessment, NSS File No: P0775/RP/001, Revision 00. We have included for completeness and convenience our preliminary report, the response by NSS and OPG, our follow up questions on any remaining issues, and their response following our June 25, 2007 meeting. These are attached as four appendices to this report.

In the probabilistic analysis of blister susceptibility, the following four elements are important:

  1. spacer movement model describing the probabilities of movement, direction of movement and magnitude of movement
  2. hydrogen ingress model combining the initial Hydrogen, time dependent Deuterium uptake and a random variable describing the difference between TSS and scrap measurement
  3. time to contact model describing the time and location of PT-CT contact in a given fuel channel over a particular time interval (in general, a time-dependent, random process)
  4. the three models are combined in a Monte Carlo simulation model, which simulates 60,000 reactor cores to estimate the probability of blister susceptibility in the evaluation period

Our review was concerned with evaluating the adequacy of NSS' methodology in addressing the probabilistic aspects of these models, their methods and the data used for obtaining estimates.

Each of these four elements is considered in our review below followed by concluding remarks.

  1. Spacer Movement Model – The NSS has clarified that the spacer movement probability model is based on maximum likelihood analysis (MLA) which is taken as the best estimate. Although the MLA provides estimates of variance associated with the model parameters, these are not utilized in the analysis. We have commented that in such cases the best estimate is generally based on an upper quantile (e.g., 95 percent fractile) of a random variable. We would have preferred that NSS include uncertainties in the MLA estimates in order to understand their sensitivity to final results, especially when the purpose of the analysis is probabilistic assessment. However, we understand that in the performance-based regulation regime it is the prerogative of the NSS analysts to define their best estimates and it is up to the CNSC to accept them. We do not wish to add to this point anymore.
  2. Hydrogen Ingress Model – We have noted that the Hydrogen excess is modelled as a time independent random variable. In the D2-uptake model, NSS calculations show that the effect of keeping a highly correlated covariate, Log HH/T, is to inflate the probability of blister formation by approximately 25 percent for 13 channels. Although it appears to be conservative in this particular case, it is not prudent to keep a highly correlated covariate in a regression model, because of the difficulty in interpreting estimated parameters. We note from Figure 3 (Appendix 4) of the response that the residuals depart significantly from normality, especially in the tail region.
  3. Time to PT-CT Contact Model – NSS developed a regression-based model to predict the time, location and spread of PT-CT contact for a given configuration of spacers in a fuel channel. We are not in a position to evaluate the statistical validity of this model without access to the CDEPTH code and the numerical results used to generate these regression models. It also not clear whether or not sampling uncertainty associated with the regression coefficients has been accounted in the analysis. We believe that NSS has not included uncertainties, as reflected by their response to our Q19 in their overhead presentation when they made a general statement: CSA N285.8-05 does not require the calculation of upper bounds taking into account the uncertainty of the input parameters”.
  4. Simulation Model – In our view, the probabilistic assessment ought to demonstrate that the probability of blister susceptibility in any fuel channel in the reactor core is below a certain limit (COG 1065). In a technical sense, the cumulative probability of the first event (of blister susceptibility) during the evaluation period should be sufficiently small, the reason being that blister susceptibility in multiple fuel channels is undesirable. The probability of at least one event can be related to the expected number of occurrences, which if taken over a reactor full power year is defined as a frequency in CSA N285.8. In response to our questions about the meaning of frequency, NSS has provided five possible interpretations in Appendix 4, incorrectly ascribing the first (a) to the reviewers. In fact, only interpretation (b) is relevant and we don't believe that any additional subjective interpretations are warranted. The CDF plotted in Figure 4 (Appendix 4) is shown as a continuous curve, though it is an interpolated curve over a few points in time. The time points used in simulation should be shown in Figure 4, and explanation provided about the method of interpolation. Figure 5 (Appendix 4) shows the PDF of the time to first blister in the core, but the PDF is a meaningless quantity. Its integral, which is the probability, is of essence in this context. In Appendix 4, it is still incorrectly stated that The probability density of the time to first blister in the core is equivalent to frequency. . . In probabilistic assessments, an upper bound on failure frequency is required to quantify the impact of uncertainties associated with input parameters. Otherwise it becomes a mean estimate, which NSS refers to as the best estimate. NSS states that CSA N285.8 does not mention a need to estimate an upper bound on the failure frequency. We believe that the standard specifies probability targets in an ideal sense. Since a probabilistic analysis is based on simplified assumptions and limited data, it is a good idea to include uncertainty bands together with best estimates. Table 2 of the original submission lists only those channels with the 13 highest probability of exceeding BFT at 181,000 EFPH (we interpret this as cumulative probability from 155,000 EFPH to 181,000 EFPH). On several occasions we requested probabilities for all 380 channels. Following the meeting of June 25, we were advised that probabilities of exceeding BFT are zero for all remaining channels. This should have been stated explicitly in the report. Even so, given the large number of remaining channels, the cumulative probability of exceeding BFT somewhere among these channels could be appreciable in view of round off error. We would have preferred to be able to confirm that this was not the case. The random number generation algorithms used in the Excel should be verified, given the known shortcomings with the simulation algorithms in Excel having a short recurrence cycle.

Concluding remarks: In addition to the above specific comments, the reviewers would recommend that NSS rewrite the report stating up front the assumptions, models, including their structure (for instance, sometimes normal assumptions are in place, other times truncated normals), data availability, simulation methods and how data were collected. Then state the precise problem, their conclusions are and then go into the analysis, making the report a self-contained document. References to external calculation notes should be reduced to a minimum and any necessary information should be included in the report. Then this would make the report easier to read and evaluate.

Finally, we should mention that the error associated with CDEPTH code has not been discussed in the report. We understand that this was not within the scope of the assessment. Nonetheless, it has an impact on the errors in spacer movement and time to contact analysis models because the code output is used as input in regression models and simulations. We believe this is an important issue that the CNSC might explore further.

Top of page

RSP-0234 – CANDU reactor core neutronic design: Design requirements and standards

It has been recognized that there are weaknesses in the standards (or equivalently policies/requirements/guidance) that apply to neutronic design and operation. New standards would help with the licensing of new reactors and might also be of help with the regulation of existing reactors. This study looks at current practice in terms of what has emanated from the regulator as well as the industry. The Canadian approach is compared with recommendations from the International Atomic Energy Agency. The study looks at the way standards for neutronic design are applied in other countries with the emphasis on the U.S. where the tradition has been to be very prescriptive. As a result of this survey it was concluded that there should be a more systematic and complete approach than the one that exists. The CNSC can do this by using the disparate parts of the current set of Canadian standards, using international standards, and by working with stakeholders. The policies/requirements/guidance for neutronic design should, at the very least, deal with shutdown systems; inherent reactivity, e.g., addressing the need for a negative power coefficient and limiting the amount of reactivity insertion possible; and the impact on fuel behavior of transients/accidents.

Top of page

RSP-0235 – Identification and evaluation of existing international standards and guidance on decommissioning

The Canadian Standards Association (CSA) in consultation with the CNSC and the Canadian nuclear industry has undertaken the development of a CSA standard on decommissioning. The standard is designated as N294. At the initial meeting of the N294 Executive Committee, the decision was made that the production of a standard on which the CNSC could rely with confidence, would benefit from an understanding of existing international standards and guidance documentation. An action was undertaken to conduct a review and evaluation of available documentation through a consulting contract.

Top of page

RSP-0236 – Piping failure frequency analysis using OECD/NEA data

The accurate estimation of piping failure frequency is an important task to support the probabilistic risk assessment and risk-informed in-service inspection of nuclear power plants. In particular, in the planning of inspection and maintenance programs, question arises about the effect of plant aging on the pipe failure frequency. Naturally, if the failure rate were found to be increasing with aging due to progressive deterioration, more stringent inspection and replacement programs would be required to maintain the plant safety.

In the literature, several probabilistic models have been reported to quantify the piping failure frequency, and they have been applied to data contained in OPDE database. In most of the previous studies, the pipe failure rate is considered as a constant and the effect of aging is not addressed in a systematic manner.

This report proposes a new stochastic point process model that incorporates both the time dependent trend and plant specific (or cohort) effects on the failure rate. A rigorous likelihood based statistical method is proposed for estimating the model parameters. The report presents a case study to analyze PWR and BWR Class 1 pipe failure data given in the OPDE database. The original results presented in the report show that the consideration of both aging and cohort effect is important.

The proposed analysis of the pipe failure frequency can be utilized by the Canadian nuclear industry as input to reliability analysis of piping systems, probabilistic safety analysis and planning for refurbishment of existing plants. It is particularly important since Canadian data contained in the OPDE database are so sparse that they preclude a rigorous statistical analysis of the failure rate in Canadian nuclear power plants (NPPs). It is recommended to augment the data collection and recording process in the Canadian NPPs in order to develop better probabilistic models of pipe failure rate and improve the risk assessment.

Top of page

RSP-0237 – Use and location of vehicle portal radiation monitors in Canada

This project was initiated by the CNSC in response to the implementation of the IAEA Code of Conduct, and the IAEA Action Plan on the Safety and Security of Radioactive Sources. There was a perceived need to become more fully aware of the number and location of radiation portal monitors (RPM) in Canada. The project set out to obtain lists of RPM users from the primary manufacturers and follow that up with a questionnaire that would bring forth data dealing with the training, experience, procedures and so on of the RPM users. However, none of the manufacturers were willing to supply their customer's identities or other sensitive information. The project was then refocused into an extensive consulting/fact finding mode. Since the author has thirty two years of experience in dealing with radioactive materials and very specifically RPMs and their operations, this was a valid course of action. The following points were the most prominent:

  1. A portion of RPM users, upon receiving an alarm simply turn the vehicle around and send it on its way, without dealing with the radioactive source in any way.
  2. There are, in effect, no standards of operation, training, response or use of RPMs in Canada. Basically, each RPM owner does what is good in his own eyes.
  3. Consideration should be given to strongly recommending that all RPM users be given radiation safety training that specifically addresses the work they will be asked to do; particularly nuclear source identification, recovery and handling.
  4. Consideration should be given to offering a fast, easy, cheap and viable method of reporting and disposing of the radioactive materials found by RPMs.

Top of page

RSP-0238 – Applicability of multidimensional methods to prediction of flow and void distributions in CANDU headers

The objective of this report was to assess the feasibility of current CFD methods to assist future safety analyses aimed at predicting header behavior during selected phases of postulated LOCA events in CANDU reactor systems. The overall findings and recommendations documented in the main body of the report can be divided into two groups:

  1. Advantages of developing CMFD technology in Canada for application in reactor safety analysis.
  2. Recommended approach to maximize the gains of using CMFD methodology for the analysis of CANDU headers during LOCA.

A summary of the major issues which have been identified for each group is given below. Detailed information in support of each issue can be found in the full text of the report. Specific technical examples shown in the report are mainly related to the modeling of LOCA events in the header/feeder system used in the CANDU reactors.

1. Advantages of Developing CMFD Technology in Canada for Application in Reactor Safety Analysis

1.A General technical issues:
  • Advance the understanding of dominant physical phenomena and provide rationale for scaling – The use of a multidimensional modeling framework for two-phase flows and heat transfer is inherently associated with proper identification of flow topology and local interfacial interactions. This, in turn, leads to identification of relative importance of individual local phenomena and the roles they are expected to play in a given class of problems or situations. Consequently, appropriate scoping studies and scaling analyses can be performed to both guide model formulation and assess the results of predictions.
  • Verify the consistency of formulation of two-phase flow models – The formulation of both the framework of the governing equations and the associated closure laws provides the modeler with multiple opportunities to investigate the correctness of various modeling concepts and assumptions and their consistency with the range of modeled physical phenomena and the mathematical framework used for this purpose.
  • Develop state-of-the-art methodology for CANDU reactor technology – Ambitious efforts and new initiatives have already been undertaken in various countries around the world to develop advanced modeling and simulation capabilities for nuclear reactor thermal-hydraulics and safety. Examples include Japan, France, South Korea and the United States. Their ranks are expected to grow quickly in the future, especially to include countries where a fast growth of the use of nuclear energy is the only solution to their energy needs (e.g., India and China). Both short-term and long-terms gains are envisioned as a result of those efforts. The availability of advanced and accurate predictive capabilities of reactor system response under accident conditions is likely to become a must for any successful nuclear energy initiatives of the future, including the continuation of, and advancements in, the CANDU technology.
1-B. Safety analyses for the current generation of CANDU reactors:
  • Improve the understanding of two-phase flow phenomena governing header/feeder response by performing CMFD simulations using the existing experimental data as a reference – Several expensive experiments were performed in Canada in the past aimed at understanding the performance of CANDU headers and feeders under abnormal conditions. However, the use to date of the data acquired from those experiments for improving our understanding of the observed fundamental two-phase flow phenomena, and for using the available data base to develop a general framework for mechanistically-based computer simulations has apparently been somewhat limited. The existing evidence worldwide (including, but not limited to, nuclear industry) clearly shows that the efforts behind developing advanced computational capabilities normally require the level of investment which is significantly lower than that associated with large-scale experiments, and that the learning curve quickly gets steeper as soon as the initial experience is acquired. Thus, starting a well-organized CMFD program has the potential of becoming a cost-effective approach to maximize the benefits of the previous experimental works.
  • Perform CMFD multidimensional simulations to assess the importance of local effects on the overall predictions. This approach can enhance the confidence behind system codes' predictions (sensitivity of results to averaging, improved methods of averaging) – The models employed by system codes typically involve several simplifying assumptions and the use of adjustable constants. Whereas such codes have served, and will continue serving, as important tools in reactor safety analyses, the validity and accuracy of their predictions are greatly limited by the uncertainties arising from the choice of the various available degrees of freedom built into the models. One way to reduce the level of such uncertainties is to perform comparisons between the system code predictions and the results obtained using CMFD models and solvers. Whereas, for obvious reasons, such comparisons are practically feasible for selected system components rather than for full systems, the usefulness of such an approach for reactor simulations has already been demonstrated quite extensively before.
  • Use CMFD to assess the effect of system aging on the Htr thermal-hydraulic and safety characteristics of the refurbished CANDU reactors – The increasing demand for electricity will likely result in bringing back to full-scale operation those CANDU reactors that have been shutdown over the recent years for inspection and maintenance. Although the refurbished reactors will not experience any major design changes, the effect of prior operation and system aging will have to be evaluated to provide the necessary input concerning the anticipated reactor performance in the future. Whereas it would be difficult to track down such issues using system codes, the CMFD method may prove useful as a tool to assess the effect of HTS aging on various thermal-hydraulic characteristics of the system (e.g., the impact of the pressure tube aging on heat transfer in general, and on CHF in particular).
1-C. Design and safety analyses for the next generation reactors (ACR)
  • Investigate local multidimensional effects in the headers and associated system components of new reactor designs – The in-depth understanding of local multidimensional two-phase flow phenomena during LOCA in the proposed ACR, and the ability to use mechanistic models to make predictions of reactor system response, have already been identified as important issues. Combining the development of mechanistic models of two-phase flow and their implementation in CMFD solvers can provide a convenient state-of-the-art framework for performing extensive virtual experiments to test safety features of the proposed ACR design concepts.
  • Use CMFD to guide new experiments – The very nature of CMFD methodology makes it a convenient tool for extensive parametric studies and rigorously defined virtual experiments. The results of such numerical experiments can be used to identify the areas where additional actual experimental evidence is needed and to guide the design of, and the instrumentation for, new experimental programs which would provide the necessary data in a manner allowing for precise and comprehensive testing of new mechanistic models of two-phase flow and heat transfer.
  • Use CMFD models and codes to enhance predictive capabilities of system codes – A very promising approach for future reactor safety studies is associated with using CFMD models of local phenomena to improve the accuracy of predictions by large system codes. Such a goal can be accomplished in several ways. One possible approach is to perform separate-effect CMFD simulations and use the results to develop improved, yet still simplified, models compatible with the modeling framework of system codes. Another approach is to develop CMFD/system-code interfaces to allow running the different codes simultaneously and effectively replace selected modules of system codes by advanced multidimensional CMFD models.

2. Recommended Approach to Maximize the Gains of CMFD Methodology for the Analysis of CANDU Headers during LOCA

2-A. Methodology
  • A systematic, multi-level approach should be used, including the following steps – scaling analysis, scoping calculations using simplified models and computational methods, parametric testing and/or experimental validation of individual CMFD models for situations pertaining to postulated LOCA events in CANDU headers, separate-effect multidimensional simulations, sensitivity analysis on the effect of physical assumptions, mathematical formulation, numerical approach, etc., CMFD simulations of individual components of the CANDU primary system using boundary conditions obtained from system code predictions.
2-B. Problem formulation for CMFD simulations
  • The combined effects of the level of modeling detail and the associated numerous potential sources of uncertainties due to both the formulation of physical models and numerical difficulties, make the issue of problem formulation for CMFD simulations a very important task. In fact, any inconsistencies and incorrect assumptions at this stage of work may dramatically diminish the advantages of the overall approach and/or lead to incorrect results. Thus, the following aspects of problem formulation should be carefully examined before any computations are performed. Select appropriate modeling framework for the range of phenomena and situations to be analyzed, including the choice of the frame of reference (typically: Eulerian or Eulerian/Lagrangian, instantaneous vs. ensemble-averaged equations), the identification of dominant governing phenomena, and the consistency of mathematical model formulation, define appropriate boundary conditions, including the interfaces with surrounding reactor components and any simplifying assumptions which may be necessary due to the lack of detailed information about local distributions of various input parameters, use state-of-the-art models of local two-phase flow phenomena, carefully examine the information about available models, including their limitations and the range of their validity.
2-C. Code and computer requirements
  • The practical usefulness of the CMFD approach to safety analysis of CANDU headers will depend not only on the theoretical aspects of model formulation, but also on the appropriate selection of numerical solvers as well as on the availability of fast computers capable of performing three-dimensional time-dependent calculations. The recommendations concerning selected critical issues are as follows – Exercise caution concerning the use of codes with a black-box approach to two-phase flow models, explore the possibility of implementing new models as parts of the source code, any models implemented via user routines should be thoroughly checked for their consistency with the main solver and their impact on code's performance, steady-state simulations should be performed first, followed by the analysis of transients and accident situations, use CFMD solvers which are capable of capturing combinations of 3-D and 1-D models.
2-D. Quality assessment and dissemination of results
  • The inherent complexity of two-phase flow phenomena requires that a thorough and consistent system be established for the evaluation of CMFD simulations aimed at assessing the consistency and accuracy of the results and at encouraging the dissemination of knowledge and expertise. In particular, the following steps are recommended: results of CMFD predictions should be reviewed at various levels – internal review, independent review, and peer review of published (archival) papers, a complete documentation of past works related to the analysis of CANDU headers (and other safety related studies) should be maintained and made available to the modelers in the future.

Top of page

RSP-0239 – Review of current practices and knowledge on performance-based design – Focus on the need of the nuclear industry

Building codes in many countries have been shifting to performance-based due to economic and social reasons, as well as progress in fire safety technologies and the development of engineering tools that are required to implement such codes. The approach used to develop performance-based codes follows a structure in which there are usually different levels: the goals, objectives, functional requirements, performance criteria and the performance quantification or verification. The number of levels used in this hierarchical structure depends on the complexity of the requirements. The middle level outlines the fire safety design methodology that should be used to satisfy the objectives. The success of performance-based codes depends on the ability to establish the necessary fire safety design methods along with verifiable and enforceable performance criteria.

The present study is a research project conducted by the Fire Research Program, Institute for Research in Construction, National Research Council of Canada, for the CNSC. The purpose of the project is to review the current practices and knowledge in performance-based design for nuclear power plants/nuclear facilities. The review includes general information on performance-based codes in different countries, the performance-based design structure for the design of fire protection systems, and the quantification and verification of performance-based designs.

In addition, this report provides a brief description of models used to determine whether the performance criteria are met by the proposed designs. The models are presented along with their limitations and availability. Furthermore, the report presents necessary performance-based design documentation.

Page details

Date modified: