ML24297A297

NUREG/IA-0550, Modelling Guidelines for Ccfl Representation During Ibloca Scenarios of PWR Reactors

ML24297A297
Person / Time
Issue date:	10/31/2024
From:	Freixa J, Martin-Gil K, Martinez-Quiroga V, Reventos F, Kirk Tien Office of Nuclear Regulatory Research, Universitat Politecnica de Catalunya, Spain
To:
References
NUREG/IA-0550
Download: ML24297A297 (1)

Text

NUREG/IA-0550

Modelling Guidelines for CCFL Representation During IBLOCA Scenarios of PWR Reactors

Prepared by:

J. Freixa, V. Martínez-Qu iroga, K. Martin-Gi l and F. Reventós

Advanced Nuclear Technologies Universitat Politcnica de Catalunya Av. Diagonal 647 Barcelona, 08028 Spain

K. Tien, NRC Project Manager

Division of Systems Analysis Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555-00 01

Manuscript Completed: February 2024 Date Published: October 2024

Prepared as part of The Agreement on Research Participation and Technical Exchange Under the Thermal-Hy draulic Code Applications and Maintenance Program (CAMP)

Published by U.S. Nuclear Regulatory Commission AVAILABILITY OF REFERENCE MATERIALS IN NRC PUBLICATIONS

NRC Reference Material Non-NRC Reference Material

As of November 1999, you may electronically access Documents available from public and special technical NUREG-series publications and other NRC records at libraries include all open literature items, such as books, NRCs Library at www.nrc.gov/reading-rm.html. Publicly journal articles, transactions, Federal Register notices, released records include, to name a few, NUREG-series Federal and State legislation, and congressional reports.

publications; Federal Register notices; applicant, Such documents as theses, dissertations, foreign reports licensee, and vendor documents and correspondence; and translations, and non-NRC conference proceedings NRC correspondence and internal memoranda; bulletins may be purchased from their sponsoring organization.

and information notices; inspection and investigative reports; licensee event reports; and Commission papers Copies of industry codes and standards used in a and their attachments. substantive manner in the NRC regulatory process are maintained at NRC publications in the NUREG series, NRC regulations, The NRC Technical Library and Title 10, Energy, in the Code of Federal Regulations Two White Flint North may also be purchased from one of these two sources. 11545 Rockville Pike

1. The Superintendent of DocumentsRockville, MD 20852-2738

U.S. Government Publishing Office These standards are available in the library for reference Washington, DC 20402-0001 use by the public. Codes and standards are usually Internet: https://bookstore.gpo.gov/ copyrighted and may be purchased from the originating Telephone: (202) 512-1800 organization or, if they are American National Standards, Fax: (202) 512-2104 from

2. The National Technical Information ServiceAmerican National Standards Institute 5301 Shawnee Road 11 West 42nd Street Alexandria, VA 22312-0002 New York, NY 10036-8002 Internet: https://www.ntis.gov/ Internet: www.ansi.org 1-800-553-6847 or, locally, (703) 605-6000 (212) 642-4900

Legally binding regulatory requirements are stated only in laws; NRC regulations; licenses, including technical A single copy of each NRC draft report for comment is specifications; or orders, not in NUREG-series available free, to the extent of supply, upon written publications. The views expressed in contractor request as follows: prepared publications in this series are not necessarily those of the NRC.

Address: U.S. Nuclear Regulatory Commission The NUREG series comprises (1) technical and adminis-trative reports and books prepared by the staff Office of Administration (NUREG-XXXX) or agency contractors (NUREG/CR-Digital Communications and Administrative XXXX), (2) proceedings of conferences (NUREG/CP-Services Branch XXXX), (3) reports resulting from international Washington, DC 20555-0001 agreements (NUREG/IA-XXXX),(4) brochures (NUREG/BR-XXXX), and (5) compilations of legal E-mail: Reproduction.Resource@nrc.gov decisions and orders of the Commission and Atomic Facsimile: (301) 415-2289 and Safety Licensing Boards and of Directors decisions under Section 2.206 of NRCs regulations (NUREG-Some publications in the NUREG series that are posted 0750),and (6) Knowledge Management prepared by at NRCs Web site address www.nrc.gov/reading-rm/doc-NRC staff or agency contractors.

collections/nuregs are updated periodically and may DISCLAIMER: This report was prepared under an differ from the last printed version. Although references to international cooperative agreement for the exchange material found on a Web site bear the date the material of technical information. Neither the U.S. Government nor any agency thereof, nor any employee, makes any was accessed, the material available on the date cited warranty, expressed or implied, or assumes any legal may subsequently be removed from the site. liability or responsibility for any third partys use, or the results of such use, of any information, apparatus, product or process disclosed in this publication, or represents that its use by such third party would not infringe privately owned rights.

NUREG/IA-0550

Modelling Guidelines for CCFL Representation During IBLOCA Scenarios of PWR Reactors

Prepared by:

J. Freixa, V. Martínez-Q uiroga, K. Martin-G il and F. Reventós

Advanced Nuclear Technologies Universitat Politcnica de Catalunya Av. Diagonal 647 Barcelona, 08028 Spain

K. Tien, NRC Project Manager

Division of Systems Analysis Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555-0 001

Manuscript Completed: February 2024 Date Published: October 2024

Prepared as part of The Agreement on Research Participation and Technical Exchange Under the Thermal-H ydraulic Code Applications and Maintenance Program (CAMP)

Published by U.S. Nuclear Regulatory Commission

ABSTRACT

Studies on pipe integrity have pointed out the relevance of Intermediate break LOCA. These scenarios are more dynamic than small break LOCA due to a faster depressurization and a stronger force exerted by the break leading to a stronger counter current flow limitation (CCFL).

The OECD/NEA ROSA-2 project conducted three IBLOCA experiments at the ROSA/large-scale test facility (LSTF) of the Japan Atomic Energy Agency. These three experiments have been analyzed with the US-NRC system code RELAP5. The attention has been focused on the representation of the CCFL phenomenology occurring simultaneously in different parts of the reactor. The post-test calculations have shown the adequate capabilities of RELAP to reproduce this type of scenarios, given that certain modelling guidelines are followed: (1) the three-dimensional representation of the core and upper plenum region (2) the correct configuration of the break nozzle and (3) the CCFL set up of locations and parameters. These three elements showed to be essential to reproduce correctly the phenomenology. The analysis has shown that it is not feasible to approach the CCFL configuration conservatively. Therefore, it is recommended to use the most realistic parameters, preferably obtained by simulating separate effect experiments, and then combine them with uncertainty analysis.

iii

FOREWORD

Thermalhydraulic studies play a key role in nuclear safety. Important areas where the significance and relevance of TH knowledge, databases, methods, and tools maintain an essential prominence, are among others:

• Assessments of plant modifications (e.g., Technical Specifications, power uprates, etc.);

• Analysis of actual transients, incidents, and/or start-up tests;

• Development and verification of Emergency Operating Procedures;

• Analytical information in support of some elements for the Probabilistic Safety Assessments (e.g., success criteria and available time for manual actions, and sequence delineation) and its applications within the risk-informed regulation framework;

• Training personnel (e.g., full scope and engineering simulators); and/or

• Assessments of new designs.

For that reason, the history of the involvement in Thermalhydraulics of CSN, nuclear Spanish Industry as well as Spanish universities, is long. It dates back to mid-80s and comes to the current days through several periods of USNRC CAMP programs. During this long history, CSN has promoted coordinated joint efforts with Spanish organizations through different periods of the so-called CAMP-Espana, the associated national program.

From the CSN perspective, we have largely achieved the objectives. Good models of our plants are in place, and reliable infrastructure of national TH experts, models, and complementary tools, just as an ample set of applications, have been created. The main task now is to maintain the expertise, consolidate it, and update the experience. We at the CSN are aware of the need of maintaining key infrastructures and expertise and see CAMP program as a good and well-consolidated example of international collaborative action implementing this goal.

Many experimental facilities have contributed to todays availability of a large thermal-hydraulic database (both separated and integral effect tests). However, there is a continuous need for additional experimental work and code development and verification, in areas where no emphasis has been made in the past. On the basis of the SESAR/SFEAR1 reports Nuclear Safety Research in OECD Countries: Major Facilities and Programmes at Risk (SESAR/FAP, 2001), Support Facilities for Existing and Advanced Reactors (SFEAR) NEA/CSNI/R(2007)6, and 2019 updated SESAR/SFEAR2 report, CSNI is promoting since the beginning of this century several collaborative international actions in the area of experimental TH research.

1 SESAR/SFEAR is the Senior Expert Group on Safety Research/Support Facilities for E xisting and Advance Reactors of NEA Committee on the Safety of Nuclear Installations (CSNI).

v These reports presented some findings and recommendations to the CSNI, to sustain an adequate level of research, identifying several experimental facilities and programmes of potential interest for present or future international collaboration within the nuclear safety community during the coming decade. The different series of PKL, ROSA, ATLAS, and RBHT projects are under these premises.

CSN, as the Spanish representative in CSNI, is involved in some of these research activities, helping in this international support of facilities and in the establishment of a large network of international collaborations. In the TH framework, most of these actions are either covering not enough investigated safety issues and phenomena (e.g., boron dilution, low power, and shutdown conditions, beyond design accidents), or enlarging code validation and qualification databases incorporating new information (e.g., multi -dimensional aspects, non-condensable gas effects, passive components).

This NUREG/IA report is part of the Spanish contribution to CAMP focused on:

• The analysis, simulation, and investigation of specific safety aspects of PKL/OECD and ATLAS/OECD experiments.

• The analysis of applicability and/or extension of the results of these projects to the safety, operation, or availability of the Spanish nuclear power plants.

Both objectives are carried out by simulating the experiments and conducting the plant application with the last available versions of NRC TH codes (RELAP5, TRACE, and/or PARCS).

An additional goal of CSN is to assure and maintain the capability of the national groups with experience in the thermalhydraulics analysis of accidents in the Spanish nuclear power plants.

Nuclear safety needs have not decreased as the nuclear share of the nations grid for the next coming years is expected to be maintained with plants of extended life and/or higher power.

This is the challenge that will require continued effort.

Javier Dies Llovera, Commisioner

Nuclear Safety Council (CSN) of Spain

vi TABLE OF CONTENTS

ABSTRACT................................................................................................................... iii

FOREWORD................................................................................................................... v

TABLE OF CONTENTS................................................................................................ vii

LIST OF FIGURES......................................................................................................... ix

LIST OF TABLES.......................................................................................................... xi

EXECUTIVE

SUMMARY

............................................................................................. xiii

ABBREVIATIONS AND ACRONYMS.......................................................................... xv

1 INTRODUCTION........................................................................................................ 1 1.1 CCFL Phenomenology and Modelling................................................................2 1.2 The NEA/OECD ROSA 2 Project........................................................................4 1.3 LSTF Facility...................................................................................................... 5

2 THE ROSA/LSTF RELAP5 NODALIZATION............................................................ 7 2.1 Reactor Pressure Vessel (RPV)......................................................................... 8 2.2 Steam Generators............................................................................................ 10 2.3 Pressurizer....................................................................................................... 10 2.4 Primary Loop Piping......................................................................................... 10 2.5 LSTF Heat Structures....................................................................................... 11

3 IBLOCA EXPERIMENTS AND RESULTS............................................................... 13 3.1 Test 1, Surge Line Break (17%)....................................................................... 13 3.2 Test 2, ECCS Line Break (17%)....................................................................... 18 3.3 Test 7, ECCS Line Break (13%)....................................................................... 23

4 SPECIFIC PHENOMENOLOGY............................................................................... 27 4.1 Possible Locations for CCFL Phenomenon (Hot Leg Break)............................ 27 4.2 Possible Locations for CCFL Phenomenon (Cold Leg Break).......................... 27 4.3 Breaking of the Top Core CCFL through CRGT Lateral Holes......................... 28

5 MODELLING GUIDELINES...................................................................................... 31 5.1 3D Representation of the Core, UP and CRGTs.............................................. 31 5.2 CCFL Locations................................................................................................ 32 5.3 Break Modelling................................................................................................ 34

6 CONCLUSIONS....................................................................................................... 35

7 REFERENCES......................................................................................................... 37

vii

LIST OF FIGURES

Figure 1-1 LSTF Test Facility (Courtesy of the OECD/NEA ROSA-2 Group)........................6 Figure 2-1 UPC RELAP5 Nodalization of ROSA/LSTF..........................................................7 Figure 2-2 UPC RELAP5 RPV Nodalization..........................................................................8 Figure 2-3 UPC RELAP5 LSTF Nodalization: Cartesian Distribution.....................................9 Figure 3-1 UPC Results of the ROSA Test 1, a Full Rupture of the Surge Line Connected to the Hot Leg (Equivalent to 17% of the Cold Leg).......................... 15 Figure 3-2 Schematic Evolution of the IBLOCA Scenario in the Hot Leg: Initial Conditions.......................................................................................................... 16 Figure 3-3 Schematic Evolution of the IBLOCA Scenario in the Hot Leg: the High-Pressure Difference at the Break Leads the Coolant towards the Break Except for the Coolant in the LSs. DC Levels Decrease as Boiling Takes Place in the Core............................................................................................... 16 Figure 3-4 Schematic Evolution of the IBLOCA Scenario in the Hot Leg: When Break Flow Switches to Mostly Steam, High Primary Depressurization Takes Place and Primary Pressure Becomes Lower than the Secondary Pressure, Steam Accumulates in the UP to U-Tubes Region. Acc Injection Starts, Core Boiling Intensifies and Core Uncovery Begins.................. 17 Figure 3-5 Schematic Evolution of the IBLOCA Scenario in the Hot Leg: with Accs Injecting Cold Water, Pressure is Locally Reduced at the Cold Leg so that the Pressure Difference between the UP and DC Increases Leading to the LS Clearing Phenomena which Replenishes the Core Effectively.

Nitrogen Flows into the Primary System............................................................ 17 Figure 3-6 UPC Results of the ROSA Test 2, a Full Rupture of the ECC Line Connected to the Cold Leg (Equivalent to 17% of the Cold Leg)........................ 20 Figure 3-7 Schematic Evolution of the IBLOCA Scenario in the Cold Leg: Initial Conditions.......................................................................................................... 21 Figure 3-8 Schematic Evolution of the IBLOCA Scenario in the Cold Leg: The Accumulator in the Broken Loop Empties to the Containment, Circulation Ceases and Condensation Takes Place in the SGs. Core Level Starts to Fall while Significant Coolant Remains in the UP............................................... 21 Figure 3-9 Schematic Evolution of the IBLOCA Scenario in the Cold Leg: The Steam Accumulated in the UP to U-Tubes Region Pushes the Water Accumulated in the Seal of the Broken Loop. The Core is Almost Empty and LS Clearing Takes Place when the Pressures of the System are Balanced............................................................................................................ 22 Figure 3-10 Schematic Evolution of the IBLOCA Scenario in the Cold Leg: Once all LSs Have Been Cleared, the Break Flow Switches to Single Vapor Phase Leading to a Faster Depressurization and Flashing Occurs in Both the DC and the Core. Acc Setpoints are Reached and Core Level Recovers. The Acc Injection Will Be Effective or not Depending on the Cladding Temperatures once Reflooding Started.............................................................. 22

ix Figure 3-11 UPC Results of the ROSA Test 7, a Full Rupture of the ECC Line Connected to the Cold Leg (Equivalent to 13% of the Cold Leg)........................ 25 Figure 4-1 Schematic Description of the Possible CCFL Locations during an IBLOCA at the Hot Leg.................................................................................................... 27 Figure 4-2 Schematic Description of the Possible CCFL Locations during an IBLOCA at the Cold Leg.................................................................................................. 28 Figure 4-3 Schematic Description of the Break-up of the CCFL Phenomenon at the Upper Core Plate............................................................................................... 29 Figure 5-1 Core and UP Levels during Test 7, Simulation of the Experiment with and without Lateral Holes in the CRGTs................................................................... 31 Figure 5-2 Break Fow and Integrated Break Flow during Test 7, Simulation of the Experiment with Different Break Nozz le Length.................................................. 34

x LIST OF TABLES

Table 2-1 Main Features of UPC RELAP5 LSTF Nodalization.............................................8 Table 3-1 Control Logic of the Test 1 Experiment.............................................................. 13 Table 3-2 Chronology of the Main Events in Test 1............................................................ 14 Table 3-3 Control Logic of the Test 2 Experiment.............................................................. 18 Table 3-4 Chronology of the Main Events in Test 2............................................................ 19 Table 3-5 Control Logic of Test 7 Experiment.................................................................... 23 Table 3-6 Chronology of the Main Events in Test 7............................................................ 24 Table 5-1 CCFL Location and Parameters Used in the Present Analysis........................... 32 Table 5-2 CCFL Location and Parameters Recommended for TRACE [23]....................... 33

xi

EXECUTIVE

SUMMARY

Studies on pipe integrity have shown that the probability of a complete rupture of a pipe depends on the pipe diameter, and in particular that this probability is higher for smaller pipes.

Therefore, the double guillotine break of smaller pipes connected to the primary side of a pressurized water reactor (PWR) such as the surge line, the emergency core cooling line or the recirculation line (cold leg to pressurizer spray) should pose a more probable scenario than an integral rupture of the cold or hot legs. The size of these piping would constitute an intermediate break loss-of-coolant-accident (IBLOCA). These scenarios are more dynamic than small break LOCA due to a faster depressurization and a stronger force exerted by the break. Yet the depressurization is not as rapid as in a LOCA situation so that high pressure differences and core uncover coexist with significant slugs of coolant in the primary system for a significant period of time. Therefore, the effects of counter current flow limitation (CCFL) are more intense, have a longer duration and are present simultaneously in several locations.

In this framework, the OECD/NEA ROSA-2 project, which dealt with key light water reactor thermal-hydraulics safety issues, conducted three IBLOCA experiments at the ROSA/large-scale test facility (LSTF) of the Japan Atomic Energy Agency (JAEA). The LSTF facility simulates a PWR Westinghouse design with a four-loop configuration and a nominal power of 3423 MWth. The facility was designed with a power to volume approach with a scaling factor of 48.

In this paper, the three IBLOCA experiments (Tests 1, 2 and 7) carried out at the ROSA/LSTF are analyzed by means of the US-NRC system code RELAP5. The attention has been focused on the representation of the CCFL phenomenology occurring simultaneously in different parts of the reactor. The post-test calculations of the cases have shown the adequate capabilities of RELAP to reproduce this type of scenarios. However, this has only been possible when applying certain modelling guidelines such as the three-dimensional representation of the core and upper plenum region, the correct configuration of the break nozzle and the CCFL set up of locations and parameters. These three elements showed to be essential to reproduce correctly the phenomenology. Details are provide so that other users may implement these guidelines in the analysis of equivalent sequences.

The analysis has shown that the intensity of CCFL occurrence in one location influences the rest of CCFL locations. A high limitation in one part of the primary side can diminish CCFL effects in other parts of the system. Using high CCFL limitations in all the places does not guarantee the most limiting case, hence, in this type of scenarios, it is difficult to define a priori what the most limiting location is in terms of peak cladding temperature. Therefore, it is recommended to use the most realistic parameters, preferably obtained by simulating separate effect experiments, and then combine them with uncertainty analysis.

xiii

ABBREVIATIONS AND ACRONYMS

Acc Accumulator AFW Auxiliary FeedWater CCF Counter Current Flow CCFL CCF Limitation CL Cold Leg CRGT Control Rod Guide Tube DBE Design Basis Events DC DownComer DEGB Double-Ended Guillotine Break ECCS Emergency Core Cooling System HL Hot Leg HPSI High Pressure Safety Injection IBLOCA Intermediate Break LOCA IET Integral Effect Tests JAEA Japan Atomic Energy Agency LBLOCA Large Break LOCA LOCA Loss of Coolant Accident LOOP Loss Of Offsite Power LP Lower Plenum LPSI Low Pressure Safety Injection LS Loop Seal LSTF Large Scale Test Facility LWR Light Water Reactor PCT Peak Cladding Temperature PORV Pressurizer Operated Relief Valve PWR Pressurized Water Reactor PZR Pressurizer RCP Reactor Coolant Pump RCS Reactor Coolant System RIDM Risk-informed decision-making ROSA Rig Of Safety Assessment RPV Reactor Pressure Vessel SBLOCA Small Break LOCA

xv SET Separate Effect Test SG Steam Generator SGTR Steam Generator Tube Rupture UH Upper Head UP Upper Plenum USNRC United States Nuclear Regulatory Commission UTP Upper Tie Plate

xvi 1 INTRODUCTION

The United States Nuclear Regulatory Commission (USNRC) has developed a Risk-informed decision-making (RIDM) framework to supplement the traditional approach for safety regulations. The term RIDM has been defined for USNRC in ref. [1] as follows:

A "risk-informed" approach to regulatory decision-making represents a philosophy whereby risk insights are considered together with other factors to establish requirements that better focus licensee and regulatory attention on design and operational issues commensurate with their importance to health and safety.

The main goal is to benefit from the use of risk insights, along with other important information, to help in make decisions for safety purposes [2]. In the framework of design basis accidents for Pressurized Water Reactor (PWR) reactors and based on recent studies on pipe integrity [3],

the USNRC proposed the Intermediate Break LOCA (IBLOCA) scenario to become a design basis accident [4]. The studies from ref. [3] reported that the probability of a complete rupture of a pipe depends on the pipe size, and in particular that this probability is higher for smaller pipes.

Therefore, the double guillotine break of smaller pipes connected to the primary side of a PWR such as the surge line, the Emergency Core Cooling System (ECCS) line or the recirculation line (cold leg to pressurizer spray) should pose a more probable scenario than an integral rupture of the cold or hot legs. The size of these pipes is mostly on the range in between Small Break LOCA (SBLOCA) and Large Break LOCA (LBLOCA) sizes and are designated as medium or IBLOCAs. These studies combined with the RIDM approach drove the USNRC to consider the IBLOCA scenario as a design basis accident.

IBLOCA scenarios are more dynamic than small break LOCA due to a faster depressurization.

Yet the depressurization is not as rapid as in a large break LOCA situation so that high pressure differences and core uncover coexist for a significant period of time. Therefore, the effects of counter current flow limitation (CCFL) are more intense and have a longer duration. During IBLOCA scenarios, counter current flow limitation (CCLF) may be expected in several locations.

The conditions for CCFL to occur are high energetic steam flows and condensate coolant flowing in the opposite direction driven by gravity. In a Light Water Reactor ( LWR) these conditions do take place during a core uncover situation and with high pressure gradients, i.e.

when the difference between containment and primary side pressures is still high. In PWRs, CCFL will impede coolant to flow back to the core and will difficult core cooling significantly. The present paper intends to address the difficulties of modelling such type of scenarios by simulations of experiments of the OECD/NEA ROSA-2 project. In particular, integral effect IBLOCA experiments from the LSTF facility.

Previous research has been carried out to simulate the experiments in discussion in this report.

Freixa et. al [5] - [7] presented simulations with the TRACE code of the IBLOCA experiments including uncertainty analysis. Later, Clifford et. al [8] continued the work with further assessment on consistent plant modelling. Takeda et al. [9] [10] presented an uncertainty analysis performed with the RELAP code for the hot and cold leg break experiments. For the CATHARE code, Carnevali et al. [11], [12] presented evaluations of the same scenarios. All these works present interesting insights on the simulation of the tests, however they do not focus on the modelling of the CCFL parameters included in their analysis which is the goal of the present report.

1 1.1 CCFL Phenomenology and Modelling

At elevated flow rates within a two-phase flow, both the gas and liquid components move cohesively in the same direction, propelled by the pressure gradient. However, at lower flow rates, the influence of gravity can cause the velocities of the two phases to differ significantly. In certain conditions, the liquid may flow downward while the lighter gas phase moves upward.

This phenomenon, known as Counter-Current Flow (CCF), can occur in vertical, inclined, and even horizontal pipes under specific flow conditions. The CCF limit represents the maximum rate at which liquid can flow down while gas flows up. Flooding, often associated with CCF, refers to the point at which this limit is reached, leading to a sudden change in flow regime.

Flooding is influenced by local geometry and can occur at specific locations within a piping system, such as flow area reduction, the entrance to a bundle of pipes, valves or plates with orifices. The occurrence of CCFL is generally connected with an increase in pressure drop at the CCFL location. It therefore does not only lead to a displacement of water from the core to the CCFL location but it may also cause a certain suppression of the core level. Hysteresis effects may also be observed in certain geometries. Additionally, CCFL plays a role in scenarios involving perforated plates, where the drainage of liquid against steam flow is crucial. Factors such as subcooling of the liquid phase and condensation further affect the flooding behavior.

Specific regions within LWRs, including the upper tie plate (UTP), channel inlet orifices, hot and cold legs, steam generator tubes, downcomer (DC), and surge line, are susceptible to CCFL.

Each region presents ins specific geometrical constraints and modelling requirements. For instance, CCFL at the UTP is a multidimensional phenomenon that can take place due to the de-entrainment of condensate water around the upper plenum internals or because ECC water flowing from the hot leg. In particular, for this last case, water is injected to the side of the cylindric Upper Plenum (UP) and onto the UTP but only covering a part of the plate. This makes this phenomenon particularly difficult to be modelled with system codes due to the tridimensional effects.

Given the current state of the art of two-phase flow modelling, a completely deterministic physical model to specify the start of flow -limiting situations for all geometrical conditions is impossible. Without a CCFL or flow limitation model, coolant distribution cannot be adequately predicted for certain situations. Specific local modeling is necessary in system codes to obtain reliable predictions. Flooding correlations may be established from experimental data, and the parameters of these correlations are part of the code input deck. This approach, strategically integrates the flooding phenomenon into the code by replacing the difference momentum equation with the flooding limit equation when warranted. This method ensures that phasic velocities adhere to the sum momentum equation, which encapsulates gravity and pressure terms.

Numerous experimental studies spanning different scales and geometries have been conducted to address CCFL. Tests involving pipes of varying sizes and liquid injection methods were instrumental in developing CCFL correlations by Wallis and Kutateladze. In particular, Wallis proposed two non-dimensional numbers to account for the superficial volumetric fluxes of gas and liquid and proposed a correlation to describe the flooding limitation.

1/2 1/2

1/2 + 1/2 =

2 In these equations, is the superficial velocity, is gravity, is the pipe diameter and is the phase density. The subindexes, stand for liquid and vapor phase respectively. Finally, and are the slope and the intercept factors of the correlation.

Alternatively, Kutateladze proposed two similar non-dimensional numbers that take into account the fluid surface tension in order to better represent the presence of waves in the flow.

2 1/8 2 1/8

- Q 

- Q Ú5 /2 + 1/2 =

In the Kutatelatdze nondimentional numbers ( ) is the surface tension. Both the Wallis and Kutateladze forms of the general flooding limit equation have yielded acceptable results when applied with constants tailored to specific geometries. Therefore, the constants m and c may need to be adapted to the specific geometrical characteristics of the location where CCFL takes place.

In RELAP5, a generalized CCFL model permits users to select the Wallis form, the Kutateladze form, or a form intermediate to the Wallis and Kutateladze forms. This overarching form, proposed by Bankoff et al., encompasses the dimensionless gas and liquid fluxes, gas intercept, and slope. The solution method involves verifying the existence of countercurrent flow and assessing if the liquid downflow exceeds the limit stipulated by the flooding limit equation. This general form has the following structure:

Hg1/2 + mHf1/2 = C

where is the dimensionless gas flux, is the dimensionless liquid flux, is the gas intercept (value of when = 0, i.e., complete flooding), and is the slope, that is the gas intercept divided by the liquid intercept (the value of when = 0). The dimensionless fluxes have the form

1/6 1/6

*  =

=

,where is given by the expression:

/ 2

= 1

can be a number from 0 to 1. For = 0, the Wallis form of the CCFL equation is obtained; and for = 1, the Kutateladze form of the CCFL equation is obtained. This formulation is general enough to allow the Wallis or Kutateladze form to appear at either small or large diameters.

The final conclusion of this review of the modelling of CCFL in RELAP5 is that the code user has to decide the particular locations where CCFL is expected and then perform research on what parameters to use for the correlation or combination of correlations.

3 It is f or this reason t hat the guide provided i n this report may be of hi gh value for o ther co de users performing similar analysis.

1.2 The NEA/OECD ROSA 2 Project

Several Integral E ffect Tests (IETs) have provided a large m atrix of t ransient sce narios that ar e used to va lidate and f urther deve lop system co des. However, the num ber of exp eriments which considered IBLOCA s cenarios i s s mall, so that t he validation and usage of system co des under these conditions is limited. I n t his f ramework the O ECD/NEA ROSA -2 project was l aunched i n 2009, for 3. 5 ye ars from April 2009 including a half -year sh ifting due to t he i nfluences from the Fukushima Daiichi Accident since March 11, 2011. The pr oject was f ocused on the validation of simulation models and methods for va rious co mplex phenom ena that m ay occur du ring design-basis eve nts ( DBE) and beyond-DBE and to increase the l evel of det ail and accuracy in the analyses of t he ke y phen omena during abnormal transients and accidents of LWRs. The OECD/NEA RO SA-2 Project deals w ith t hree subject s: (1) I BLOCA, (2 ) S team G enerator Tube Rupture (SGTR) and ( 3) counterpart t esting t o the P KL tests as a r equest from t he Project participants. Seven LS TF experiments, Tes t 1 through Test 7, were per formed as ei ther IET or Separate Effect Tests (SET). As r egards I BLOCA sce nario, t hree IBLOCA exp eriments at t he Large Scale Test Facility ( LSTF) of t he Japan A tomic E nergy A gency (JAEA) [13], [ 14] were conducted:

Test 1 H ot Leg IBLOCA

This t est was pr oposed to clarify t hermal-hydraulic r esponses du ring an I BLOCA due to D ouble-Ended G uillotine B reak ( DEGB) o f pressurizer (PZR) su rge l ine connected to th e hot l eg. The IBLOCA w as se lected t o address a safety i ssue concerning risk -informed changes t o L OCA technical requirements (break si ze definition) and to well understand t he thermal-hydraulic phenomena that may oc cur du ring the transition break size L OCAs.

The LSTF experiment w as pl anned to perform an I ET that simulate d IBLOCA f rom the nominal operating condition for P WR w ith assumptions o f total failur e of bo th high-pressure injection system (HPIS) and auxiliary feedwater.

A bl ind analysis act ivity was l aunched where par ticipants simulated t he ex periment bef ore its performance.

Test 2 C old Leg IB LOCA

This t est was int ended to clarify t hermal-hydraulic responses dur ing an I BLOCA due to D EGB o f one of t he four E CCS pi ping nozzles co nnected to f our co ld legs.

The LSTF experiment w as pl anned to perform an I ET that si mulates t he IBLOCA f rom t he nominal operating condition for P WR w ith assumptions of single-failure of t he diesel generators related t o flow r ates of b oth HPIS and low-pressure injection systems (LPI S) and total f ailure of auxiliary f eedwater. The employed break si ze was equi valent t o that for T est 1.

A second blind analysis activity w as al so launched for this exp eriment.

4 Test 7 Cold Leg IBLOCA

This test was intended to clarify thermal-hydraulic responses during IBLOCA due to DEGB of one of the four ECCS piping nozzles connected to four cold legs.

The LSTF experiment was planned to perform an IET that simulates the IBLOCA from the nominal operating condition for PWR with assumptions of full injection of both HPIS and LPIS systems, and total failure of auxiliary feedwater. The test was similar to Test 2, some modifications to the boundary conditions were applied in order to avoid the shutdown of the simulated core due to too high cladding temperatures as it happens in Test 2.

Again, a blind analysis activity was performed among the Project participants.

1.3 LSTF Facility

LSTF (see Figure 1-1) is an experimental facility designed to simulate a Westinghouse-type 4-loop 3,420 MWth PWR under accidental conditions. It is a full-height and 1/48 volumetrically-scaled two-loop system with a maximum core power of 10 MW (14 % of the scaled PWR nominal core power) and pressures scaled 1:1. Loops are sized to conserve volumetric factor (2/48) and to simulate the same flow regime transitions in the horizontal legs (respecting L/D factor).

There is one steam generator (SG) for each loop respecting the same scaling factors. They have 141 full-size U-tubes, inlet and outlet plenum, steam separator, steam dome, steam dryer, main steam line, four downcomers and other internals.

All emergency systems are represented and have a big versatility referred to their functions and positions. Many break locations (20) are available too.

LSTF test facility has about 1,760 measurement points that allow an exhaustive analysis of the tests. There are two types of data or measurements of interest: directly measured quantities (temperature, pressure, differential pressure), and derived quantities (from the combination of two or more direct measured quantities-coolant density, mass flow rate...-)

5 Figure 1-1 LSTF Test Facility ( Courtesy of the OECD/ NEA ROSA-2 Group)

6 2 THE ROSA/LSTF RELAP5 NODALIZATION

The calculations presented in this report were performed with the US-NRC code RELAP5 MOD3.3 patch4 with an additional modification on the CCFL subroutine implemented by Kim et al. [15]. All three post-test calculations have been performed with the same nodalization and model coefficients in all parts of the system. The input deck was previously qualified for the OECD/NEA ROSA-1 Tests 3.2 [16] and OECD/NEA ROSA-2 Test 3 [17].

All the components of the nodalization are shown in Figure 2-1 and the main characteristics are described in Table 2-1. The input deck consists on a pseudo-3D reactor pressure vessel with two identical coolant loops, Loop A and Loop B, each containing a reactor coolant pump, a steam generator and the connecting piping. The primary loop piping is divided into three segments: hot leg, loop seal, and cold leg. The pressurizer is connected to the hot leg of Loop A through the surge line. The pressurizer spray draws water from the cold leg of Loop A. An accumulator is connected to each of the cold legs. The pump model was obtained from a supplied JAEA RELAP5 ROSA/LSTF input deck [9] although the pump coastdown curve were adapted to match with the IBLOCA ROSA-2 test conditions.

Figure 2-1 UPC RELAP5 N odalization of ROSA/LSTF

7 Table 2-1 Main Features of UPC RELAP5 LSTF N odalization

Parameter Quantity Hydraulic nodes 837 Hydraulic junctions 1367 Heat Structure mesh points 10304 Control block 347 Trips 104

2.1 Reactor Pressure Vessel (RPV)

The RPV is modelled using a pseudo-3D approach (Figure 2-2) by defining parallel channels and cross flows to allow multi-dimensional flow paths within the vessel.

Figure 2-2 UPC RELAP5 RPV Nodalization

As r egards the core r egion, i t is di vided into 13 channels of 2 0 axial l evels. Each channel represents each one of t he actual fuel as semblies of t he f acility, and t hey ar e radially distributed by using C artesian crossflows as r eported in the LS TF facility r eport [8] (see Figur e 2-3 ). The last two ce lls of t he core channels r epresent the r egion between the exit of t he ac tive core and the inlet o f the upper co re plate. I n addi tion, 13 heat s tructures have been i ncluded to r eproduce the thermocouples t hat provide the CET measurement.

8 The 8 Control Rode Guide Tubes ( CRGTs) are also modeled with 8 distinct pipes connected to the corresponding channel at the bottom and all connected to the same volume that represents the Upper Head (UH).

Figure 2-3 UPC RELAP5 LSTF N odalization: Cartesian Distribution

As regards the downcomer region, it is split in two vertically interconnected annulus components, in which the transversal lengths, crossflows and momentum equations are activated to reproduce any potential ECCS bypass phenomenon during accident conditions.

DC-UH and DC-HL bypasses are also simulated by modelling the 8 spray nozzles that connect the top of the downcomer to the UH as well as the two nozzles that simulate the leakage between the hot leg and the downcomer. User defined form losses have been adjusted in those connections to fit the mass flows to the values reported in the LSTF facility report. Pipes were divided in 15 axial levels. 12 below the cold leg inlet nozzles and 3 above such connections.

The first 10 axial levels preserve a symmetry with the relative heights of the lower plenum and core components to fit well the surface areas and the meshes of the core barrel passive heat structures. Nodes 11 and 12 have been adjusted to correctly reproduce the elevation of the cold leg inlet nozzles and the DC-HL bypasses. Finally, the height of the volumes above the inlet connections have been selected based on the actual height of the DC-UH and DC-UP bypasses.

As regards the UH, it is split in two parallel channels of 6 axial levels. Parallel channels were selected to avoid a flow end path at the top of the vessel and to enhance the mixing and the circulation in the UH during steady state and transient conditions. The heights and the number of axial levels were selected based on the actual location of the DC-UH and CRGTs nozzles.

The Upper Plenum is modelled by one channel of 7 axial components. Finer meshing has been used at the bottom of the UP to better reproduce CRGTs inlet connections and potential multi-flow paths during RPV flow stagnation in loss of coolant accidents. At the top of the UP it has been also modelled an annulus component around the UP plate to better reproduce the DC-UP bypass. Finally, the Lower Plenum has been modelled by using three axial hydrodynamic components to correctly reproduce the actual elevation of the core inlet and the DC-LP connection.

9 2.2 Steam Generators

The SG model is derived from the one supplied by JAEA in [9]. 5 heat structures are modelled to simulate primary to secondary heat exchange between the U-tubes and the SG riser. The U-tubes are divided into 5 groups (see Figure 2-1) to reproduce the 9 LSTF U-tube types: the first group simulates the shortest type (this length was kept in order to reproduce the moment in which natural circulation is completely lost), and the other 4 are pair-wise averages of the rest of the lengths. In the secondary side, the number of volumes of the riser was set to 16 by assessing the DTs along the U-Tubes. The downcomer is modeled to reproduce the annulus parts at the top and the bottom of the component. In addition, form losses have been adjusted at the inlet of the riser and the outlets of the separator to correctly reproduce the SG recirculation ratio reported in the LSTF facility report.

The main and the auxiliary F eedwater systems are represented by time dependent components.

Feedwater conditions are specified according to test procedures and requirements. The steam from the SGs flows through the main steam lines and into a steam header, where the heat is dumped into a boundary condition volume.

2.3 Pressurizer

The pressurizer is modeled by five separate components: the surge line, the vessel, the spray line, the Pressurizer Operated Relief Valve (PORV) line and the heaters. For the surge line, the lengths, heights and form losses have been adjusted to correctly reproduce the pressure drops between the PZR and the hot leg. The off -take model is active in the inlet nozzle of the surge line to reproduce the effect that hot leg stratified levels can have in the final mass flows through the line. As regards the PZR vessel, it is modelled with a pipe of 10 axial levels.

The height of the volumes at the top and the bottom of the vessel has been fit to reproduce the actual height of the PZR heaters and the spray line. A 5 components pipe has been used to reproduce the spray line from the cold leg to the PZR. Finally, the line between the pressurizer and the PORV has been also modelled with detail (42 nodes) in order to correctly determine the density and the pressure in the upstream volume of the PORV valve.

2.4 Primary Loop Piping

Loop A and Loop B are identical except for some nozzles that could be used for different purposes in the two loops. In particular, Loop A reproduces the intact loop, hence inlet connections of the spray and surge lines are simulated in this loop. The dimensions of the piping are based on the system description for the ROSA-IV facility [8]. Two criteria were followed on the modelling: the first one was to preserve the relative distances and heights between the RPV inlet/outlet connections and the main components of the loops: RCPs, SGs inlet/outlet connections, surge and spray lines, ECCSs nozzles and break connections; the second one was to qualify the differential pressures simulated by the code with those reported in steady state conditions.

For the reactor coolant pumps, the specifications (rated velocities, rated head, rated torque, rated density, moment of inertia, rated friction torque coefficients, homologous curves, two-phase multiplier tables and two-phase and torque difference tables) are set from the data supplied in the LSTF facility report [8]. As regards the pump velocity, for both steady state and transient simulations the angular velocity of the pump is defined by using control block cards.

10 This velocity is constant for steady state and follows a coastdown curve after SCRAM signal.

The coastdown curve is set to the boundary conditions of the tests.

ECCS systems are modelled from the layouts reported in the LSTF facility report [8]. Despite in LSTF nodalization the LPIS, HPIS and Acc s are not connected to a unique cold leg injection line (as in the LSTF facility), all subsystems have been connected at the same position within the cold leg. In addition, the relative distances between such connections and the RCPs and the RPV have been preserved compared to the experimental data. LPIS and HPIS are modelled by using time dependent components. For the Acc s, special attention has been taken to set the relative height of the tanks as well as the pressure drops and the fractions of stored liquid and nitrogen. Such conditions are slightly different depending on the test.

2.5 LSTF Heat Structures

The simulation of metal structures and heater rods has a considerable impact on the final evolution of the postulated experiments. The nodalization represents all the metal structures to simulate both heat losses and heat transfer within the facility. In total there are 10304 heat structure mesh points.

For the active heat structures, the nodalization simulates the 13 Fuel Assemblies heating profiles of the core as well as the PZR proportional and backup heaters and the RCPs cooling systems. Figure 2-3 displays the radial heat structure arrangement of the active core. It reflects the core power zoning reported in LSTF facility report [8]. There are 4 mean power bundles at the center followed by eight high power bundles, 4 low power bundles and 4 peripheral power bundles. The three different core power profiles (LOW, HIGH, MEAN) of the LSTF facility are also modelled in the input deck. They are chopped-cosine axial profiles with identical nine divisions but different peaking factors.

For the passive heat structures, the walls of the whole primary and secondary piping are modelled as well as the rockwool insulator materials that are used to minimize the environmental heat losses. In addition, the passive metal structures of the RPV (core barrel, shroud, upper core plate, instrumentation rods, UP internals, CRGTs) are also included in the model.

11

3 IBLOCA EXPERIMENTS AND RESULTS

3.1 Test 1, Surge Line Break (17%)

Double-ended guillotine break of Pressurizer (PZR) surge line was simulated by isolating the PZR before the break and connecting the break unit to the Hot Leg (HL) in one of the two primary loops. The control logic of the experiment is summarized in Table 3-1 and the chronology of the events in comparison with the RELAP simulation are displayed in Table 3-2.

The main time trends of the experiment compared with the post-test calculation are displayed in Figure 3-1. A schematic depiction of the evolution of the scenario is shown in Figures 3-2 to 3-5.

The primary system was rapidly depressurized reaching soon the low pressure SCRAM signal.

Loss Of Offsite Power (LOOP) was assumed to occur concurrent with the SCRAM signal and in addition the HPIS and auxiliary feed water (AFW) were unavailable. Further depressurization led to the actuation of the Accumulators (Accs) at approximately 150 seconds. During this period of time, the inventory in the primary system decreased continuously so that the circulation flow in the primary loops was not maintained. The core uncovery occurred due to sudden decrease of the liquid level in the core immediately after the actuation of Acc system.

This implied that the steam condensation on cold cold A water within the volume of CL resulted in decreasing CL pressure and drawing core and Loop Seal (LS) water toward CL. The Peak Cladding Temperature (PCT) observed was approximately 610 K. LS clearing, which took place soon after the occurrence of the core uncover, brought the recovery of the liquid level in the core.

Further information on the experiment can be found in ref. [14].

Table 3-1 Control Logic of the Test 1 E xperiment

Event Condition Close of PZR spray line valves (PZR heaters used only to 30 minutes before break compensate heat losses)

Isolation of PZR by closing surge line valve 1 minute before break Break Time zero Generation of scram signal 4 seconds Initiation of primary coolant pump coastdown Initiation of core power decay simulation Turbine trip (closure of stop valve)

Closure of main steam isolation valve Termination of main feedwater Opening and closing of the SG relief valves SG secondary-side pressure = 8.03/7.82 MPa Initiation of accumulator system Primary pressure = 4.51 MPa Initiation of Low Pressure Injection System (LPIS) RPV lower plenum pressure = 1.24 MPa

13 The calculated results are in very close agreement with the experiment as shown in Figure 3-1.

The mass discharged through the break and the system pressures are very similar in both the experiment and the simulation, hence RPV levels are quite well reproduced. The only discrepancy that can be observed is a shorter duration of the core uncovery leading to a very small PCT increase in the calculation. The reason for this discrepancy is a slight overestimation of the DC level at around 150 seconds. In the experiment, the DC is completely depleted for a short period of time allowing for the hot core steam to be evacuated through the DC. This allows for an early drop of the UP water into the core but means that core uncovery starts a bit earlier in comparison to the Acc injection time.

Table 3-2 Chronology of the M ain Events in Test 1

Event Experimental (s) UPC (s)

Closure of PZR isolation valve -60 -64 Onset of break 0 0 Generation of scram signal 1 0 Turbine trip (closure of stop valve) 1 1 Initiation of decrease in main steam flow rate 1 1 Initiation of decrease in main feedwater flow rate 2 3 Completion of main steam isolation 2 5 Initiation of primary coolant pump coastdown 4 7 Completion of main feedwater isolation 8 5 Initiation of decay heat simulation 20 24 Actuation of Acc 154 157 Initiation of cladding temperature increase 164 173 Peak cladding temperature 182 178 Actuation of LPIS system 504 425 Termination of core power supply 1533 N/A

14 Figure 3-1 UPC Results of the ROSA Test 1, a Full Rupture of the Surge Line Connected to the Hot Leg (Equivalent to 17% of the Co ld Le g)

15 Figure 3-2 Schematic Evolution of the IBLOCA Scenario in the Hot Leg:

Initial Conditions

Figure 3-3 Schematic Evolution of the IBLOCA Scenario in the Hot Leg: the High-Pressure Difference a t the Break Leads the Coolant t owards the Break Except for the Coolant in the LS s. DC Levels Decrease as Boiling Takes Place in the Core

16 Figure 3-4 Schematic Evolution of the IBLOCA Scenario in the Hot Leg: When Break Flow S witches t o M ostly Steam, High Primary Depressurization T akes Place and Primary P ressure Becomes Lower t han t he Secondary Pressure, Steam Accumulates i n t he UP t o U-Tubes R egion. Acc Injection Starts, C ore B oiling Intensifies a nd C ore Uncovery Begins

Figure 3-5 Schematic Evo lution of the IBLOCA Sce nario in the Hot Leg: wit h Accs Injecting Cold Water, Pressure is Locally Reduced at the Cold Leg so th at the Pressure Difference be tween th e UP an d DC Increases Leading to the LS Clearing Phenomena wh ich Replenishes th e Core Effectively. Nitrogen Flows into th e Primary System

17 3.2 Test 2, ECCS Line Break (17%)

Test 2 simulated t he thermal-hydraulic responses during a PWR 17% c old leg IBLOCA using a long break nozzle upwardly mounted at the cold l eg. S ingle-failure of bot h HPIS and LPIS ( 3/8 of t he total capacity) and total f ailure of auxi liary feedwater w ere assumed. 3 out of 4 accumulators w ere available. The control l ogic of t he exp eriment i s su mmarized in Tabl e 3-3 and the ch ronology of t he events i n comparison with the R ELAP s imulation are di splayed in Table 3-4. The mai n time t rends o f the experiment co mpared with t he pos t-test calculation are displayed in Figur e 3-6. A sch ematic depi ction of the evolution of t he sce nario is shown i n Figures 3-7 to 3-10. As s hown in Figur e 3-6 the rather l arge size of t he br eak ca used a fast primary depressurization. Break flow turned from si ngle-phase liquid to two-phase flow i n a very short time after the break, enhancing the deca y power heat r emoval by reflux and condensation and also reducing the pr imary sys tem dep ressurization. Thi s phenom enon kept the primary pressure hi gher t han t he secondary pr essure from ar ound 30 seconds to 60 seconds. D uring this t ime interval, despite t he co oling-down induced by t he condensati on process, core dr yout took pl ace because part of the water liquid condensed in the U -tubes w as retained i n the U P, SG U-tube upflow-side and hot l egs by a CCFL phenomenon. H PIS w as s tarted almost simultaneously w ith t he c ore dryout, but was i neffective for co re co oling because the water injected i n the i ntact loop w as mainly dr agged t o t he broken l oop by E CCS by pass phenomenon. A t around 60 seconds, the DC w ater l evel dr opped below R PV i nlet co nnections, stopping the E CCS byp ass phenomenon and al lowing the circulation of the vapor r etained in the loop seal t o t he br eak. This phenomenon i nduced the depr essurization of t he pr imary sys tem as well as t he entrance of the Acc s. At this t ime, co re r eflooding occurred and bottom -up quench started. A nyhow, Acc injection was no t enough to r educe the hi gh cladding temperatures and LSTF core pr otection sys tem w as au tomatically a ctivated at ar ound 150 seconds. The experiment was terminated when the p rimary and S G se condary-side pressures r eached nearly-equilibrium co ndition with well-cooled core after the actuation of t he LPIS. Fu rther information can be found in ref.[13].

Table 3-3 Control Logic of the Test 2 Experiment

Event condition Break Time zero Generation of scram signal Primary pressure = 12.97 MPa.

PZR heater off Generation of scram signal or PZR liquid level below 2.3 m Initiation of core power decay curve simulation Initiation of primary coolant pump coastdown Turbine trip (closure of stop valve) Generation of scram signal Closure of main steam isolation valve Termination of main feedwater Generation of SI signal Primary pressure = 12.27 MPa Initiation of HPIS in the loop with PZR only 12 s after SI signal Initiation of accumulator system Primary pressure = 4.51 MPa Initiation of LPIS in the loop with PZR only RPV lower plenum pressure = 1.24 MPa Opening and closing of the SG relief valves SG secondary-side pressure = 8.03/7.82 MPa

18 The calculated results are in very close agreement with the experimental values as shown in Figure 3-6. The mass discharged through the break as well as the system pressures and temperatures are very similar in both the experiment and the simulation. There is although a slower depletion of the core inventory in the experiment between 50 and 125 seconds into the transient. In addition, CCFL phenomenon and UP water accumulation is slightly overpredicted if compared to the experimental data. This leads to slightly lower PCT temperatures which in the experiment triggered to core power reduction. Afterwards, the core power in the experiment and in the calculation are different and no further assessment can be performed.

Table 3-4 Chronology of the M ain Events in Test 2

Event Experimental UPC Break valve opened 0 0 Scram signal (primary pressure = 12.97 MPa) 7 1 SI signal (primary pressure = 12.27 MPa) 7 8 Turbine trip and closure of SG MSIVs 10 5 Initiation of decrease in liquid level in SG U-tube About 10 About 13 Initiation of coastdown of primary coolant pumps 11 7 Termination of SG main feedwater 13 9 Initiation of decrease in liquid level in crossover leg downflow-side About 25 About 26 Open of SG relief valves About 27-57 About 21-59 Initiation of core power decay 29 24 Initiation of HPIS system in loop with PZR About 35 35 Loop seal clearing took place About 40 About 40 Primary pressure became lower than SG secondary-side pressure About 55 55 Initiation of Acc system in loop with PZR only About 110 116 Core power decrease by LSTF core protection system when PCT reached 958 K About 140 N/A Maximum fuel rod surface temperature About 150 129 Whole core was quenched About 180 178 Primary coolant pumps stopped 260 257 Termination of Acc system in loop with PZR only About 280 About 253 Initiation of LPIS system in loop with PZR only About 290 N/A Break valve closure 1212 N/A

19 Figure 3-6 UPC Re sults of the ROSA Test 2, a Fu ll Ru pture of the ECC Li ne Co nnected to the Co ld Le g (Eq uivalent to 17% of the Co ld Le g)

20 Figure 3-7 Schematic Ev olution of the IBLOCA Sc enario in the Co ld Le g: Initial Co nditions

Figure 3-8 Schematic Ev olution of the IBLOCA Sc enario in the Cold Leg: The Accumulator in the Broken Loop Empties to the Containment, Circulation Ceases and Condensation Takes Place in the SGs. Core Level Starts to Fall while Significant Coolant Remains in the UP

21 Figure 3-9 Schematic Evolution of the IBLOCA Scenario in the Cold Leg: The Steam Accumulated in the UP to U-Tubes Region Pushes the Water Accumulated in the Seal of the Broken Loop. The Core is Almost Empty and LS Clearing Takes Place when the Pressures of the System are Balanced

Figure 3-1 0 Schematic Evolution of the IBLOCA Scenario in the Cold Leg: Once all LSs Have Been Cleared, the Break Flow Switches to Single Vapor Phase Leading to a Faster Depressurization and Flashing Occurs in Both the DC and the Core. Acc Setpoints are Reached and Core Level Recovers. The Acc Injection Will Be Effective or not Depending on the Cladding Temperatures once Reflooding Started

22 3.3 Test 7, ECCS Line Break (13%)

Test 7 simulated the thermal-hydraulic responses during a PWR 13% cold leg IBLOCA using a long break nozzle upwardly mounted at the cold leg. Injection of 3/4 of the total capacity of the HPIS and LPIS systems and total failure of auxiliary feedwater were assumed. 3 out of 4 accumulators were available. The control logic of the experiment is summarized in Table 3-5 and the chronology of the events in comparison with the RELAP simulation are displayed in Table 3-6.The transient evolution is similar as for Test 2, so Figures 3-7 to 3-10 are also applicable for Test 7. The main time trends of the experiment compared with the post-test calculation are displayed in Figure 3-11. Again, the experiment was initiated with a rather fast primary depressurization (Figure 3-11). HPIS started before the core dryout and was enough to mitigate the first core uncover situation that took place due to the rapid loss of coolant. The dryout was soon recovered due to the occurrence of the LS clearing in combination with the HPIS. The primary pressure soon became lower than SG secondary-side pressure as a result that the seal of the broken loop was emptied and vapor generated at the core could freely flow from the hot leg to the break. Since the coolant injected by the HPIS system was lower than the break flow, water levels in the primary system continued dropping. The second core dryout occurred with small temperature excursion at the incipience of Acc coolant injection that pushed water towards the ECC injection points. Further information on this test can be found in ref. [18].

Table 3-5 Control Logic of Test 7 Experiment

Event condition Break Time zero Generation of scram signal Primary pressure = 12.97 MPa.

PZR heater off Generation of scram signal or PZR liquid level below 2.3 m Initiation of core power decay curve simulation Initiation of primary coolant pump coastdown Turbine trip (closure of stop valve) Generation of scram signal Closure of main steam isolation valve Termination of main feedwater Generation of SI signal Primary pressure = 12.27 MPa Initiation of HPIS in the loop with PZR only 12 s after SI signal Initiation of accumulator system Primary pressure = 4.51 MPa Initiation of LPIS in the loop with PZR only RPV lower plenum pressure = 1.24 MPa Opening and closing of the SG relief valves SG secondary-side pressure = 8.03/7.82 MPa

This scenario was the most challenging to simulate, and was the one that gave more insight on the CCFL phenomenology, particularly because of the very abrupt core level increase at around 50 seconds which did not appear with the first post-test approach. In the final post-test calculation this rise was well reproduced. However, the first core drop was slightly deeper in the experiment, the short PCT increase was not reproduced in the calculation. As the scenario continued the core level slowly decreased again, and this reduction was well captured in the calculation with a second core uncovery slightly overpredicted by the simulation.

23 The level rise in t he D C due to t he A cc injection was overpredicted. Overall, t he time t rends show a very good agreement with the experiment.

Table 3-6 Chronology of the Main Events in Test 7

Event Experimental UPC Break valve opened 0 0 Highest rotation speeds of primary coolant pumps 4-11 4-11 Scram signal (primary pressure = 12.97 MPa) 9 5 Turbine trip (Closure of SG main steam stop valve) 9 5 Initiation of decrease in liquid level in SG U-tube 10 About 13 SI signal (primary pressure = 12.27 MPa) 11 8 Initiation of coastdown of primary coolant pumps 11 11 Termination of SG main Feedwater 12 9 Closure of SG MSIVs 16 9 Initiation of decrease in liquid level in crossover leg downflow-side 25 About 27 Open of SG relief valves 25-60 19-56 Initiation of core power decay 30 29 Initiation of HPIS system in loop with PZR 27 21 Loop seal clearing took place 60 About 60 Maximum fuel rod surface temperature 67 169*

Primary pressure became lower than SG secondary-side pressure 75 71 Initiation of Acc system in loop with PZR only 150 154 Primary coolant pumps stopped 259 260 Termination of Acc system in loop with PZR only 350 290 Initiation of LPIS system in loop with PZR only 800 N/A Break valve closure 1212 N/A

• Second peak

24 Figure 3-1 1 UPC Re sults of the ROSA Test 7, a Fu ll Ru pture of the ECC Li ne Co nnected to the Co ld Le g (Eq uivalent to 13% of the Co ld Le g)

25

4 SPECIFIC PHENOMENOLOGY

4.1 Possible Locations for CCFL Phenomenon (Hot Leg Break)

When the break is located at the hot leg, the path from the core to the break is much shorter and with less obstacles than in the case of the cold leg side break. In this situation CCFL will appear as soon as core uncover begins mainly in the upper core plate and hot leg connection to the UP. Figure 4-1 displays the situation at the time CCFL phenomena may occur with indications on the possible locations. In Test 1, the window of time when CCFL may play a role is short because core uncovery only starts when primary pressure falls below secondary pressure and the break transits to single phase vapor flow, at this moment the core starts to boil off but pressure starts to decrease faster leading the system to the Acc setpoint. In Test 1, the core uncover only occurred with the actuation of the Accs which pulled the core level down. At the same time, the Acc injection triggered the loop seal clearing. The time window between the start and the inevitable core level recovery is very short and this is the reason why the increase of PCT is very small. During this time the CCFL at the top of the core plays a very significant role.

Figure 4-1 Schematic De scription of the Po ssible CCFL Lo cations during an IBLOCA at the H ot Leg

4.2 Possible Locations for CCFL P henomenon (Cold Leg Break)

When the break is located at the cold leg side, one can expect CCFL phenomenon taking place in several locations. The characterization of the phenomenon is that highly energetic steam formed in the core will try to flow towards the most intensive momentum sink available, which in this case is the break location. Figure 4-2 displays the situation around the UP and hot leg area when the intensity of the CCFL is expected to be more intense. The drawing shows the steam flow paths and the particular locations of CCFL. During the blowdown phase (first 50 seconds) the break performs a suction of the core water from the bottom and the steam generated flows towards the U-tubes while the secondary pressure is still below the primary pressure. In this phase, CCFL may occur at the upper core plate, along the hot leg, at the SG inlet plenum and at the U-tube entrance. As primary pressure falls below secondary pressure, the heat transfer reverses and the SG U-tubes will switch from a momentum sink phenomenon to steam binding

27 effect leading to the clearing of the loop seals. During this phase steam will flow through the CRGTs, UH and then to the DC and break. Once the loop seals have been cleared, the easiest way for the steam to reach the break will be circulating all along the loop. Hence, CCFL can take place again in the upper core plate, hot leg, SG plena and U-tubes entrance. However, very limited coolant will remain in those parts because the secondary side does not longer act as a heat sink, so that CCFL will be negligible. Most of the remaining liquid will be in the hot leg.

If the break is intense enough to fully deplete the coolant in the core, steam may try to find a path to the break through the lower plenum and DC, in this situation CCFL may take place at the DC.

Figure 4-2 Schematic Description of the Possible CCFL Locations during an IBLOCA at the Cold Leg

4.3 Breaking of the Top Core CCFL through CRGT Lat eral Holes

A particular phenomenon has been obs erved in both tests with the break in the cold side, where the core swell level highly increas ed when the break flow switched to s ingle phase vapor. This phenomenon was very intense in Test 7 and prevented a high PCT taking place during the transient. In the following lines we try to explain the phenomenon with the help of RELAP5 calculations. The governing force for this level swell is the end of the clearing of the loop seals which permits s team to flow along the loop towards the break. This will push the core level up, however, since there is a significant layer of coolant in the UP the steam in the core is prevented to flow through and c ontinues to push the core level down. It is important here to notice that there are some lateral holes in the CRGTs at around 0.6m above the UTP. When the lev el in the UP is going down, there is a moment when these holes become uncovered, at this precise instant, there is no more liquid addition into the CRGTs and the level there decreases faster than in the UP. It does not take long until the coolant in the CRGTs is depleted and a path is established between the core and the UP. In addition, if the UH and DC levels are low enough, s team may flow through the UH into the UP through the UH-D C and DC-UP bypasses. The two phenomena favor a balancing of pres sures between the UP and the c ore that lead to the fall of the UP coolant into the core and the core swell level can be pushed up by the force of the DC c olumn of water. The phenomenon is illustrated in Figure 4-3. A sensitivity study has been performed to c haracterize the importance of the CRGT lateral holes. Figure 5-1 shows the results with and without the representation of the lateral holes. In addition, it has been observed that the rapid core swell level increase takes place after the voiding of both the CRGT bottom and the flow path through the two bypasses.

28 Figure 4-3 Schematic Description of the Break-up of the CCFL Phenomenon at the Upper Core Plate

29

5 MODELLING GUIDELINES

5.1 3D Representation of the Core, UP and CRGTs

The steam velocities at the core outlet during boil-off will have a heterogeneous radial distribution. This is a consequence of the radial power profiles, the possible liquid flow-back from the hot legs and the non-homogeneous size of openings through the core. In order to capture these radial heterogeneities, a 3-D representation of the simulation domain is essential.

In the opinion of the authors, the pseudo 3-D capabilities of current two-phase thermal-hydraulic system codes, based on inviscid flow assumption (except for interactions with wall and phase-interfaces) should be able to capture these details. The hydraulic simulation domain is indeed densely populated with internals (fuel rods, support columns and other structures) thus reducing the importance of shear stress forces in the region. In the lower part of the UP, local secondary flow patterns should also be outplayed by the pressure gradient induced by the break and the SG heat sink, by the stack effect due to the CRGTs, and by the liquid flowing back from the hot legs. The nodalization must however be sufficiently detailed to take into account the different power zones (or most relevant power zones), possible differences in hole sizes (and shape dependent singular pressure loss coefficients) in the axial planes and non-uniform distribution of passive heat structures. In addition, important details like lateral holes in CRGTs and positioning of the CRGTs play an important role in the scenarios studied.

Figure 5-1 Core and UP L evels during Test 7, S imulation of the Experiment with and without Lateral Holes in the CRGTs

31 5.2 CCFL L ocations

As i t has been m entioned in the i ntroduction, sp ecial at tention needs t o be pai d to t he modelling of the CCFL. I n par ticular, the use r m ust decide the locations where it might take place and define proper co efficients. This ca n only be done after a t horough evaluation of t he phenomenology. For t he present anal ysis, CCFL has been set up in all the l ocations mentioned in the previous se ctions. Because CCFL depends hi ghly on geometry, each location can have different coefficients. Table 5-1 specifies the coefficients use d i n each location for t he present analysis. The coefficients w ere ext racted from references w hen possi ble. Specific se nsitivities on the co efficients f or ea ch location have not been performed. S ince the coefficients do not need to be the same i n each location and beca use the ef fect in one l ocation has an i nfluence in the other l ocations, a very bi g matrix of ca lculations sh ould be c onfigured and the anal ysis o f the results w ould bring probably l ittle insight on t he phenomenology. The complexity of t he experiment and t he i nteraction with other pheno mena makes it impossible to adj ust all the parameters. The au thors w ould rather su ggest adding CCFL values t ypical for each location preferably obt ained f rom testing in se parate effect test facilities and then apply unce rtainties to each coefficient. Another co nclusion of t he present study is that a co nservative approach to account for C CFL phenomenon in a co ld leg side IBLOCA sh ould be taken w ith care because of the feedback between different C CFL locations an d the i nteraction may lead to non-conservative results. For instanc e, a strong CCLF limitation in the hot leg c ould mean a less intense CCFL in the upper core plate and lead to lower core temperatures.

Table 5-1 CCFL Location and Parameters Used in the Present Analysis

Correlation m c Ref.

Hot leg Wallis 0.614 0.62 [15]

Upper core plate Wallis 0.8625 1.0 [19]

RPV to hot leg Wallis 0.614 0.62 [15]

SG plena entrance Wallis 0.725 1.0 [20]

U-tube inlet Wallis 0.725 1.0 [20]

CRGT entrance Wallis 0.725 1.0 -

CRGT holes Wallis 0.725 1.0 -

CRGT middle Wallis 0.725 1.0 -

Downcomer Wallis 1.0 1.0 -

Intermediate leg - - - -

The original RELAP5 MOD3.3 patch 4 distribution does not allow CCFL phenomenon in fully horizontal pipes. However, as mentioned before, CCFL can take place in the horizontal part of the hot leg. This has been experimented and characterized by several researchers [21]. Kim and No [15], [22] investigated the performance of the RELAP5 code for the simulation of CCFL in horizontal pipes, to do so, they proposed a modification of the code. The modification is fully described in a NUREG report [15] and has been added in the code used for the present work in order to analyze the effect of CCFL in horizontal pipes. The correlation and the coefficients were the ones suggested in [15].

32 In summary, we suggest the following approach in order to configure CCFL parameters in complex scenarios such as an IBLOCA case:

1. Thorough evaluation of the phenomenology of the scenario to identify the locations where CCFL might take place

2. Perform a literature review to find correlations that describe the CCFL phenomenon corresponding to the geometrical aspects of each important locations. There are many experiments available with different geometrical aspects This step requires scaling analysis and an evaluation of the overall conditions of the experiments

3. (Optional) Validation of the parameters for the particular location with separate effect experiments and assessment of the uncertainties. In principle, it is possible to take the values of the correlation directly from the experimental results. The CCFL correlations characterize the phenomenon of flooding for a given geometry and thus can be used directly in the model. However, it is advisable to confirm the validity of the parameters by performing post-test calculations of the separate effect experiments. In addition, this step can be extended to generate the uncertainties on the parameters by using Inverse Uncertainty Quantification (IUQ) techniques. This step may require significant workload that cannot always be performed.

4. Perform uncertainty propagation of the CCFL results to see the effect of the parameters.

On the other hand, it is interesting to bring forth a guide for the configuration of the CCFL locations and parameters for the TRACE code that is provided in ref. [23]. The values provided are very similar to the findings used in the present analysis except for the Upper core plate where a special model for plates is used. The values are shown in Table 5-2. There are some other differences in the selection of locations and parameters, however, the report does not provide references to the values implemented or how they were derived and therefore they are not considered in the present analysis.

Table 5-2 CCFL Location and P arameters Recommended for TRACE [23]

Correlation m c Ref.

Hot leg Wallis 0.74 0.38 -

Upper core plate Bankoff 1.07 0.00433 -

RPV to hot leg - - - -

SG plena entrance Wallis 0.725 1.0 -

U-tube inlet Wallis 1.0 0.8 -

CRGT entrance Wallis 0.85 0.8 -

CRGT holes - - - -

CRGT middle - - - -

Downcomer - - - -

Intermediate leg Kutateladze 1.0 1.79 -

33 5.3 Break Modelling

The results of the simulations are very sensitive to the break nodalization. In the presented experiments the break was represented by an upward oriented nozzle connected to the primary piping. In the calculations, the break was modelled with an upwardly connected cell with the offtake model activated to account for variations of the void distribution in the horizontal plane of the primary piping. The choked flow model was the Henry-Fauske model with a discharge coefficient of 0.85 and the frozen option of the model which is suitable for nozzle configurations.

It has been detected that the length of the horizontal nozzle has a strong effect on the break flow during the subcooled phase and that the most appropriate value might not be coincident with the actual length of the nozzle. Figure 5-2 shows results of 3 calculations of Test 7 where the only difference is the nozzle length between the cold leg and the choking plane.

Figure 5-2 Break Fow and Integrated Break Flow during Test 7, S imulation of the Experiment with Di fferent Br eak Nozzle Le ngth

34 6 CONCLUSIONS

Three different experiments dealing with IBLOCA scenarios for PWR reactors have been performed. The experiments were part of the NEA/OECD ROSA2 project and had been carried out at the LSTF facility operated by JAEA. The IBLOCA experiments of the project have been simulated to assess the capabilities of the RELAP5 code in this type of scenarios and to analyse the possible nodalization approaches and understand the system behaviour of this kind of scenarios. Especial attention has been placed to the representation of the CCFL phenomenology.

The post-test calculations of the cases have shown the adequate capabilities of RELAP to reproduce this type of scenarios. However, this has only been possible when applying certain modelling guidelines such as the three-dimensional representation of the core and upper plenum region, the correct configuration of the break nozzle and the CCFL set up of locations and parameters. These three elements showed to be essential to reproduce correctly the phenomenology.

IBLOCA scenarios are very dynamic due to its continuous change in momentum driving forces making the understanding of the system behavior a challenge by itself. Detailed explanations on the evolution of the scenarios have been provided with support from post -test calculations. The configuration of the CCFL phenomena in different location has been observed to be of main importance in the evolution of the scenario. A set of locations and parameters is proposed. In addition, the intensity of CCFL occurrence in each location is influenced by the other CCFL locations. In this sense, it is difficult to set up a conservative configuration of the CCFL parameters because a high limitation in one part of the primary side can diminish CCFL effects in other parts of the system leading to possible cliff-edge phenomena. Using high CCFL limitations in all the places does not guarantee the most limiting case, hence, in this type of scenarios, it is difficult to define a priori what the most limiting location is in terms of PCT. In this sense, a best estimate approach with the added estimation of uncertainties is recommended.

Therefore, it is recommended to use the most realistic parameters, preferably obtained by simulating separate effect experiments, and then combine them with uncertainty analysis.

35

7 REFERENCES

[1] USNRC and US-NRC, Regulatory Guide 1.157 - Best-Estimate Calculations of Emergency Core Cooling System Performance, 1989. Accessed: Sep. 17, 2019.

[Online]. Available: https://www.nrc.gov/docs/ML0037/ML003739584.pdf

[2] USNRC, Risk -Informed Decision-Making for Nuclear Material and Waste Applications Draft for Trial Use, 2005.

[3] R. L. Tregoning, L. R. Abramson, P. M. Scott, and N. Chokshi, LOCA frequency evaluation using expert elicitation, Nucl. Eng. Des., vol. 237, no. 12, pp. 1429-1436, 2007, doi: http://dx.doi.org/10.1016/j.nucengdes.2006.09.036.

[4] USNRC, 10CFR part 50. Domestic Licensing of Production and Utilization Facility, Appendix K ECCS Evaluation Model. 1998.

[5] J. Freixa and A. Manera, Remarks on Consistent Development of Plant Nodalizations:

An Example of Application to the ROSA Integral Test Facility, Sci. Technol. Nucl. Install.,

vol. 2012, no. Article ID 158617, 2012, doi: 10.1155/2012/158617.

[6] J. Freixa, T.-W. Kim, and A. Manera, Thermal -hydraulic analysis of an intermediate LOCA test at the ROSA facility including uncertainty evaluation, Nucl. Eng. Des., vol.

264, no. 2013, pp. 153-160, Nov. 2013, doi: 10.1016/j.nucengdes.2013.02.023.

[7] J. Freixa, T.-W. Kim, and A. Manera, Post -test thermal-hydraulic analysis of two intermediate LOCA tests at the ROSA facility including uncertainty evaluation, Nucl. Eng.

Des., vol. 264, 2013, doi: 10.1016/j.nucengdes.2013.02.023.

[8] I. Clifford, O. Zerkak, A. Pautz, H. Yao, and J. Freixa, System Code Validation Series based on a Consistent Plant Nodalisation of the ROSA / LSTF Integral Test Facility using TRACE v5. 0 Patch 4, in NUTHOS-11, 2016, pp. 1-15.

[9] T. Takeda and I. Ohtsu, Uncertainty analysis of ROSA/LSTF test by RELAP5 code and PKL counterpart test concerning PWR hot leg break LOCAs, Nucl. Eng. Technol., vol.

50, no. 6, pp. 829-841, 2018, doi: 10.1016/j.net.2018.05.005.

[10] T. Takeda and I. Ohtsu, RELAP5 uncertainty evaluation using ROSA/LSTF test data on PWR 17% cold leg intermediate-break LOCA with single-failure ECCS, Ann. Nucl.

Energy, vol. 109, pp. 9-21, 2017, doi: 10.1016/j.anucene.2017.05.007.

[11] S. Carnevali and P. Bazin, Comparison of CATHARE code using a 3D Reactor Pressure Vessel modelling approach and experimental results on intermediate break LOCAs of ROSA 2 program, in NUTHOS-11, 2016, pp. 1-11.

[12] S. Carnevali and P. Bazin, Validation of cathare code on the 3D ROSA-LSTF pressure vessel, Int. Top. Meet. Nucl. React. Therm. Hydraul. 2015, NURETH 2015, vol. 3, pp.

2103-2116, 2015.

[13] T. Takeda, M. Suzuki, H. Asaka, and H. Nakamura, Quick-look Data Report of ROSA-2/LSTF Test 2 (Cold Leg Intermediate Break LOCA IB-CL-03 in JAEA, 2010.

[14] T. Takeda, M. Suzuki, H. Asaka, and H. Nakamura, Quick-look Data Report of Test 1, Test for Hot Leg Intermediate Break Loss of Coolant Accident with Break Size Equivalent to 17% of Cold Leg Flow Area, 2010.

[15] h. T. Kim and H. C. No, NUREG_IA-0203, Assessment of RELAP5_MOD3.2.2y Against Flooding Database in Horizontal-to-Inclined Pipes, no. July 2001. 2001.

[16] V. Martínez-quiroga, F. Reventós, and C. Pretel, Post-Test Calculation of the ROSA/LSTF Test 3-1 Using RELAP5/Mod3.3, 2012. [Online]. Available:

http://pbadupws.nrc.gov/docs/ML1208/ML12083A012.pdf

[17] J. Freixa, V. Martínez-quiroga, O. Zerkak, and F. Reventós, Modelling guidelines for core exit temperature simulations with system codes, Nucl. Eng. Des., vol. 286, pp. 116-129, 2015, doi: 10.1016/j.nucengdes.2015.02.003.

37

[18] J. A. E. A. (JAEA), Final Data Report of ROSA-2/LSTF Test 7 (Cold Leg Intermediate Break LOCA IB-CL-05 in JAEA), 2013.

[19] NEA-CSNI, Bemuse Phase V Report, Uncertainty and Sensitivity Analysis of a LB-LOCA in ZION Nuclear Power Plant, 2011.

[20] T. Yonomoto, CCFL characteristics of PWR steam generator U-tubes, Portland, USA, 2001.

[21] S. Al Issa and R. Macián-Juan, A review of CCFL phenomenon, Ann. Nucl. Energy, vol.

38, no. 9, pp. 1795-1819, Sep. 2011, doi: 10.1016/j.anucene.2011.04.021.

[22] H. T. Kim and H. C. No, assessment of RELAP5 against flooding database in horizontal-to-inclined pipes, Ann. Nucl. Energy, vol. 29, pp. 835-850, 2002.

[23] TRACE Pressurized Water Reactor Modeling Guidance, 2015.

38 NUREG/IA-0550

Modelling Guidelines for CCFL Representation during IBLOCA Scenarios of October 2024 PWR Reactors

J.Freixa, V. Martínez-Q uiroga, K. Martin-G il and F. ReventósTechnical

Universitat Politcnica de Catalunya Av Diagonal 647, Barcelona 08028 (Spain)

Division of Syatems Analysis Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, D.C. 20555-0001

K. Tien, NRC Project Manager

Studies on pipe integrity have pointed out the relevance of Intermediate break LOCA. These scenarios are more dynamic than small break LOCA due to a faster depressurization and a stronger force exerted by the break leading to a stronger counter current flow limitation (CCFL). The OECD/NEA ROSA-2 project conducted three IBLOCA experiments at the ROSA/large-scale test facility (LSTF) of the Japan Atomic Energy Agency.

These three experiments have been analyzed with the US-NRC system code RELAP5. The attention has been focused on the representation of the CCFL phenomenology occurring simultaneously in different parts of the reactor. The post-test calculations have shown the adequate capabilities of RELAP to reproduce this type of scenarios, given that certain modelling guidelines are followed: (1) the three-dimensional representation of the core and upper plenum region (2) the correct configuration of the break nozzle and (3) the CCFL set up of locations and parameters. These three elements showed to be essential to reproduce correctly the phenomenology. The analysis has shown that it is not feasible to approach the CCFL configuration conservatively. Therefore, it is recommended to use the most realistic parameters, preferably obtained by simulating separate effect experiments, and then combine them with uncertainty analysis.

Thermal-hydraulics, CCFL, IBLOCA, RELAP