Preamble

Verification and Research of SGTR Accident Based on the PKL Test Facility

Dechen Zhang¹, Xiangbin Li¹ (Corresponding author), Xiaojun Qiu¹, Yusheng Liu², Chengshen Wang¹

¹School of Nuclear Science and Engineering, North China Electric Power University, Beijing 102206, China
²[Affiliation not provided]

Abstract

To assess steam generator tube rupture (SGTR) accident scenarios under compromised safety systems (e.g., post-earthquake conditions), as specified in the OECD/NEA/CSNI International Standard Problem (ISP) No. 52 project, a full-scale facility model was developed using the RELAP5 thermal-hydraulic code for numerical simulation. The computational results were systematically validated against experimental data obtained from the PKL test facility. The analysis demonstrates that RELAP5 accurately reproduces the primary progression of the SGTR accident, with the simulated accident sequence and key thermal-hydraulic parameters exhibiting strong agreement with empirical observations. During the early phase of the transient process, depressurization of the primary circuit facilitates secondary-side coolant ingress into the reactor core, thereby maintaining adequate core cooling throughout the initial and intermediate stages. Subsequent depletion of the steam generator inventory results in partial core uncovery, precipitating a rapid escalation in core exit temperature. However, the coordinated actuation of the high-pressure safety injection (HPSI) system, accumulator (ACC), and low-pressure safety injection (LPSI) system effectively preserves core integrity, ultimately enabling a transition to the long-term cooling phase. These findings confirm that adherence to appropriate operational protocols can ensure reactor safety even under degraded safety system conditions.

Keywords: SGTR, PKL, RELAP5, Numerical simulation

Abbreviations:
SGTR – Steam Generator Tube Rupture
OECD – Organization for Economic Co-operation and Development
CSNI – Committee on the Safety of Nuclear Installations
NEA – Nuclear Energy Agency
ISP – International Standard Problem
RELAP – Reactor Excursion and Leak Analysis Program
HPSI – High Pressure Safety Injection System
ACC – Accumulator
LPSI – Low Pressure Safety Injection System
CET – Core Exit Temperature
SOT – Start of Test
EOT – End of Test

1. Introduction

Steam generator tube failure incidents have been documented during the long-term commercial operation of pressurized water reactors (PWRs). In response, the nuclear industry has introduced a range of mitigation strategies to minimize the likelihood of such occurrences \cite{1-3}. These measures include enhancements in steam generator design, rigorous monitoring of secondary-side water chemistry, and the adoption of advanced non-destructive examination techniques. While these improvements have substantially increased steam generator reliability, extreme external events may still induce tube ruptures due to mechanical stresses surpassing design-basis limits \cite{4}. Consequently, SGTR accidents and their resultant thermal-hydraulic effects on the reactor coolant system remain a critical area of experimental research.

The ISP program represents a collaborative multinational framework in which participating organizations employ standardized experimental facility data to benchmark reactor safety analysis codes, thereby evaluating their predictive validity and computational accuracy \cite{5}. Aligned with preceding findings, OECD/NEA/CSNI endorsed ISP-52, which leverages experimental data from the PKL III J5.1 experiment \cite{6}. As a continuation of the OECD/NEA-PKL4 project’s research scope, ISP-52 specifically investigates thermal-hydraulic phenomena under design extension conditions, including two-phase flow dynamics during medium-break loss-of-coolant (LOCA) accidents, temporal evolution of the "cold dome" configuration across varied accident sequences, and transient thermal-hydraulic responses during small-break LOCA occurring at the reactor pressure vessel head. Additionally, the project validates innovative safety technologies through two key assessments: core cooling efficacy under localized flow blockage scenarios at the reactor pressure vessel’s flow distribution plate, and operational performance of passive residual heat removal systems.

To establish a comprehensive understanding of the experimental configurations and simulation parameters associated with the PKL test facility and SGTR accident scenarios, extensive prior research has been performed. Umminger et al. \cite{7} reviewed PKL experiments on PWR accident thermal-hydraulic behavior, covering LOCA, boron dilution, and main steam line break scenarios. The experimental findings validated core cooling mechanisms and provided data for code verification, emergency procedures, and operator training. Al-Yahia et al. \cite{8} simulated passive safety condenser performance during station blackout (SBO) at PKL/SACO using TRACE, finding it effectively removed core decay heat, ensuring safety within 72 hours. Bousbia Salah \cite{9} simulated natural circulation experiments at PKL using CATHARE. Comparing 1D single/multi U-tube and 3D inlet/inlet-outlet multi U-tube models, the 1D multi U-tube showed better overall predictions, while 3D modeling better reproduced key physical phenomena. Carlos et al. \cite{10} analyzed loss of residual heat removal system (RHRS) during refueling shutdown at PKL using RELAP5/MOD3.3. The code demonstrated predictive capability for mid-loop conditions but required improvements in complex flow and boron concentration modeling. Mukin et al. \cite{11} simulated OECD/NEA PKL-3 SBO experiments using TRACE. Secondary side depressurization showed limited cooling during SG drain-out, while primary side depressurization with accumulator injection delayed core overheating. Freixa et al. \cite{12} simulated three SBO scenarios under different boundary conditions at PKL using RELAP5. The code reproduced thermal-hydraulic phenomena well but exhibited event sequence deviations, attributed to uncertainties in heat loss estimation, pressurizer entrainment, condensation modeling, and boundary conditions. Martinez-Quiroga et al. \cite{13} summarized semi-blind benchmark results from 16 international institutions using system codes for OECD/NEA PKL-4 intermediate break loss-of-coolant accident (IBLOCA) experiments. The codes demonstrated reasonable prediction capability for key phenomena like break flow and natural circulation termination, while discrepancies were observed in complex scenarios involving emergency core cooling bypass, countercurrent flow limitation, and condensation modeling. Freixa et al. \cite{14} simulated IBLOCA experiments from OECD/NEA PKL-4 using RELAP5/MOD3.3 and validated the methodology via best estimate plus uncertainty analysis. This approach accurately predicted key parameters like core uncovery time and peak cladding temperature, with most experimental results residing within uncertainty bands. Jonnet et al. \cite{15} simulated a main steam line break accident at PKL using RELAP5/MOD3.2. The model reproduced key accident phenomena well, though calculated minimum coolant temperatures in the core exceeded experimental values, and SG-2 depressurization was slower than observed.

Xu et al. \cite{16} simulated the PKL I2.2 IBLOCA experiment from OECD/PKL4 using ATHLET and developed a two-layer sensitivity analysis method. The critical flow model exhibited high sensitivity for most response parameters, particularly affecting secondary pressure and peak cladding temperature. Yang et al. \cite{17} simulated SBLOCA superposed SBO using RELAP5 to investigate primary circuit characteristics under steam bypass discharging system operation. The system played a critical role in the initial phase, effectively reducing primary circuit pressure through rapid steam discharge. Bryk et al. \cite{18} performed parametric studies on PKL SG heat transfer mechanisms during RHRS failure under PWR cooldown conditions with nitrogen/steam/liquid multiphase presence. Coolant inventory variations significantly affected collapsed liquid level distribution and condensation zones, thereby influencing core temperature and pressure stability. Umminger et al. \cite{19} investigated natural circulation recovery after SBLOCA in PKL using S-RELAP5. Despite symmetric ECC injection, natural circulation exhibited significant asymmetries across loops. Sui et al. \cite{20} simulated SGTR accidents in HPR1000 using RELAP5, comparing two mitigation strategies. Balancing steam generator overflow risk against radioactive release was essential. Martin et al. \cite{21} simulated SGTR accidents using RELAP5 based on NEA/OECD ROSA-2 LSTF Test 4 and scaled the model with power-volume scaling tool. The inconsistent length-to-diameter ratio of SGTR components was the primary cause of scaling distortion.

In summary, numerical simulation methodologies for SGTR accident analysis have garnered significant research interest and demonstrated considerable advancement. Nevertheless, the coupled parameter interactions between the residual primary and secondary circuits under degraded safety system conditions necessitate further systematic investigation.

In this study, to investigate SGTR accident progression with restricted safety systems, a comprehensive RELAP5 model of the PKL test facility was developed. Through numerical simulation of SGTR scenarios validated against experimental data, this work characterizes the transient response of critical thermal-hydraulic parameters, including system pressures, mass flow rates, temperature distributions, and differential pressure across key components.

2.1 Test Facility Description

The PKL test facility simulates the nuclear steam supply system of a Siemens KWU-designed 1300 MWe pressurized water reactor, with the German Philippsburg Unit 2 nuclear power plant serving as the reference plant \cite{22}. This facility employs scaling ratios of 1:1 in height, 1:12 in diameter, and 1:145 in volume/power. With a maximum power capacity of 2.5 MW (equivalent to 10% of scaled full power), it operates at primary/secondary circuit pressures up to 46 bar and 60 bar, respectively.

The electrically heated core comprises 314 fuel rod simulators, each featuring three independently controlled power zones. The PKL test facility comprehensively models the reactor primary coolant system and replicates all relevant safety and auxiliary systems, including HPSI, ACC, LPSI, and emergency feedwater system. Critical secondary side components such as steam generators, main steam lines, and feedwater piping are also integrated \cite{23}.

2.2 Nodalization Model

The overall nodalization of the PKL test facility is shown in Figure 1 [FIGURE:1]. The model simulates the complete primary circuit system, including the pressure vessel and four symmetrically arranged independent loops. Each loop comprises one steam generator, cold leg, hot leg, and associated valves and piping. The surge line connects to the pressurizer at the mid-section of the hot leg of Loop 2. For the secondary side, the downcomer, steam separator, feedwater line, steam discharge line, steam discharge volume, and the overpressure protection system are modeled. Regarding safety systems, the HPSI, ACC, and LPSI are modeled. The remaining safety systems are not modeled as they are assumed unavailable during the simulated accident scenario.

Components in the diagram are classified using alphabetic prefixes: P for pipe components, B for branch components, V for valve components, TDV for time-dependent volume components, and PP for pump components. Elements outlined with bold black borders specifically indicate the placement of heat structures at these locations.

Fig. 1 PKL nodalization

The pressure vessel of the PKL test facility features a cylindrical design. Components 103 and 104 represent the inlet ring segments of the pressure vessel. The inlet sections for Loops 1&2 and Loops 3&4 are simplified, consolidating the original four inlets into two. Components 106 and 107 model the downcomer segments. Coolant flows through the downcomers, converges in the lower head 101 and lower support plate 102 sections, and subsequently enters the core section 108 for heating. A core bypass segment 109 is also modeled. The RELAP5 code employs nodalization to model and calculate each control volume. The number of nodes assigned significantly impacts the final calculation results, necessitating sensitivity analysis on nodal division. Following comprehensive computational verification, the optimal number of nodes was determined. The core critical segment is divided into 18 nodes.

The coolant heated in the core flows through the upper unheated core section 111 to the core outlet segment 112, forming the pressure vessel internal loop. The core section 108 incorporates heat loss components to simulate thermal losses during experiments. Bypass lines connect the upper head to the pressure vessel inlet, running parallel to the cold legs of all four loops. These lines feature 2.9 mm orifice plates to model internal flow resistance. The bypass mass flow constitutes 0.35% of the total core mass flow rate. Due to the simplification of the inlet sections, the number of bypass lines was correspondingly reduced to two.

The PKL test facility comprises four identical loops. Given this symmetry, only one loop (Loop 1) will be described in detail. Pipe component 201 models the primary system hot leg. Component 203 models the U-tube section. Component 204 represents the U-tube outlet downcomer, simulating the inlet section at the break location. Components 205 and 206 model the transition section and the main coolant pump, respectively. The cold leg, modeled by component 207, connects back to the pressure vessel to complete the primary circuit circulation path.

The feedwater system is simplified in the model, utilizing a time-dependent volume 221 to simulate feedwater flow. Feedwater combines with water separated by the steam separator in the upper ring header at component 218. This mixture flows through the secondary side downcomer 219 into the steam generator cylinder section 215. For modeling purposes, the two original downcomers were consolidated into one. The combined feedwater is heated within the steam generator cylinder section before entering the steam separator 216. Following separation, steam flows into the upper steam dome 217. Accumulated steam discharges through steam line 223, passes through the steam discharge valve 226, and enters the steam discharge volume 227. A time-dependent volume controls the secondary side pressure boundary in the simulation. A secondary-side steam-generator safety-relief valve 224 and a relief-discharge volume 225 are installed on the opposite end of the steam discharge line. This ensures the steam generator operates without overpressure during normal conditions. A steam generator bypass heater 232 is modeled at the lower section of the steam generator cylinder to compensate for thermal losses during operation.

The pressurizer connects to the mid-section of the Loop 2 hot leg via surge line segment 341. The pressurizer vessel body is modeled by component 343, with its upper head section 344 connected at the top. The upper head connects to the primary circuit pressure relief system, comprising safety valve 351 and relief valve 353. Pressure relief line 348 routes steam from the upper head to prevent primary circuit overpressurization. This pressure relief system also enables depressurization during accidents to ensure core safety. A pressurizer bypass heater 361 is modeled in the lower section to compensate for operational heat losses, which is shut down during accident transients.

For this specific accident scenario, the auxiliary feedwater system, chemical and volume control system, and secondary-side cooling system were isolated due to experimental constraints. Only the HPSI, ACC, and LPSI were activated to explore the test facility's safety performance under restricted safety system availability. Each loop contains two ACCs, originally connected to both the cold leg and hot leg. However, during this accident, the ACC connected to the hot leg was unavailable. Consequently, the model exclusively incorporates the cold leg ACC vessel 601, with injection routed through the upstream piping 602, isolation valve 603, downstream piping 604, and converging line 606. The HPSI 608 and LPSI 612 connect to the converging line via their respective isolation valves 609 and 615, enabling integrated injection into the cold legs.

2.3 Break Model

ISP-52 is developed based on the PKL J5.1 RUN 2 experiment, which simulated a scenario of SGTR occurring in all four steam generators. Specifically, RUN 2 assumed a 4A-class break in each affected SG, corresponding to a double-ended guillotine break of two heat transfer tubes per impacted steam generator.

The study simulated a break size of 3.7 mm, located at the primary-side outlet and secondary-side shell bottom of all four steam generators. Since the break configurations were identical across steam generators, the nodalization for only one break (Loop 1) is presented as shown in Figure 2 [FIGURE:2]. The break system consists of motor-operated valve and pipe components. During accident conditions, the valve activates to connect the primary and secondary sides. As illustrated in the diagram, component 204 represents the primary-side outlet section of the SG, 203 denotes the U-tube bundles, and 215 indicates the secondary-side shell of the SG. Component 618 refers to the motor-operated valve, which simulates multiple SGTRs. The outlet of this valve is connected to the bottom of the SG shell.

Fig. 2 Break nodalization

3.1 Steady-State Calculation Results

The RELAP5/MOD3.4 code was first employed to calculate the steady-state process based on initial conditions. As demonstrated in Table 1 [TABLE:1], the computational results exhibit strong agreement with experimental parameters. The RELAP5 input deck (excluding break and safety injection systems) confirms stabilized primary and secondary side parameters, with consistent heat transfer rates between both circuits. Computed temperatures and power correspond closely to experimental initial values, with relative deviations remaining within acceptable limits. This validates the model’s reliability and operational realism.

Table 1 Steady-state parameters comparison

Variable Experimental value Numerical value Relative error Core power Pressurizer pressure Upper plenum pressure SG 1 outlet pressure SG 2 outlet pressure SG 3 outlet pressure SG 4 outlet pressure Pressurizer bottom temperature Core inlet temperature Core exit temperature LOOP 1 mass flow rate LOOP 2 mass flow rate LOOP 3 mass flow rate LOOP 4 mass flow rate Pressurizer collapsed level Core collapsed level SG 1 collapsed level SG 2 collapsed level SG 3 collapsed level SG 4 collapsed level

3.2 Transient Calculation Results

In the PKL J5.1 RUN2 test, partial activation of safety systems was implemented as follows: (1) Secondary-side pressure was limited to 43 bar; (2) Secondary-side auxiliary feedwater system became inoperative post-accident; (3) Primary loop depressurization initiated via PRZ safety valves after 1000s; (4) HPSI injection triggered when CET exceeded 350°C; (5) ACC injection activated at primary pressure below 23.6 bar; (6) LPSI injection engaged at primary pressure below 12.4 bar; (7) Post-rupture valve opening, core power progressively declined with its profile shown in Figure 3 [FIGURE:3]; (8) SG bypass heaters deactivated after SG emptying; (9) PRZ bypass heaters were shut down following accident initiation.

The sequence of events during the accident is summarized in Table 2 [TABLE:2]. After the accident occurred, core power evolved according to Figure 3, dropping to approximately 550 kW at 2500s and further declining to 350 kW around 12500s. Beyond 12500s, the power reduction rate decelerated, reaching 287 kW at termination.

Table 2 Sequence of events during the accident

Event Experimental value (s) Numerical value (s) SG 1-4 feedwater isolation PRZ bypass heater shut down Start of primary side depressurization Final isolation of the main steam system Secondary side of SG1-4 empty HPSI start injection ACC start injection LPSI start injection

Fig. 3 Core heating rod power profile

Figure 4 [FIGURE:4] shows the total mass flow rate at the break. Positive flow indicates fluid transfer from the primary to secondary circuit, while negative flow signifies reverse flow from secondary to primary circuit. Figure 5 [FIGURE:5] shows primary and secondary pressure. Following break valve opening, higher primary pressure drove coolant leakage from primary to secondary circuit. The mass flow rate increases rapidly, peaking at about 0.75 kg/s. Continuous coolant loss through the break valve reduced primary pressure. Meanwhile, the main steam system maintained constant secondary pressure at 43 bar. As the primary and secondary sides pressure differential progressively narrowed, the break location exhibited a corresponding decline in mass flow rate.

Figure 6 [FIGURE:6] shows void fraction and volume equilibrium quality distribution in the core. Figure 7 [FIGURE:7] shows CET and core collapsed level. Figure 8 [FIGURE:8] gives the distribution of fuel cladding temperatures in the core. At 1000 seconds, the opening of the PRZ depressurization valve triggered primary pressure reduction, during which the mass flow rate at the break showed good agreement. However, around 2500s, as the break flow had just reversed and failed to fully compensate for the primary coolant loss, the void fraction and volume equilibrium quality in the upper core nodes 10-18 gradually increased, leading to partial core uncovering and a decline in core collapsed level, which induced fluctuations in the break mass flow. As the pressure differential between the primary and secondary sides continuously narrowed, secondary-side water migrated into the primary circuit, facilitating core reflooding. Under this transient condition, the maximum CET stabilized at ~280°C, remaining below the HPSI activation threshold. The cladding temperature peaked at approximately 400°C, still within the operational safety margin. The core collapsed level reached its first minimum at 3750s due to intense steam generation within the fuel assemblies.

Subsequent continuous secondary-side fluid influx resulted in a simulated mass flow rate increase of ~0.35 kg/s. At ~15000s, the break flow diminished to zero following SG depletion. By ~25000s, LPSI initiation, coupled with reduced PRZ depressurization flow, induced a significant fluid reflux from the primary to the secondary side, thereby amplifying the simulated break flow rate. As the primary circuit pressure continued decreasing, secondary-side water refluxed into the primary side. Around 14700 seconds, the SGs progressively emptied. This SG emptying terminated coolant replenishment to the primary system, causing the void fraction and volume equilibrium quality in the upper core nodes (10-18) to increase again. Consequently, the core collapsed level began declining, leading to inevitable fuel uncovery, with the minimum level reaching approximately 3 m. This resulted in a rise in CET, accompanied by a rapid increase in the fuel cladding temperature in the upper core nodes (10-18), peaking at about 600°C, which remained within safe operating limits.

Fig. 4 Total mass flow rate at break
Fig. 5 Primary and Secondary side Pressure
Fig. 6 (a) Void fraction distribution in the core, (b) Volume equilibrium quality distribution in the core
Fig. 7 (a) Core exit temperature, (b) Core collapsed level
Fig. 8 Fuel cladding temperature distribution in the core

Figure 9 [FIGURE:9] presents the mass flow rate and pressure differential across the PRZ depressurization valve. Figure 10 [FIGURE:10] shows the PRZ collapsed level. At 1000 seconds, the PRZ depressurization valve opened, and due to the high primary circuit pressure, the mass flow rate through the valve increased rapidly, reaching a peak of approximately 0.55 kg/s. The simulated value was 0.9 kg/s. As illustrated in Figures 5 and 10, the opening of the PRZ depressurization valve caused a rapid pressure drop in the primary circuit. During the early accident phase, the low PRZ collapsed level facilitated significant coolant transfer from the primary circuit to the PRZ, resulting in a rapid rise of the PRZ collapsed level to approximately 10 m. As the PRZ collapsed level stabilized, the primary circuit pressure decline trend gradually ceased. Around 17000 seconds, the PRZ collapsed level increased due to HPSI injection. At this stage, the simulated PRZ collapsed level stabilized at approximately 10 m due to the flow rate discrepancy in the ACC injection. Around 22500 seconds, fluctuations in the mass flow rate of the PRZ depressurization valve were observed, which were attributed to the introduction of the safety injection system that caused significant coolant water influx into the PRZ. Around 25000 seconds, the depressurization flow rate of the PRZ decreased due to the reduction in the pressure differential across the PRZ depressurization valve. When the pressure differential approached zero, the mass flow rate through the PRZ depressurization valve also decreased to zero. At this stage, the simulated PRZ collapsed level rose again due to HPSI injection, eventually stabilizing at approximately 13 m.

Fig. 9 (a) PRZ depressurization valve mass flow rate, (b) PRZ depressurization valve pressure differential
Fig. 10 PRZ collapsed level

Figure 11 [FIGURE:11] shows the SG collapsed level. From Figures 4 and 11, it can be seen that the collapsed level of SG began to decrease after the accident. Around 15000 seconds, the SG collapsed level dropped to zero, resulting in complete SG depletion and causing the break mass flow rate to reach zero. At approximately 25000 seconds, as the PRZ depressurization valve flow rate decreased, a significant amount of coolant water from the primary circuit was redirected back to the SG through the break valve, leading to a simulated break mass flow rate of 0.4 kg/s. This coolant reflux caused the SG liquid level to rise again, reaching approximately 3.4 m by 27500 s.

Fig. 11 SG collapsed level

Figure 12 [FIGURE:12] shows the total mass flow rate of HPSI across four loops. When the CET exceeded 350°C, HPSI initiated injection into the cold leg of each loop at a flow rate of 0.12 kg/s. As seen in Figure 7(b), the injected cold water induced a local pressure drop due to condensation effects at the injection point, causing a decline of the core collapsed level. Observations from Figures 5, 6, and 7 indicate that HPSI injection triggered core quenching, resulting in localized violent evaporation and continuous water injection that causes a temporary pressure recovery. However, gradual pressure decline in the primary circuit continued. In the simulation, the initial ACC injection flow rate was relatively high, which accelerated the primary circuit pressure drop and consequently diminished the visibility of the pressure recovery phenomenon.

Fig. 12 HPSI total mass flow rate

Figure 13 [FIGURE:13] shows the liquid level and total mass flow rate of the ACC. As HPSI activation caused a pressure drop in the primary circuit and reached the ACC trigger threshold, the ACC initiated operation and injected water into the primary circuit, peaking at a flow rate of 0.9 kg/s. Figures 5 and 7 indicate that the combined loss from primary-side injection of HPSI and ACC compensated for coolant depressurization, sustaining core collapsed level. The cumulative effect of steam pressure in the core overwhelmed the condensation effect from coolant injection, causing pressure recovery in the primary circuit. This resulted in flow blockage of ACC injection. When the primary pressure again fell below the ACC pressure threshold, ACC resumed injection with an average mass flow rate of about 0.15 kg/s. In the simulation, the peak injection mass flow rate of ACC is higher than the experimental value, about 1.7 kg/s. This is because ACC shares the same injection pipeline with HPSI and LPSI, and there is mutual influence during the continuous injection process. To maintain continuous ACC injection, a larger initial pressure of ACC is required, which will cause the liquid level to rapidly decrease during the initial injection of ACC and result in a high peak in injection flow rate. In the later stage of simulation, due to the injection of HPSI and LPSI, the injection of ACC was blocked, and its injection flow rate decreased to 0.

Fig. 13 (a) ACC liquid level, (b) ACC total mass flow rate

Figure 14 [FIGURE:14] shows the total mass flow rate of LPSI across the four loops. As the primary pressure continued decreasing, LPSI is put into operation at 26475 seconds with an injection rate of 0.07 kg/s. All loops exhibit a symmetric water injection pattern, replenishing the primary circuit water inventory and ensuring the system enters a long-term cooling phase, thereby maintaining the overall safety of the reactor. In the simulation, due to the faster depressurization rate of primary circuit compared to the experiment, the injection time of LPSI has been advanced to some extent.

Fig. 14 LPSI total mass flow rate

4. Conclusion

This paper presents a RELAP5 simulation of the multiple SGTR accident conducted on the ISP-52 PKL test facility, with comparative analysis focusing on accident sequences, key parameters, and critical phenomena. The main conclusions are as follows:

(1) RELAP5 effectively reproduced the primary accident progression of the multiple SGTR scenario in ISP-52, demonstrating consistency with experimental data in both event chronology and key thermal-hydraulic phenomena.

(2) Following accident initiation, the main feedwater system and the pressurizer bypass heaters tripped. The coolant underwent primary-to-secondary leakage and countercurrent flow. The core temperature rose rapidly after the steam generator was depleted. Subsequent staged activation of HPSI, ACC, and LPSI maintained core integrity throughout the transient process, proving that the residual safety systems can ensure reactor safety under safety system constraints. The simulation results confirm that primary side depressurization effectively prevents large-scale radioactive coolant release to the secondary side while enabling reverse inventory transfer from secondary to primary systems, thereby mitigating potential core uncovery accidents.

In this study, RELAP5 modeling employed simplified representations of core components, U-tubes, and secondary circuit systems, introducing inherent limitations. Subsequent research will refine modeling for the core and U-tubes, including implementation of multichannel core models to evaluate the impact of channel quantity on simulation accuracy.

Acknowledgments: This study is part of the ISP project under the WGAMA task framework, with particular acknowledgment to the support of NEA/OECD.

References

[1] M. Liu, L. Wang, Y. Lee, Diagnosis of break size and location in LOCA and SGTR accidents using support vector machines, Prog. Nucl. Energy 140 (2021) 103902. doi: 10.1016/j.pnucene.2021.103902.

[2] A. Ayodeji, Y. Liu, SVR optimization with soft computing algorithms for incipient SGTR diagnosis, Ann. Nucl. Energy 121 (2018) 89–100. doi: 10.1016/j.anucene.2018.07.011.

[3] J.M. Izquierdo-Rocha, M. Sánchez-Perea, Application of the integrated safety assessment methodology to the emergency procedures of a SGTR of a PWR, Reliab. Eng. Syst. Saf. 45 (1994) 159–173. doi: 10.1016/0951-8320(94)90083-3.

[4] T. Lind, S. Campbell, L.E. Herranz, M. Kissane, J. Song, A summary of fission-product-transport phenomena during SGTR severe accidents, Nucl. Eng. Des. 363 (2020) 110635. doi: 10.1016/j.nucengdes.2020.110635.

[5] H. Karwat, CSNI international standard problems (ISP): brief descriptions (1975-1999), OECD Nuclear Energy Agency, 2000.

[6] Framatome, "ISP52 – Information on Test OECD/ETHARINUS J5.1", Excerpt from N° D02-ARV-01-208-128 Rev. 0, Framatome GmbH, Erlangen, Germany.

[7] K. Umminger, L. Dennhardt, S. Schollenberger, B. Schoen, Integral test facility PKL: experimental PWR accident investigation, Sci. Technol. Nucl. Install. 2012 (2012) 891056. doi: 10.1155/2012/891056.

[8] O.S. Al-Yahia, I. Clifford, H. Ferroukhi, Assessment of TRACE code for modeling of passive safety system during long transient SBO via PKL/SACO facility, Nucl. Technol. (2024) 10.1016/j.net.2024.02.050.

[9] A. Bousbia Salah, CATHARE simulation results of the natural circulation characterisation test of the PKL test facility, Nucl. Eng. Technol. 53 (2021) 1446–1453. doi: 10.1016/j.net.2020.11.016.

[10] S. Carlos, J.F. Villanueva, S. Martorell, V. Serradell, Analysis of a loss of residual heat removal system during mid-loop conditions at PKL facility using RELAP5/Mod3.3, Nucl. Eng. Des. (2008) 10.1016/j.nucengdes.2008.04.013.

[11] R. Mukin, I. Clifford, O. Zerkak, H. Ferroukhi, Modeling and analysis of selected organization for economic cooperation and development PKL-3 station blackout experiments using TRACE, Nucl. Eng. Technol. 50 (2018) 356–367. doi: 10.1016/j.net.2017.12.005.

[12] J. Freixa, V. Martinez-Quiroga, F. Reventós, Modelling guidelines for safety analysis of station black out sequences based on experiments at the PKL test facility, Nucl. Energy (2020) 10.1016/j.anucene.2019.107179.

[13] V. Martinez-Quiroga, M. Szogradi, S. Schollenberger, M. Sanchez-Perea, N. Sandberg, J. Zhongyun, P. Freydier, M. Suslov, H. Austregesilo, T. Glantz, J.H. Lee, Z. Li, I. Shvetsov, R. Mukin, J. Holmstrom, J.F. Villanueva, E. Virtanen, J. Freixa, OECD/NEA PKL-4 benchmark activity. Code assessment of the relevant phenomena associated to a blind IBLOCA experiment, Nucl. Eng. Des. 389 (2022) 111632. doi: 10.1016/j.nucengdes.2021.111632.

[14] J. Freixa, V. Martínez-Quiroga, M. Casamor, F. Reventós, R. Mendizábal, M. Sánchez-Perea, On the validation of BEPU methodologies through the simulation of integral experiments: application to the PKL test facility, Nucl. Eng. Des. 379 (2021) 111238. doi: 10.1016/j.nucengdes.2021.111238.

[15] J.R. Jonnet, M.M. Stempniewicz, A. de With, P.H. Wakker, RELAP5 analysis of PKL, main steam line break test, Nucl. Eng. Des. 265 (2013) 755–764. doi: 10.1016/j.nucengdes.2013.09.010.

[16] H. Xu, A.F. Badea, X. Cheng, Sensitivity analysis of thermal-hydraulic models based on FFTBM-MSM two-layer method for PKL IBLOCA experiment, Ann. Nucl. Energy 147 (2020) 107732. doi: 10.1016/j.anucene.2020.107732.

[17] S. Yang, X.-B. Li, Y.-S. Liu, J. Xu, D. Zhang, Response characteristics of PWR primary circuit under SBLOCAs considering steam bypass discharging, Nucl. Sci. Tech. 35 (2024) 101. doi: 10.1007/s41365-024-01457-7.

[18] R. Bryk, S.P. Schollenberger, L. Dennhardt, K. Świrski, Parameter study at the PKL test facility on heat transfer mechanisms in the steam generator in presence of steam and nitrogen, liquid, Energy (2020) 10.1016/j.energy.2020.118612.

[19] K. Umminger, R. Mandl, R. Wegner, Restart of natural circulation in a PWR–PKL test results and s-relap5 calculations, Nucl. Eng. Des. 215 (2002) 39–50. doi: 10.1016/S0029-5493(02)00040-7.

[20] D. Sui, D. Lu, C. Shang, Y. Wei, X. Xu, Investigation on response of HPR1000 under different mitigation strategies after SGTR accident, Ann. Nucl. Energy 112 (2018) 328–336. doi: 10.1016/j.anucene.2017.10.024.

[21] K. Martin, V. Martinez-Quiroga, J. Freixa, Full scope scaling analysis and recommendations for PWR SGTR scenarios, Nucl. Eng. Des. 418 (2024) 112864. doi: 10.1016/j.nucengdes.2023.112864.

[22] H. Xu, Review and outlook of the integral test facility PKL III corresponding studies, Kerntechnik 86 (2021) 391–399. doi: 10.1515/kern-2021-1011.

[23] K. Umminger, L. Dennhardt, S. Schollenberger, B. Schoen, Integral test facility PKL: experimental PWR accident investigation, Sci. Technol. Nucl. Install. 2012 (2012) 1–16. doi: 10.1155/2012/891056.