Research on the Activation Strategies of Passive Decay Heat Removal Systems in a Pool-Type SFR by Three-Dimensional Numerical Simulation

Abstract

A Decay Heat Removal System (DHRS) is an essential passive safety feature in pool-type Sodium-Cooled Fast Reactors (SFRs), maintaining core temperatures within design limits via natural circulation after reactor scram. Operation of the DHRS is regulated by the damper of the Air Heat Exchanger (AHX), which controls its activation and shutdown. In the current design guidelines, it is typically recommended to initiate the Decay Heat Exchanger (DHX) at 600 s after a Station Blackout (SBO) event. However, this activation timing requires minor dynamic adjustment based on the transient response of the system, which can be obtained by either real-reactor experiments or numerical simulations. Since full-scale real-reactor experiments are not easy to conduct, numerical simulations are effective ways to enhance the passive safety performance of pool-type SFRs under SBO conditions, clarify the regulatory mechanism of DHX activation timing on system behavior, and optimize DHRS operational strategies. This study developed an integrated full-reactor three-dimensional numerical model that comprehensively incorporated key components such as the core, sodium pools, and DHX. Transient variations in power and boundary conditions were precisely controlled via User-Defined Functions (UDFs). The impact of different DHX activation strategies on the reactor’s decay heat removal capability was systematically analyzed. Three-dimensional numerical simulations were performed for three representative DHX operational strategies, immediate activation post-accident (0 s), delayed activation per the standard strategy (600 s), and complete DHX non-activation, yielding detailed temperature and flow field distributions within the reactor. Results demonstrate that under the standard strategy, not only can the temperature in the pool be controlled below the safety limit (550 °C) in the early stage but the temperature can also drop in the subsequent stage while retaining a 600 s safe operation threshold. Notably, the results reveal that “sooner is not always better”. Immediate DHX activation accelerates internal circulation and drives hot fluid downwards, paradoxically heating the cold pool faster than delayed activation, thereby resulting in a higher core outlet temperature. This study contributes to enhancing the credibility of passive safety in SFRs and provides reliable data to support the development of optimized reactor operation protocols.

1. Introduction

SFRs represent one of the pivotal Generation IV reactor types, distinguished by their advantages in high-temperature operation and high thermal efficiency, inherent safety features, and fuel breeding capability. They effectively address the resource limitations and waste management challenges associated with traditional nuclear energy, positioning themselves as a core technology within sustainable energy systems. Their engineering maturity and continuous innovation have established them a leading role among Generation IV reactor designs [1,2,3,4]. Numerous nations have invested in SFR research and deployment, with notable examples including China’s CEFR [5], Russia’s BN-600, France’s Phenix, Japan’s Monju, the U.S. EBR-∥, and India’s PFBR [6].

SBO is one of the key accident conditions extensively studied in nuclear power plant operation. The DHRS is a crucial system for accident mitigation during SBO, serving as the ultimate barrier for achieving passive safety. Its reliability is directly related to reactor safety. The DHRS functions to remove decay heat generated in the core following the loss of off-site power, the unavailability of the main pumps, and the consequent loss of external driving force for the sodium coolant, thereby preventing core meltdown and more severe accidents [7]. The effectiveness of the DHRS relies on the establishment of natural circulation following the accident. The DHX, as the core component of the DHRS, has an activation timing that is a key controllable parameter influencing both the natural circulation establishment process and the decay heat removal efficiency. The general schematic diagram of DHRS is illustrated in Figure 1. Hot sodium from the upper region of the hot pool enters the DHX through the inlet grill, subsequently flowing downward through the annular channel formed between the outer tube and the exchanger shell. During this downward flow, heat is transferred to the secondary sodium on the tube side, before the cooled sodium returns to the lower section of the hot pool through the bottom outlet grill. Simultaneously, cooler secondary sodium enters the DHX through the top inlet, descends along the central downcomer, reverses direction at the bottom elliptical head, and flows upward through the heat exchanger tube bundle arranged on the shell side, where efficient heat exchange occurs. The heated secondary sodium then flows to the AHX, where the residual heat is ultimately rejected to the atmospheric environment.

In pool-type SFRs, major components—including the core, primary pumps, Intermediate Heat Exchangers (IHXs), and DHX—are fully immersed within a unified sodium pool, resulting in a structurally integrated design. Together with the core housed in the main vessel, these constitute the reactor’s primary circuit. Consequently, conducting experimental studies on in-reactor thermal–hydraulic phenomena is highly challenging. One-dimensional system codes or three-dimensional numerical simulation methods are typically employed to identify thermal–hydraulic phenomena within the reactor. Among these, one-dimensional system codes are widely used to simulate system responses during accident scenarios and reflect variations in key parameters, such as FLOW++ [8], MARS-LMR [9], SAC-CFR [10], THACS [11,12] and ACENA [13,14]. Nevertheless, as nuclear reactor safety criteria become increasingly stringent, reliance solely on one-dimensional parameter-based assessments is now considered inadequate for delivering a holistic safety evaluation. Three-dimensional numerical simulations, in contrast, offer the capability for high-fidelity modeling of specific internal components alongside full-reactor global analysis. This approach successfully captures the intricate three-dimensional thermal–hydraulic features present in local regions, key equipment, and the entire system under diverse operating states. Zaryugin et al. developed a quarter model of the BN-1200 reactor and performed three-dimensional numerical simulations of the process from rated power operation to long-term cooling, obtaining transient variations in key parameters and temperature field distributions. This work allowed for analysis and validation of the effectiveness of the DHX in the DHRS [15]. Jiang et al. present a systematic numerical validation framework for thermal–hydraulic analysis of helical-clad seven-rod bundles using STAR-CCM+, demonstrating exceptional predictive accuracy across laminar, transitional, and turbulent regimes. Particularly noteworthy are the model’s robust capabilities in capturing flow resistance characteristics, complemented by dimensional similarity analyses establishing critical scaling relationships for bundle simulations. The comprehensive evaluation methodology and stringent validation benchmarks developed in this work provide essential guidance for high-fidelity Computational Fluid Dynamics (CFD) modeling of advanced nuclear fuel assemblies [16]. Through a detailed 3D-CFD analysis of a pool-type SFR’s main vessel cooling system, Vivek et al. analyzed flow uniformity and circumferential temperature variations. Their findings established a basis for determining the requisite sodium flow rate in such systems [17]. Zhou et al. develop a novel Open FOAM-based deterministic neutron transport solver implementing the SP₃ approximation method to overcome limitations of diffusion theory in advanced reactor applications. The solver demonstrates superior accuracy. The implementation features innovative Marshak boundary conditions and non-orthogonal mesh corrections, validated across 1D-3D problems including BWR fuel assemblies and fast reactor cores [18]. A research team from North China Electric Power University adopted a modular modeling and integrated calculation approach to perform numerical simulations of steady-state and long-term transient conditions for typical design scenarios of the CEFR. These included normal operation (rated power operation), anticipated transients without scram (such as loss of feedwater in the steam generator [19], emergency shutdown at rated power [20], and trip of one primary circulation pump [21]), as well as emergency transients including seizure of one primary circulation pump [22] and seizure of one secondary circulation pump [23]. The simulations replicated the global flow phenomena within the sodium pool under anticipated operational occurrences, yielding three-dimensional temperature distributions of the sodium pool and key internal components. This provides valuable numerical references for the structural design and safety assessment of the reactor vessel and internal structures in pool-type SFRs.

Currently, two relatively mature decay heat removal strategies have been established for fast reactors: the secondary loop cooling approach and the direct cooling configuration with the DHX immersed in the hot pool [24]. The secondary loop cooling method represents a conventional yet well-developed design approach, implemented in reactors such as the French Phénix [25], Japanese Monju [26], and Russian BN-600 and BN-800 [27]. This configuration positions the ultimate heat sink on the secondary side of the main heat transport system, facilitating heat transfer through the IHX. It offers advantages of structural simplicity and well-defined fluid flow paths. However, this system’s effectiveness is inherently dependent on the configuration of the main heat transport system’s secondary loop and the IHX design, necessitating the classification of the secondary loop as a safety-grade system. This approach demonstrates relatively lower safety margins and limited decay heat removal capacity. An alternative design increasingly adopted in modern reactors employs the direct cooling configuration with the DHX located in the hot pool, as implemented in the EFR, CEFR [28], and DSFR. This design integrates the independent heat exchanger directly within the reactor vessel, connected via piping to elevated air coolers outside the vessel to form a dedicated passive loop. The direct immersion of the DHX within the reactor vessel enables immediate core cooling following reactor scram, providing significant advantages in rapid system response and enhanced decay heat removal capability. Correspondingly, this configuration introduces greater complexity in both structural design and flow path arrangements.

The DHX serves as the core component of the DHRS. Its design and operational deployment during accidents critically influence the natural circulation and decay heat removal capability of the reactor. Consequently, numerous scholars have conducted research in this area. Researchers at the KAERI systematically validated the heat transfer models for the IHX and DHX using the developed MARS-LMR code. By integrating experimental data from Japan’s JOYO fast reactor and Korea’s STELLA-1 facility, they performed a comparative analysis of various heat transfer correlations, providing a significant foundation for the safety analysis of prototype Generation IV SFRs [29]. Vikram G integrated a complete system model, including key components such as a steam separator and a decay heat removal condenser, into the self-developed DYANA-P code. Simulations of typical fast reactor scram scenarios verified that two decay heat removal schemes—operating at 170 bar (high pressure) and 15 bar (low pressure)—could achieve cooldown rates of 20 K/h and 60 K/h, respectively [30]. Mente et al. established a 1:4 scale primary circuit test facility named SAMRAT, using water as the working fluid, to analyze the role of the DHX in core decay heat removal. They conducted multiple experimental studies investigating various pool temperatures, different core flow paths, and different DHX operational states [31]. Cheng performed a three-dimensional steady-state simulation using FLUENT on a quarter-section of the cold sodium pool. The study obtained the flow and temperature distributions near the DHX in the cold pool, thereby validating the thermal–hydraulic design of the cold pool DHX under standby conditions. The results indicated that inter-subassembly flow plays a significant role in decay heat removal. Under DHX failure conditions, the core outlet and hot pool temperatures were 13 °C higher than in normal operation, and the degree of thermal stratification in the hot pool was reduced [32]. Andre et al. introduced a novel concept for a passive decay heat removal system tailored for liquid metal fast reactors. Utilizing CFD analysis, they developed a direct heat exchanger based on radiative heat transfer. This DHX features a triple-walled tube structure, achieving complete isolation between the primary and secondary side fluids via a vacuum interlayer [33]. In previous research by our team, Liang identified the natural circulation flow paths and analyzed thermal stratification within the sodium pool of the CEFR under SBO conditions [34]. Zhao et al. employed numerical simulation methods to systematically study the impact of different DHX placement schemes on natural circulation characteristics in a pool-type SFR during an SBO accident. Comprehensive three-dimensional CFD simulations confirmed that a dual-DHX arrangement (located in both the hot and cold pools) significantly enhances natural circulation capacity, and accelerates the transition to the long-term cooling phase compared to the traditional single-DHX (hot pool only) arrangement [35].

However, the previous studies have primarily concentrated on the design or layout aspects of the DHX, largely overlooking a critical engineering practice issue: during an actual transient, the activation and deactivation of the DHX constitute an active control strategy requiring dynamic adjustment. Currently, there is a lack of systematic, high-fidelity numerical research on how the DHX activation timing influences the reactor’s three-dimensional transient thermal–hydraulic response, particularly regarding the controlling mechanism over the natural circulation establishment process. This constitutes the core knowledge gap that this paper aims to fill. To address this, the present study employs a high-fidelity three-dimensional CFD methodology to construct a full-core model of the CEFR. For the first time, it systematically compares three typical (including extreme) DHX operational strategies: immediate activation post-accident (0 s), delayed activation per standard strategy (600 s), and complete non-activation of the DHX. The study reveals the regulation effect of DHX activation timing on the thermal driving force of natural circulation and the associated three-dimensional flow field structures. The study breaks through the limitations of traditional simplified models, achieving a full-reactor complex spatial simulation of the system-wide three-dimensional transient thermal–hydraulic characteristics under an SBO scenario. This research provides crucial numerical evidence for formulating and optimizing passive operational strategies for subsequent SFRs. Furthermore, it offers a theoretical foundation and safety guidance for nuclear power plant operational protocols, underscoring its significant engineering value and safety implications.

2. Geometry and Simulation Methods

2.1. Geometry and Meshing

The CEFR comprises two parallel primary loops, each equipped with a primary main circulation pump, an IHX, and associated piping and pressure boundary components. The main vessel is filled with liquid sodium, and an in-vessel support plate divides it vertically, forming an upper hot sodium pool and a lower cold sodium pool. The main pump draws sodium from the cold pool and directs it into the core inlet via the grid plate. After being heated in the core, the sodium flows upward into the hot sodium pool. It then enters the IHX, located in the upper section of the hot pool, and flows downward while transferring heat to the secondary sodium. The cooled primary sodium exits the IHX into the cold pool and returns to the main pump suction, thereby completing the primary circuit. Figure 2 illustrates the schematic structure of the CEFR.

In our previous work, a full-scale integrated model of the CEFR was developed, and comprehensive mesh generation and sensitivity analysis were performed. A modular meshing strategy was adopted, dividing the reactor into three distinct regions: the hot pool, cold pool, and core. Owing to their relatively regular geometries, the hot pool and core modules were discretized using structured meshes. In contrast, the complex geometry of the cold pool necessitated the use of an unstructured mesh. A detailed mesh sensitivity study and quality verification have been conclusively conducted in our prior research [34]. Balancing computational accuracy and resource requirements, a final model comprising approximately 58 million cells was selected for the simulations, as depicted in the corresponding Figure 3.

While ensuring numerical accuracy, the model incorporates geometric simplifications of internal structures and relevant components. The core assemblies are simulated using a homogenized porous medium approach, with power transients controlled via UDF. The DHX and IHX contain numerous heat transfer tubes with complex geometries, serving as critical components for both heat transfer and flow resistance. These exchangers are geometrically simplified while maintaining their overall configuration and dimensions, and are represented using representative porous media models to simulate their flow and heat transfer characteristics. The viscosity and inertia resistance coefficients of porous media in various parts have been determined by numerical experiments in the previous work [35]. Therefore, they are not repeated in this paper. To strictly justify the accuracy of the FLUENT-based methodology applied in this study, rigorous Verification and Validation procedures were executed and documented in our preceding work. Specifically, the current numerical computational scheme, including the development of physical models, the selection of turbulence models, and the basic settings of boundary conditions, has been systematically benchmarked against full-scale natural circulation experimental data from the French PHENIX reactor and the Japanese PLANDTL experimental facility [36,37,38]. The quantitative agreement between our numerical predictions and the experimental measurements conclusively verified the reliability and high fidelity of this calculation method for resolving complex thermal–hydraulic phenomena in sodium pool systems.

2.2. Model

The three-dimensional numerical simulation was performed using the commercial CFD software FLUENT(2023R1). FLUENT is based on the Navier–Stokes equations, simulating the dynamic behavior of fluid flow by solving the governing equations for mass, momentum, and energy conservation [34].

$${\displaystyle \frac{\partial \rho }{\partial t}}+\nabla ·\left(\rho \overrightarrow{\nu }\right)=S_m$$

(1)

$${\displaystyle \frac{\partial }{\partial \mathrm{t}}}\left(\mathrm{\rho }\overrightarrow{\mathrm{\nu }}\right)+\nabla ·\left(\mathrm{\rho }\overrightarrow{\mathrm{\nu }}\overrightarrow{\mathrm{\nu }}\right)=-\nabla \mathrm{p}+\nabla ·\left(\stackrel{̿}{\tau }\right)+\overrightarrow{\mathrm{G}}+\overrightarrow{\mathrm{F}}$$

(2)

$${\displaystyle \frac{\partial \left(\mathrm{\rho }\mathrm{E}\right)}{\partial \mathrm{t}}}+\nabla ·\left(\overrightarrow{\mathrm{\nu }}\left(\mathrm{\rho }\mathrm{E}+\mathrm{p}\right)\right)=-\nabla \left(\sum _{\mathrm{j}}{\mathrm{h}}_{\mathrm{j}}{\mathrm{J}}_{\mathrm{j}}\right)+{\mathrm{S}}_{\mathrm{h}}$$

(3)

Considering the intricate geometry of sodium pool and the predominantly turbulent flow regime, combined with the previous work of our team and the relevant references [37,38,39,40,41], the standard k-ε turbulence model is selected for this study. The primary governing equations are as follows:

$${\displaystyle \frac{\partial }{\partial t}}\left(\rho k\right)+{\displaystyle \frac{\partial }{\partial x_i}}\left(\rho k{u}_i\right)={\displaystyle \frac{\partial }{\partial x_j}}\left[\left(\mu +{\displaystyle \frac{{\mu }_t}{{\sigma }_k}}\right){\displaystyle \frac{\partial k}{\partial x_j}}\right]+G_k+G_b-\rho \epsilon$$

(4)

$${\displaystyle \frac{\partial }{\partial t}}\left(\rho \epsilon \right)+{\displaystyle \frac{\partial }{\partial x_i}}\left(\rho \epsilon u_i\right)={\displaystyle \frac{\partial }{\partial x_j}}\left[\left(\mu +{\displaystyle \frac{{\mu }_t}{{\sigma }_{\epsilon }}}\right){\displaystyle \frac{\partial \epsilon }{\partial x_j}}\right]+C_{1\epsilon }{\displaystyle \frac{\epsilon }{k}}\left(G_k+C_{3\epsilon }{G}_b\right)-C_{2\epsilon }\rho {\displaystyle \frac{{\epsilon }^2}{k}}$$

(5)

${\mathsf{\mu }}_{\mathrm{t}}$ is turbulent viscosity, and ${\mu }_t=\rho C_{\mu }{\displaystyle \frac{k^2}{\epsilon }}$

The model constants are determined from extensive experimental data, exhibit broad applicability for turbulent flows and are assigned as follows [19]:

$C_{1\epsilon }=1.44$; $C_{2\epsilon }=1.92$; $C_{\mu }=0.09$; ${\sigma }_k=1.0$; ${\sigma }_{\epsilon }=1.3$.

The homogenized porous medium simplification requires the incorporation of an equivalent flow resistance coefficient to replicate the flow resistance imposed by solid structures in the actual reactor. This is achieved by introducing a momentum source term dependent on velocity, expressed as:

$$S_i=-\left(D\mu {\upsilon }_i+C_2{\displaystyle \frac{1}{2}}\rho \left|\upsilon \right|{\upsilon }_j\right)$$

(6)

where $D$ is the viscous resistance coefficient; which can be reformulated as $1/\alpha$, $\alpha$ is permeability; $C_2$ is the inertial resistance coefficient, $\upsilon$ is the velocity.

2.3. Boundary Settings

SIMPLE algorithm was used to calculate. The second-order upwind scheme was used for the discretization of pressure, momentum and energy spaces. The Courant number is set to 5. Steady state was taken as the initial state, followed by transient calculations from 0 to 2500 s. The selection of time step considers both the calculation accuracy and calculation economy.

The model configurations and associated settings for the three DHX activation strategies remain identical to ensure comparability, including all boundary conditions (except DHX power) and solution methodologies. Porous media regions incorporate volumetric source terms to represent heating and cooling effects.

3. Results and Discussion

3.1. Steady Simulation

3.1.1. Boundary Conditions for Steady Simulation

The numerical methodology employed in this study has been previously validated in our earlier work. The boundary conditions for the transient simulations were defined using the 100% power steady-state operational parameters of the CEFR.

The boundary conditions applied for the steady-state baseline are specified in the

Table 1. Boundary conditions for steady simulation.

Item	Value
Reactor core thermal power/(MW)	65
cooling power of each IHX/(MW)	16.25
cooling power of each DHX/(MW)	0.0525
Mass flow of core/(kg/s)	301
Mass flow through RVCS/(kg/s)	40

below:

3.1.2. Steady Simulation Results

The key computational results for the CEFR operating at 100% power under steady-state conditions are summarized as follows:

• Flow Distribution

The overall flow field distribution within the CEFR primary sodium pool is depicted in Figure 4. Under normal operating conditions, the primary circulation pumps draw sodium from the cold pool and deliver it to the grid plate plenum. Following flow distribution by the grid plate, the sodium enters the core from the bottom. The heated coolant exits the core into the hot sodium pool, subsequently passes through the overflow window, and enters the IHX. While flowing downward within the IHX shell side, the primary sodium transfers heat to the secondary coolant. The primary sodium temperature decreases gradually along the flow path, and the cooled sodium exits the IHX into the cold pool, thereby completing the forced circulation loop of the primary system. The calculated flow rates at key locations show good agreement with the design values, confirming the model’s capability to simulate the essential flow characteristics during steady-state operation.

2.

Key Parameters

The key temperature parameters obtained from the steady-state simulation at rated power operation demonstrate good agreement with the design values as shown in

Table 2. Comparison of Key Parameters.

Item	Simulation Values	Reference Values	Deviation
Core flow rate/(kg/s)	301	301	0
RVCS flow rate/(kg/s)	35.54	40	0.40%
Core inlet temperature/°C	360.54	360	0.15%
Core outlet temperature/°C	522.79	530	1.36%
IHX inlet temperature/°C	514.1	516	0.37%
IHX outlet temperature/°C	350.8	353	0.62%
DHX inlet temperature/°C	516.1	516	0.02%
DHX outlet temperature/°C	486.2	490	0.78%

, with an overall discrepancy not exceeding 2%. This close correlation validates the reasonableness of the physical model simplifications adopted in this study. Consequently, the steady-state results provide reliable initial conditions for subsequent natural circulation transient calculations.

3.2. Transient Simulation

3.2.1. Boundary Conditions for Transient Simulation

The SBO accident scenario comprises two distinct phases. The initial phase is the pump coast-down period, which spans the first 110 s following the loss of electrical power to the main circulation pumps. During this phase, the pumps continue to rotate by inertia, maintaining forced circulation within the reactor system. Driven by the declining pump rotational speed, the primary circuit flow rate decreases rapidly throughout this period until the pumps come to a complete stop at 110 s. The subsequent phase, initiating at 110 s, is characterized by the establishment of natural circulation. At this stage, the sodium coolant completely loses external driving force, and its circulation relies solely on density gradients induced by temperature differences within the gravitational field. This density-driven flow mechanism facilitates the removal of decay heat from the core.

The transient variations in relative core power and the primary circuit flow rate are illustrated in Figure 5. Upon the occurrence of an SBO accident, the reactor protection system triggers a reactor scram. Control rods are inserted into the core, leading to a rapid reduction in core power. In comparison, the cooling capacity of the IHX decreases more gradually. This ensures heat removal during the early phase of the SBO. The IHX power declines rapidly as the accident progresses, completely losing its cooling capability by 110 s.

Under the reference natural circulation scenario, the DHRS initiates operation 600 s after the accident begins. The cooling power of the DHX is determined by the AHX. The externally located AHX activates its fan upon accident initiation, starting from the 600 s, and requires approximately 20 s to reach fully open status. Accordingly, the DHX power also begins increasing at 600 s, rising from the standby power of 0.0525 MW to the full power of 0.525 MW. This paper investigates three distinct DHX activation strategies:

• Immediate DHX activation upon accident initiation (0 s);
• Standard DHX activation (600 s after accident initiation);
• DHX remaining closed throughout the transient.

The simulation assumes a linear variation in DHX power over time. UDF is implemented in FLUENT to control the DHX power level and activation timing.

3.2.2. Transient Simulation Results

• Flow Field Analysis

Figure 6 illustrates the sodium flow patterns across the DHX-core section under three distinct operational scenarios during the natural circulation process. As observed, the fluid exiting the core flows upward through the sodium overflow window into the hot pool. It is subsequently drawn into the DHX, where it undergoes downward flow, ultimately reaching the bottom region of the hot pool. This flow path completes the heat extraction process from the hot pool. As observed from the magnified view of the DHX region, the downward sodium flow is most prominent under the immediate activation scenario, followed by the 600 s delayed activation scenario. In the DHX-closed scenario, driven solely by the standby power, a downward flow still exists but at a significantly lower flow rate.

Figure 7 and Figure 8 present the temperature contours under three operational scenarios. A longitudinal comparison of the contours reveals a notable temperature decrease within the reactor vessel over time across all scenarios—particularly in the upper region of the hot sodium pool—accompanied by an expansion of the low-temperature zone at the bottom of the cold pool. A transverse comparison indicates that, as expected, the temperature in the upper hot pool decreases progressively with earlier DHX activation, with the highest temperatures observed when the DHX remains closed throughout. Under the scenario with immediate DHX activation upon accident initiation, the low-temperature region within the cold pool is significantly larger compared to the reference case. In contrast, the temperature distribution in the DHX-closed scenario shows little deviation from the reference case. This phenomenon is attributed to the influence of internal circulation flow: early activation of the DHX accelerates the internal fluid circulation, thereby reducing the overall fluid temperature more effectively and enhancing the removal of thermal energy from the reactor. Nevertheless, even in the scenario where the DHX remains inactive, the core temperature exhibits a steady decline. This can be primarily attributed to the large thermal inertia of the substantial sodium inventory within the pool, coupled with the establishment of a core heat-driven natural circulation flow, which collectively contribute to decay heat removal and ensure reactor safety.

Quantitative monitoring of temperature characteristics was conducted for the cold and hot sodium pools, the IHX, and the DHX. The variation in temperature with elevation at different time points is illustrated in Figure 9. The temperature monitoring points in the sodium pool were horizontally positioned at the interface between the two loops in the outer upper hot pool, as shown in Figure 9a. Under all three scenarios, the temperature in the sodium pool increases with height. Earlier activation of the DHX results in a narrower temperature gradient and more uniform temperature distribution as natural circulation progresses. The IHX, functioning solely as a flow path for natural circulation in the absence of a cold sink, exhibits a temperature profile trend similar to that of the sodium pool, as shown in Figure 9b. However, as the upper section of the DHX is located in the inner hot pool closer to the core, the temperature at corresponding heights is higher than in the sodium pool, and the associated temperature gradient is accordingly steeper. Figure 9c clearly demonstrates the cooling effect exerted by the activated DHX on the flowing sodium coolant.

The core power decreases from 100% to approximately 7% within 1 s, during which the pumps undergo coast-down and the coolant flow rate through the core experiences a rapid overall decline. Natural circulation gradually initiates after approximately 200 s. The core flow rate profiles under different scenarios are presented in Figure 10.

As shown, the scenario with immediate DHX activation upon accident initiation exhibits the highest natural circulation flow rate through the core, while the scenario with the DHX remaining closed throughout results in the lowest flow rate. The core flow rates under the three scenarios demonstrate a descending trend corresponding to the delay in DHX activation, which aligns with expected behavior. In the case of immediate DHX activation, the high-temperature fluid in the upper hot pool is cooled by the DHX and discharged to the bottom region, creating a greater temperature differential between the upper and lower sections of the hot pool compared to the reference case. This enhanced thermal driving force promotes faster fluid circulation from the upper hot pool through the IHX into the cold pool, consequently yielding a notably higher core flow rate than the other two scenarios. Furthermore, although the DHX-closed scenario lacks active cooling, a core heat-driven natural circulation regime still develops through the mechanism of “core heating and cold fluid replenishment,” albeit with a relatively lower flow rate, leading to a gradual reduction in the temperature difference between the hot and cold pools.

Figure 11 presents a comparative analysis of the DHX inlet and outlet temperatures under the three operational scenarios. Analysis of the temperature evolution reveals that the DHX inlet temperature increases correspondingly with delays in the DHX activation time. For both the reference case (DHX activation at 600 s) and the DHX close scenario, the initial temperature rise at the DHX outlet within the first 110 s is attributable to the declining pumping capacity during the coast-down phase. Upon DHX activation, the fluid entering the exchanger experiences a temperature drop of approximately 40 °C. As natural circulation progresses, both the DHX inlet and outlet temperatures demonstrate a continuous decreasing trend across all scenarios.

Figure 12 illustrates the comparative IHX inlet and outlet temperatures under the three scenarios. The IHX inlet temperature exhibits a trend from low to high corresponding to later DHX activation times, indicating effective heat removal from the sodium in the hot pool. In the scenario with immediate DHX activation, the increased flow rate of upper hot pool fluid through the IHX results in a relatively higher IHX outlet temperature. The DHX closed scenario shows a slightly dropped IHX outlet temperature compared to the reference case.

The core inlet and outlet temperatures are shown in Figure 13. In the immediate DHX activation scenario, the core outlet temperature remains lower than the reference case for approximately the first 600 s, but subsequently exceeds it.

The characteristics of the core outlet temperature indicate that while early DHX activation provides enhanced initial cooling—resulting in the lowest core outlet temperature initially—the temperature differences among the three scenarios remain modest due to the substantial thermal inertia of the large sodium inventory. As natural circulation develops further, the core outlet temperature in the immediate DHX activation scenario exceeds that of the other two cases. Correlating with the flow rate data, this scenario achieves the highest natural circulation flow rate (exceeding the 600 s activation case by over 10%), which promotes faster circulation of hot fluid from the upper hot pool through the IHX into the cold pool. This leads to an earlier temperature increase in the cold pool sodium, consequently raising the temperature at the core inlet and subsequently at the core outlet. It should be noted that this difference in core outlet temperature remains relatively small.

Computational data further substantiate these findings at 1600 s, where the average core temperature in the immediate DHX activation scenario is approximately 376.85 °C, compared to about 368 °C in both the DHX-closed and reference scenarios. The core inlet temperature in the immediate activation case is approximately 8.85 °C higher than in the other scenarios, while the core outlet temperature is only about 5.87 °C higher. This reduced core temperature differential indicates that the increased flow rate through the core enhances the decay heat removal effectiveness of the natural circulation process.

4. Conclusions

A three-dimensional full-reactor integrated model of the pool-type sodium-cooled fast reactor CEFR has been developed in this study, incorporating key components such as the hot pool, cold pool, and core. To systematically evaluate the impact of DHX activation strategies on the decay heat removal process under SBO conditions, three-dimensional numerical simulations were conducted for two extreme scenarios: immediate DHX activation at 0 s and DHX remaining closed throughout the transient. The temperature and flow fields within the reactor under these conditions were obtained. These were compared with the existing natural circulation operation scheme of CEFR (DHX activation at 600 s after accident initiation), with a focus on elucidating the regulatory effect of DHX activation timing on sodium flow behavior and in-reactor temperature distribution. The main conclusions are as follows:

• The established CFD model demonstrates a high degree of rationality, with a deviation of less than 2% at the 100% rated power baseline of the CEFR. This good agreement provides a reliable initial condition and a fundamental baseline for the subsequent transient accident analysis.
• The DHX activation timing exerts significant active regulatory effects on the establishment process of natural circulation. Regarding DHX activation, “sooner is not always better”. Compared with standard strategies, immediate activation of the DHX enhances fluid circulation and mixing between the hot and cold sodium pools at an earlier stage, thereby increasing core flow by over 10%. This enhanced circulation mechanism effectively reduces spatial temperature gradients and improves temperature field uniformity within the sodium pool. However, it simultaneously causes an accelerated return of hot fluid to the cold pool, which leads to a slight elevation in the core inlet temperature and, consequently, an increase in the core outlet temperature.
• Although the DHX activation strategy influences system transient response characteristics, under all simulated conditions, the core outlet temperature can be successfully reduced below the safety limit (550 °C) by leveraging the substantial thermal inertia of sodium coolant and the established natural circulation. This fully demonstrates the inherent passive safety capability of pool-type SFRs.
• Among the evaluated scenarios, under the standard strategy of activating the DHRS 600 s after an SBO, the sodium pool temperature is not only controlled below the safety limit during the initial stage, but also continues to decrease in the subsequent phases. Furthermore, reserving this 600 s grace period allows the reactor protection system and on-site personnel to assess the accident conditions and determine whether DHRS activation is necessary. This effectively prevents the spurious actuation of the DHRS during brief transient fluctuations. Moreover, in the event of an automatic control system failure, it ensures operators still have time to manually activate the system.

For the first time, through high-precision three-dimensional numerical simulation, this study reveals the influence of DHX activation timing on the regulation mechanism of SFR natural circulation transient characteristics from both numerical and three-dimensional phenomena. It is different from the previous research on the spatial design level of DHRS, and makes a supplement to it on the temporal level. The research outcome provides critical data support for formulating and optimizing passive operation strategies for fast reactors, offering substantial engineering guidance for enhancing reactor operational safety and reliability.