High-fidelity simulation method for nuclear reactors based on multi-scale coupling and uncertainty quantification

Abstract

The application discloses a reactor high-fidelity simulation method based on multi-scale coupling and uncertainty quantification, and belongs to the technical field of nuclear reactor safety analysis. The application constructs a one-dimensional system-level model, a three-dimensional macroscopic porous medium model and a three-dimensional refined fuel assembly model, wherein the one-dimensional system-level model is packaged through a standardized interface FMU; firstly, the one-dimensional model is coupled with the three-dimensional macroscopic model to complete system simulation and carry out uncertainty quantification, and a high-uncertainty target region is located; then, dynamic bidirectional coupling is performed in the target region: the macroscopic model provides boundary conditions for the refined model, and the refined model feeds back correction coefficients to correct the macroscopic model, and iteration is carried out until convergence is achieved. The application solves the contradiction between insufficient precision of a traditional one-dimensional program and excessively high cost of a three-dimensional CFD, intelligently allocates calculation resources through uncertainty quantification, significantly improves simulation precision by relying on bidirectional coupling, and can provide an efficient and reliable technical approach for advanced reactor safety analysis.

Description

Technical Field

[0001] This invention relates to the field of nuclear reactor safety analysis technology, and in particular to a high-fidelity simulation method for reactors based on multi-scale coupling and uncertainty quantification. Background Technology

[0002] Nuclear reactor safety analysis is a core technical component of reactor design verification, operational evaluation, and accident condition assessment. Simulation is the core method of safety analysis, and its computational efficiency and simulation accuracy directly determine the reliability and engineering applicability of reactor safety evaluation. Currently, reactor thermal-hydraulic and multiphysics coupled simulation mainly employ two technical approaches, both of which suffer from an irreconcilable trade-off between efficiency and accuracy: One type is the one-dimensional system-level simulation program represented by RELAP5. Based on one-dimensional flow and lumped parameter assumptions, this type of program can quickly simulate the macroscopic dynamic characteristics of the reactor's primary and secondary loop systems. It boasts low computational cost and short processing time, making it suitable for the efficiency requirements of long-term transient system analysis. However, this type of program relies on numerous simplifying assumptions and cannot resolve intricate physical phenomena such as three-dimensional asymmetric flow and localized heat transfer distortion within the reactor core. Under critical accident conditions such as loss of current and sudden power surges, local physical details of the reactor core dominate the accident's progression. The simplified models of one-dimensional programs lead to large deviations and insufficient accuracy in the calculation results, failing to accurately characterize the safety margins of critical areas.

[0003] Another type is a three-dimensional high-resolution simulation technology based on computational fluid dynamics (CFD). This technology can construct detailed geometric models of fuel rod bundles, component flow channels, etc., and simulate turbulent flow, convective heat transfer, and neutron physics with high fidelity. The simulation of complex physical processes such as thermal-hydraulic coupling is highly faithful and complete in terms of local details. However, the large mesh size and complex physical coupling of 3D CFD simulation result in high computational resource consumption for high-fidelity calculations of the entire reactor core. For full-scale simulations of long-term transient reactor accidents, existing computing resources are insufficient to support the real-time requirements of engineering.

[0004] Currently, the industry lacks an integrated simulation method that can balance the overall computational efficiency of the system with the simulation accuracy of key local parts of the reactor core. Traditional technologies cannot simultaneously meet the dual requirements of efficiency and accuracy for advanced reactor safety analysis. There is an urgent need to innovate simulation technologies with multi-scale coupling and intelligent resource allocation to improve the balance between computational efficiency and local fine analysis in traditional methods. Summary of the Invention

[0005] The purpose of this invention is to provide a high-fidelity reactor simulation method based on multi-scale coupling and uncertainty quantization, which overcomes the contradiction between insufficient accuracy of one-dimensional system programs and excessive computational cost of three-dimensional CFD models in existing reactor simulations. It establishes a multi-scale, bidirectional coupling framework and introduces uncertainty quantization to adaptively allocate computational resources, thereby achieving dynamic simulation of reactor systems (especially key areas of the reactor core) with both high fidelity and high efficiency.

[0006] To achieve the above objectives, this invention provides a high-fidelity reactor simulation method based on multi-scale coupling and uncertainty quantification, comprising the following steps: S1: Construct a multi-scale model, establish one-dimensional system models of the primary and secondary loops of the reactor and derive them as functional model units (FMUs), a three-dimensional macroscopic porous medium model of the reactor core, and a three-dimensional refined fuel assembly model of the local fuel assembly. S2: Couple the one-dimensional FMU model with the three-dimensional macroscopic porous medium model to carry out system-level macroscopic simulation and obtain the preliminary physical field distribution of the reactor core; S3: Perform uncertainty quantification analysis on the simulation results of the three-dimensional macroscopic porous medium model, and locate the region with the highest uncertainty as the target simulation region; S4: Perform dynamic bidirectional coupling calculation within the target simulation area, passing boundary conditions from the three-dimensional macroscopic porous medium model to the three-dimensional refined fuel assembly model. After the three-dimensional refined fuel assembly model completes high-fidelity calculation, it feeds back correction coefficients to the three-dimensional macroscopic porous medium model. Iterate through S3 and S4 until the calculation results of the entire multi-scale coupling model meet the preset convergence criteria.

[0007] Preferably, in S2, the one-dimensional FMU model provides the three-dimensional macroscopic porous medium model with system-level mass flow rate, pressure, or temperature as boundary conditions; the three-dimensional macroscopic porous medium model feeds back the total core pressure drop to the one-dimensional FMU model.

[0008] Preferably, the three-dimensional macroscopic porous medium model is built based on CFD software. The core region adopts the porous medium assumption to simplify the calculation. The macroscopic flow and heat transfer behavior of the core is simulated by setting anisotropic flow resistance parameters and volumetric heat source terms.

[0009] Preferably, the three-dimensional refined fuel assembly model is a refined geometric model for one or more fuel rod bundles, including the geometry of the fuel rod bundles and the winding wires.

[0010] Preferably, the boundary conditions passed from the three-dimensional macroscopic porous medium model to the three-dimensional refined fuel assembly model include local flow distribution and temperature field distribution.

[0011] Preferably, the correction coefficient fed back from the three-dimensional refined fuel assembly model to the three-dimensional macroscopic porous medium model is the porous medium resistance coefficient of the target simulation region.

[0012] Preferably, the position of the target simulation region changes dynamically with the iteration process, and the region with the highest uncertainty is repositioned each time S3 is executed.

[0013] Preferably, the high-fidelity calculation of the three-dimensional refined fuel assembly model includes coupled calculation of neutron physics and thermal-hydraulic processes: the neutronics program calculates the power distribution and passes it to the CFD solver, and the CFD solver calculates the temperature field and returns it to the neutronics program, thereby realizing nuclear-thermal coupling at the fuel rod bundle level.

[0014] Preferably, the preset convergence criteria include the relative change rate of the core outlet temperature, fuel hot spot temperature, or total core pressure drop being less than a set threshold.

[0015] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the reactor high-fidelity simulation method based on multi-scale coupling and uncertainty quantification as described above.

[0016] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects: 1. Balancing global and local aspects: Through multi-scale coupling of one-dimensional, three-dimensional macroscopic and three-dimensional fine-scale simulation, it retains the efficient computational capability of system-level simulation for long-term transients, while accurately capturing the three-dimensional and asymmetric local physical details of key areas of the reactor core, thus achieving high-fidelity system-level simulation.

[0017] 2. Intelligent allocation of computing resources: By introducing uncertainty quantification technology, high uncertainty areas are dynamically identified, and the most computationally expensive fine simulations are deployed only in necessary locations, avoiding full-core fine calculations, greatly improving overall computing efficiency, and making high-fidelity analysis practical for engineering applications.

[0018] 3. Two-way correction improves accuracy: By feeding back correction coefficients (such as the resistance coefficient of porous media) to the macroscopic model through the refined model, a bottom-up closed-loop correction of the model is achieved, breaking the limitation of one-way information transmission in traditional multi-scale coupling and significantly improving the physical fidelity of the macroscopic model in key areas.

[0019] 4. Dynamic adaptive coupling: The target simulation area changes dynamically with the iteration process, always pointing to the area with the highest uncertainty, so that computing resources are continuously focused on the parts that need the most detailed analysis, further enhancing the adaptability of the simulation.

[0020] 5. Wide applicability: This invention is applicable to transient safety analysis of advanced reactors such as lead-bismuth fast reactors under accident conditions, and can also be extended to multi-scale high-fidelity simulation of other complex thermal-fluid systems, providing a new technical path for reactor design optimization, operation evaluation and safety review.

[0021] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart illustrating the implementation of the reactor high-fidelity simulation method based on multi-scale coupling and uncertainty quantification in an embodiment of the present invention. Figure 2 This is a three-dimensional macroscopic reactor model diagram according to an embodiment of the present invention; Figure 3 This is a three-dimensional refined fuel assembly model diagram according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the multi-scale model hierarchy and relationships in an embodiment of the present invention. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0025] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0026] Example This embodiment uses a lead-bismuth fast reactor as an example to provide a detailed explanation of the high-fidelity reactor simulation method based on multi-scale coupling and uncertainty quantification proposed in this invention. Figure 1 As shown, the method includes the following implementation steps: I. Construction of Multi-Scale Model First, execute S1 to construct three simulation models at different scales, with the following hierarchical relationship: Figure 4 As shown.

[0027] 1. Establishment of a one-dimensional system model (1D-System) and FMU derivation Using system simulation software such as Modelica or Numap, a reactor primary and secondary loop system model was built, including equipment such as the reactor pressure vessel, main pumps, heat exchangers, and piping. This model employs a lumped parameter method to describe the macroscopic thermal-hydraulic dynamics of the system, including coolant flow rate, pressure, and temperature, as well as pump speed and heat exchanger heat transfer characteristics. After the model was built, it was exported as an FMU file according to the Functional Mock-up Unit (FMU) standard interface specification. This FMU file encapsulates the model's dynamic equations, input and output variables (such as core inlet mass flow rate, core outlet temperature, and total core pressure drop), and solver information for subsequent calling and coupling within the Python main control script.

[0028] 2. Establishment of a three-dimensional macroscopic porous medium model (3D-Macro) A three-dimensional model of the reactor core is created using computational fluid dynamics (CFD) software such as ANSYS Fluent. Figure 2 As shown, the model covers the entire active region of the reactor core and the upper and lower chambers. To reduce computational complexity, the active region of the core is treated as a porous medium region. Specifically, a porous medium model is set up in Fluent, defining viscous drag coefficients and inertial drag coefficients in three orthogonal directions (axial, radial, and circumferential) to equivalently simulate the obstruction of coolant flow by the fuel rod bundles. Simultaneously, based on the core power distribution, a volumetric heat source term is added to each grid layer to simulate the heat released by nuclear fission. This three-dimensional macroscopic porous medium model does not analyze individual fuel rods or filaments; the number of grids is typically on the order of hundreds of thousands, enabling rapid solutions for the three-dimensional velocity, temperature, and pressure fields at the core scale.

[0029] 3. Establishment of a three-dimensional refined fuel assembly model (3D-Fine) like Figure 3 As shown, a CFD model is established for a single fuel assembly or a group of adjacent fuel assemblies, including detailed geometries such as fuel rod bundles, filament windings, and positioning grids. This model employs unstructured or polyhedral meshes, with localized refinement of areas such as fuel rod gaps and filament contact regions. The number of meshes can reach millions to tens of millions. This model is used to solve for refined flow and heat transfer characteristics within the assembly and can be coupled with neutron physics programs (such as OpenMC) for high-fidelity nuclear thermal coupling calculations.

[0030] II. Macroscopic System-Level Coupling Simulation In the Python main control script, the FMU file exported by S1 is loaded through FMPy or a similar FMU calling library, and the solution of the three-dimensional macroscopic porous medium model is controlled by PyFluent or Fluent's batch processing interface.

[0031] Execution S2: Within a transient simulation time step, the co-simulation controller first invokes the one-dimensional FMU model to calculate the total mass flow rate, temperature, and pressure entering the reactor upper chamber based on the current system state (such as pump speed, valve opening, etc.), and passes these parameters as boundary conditions to the three-dimensional macroscopic porous media model. Subsequently, the co-simulation controller initiates Fluent solving, and the three-dimensional macroscopic model calculates the three-dimensional flow field and temperature field of the entire core under the given inlet boundary conditions. After the calculation converges, parameters such as the core outlet temperature and total core pressure drop are extracted from the three-dimensional macroscopic model and fed back to the one-dimensional FMU model, completing the closed-loop coupling for this time step. The above process is repeated until the entire system-level transient simulation is completed, obtaining the preliminary physical field distribution of the core.

[0032] III. Uncertainty Quantification and Target Area Positioning Execution S3: After obtaining the initial physical field through macroscopic system-level simulation, the co-simulation controller invokes the Uncertainty Quantization (UQ) module. This module performs spatial uncertainty analysis on the calculation results of the three-dimensional macroscopic model. One option is to calculate the physical quantity gradient and local deviation within each grid cell or component region. For example, for the temperature field, its spatial gradient magnitude is calculated; regions with large gradients often correspond to areas with drastic changes in thermal-hydraulic parameters, and the simplification assumptions of the macroscopic porous media model in these regions may lead to significant uncertainties. Another approach is to pre-establish a database of the result deviations between the macroscopic model and the refined model, and quickly assess the uncertainty through a surrogate model. This embodiment uses a method combining temperature gradient and local power density to quantify the uncertainty index of each fuel assembly location. Then, one or more component locations with the highest uncertainty index are automatically identified and marked as target simulation regions. The location information of this target region (such as component number and spatial coordinate range) is dynamically updated in subsequent iterations.

[0033] IV. Dynamic Two-Way Coupling and Model Correction Execute S4 to perform dynamic bidirectional coupling calculations within the target simulation region defined in S3.

[0034] Boundary condition transfer (macro to fine): The co-simulation controller extracts the velocity components and temperature values ​​of each grid point on the inlet section of the target simulation region from the calculation results of the 3D macroscopic porous medium model, forming detailed velocity and temperature distribution profiles. Simultaneously, the outlet pressure of this region is extracted as the pressure outlet boundary condition. These boundary conditions are then written into the input file of the 3D fine fuel assembly model via Profile files or direct data transfer.

[0035] Local high-fidelity simulation: The co-simulation controller initiates the solver of the 3D refined fuel assembly model (e.g., a CFD solver coupled with a neutron transport program). During the solution process, OpenMC first performs neutron transport calculations based on the current geometry and materials to obtain the power distribution inside the fuel rod bundle; this power distribution is then mapped onto the CFD mesh as a volumetric heat source; the CFD solver performs flow heat transfer calculations to obtain new temperature and density fields; the updated temperature field is then returned to OpenMC to update the neutron cross-section data. This process iterates until the nuclear-thermal coupling converges. Finally, detailed temperature distribution on the surface of the fuel rod bundle, coolant flow and temperature distribution, and pressure loss are obtained within the target region.

[0036] Model parameter feedback (refined to macroscopic): Based on the calculation results of the 3D refined fuel assembly model, the correction coefficients that should be used in the macroscopic porous media model for the target simulation region are calculated. Specifically, the viscous drag coefficient and inertial drag coefficient of the region are refitted using the pressure drop and mass flow rate obtained from the refined model and the drag formula of the porous media model (such as the Forchheimer equation). The co-simulation controller extracts these correction coefficients and updates the drag coefficient parameters of the corresponding region in the 3D macroscopic porous media model.

[0037] V. Iteration and Convergence Judgment After completing the bidirectional coupling described above, return to S3 and re-perform uncertainty quantification analysis of the entire core. Since the macromodel of the previous target region has been corrected, the uncertainty distribution will change, potentially identifying a new region with the highest uncertainty. The co-simulation controller continues to execute the dynamic bidirectional coupling in S4 within the new target region. Repeat S3 and S4 until the key parameters of the entire coupled system (such as core outlet temperature and hot fuel temperature) no longer change significantly, achieving final convergence. When the convergence condition is met, the master script terminates the iteration and outputs the final simulation results, including the dynamic response of the system loop and the high-fidelity physical field of the key core regions.

[0038] VI. Full-process management and implementation results The entire coupled iterative process is uniformly managed by a Python main control script, automating functions such as FMU invocation, CFD solution control, data transfer, uncertainty quantification analysis, and iterative convergence judgment. Through the above implementation methods, this invention can intelligently allocate high-fidelity computing resources to high-uncertainty regions and continuously improve the accuracy of the macroscopic model through bidirectional correction, ultimately achieving reactor safety analysis that balances efficiency and accuracy.

[0039] The remaining technical features in the above embodiments can be flexibly selected by those skilled in the art to meet different specific practical needs according to actual circumstances. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims. In the above description, numerous specific details have been set forth to provide a thorough understanding of the present invention. However, it will be apparent to those skilled in the art that these specific details are not necessary to implement the present invention. In other instances, to avoid obscuring the present invention, well-known techniques, such as specific construction details, operating conditions, and other technical conditions, have not been specifically described.

[0040] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A high-fidelity reactor simulation method based on multi-scale coupling and uncertainty quantification, characterized in that, Includes the following steps: S1: Construct a multi-scale model, establish one-dimensional system models of the primary and secondary loops of the reactor and derive them as functional model units (FMUs), a three-dimensional macroscopic porous medium model of the reactor core, and a three-dimensional refined fuel assembly model of the local fuel assembly. S2: Couple the one-dimensional FMU model with the three-dimensional macroscopic porous medium model to carry out system-level macroscopic simulation and obtain the preliminary physical field distribution of the reactor core; S3: Perform uncertainty quantification analysis on the simulation results of the three-dimensional macroscopic porous medium model, and identify the target simulation region based on preset evaluation indicators. These evaluation indicators include one or more of the following: local temperature gradient, power density distribution, pressure change rate, or model prediction deviation. S4: Perform dynamic bidirectional coupling calculation within the target simulation area, transferring boundary conditions from the three-dimensional macroscopic porous medium model to the three-dimensional refined fuel assembly model. After the three-dimensional refined fuel assembly model completes high-fidelity calculation, it feeds back equivalent parameters to update the flow and heat transfer parameters of the corresponding region of the three-dimensional macroscopic porous medium model. S3 and S4 are executed iteratively until the calculation results of the entire multi-scale coupling model meet the preset convergence criteria.

2. The reactor high-fidelity simulation method based on multi-scale coupling and uncertainty quantification according to claim 1, characterized in that: In S2, the one-dimensional FMU model provides the three-dimensional macroscopic porous medium model with system-level mass flow rate, pressure, or temperature as boundary conditions; the three-dimensional macroscopic porous medium model feeds back the total core pressure drop to the one-dimensional FMU model.

3. The reactor high-fidelity simulation method based on multi-scale coupling and uncertainty quantification according to claim 1, characterized in that: The three-dimensional macroscopic porous medium model is built based on CFD software. The core region adopts the porous medium assumption to simplify the calculation. The macroscopic flow and heat transfer behavior of the core is simulated by setting anisotropic flow resistance parameters and volumetric heat source terms.

4. The reactor high-fidelity simulation method based on multi-scale coupling and uncertainty quantification according to claim 1, characterized in that: The three-dimensional refined fuel assembly model is a refined geometric model for one or more fuel rod bundles, including the geometry of the fuel rod bundles and the winding wires.

5. The reactor high-fidelity simulation method based on multi-scale coupling and uncertainty quantification according to claim 1, characterized in that: The boundary conditions passed from the three-dimensional macroscopic porous medium model to the three-dimensional refined fuel assembly model include local flow distribution and temperature field distribution.

6. The reactor high-fidelity simulation method based on multi-scale coupling and uncertainty quantification according to claim 1, characterized in that: The correction coefficient fed back from the three-dimensional refined fuel assembly model to the three-dimensional macroscopic porous medium model is the porous medium resistance coefficient of the target simulation region.

7. The reactor high-fidelity simulation method based on multi-scale coupling and uncertainty quantification according to claim 1, characterized in that: The position of the target simulation region changes dynamically with the iteration process, and the region with the highest uncertainty is repositioned each time S3 is executed.

8. The reactor high-fidelity simulation method based on multi-scale coupling and uncertainty quantification according to claim 1, characterized in that: The high-fidelity calculation of the three-dimensional refined fuel assembly model includes coupled calculations of neutron physics and thermal hydraulics: the neutronics program calculates the power distribution and passes it to the CFD solver, the CFD solver calculates the temperature field and returns it to the neutronics program, realizing nuclear thermal coupling at the fuel rod bundle level.

9. The reactor high-fidelity simulation method based on multi-scale coupling and uncertainty quantification according to claim 1, characterized in that: The preset convergence criteria include the relative change rate of core outlet temperature, fuel hot spot temperature, or total core pressure drop being less than a set threshold.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the reactor high-fidelity simulation method based on multi-scale coupling and uncertainty quantification as described in any one of claims 1 to 9.

Images