A multi-level multi-physics coupling method and system for coupling multiple programs

Abstract

The application discloses a kind of multilevel multi-physical coupling method and system for multiple program coupling, it is related to the field of multi-physical field coupling simulation calculation.Its system includes five hierarchical structures of coupling model module encapsulation layer, coupling problem manager encapsulation layer, coupling problem data interaction layer, coupling multilevel scene implementation layer and coupling infrastructure layer.The application can effectively solve the problems such as rigid coupling framework of existing multiple program coupling coupling technical scheme, difficult to unify coupling paradigm, lack of efficient multi-physical field coupling management, coupling decision and close coupling pedigree, poor adaptability of complex coupling scene, etc.by modularizing existing multiple program, systematizing coupling problem, standardizing coupling data interaction and hierarchical coupling management.

Description

Technical Field

[0001] This invention belongs to the field of multiphysics coupling simulation calculation, and more specifically, relates to a multi-level multiphysics coupling method and system for coupling multiple programs. Background Technology

[0002] Interdisciplinary collaboration and multiphysics coupling are prevalent in scientific research and engineering. With the development of high-performance numerical simulation and the increasing demand for exploring the essence of multiphysics coupling phenomena, numerical simulation has gradually evolved from single-discipline, single-physics simulation to multiphysics coupling in complex scenarios. Examples include the multiphysics coupling of neutron transport, thermal heat transfer, and structural mechanics in advanced nuclear reactor cores, and the multiphysics coupling of fluid-structure interaction and radiation heat transfer in microelectronic devices. It can be said that multiphysics coupling numerical simulation has become a core challenge and a key area for breakthroughs in many important fields, including aerospace, microelectronic systems, and advanced energy and power.

[0003] Currently, mainstream multiphysics coupling platforms and systems, such as MOOSE and COMSOL in the United States, generally adopt a coupling approach based on the coupling modification or encapsulation of multiple well-developed single-physics simulation programs. This involves coupling and modifying existing single-physics simulation programs or encapsulating them into mutually identifiable coupled subroutine modules. This enables the transmission of coupling information between multiple coupled subroutine modules, and allows for the coupled assembly of these modules to build a multiphysics coupling process or framework, thereby achieving multiphysics coupling. This coupling method and related multiphysics coupling platforms and systems can fully utilize existing programs, thus reducing the difficulty and workload of multiphysics coupling.

[0004] However, the aforementioned multiphysics coupling processes and framework construction generally rely on manual implementation by coupling developers, requiring a high level of expertise. Furthermore, the coupling frameworks are rigid, making it difficult to unify coupling paradigms and limiting the applicability of coupling. Simultaneously, the lack of efficient multiphysics coupling management, coupling method decision-making, and tight coupling genera poses significant challenges to the reusability, scalability, and applicability to complex scenarios of multiphysics coupling in complex multiphysics coupling scenarios, especially multi-level coupling scenarios where multiple similar programs exist within a single physics field. This necessitates comprehensive consideration of coupling problems and decisions on suitable coupling methods. Summary of the Invention

[0005] To address the aforementioned shortcomings and improvement needs of existing coupling methods and systems, this invention provides a multi-level multi-physics coupling method and system for coupling multiple programs. This addresses the problems of rigid coupling frameworks, difficulty in unifying coupling paradigms, and lack of efficient multi-physics coupling management, coupling decision-making, and tight coupling spectrum in existing coupling methods and systems based on multiple programs.

[0006] To achieve the above objectives, according to one aspect of the present invention, a multi-level multi-physical coupling system for coupling multiple programs is provided, comprising: a coupling model module encapsulation layer, a coupling problem manager encapsulation layer, a coupling problem data interaction layer, a coupling multi-level scenario implementation layer, and a coupling infrastructure layer;

[0007] The coupling infrastructure layer includes preprocessing tools, numerical algebra libraries, grid data tools, and postprocessing tools required for coupling. Among them, the preprocessing tools are used for parameter configuration and computational domain initialization before multiphysics coupling calculations; the numerical algebra library is used for numerical solutions in multiphysics coupling; the grid data tools are used for low-level grid topology management and data interaction between different physics fields in multiphysics; and the postprocessing tools are used for the storage, transformation, and visualization of calculation results.

[0008] The coupling model module encapsulation layer is used to decompose and encapsulate multiple computation programs that need to be coupled into coupling modules in a unified manner.

[0009] The Coupling Problem Manager encapsulation layer further encapsulates and adds the encapsulated coupling model in the Coupling Model Module encapsulation layer into a dedicated Coupling Problem Manager, based on the multiphysics coupling scenario it faces.

[0010] Based on the multi-physics coupling scenario faced, the coupling problem data interaction layer develops, builds, and encapsulates coupling data interaction modules between coupling problem managers from the coupling problem manager encapsulation layer;

[0011] Based on the multi-physics coupling scenario faced, the multi-level coupling implementation layer assembles and encapsulates the coupling problem manager from the coupling problem manager encapsulation layer and the coupling data interaction module from the coupling problem data interaction layer in multiple levels, and finally forms a multi-level coupling framework unified in the main coupling scenario manager.

[0012] The main coupled scene manager includes a preprocessing layer, an execution layer, and a postprocessing layer. The preprocessing layer coordinates the initialization of each sub-manager, reads the parameters of each model, and allocates memory space. The execution layer is used for coupled iteration, coordinates the solution order of each manager, triggers the data interaction layer's transfer operations, and monitors the global coupled residuals to determine convergence. The postprocessing layer is used to release memory resources and output the calculation results after the coupled calculation is completed.

[0013] Furthermore, the coupled model module encapsulation layer uses an object-oriented coupling encapsulation paradigm based on the division and encapsulation of three modules: program preprocessing, program core calculation, and program postprocessing. This encapsulates the computational program that needs to be coupled into a coupled model with a unified encapsulation paradigm.

[0014] Furthermore, the coupled model includes a model preprocessing submodule, a model core computation submodule, and a model postprocessing submodule.

[0015] Furthermore, the coupling problem manager in the coupling problem manager encapsulation layer is used to represent single physics problems, complex scene coupling problems, and multi-level sub-coupling problems;

[0016] The coupling problem manager encapsulation layer assigns coupling model numbers to the added coupling models according to the order in which the coupling models are added after being encapsulated in the coupling model module encapsulation layer, or the coupling model number can be directly specified through the interface provided by the coupling manager.

[0017] Furthermore, the coupled data interaction module in the coupled problem data interaction layer is also a public sub-function belonging to one or more coupled problem managers;

[0018] The aforementioned coupling problem data interaction layer is used to complete the coupling data interaction from different coupling problem managers in the coupling problem manager encapsulation layer.

[0019] Furthermore, the coupling problem data interaction layer specifies the coupling model from different coupling problem managers in the coupling problem manager encapsulation layer through the coupling model number, so as to realize the coupling data interaction of different coupling models in a single or different coupling problem managers.

[0020] Furthermore, the multi-level coupling scenario implementation layer will add and encapsulate the lower-level coupling managers into the upper-level coupling problem manager based on the object-oriented derivation inheritance mechanism according to the multi-level coupling scenario; the multi-level coupling scenario implementation layer will assign coupling manager numbers to the added coupling managers according to the order in which the coupling managers encapsulated in the coupling problem manager encapsulation layer are added, or the numbers can be directly specified through the interface provided by the coupling problem manager;

[0021] The multi-level coupled scenario implementation layer is also used to encapsulate the main coupled scenario manager preprocessing layer, the main coupled scenario manager execution layer, and the main coupled scenario manager postprocessing layer. The main coupled scenario manager, as the top-level coupling manager, is composed of multiple sub-coupled managers under a multi-level coupling relationship tree. The main coupled scenario manager execution layer executes the coupled scenario sequentially according to the pre-compiled coupling manager number and the coupling model number in each sub-coupled manager, and simultaneously performs coupling data exchange between single or different coupling problem managers to complete the coupling data interaction.

[0022] Furthermore, the multi-level scenario implementation layer uses the input files provided by the coupling problem manager for each sub-level to further specify the coupling iteration control, coupling method, and / or coupling time step.

[0023] Furthermore, the main coupled scene manager preprocessing layer and the main coupled scene manager postprocessing layer described in the multi-level coupled scene implementation layer will complete the preprocessing and postprocessing of all sub-coupled managers at once or optionally.

[0024] According to another aspect of the present invention, a multi-level multi-physical coupling method for coupling multiple programs is provided, implemented based on a multi-level multi-physical coupling system as described in any of the preceding claims, comprising:

[0025] First, in the coupled infrastructure layer and the coupled model module encapsulation layer, the corresponding basic numerical library and mesh tools are called according to the calculation request, the required underlying physical calculation program is loaded and initialized and transformed into a standardized coupled model, thereby completing the transformation from the original code to the basic data unit that can be called by the system.

[0026] Subsequently, based on the physical properties, the standardized coupled model is injected into the corresponding proprietary manager for containerized assembly. Based on the correlation between physical fields, the underlying mesh mapping tool is called to establish a data interaction module, thereby completing the integration from single model data to local physical field data and establishing a bridge for data flow.

[0027] Finally, at the top-level coupled multi-level scene implementation layer, the various local managers and interaction modules are further integrated upwards into a complete main coupled scene manager;

[0028] After the main coupled scene manager takes over control, it drives the flow of underlying data between different levels: the physical data generated by the underlying model calculations is processed by the coupled problem data interaction layer and then passed between the managers; the calculation status of each sub-manager is fed back through the coupled problem manager encapsulation layer and converges to the main coupled scene manager level by level; the main coupled scene manager iteratively processes the data according to the convergence status, and finally integrates the data calculated at each level into a complete multiphysics solution, and outputs the final calculation result through the post-processing interface of the main coupled scene manager.

[0029] Specifically, the method and system provided by the present invention have the following advantages compared with the prior art:

[0030] 1. Existing programs are modularized through a coupling model module encapsulation layer, unifying different programs into mutually recognizable coupling models. This provides a unified module paradigm for the coupling model, allowing core modules to be plug-and-play. The clear and concise "three-module division-encapsulation" effectively ensures the accuracy of coupling logic and promotes collaborative coupling.

[0031] 2. The coupling problem system is encapsulated through a coupling problem manager layer, unifying different coupling models (programs) based on object-oriented derivation and inheritance mechanisms into a unified system management system. Users can add / encapsulate new coupling models (subroutines) that they need or are concerned with in the coupling problem manager, or directly use the coupling models in the existing coupling problem manager. This realizes the software engineering principle of separating coupling concerns. Each coupling problem manager is plug-and-play, reducing the threshold, difficulty, and workload of creating couplings.

[0032] 3. By standardizing coupled data interaction through a coupled problem data interaction layer, and based on the encapsulation of coupled model modules and coupled problem managers, complex data from different programs are formed into a unified and standardized coupled problem manager-coupled model-original program derived inheritance data type structure. Based on the principle of object-oriented derived inheritance, the data of each program has a unified calling form, with strong data compatibility and low memory management risk. This overcomes the shortcomings of existing technical solutions, which are difficult to unify and manage data types of different programs, and greatly reduces the development difficulty and workload of coupled data interaction.

[0033] 4. By implementing multi-level coupling scenarios, the coupling management layer is hierarchically structured. Through multi-level derivation and assembly of different coupling problem managers, an efficient multi-physics coupling management system and a tightly coupled spectrum are formed. Complex multi-level coupling relationship trees can be constructed. For complex multi-level multi-physics coupling scenarios, each coupling level (sub-coupling problem manager) possesses the characteristics of dynamic multi-level combination, dynamic scheduling and control, and flexible decision-making. The implementation of coupling scenarios is no longer locked into the rigid single coupling process or single strategy paradigm of existing technical solutions, and has applicability to complex multi-level coupling scenarios. At the same time, the coupling scenarios are defined within a unified main coupling scenario manager framework. Complex coupling strategies can be executed through the coupling manager number and the coupling model number in each sub-coupling manager, which can efficiently cope with different coupling scenarios and avoid the inefficient mode of existing technical solutions that require cumbersome coupling of multiple heterogeneous legacy programs for different coupling scenarios. Attached Figure Description

[0034] Figure 1 This is a framework diagram of a multi-level, multi-physical coupling system for coupling multiple programs according to the present invention.

[0035] Figure 2This is a schematic diagram of the multi-level, multi-physical coupling method for coupling multiple programs according to the present invention.

[0036] Figure 3 This is a schematic diagram of a multi-level, multi-physics coupling embodiment of the present invention for multiple reactor coupling procedures.

[0037] Figure 4 This is a diagram showing the power distribution calculation results of a multi-level multi-physics coupling embodiment for multiple reactor coupling procedures according to the present invention. Detailed Implementation

[0038] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0039] This invention addresses the rigid coupling framework and difficulty in unifying coupling paradigms in existing coupling methods and systems based on multiple programs. It also addresses the lack of efficient multiphysics coupling management, coupling decision-making, and tight coupling systems. The invention provides a multi-level multiphysics coupling method and system for coupling multiple programs.

[0040] in Figure 1 The diagram shows a multi-level, multi-physical coupling method and system framework for multiple program coupling according to the present invention. The system comprises five layers: a coupling model module encapsulation layer, a coupling problem manager encapsulation layer, a coupling problem data interaction layer, a multi-level coupling scenario implementation layer, and a coupling infrastructure layer.

[0041] The coupling infrastructure layer includes preprocessing tools, a numerical algebra library, mesh data tools, and post-processing tools required for coupling. The preprocessing tools are primarily responsible for parameter configuration and computational domain initialization before multiphysics coupling calculations, capable of identifying and resolving complex physical boundary conditions and material properties. The numerical algebra library is mainly responsible for numerical solutions in multiphysics coupling, integrating high-performance linear and nonlinear equation solvers for sparse matrices in multiphysics coupling (such as the Krylov subspace iteration method, operator splitting method, and fixed-point iteration method), ensuring the convergence and efficiency of numerical calculations in multiphysics coupling. The mesh data tools are mainly responsible for managing the underlying mesh topology and data interaction between different physics fields, resolving mesh mismatch issues between different physical modules. The post-processing tools are mainly responsible for storing, converting, and visualizing calculation results, supporting the export of multiphysics coupling calculation data in specific formats and statistical analysis of key physical quantities.

[0042] The coupling model module encapsulation layer encapsulates computational programs requiring coupling into a coupled model with a unified encapsulation paradigm. It encapsulates heterogeneous subphysics solvers according to a standard three-module structure: pre-processing submodule encapsulation functions, model core computation submodule encapsulation functions, and post-processing submodule encapsulation functions. The pre-processing submodule encapsulation function is primarily responsible for initializing the physics solver during the coupling computation startup phase, mainly performing model parameter reading and memory allocation. The model core computation submodule encapsulation function abstracts the specific solution process into a standard execution interface, supporting calls within the coupling framework for exchanging and updating multiphysics coupling data. The post-processing submodule encapsulation function is primarily responsible for outputting the computation results and releasing memory resources after the coupling computation is complete.

[0043] The Coupling Problem Manager encapsulation layer is used to encapsulate / add coupled models from the Coupled Model Module encapsulation layer into a dedicated Coupled Problem Manager. The added coupled models are numbered according to the order in which they were added after encapsulation; this numbering order can also be directly specified by the Coupled Model Module encapsulation layer. During computation, this layer coordinates the execution queues of each sub-model in the dedicated coupled problem, sequentially calling the model's "core computation" interface according to physical logic, and controlling its solution at the current time step.

[0044] The Coupled Problem Data Interaction Layer is used to develop, build, and encapsulate the coupled data interaction modules between coupled problem managers. Based on the underlying grid data tools, this layer accurately indexes the source and target models through coupled model numbers, and transmits the calculation results of the source physics field to the target physics field with high fidelity, thereby realizing bidirectional data communication between models in different coupled problem managers;

[0045] The multi-level coupled scenario implementation layer, based on the actual multiphysics coupled scenario, assembles the lower-level coupled problem manager and data interaction module into multiple levels. It completes the construction of the main coupled scenario manager, including the encapsulation of the main coupled scenario manager pre-processing layer, main coupled scenario manager execution layer, and main coupled scenario manager post-processing layer. It supports recursive hierarchical management, meaning that an upper-level manager can consist of multiple lower-level sub-coupled managers, forming a tree-like coupling hierarchy. Specifically, the main coupled scenario manager pre-processing layer is responsible for coordinating the initialization of each sub-manager, mainly performing the reading of model parameters and memory allocation. The main coupled scenario manager execution layer runs the coupled iterative algorithm, coordinates the solution order of each manager, triggers the transmission operation of the data interaction layer, and monitors the global coupling residual to determine convergence. The main coupled scenario manager post-processing layer is responsible for outputting the calculation results and releasing memory resources after the coupled calculation is completed.

[0046] like Figure 2As shown, the multi-level multi-physical coupling system for multiple program coupling provided by the present invention specifically includes:

[0047] The coupling model module encapsulation layer unifies and encapsulates multiple computational programs that need to be coupled, forming a coupled model with a unified encapsulation paradigm. Specifically, based on the object-oriented coupling encapsulation paradigm of "three-module division-encapsulation" (program preprocessing - core computation - postprocessing), the computational programs that need to be coupled are encapsulated into a coupled model with a unified encapsulation paradigm. The coupled model includes a model preprocessing submodule, a model core computation submodule, and a model postprocessing submodule. The encapsulation functions of the model preprocessing submodule are mainly responsible for initializing the physics solver during the coupled computation startup phase, primarily performing model parameter reading and memory allocation. The encapsulation functions of the model core computation submodule are mainly responsible for abstracting the specific solution process into a standard execution interface, supporting calls within the coupling framework for exchanging and updating multiphysics coupling data. The encapsulation functions of the model postprocessing submodule are mainly responsible for outputting the computation results and releasing memory resources after the coupled computation is completed. Each individual coupled model has the complete computational functionality of the original program, and specific coupled models can also be developed based on the coupling infrastructure layer provided by this invention.

[0048] The coupling problem manager encapsulation layer, based on the multi-physics coupling scenario, encapsulates / adds the coupling models encapsulated in the coupling model module encapsulation layer into a dedicated coupling problem manager according to the type or specific function of the physics field being solved, based on the object-oriented derivation and inheritance coupling manager encapsulation paradigm. The coupling problem manager can represent single-physics problems, complex scenario coupling problems, and multi-level sub-coupling problems.

[0049] The coupling problem manager encapsulation layer will assign coupling model numbers to the added coupling models according to the order in which the coupling models encapsulated in the coupling model module encapsulation layer are added. This numbering order can also be directly specified through the interface provided by the coupling manager. During the calculation process, this layer coordinates the execution queue of each sub-model in the dedicated coupling problem and calls the "core calculation" interface of the model in sequence according to the physical logic to control its solution at the current time step.

[0050] The coupling problem data interaction layer, based on the multiphysics coupling scenario, develops, constructs, and encapsulates coupling data interaction modules between coupling problem managers from the coupling problem manager encapsulation layer. The developed coupling data interaction modules are also public sub-functions belonging to the single or multiple coupling problem managers.

[0051] The aforementioned coupling problem data interaction layer is used to complete the coupling data interaction between different coupling problem managers from the coupling problem manager encapsulation layer. Considering the differences in grid data structures and other aspects of different programs, the coupling problem data interaction layer can further specify the coupling model (from the coupling model module encapsulation layer) in different coupling problem managers from the coupling problem manager encapsulation layer through the coupling model number, thereby realizing the coupling data interaction between different coupling models in a single / different coupling problem manager.

[0052] The multi-level coupling scenario implementation layer, based on the multi-physics coupling scenario encountered, assembles and encapsulates the coupling problem manager from the coupling problem manager encapsulation layer and the coupling data interaction module from the coupling problem data interaction layer in multiple levels, ultimately forming a unified multi-level coupling system under the main coupling scenario manager. The multi-level coupling scenario implementation layer adds and encapsulates lower-level coupling managers into the upper-level coupling problem manager based on an object-oriented inheritance mechanism, according to the multi-level coupling scenario. The multi-level coupling scenario implementation layer assigns coupling manager numbers according to the order in which the coupling managers encapsulated in the coupling problem manager encapsulation layer are added; this numbering order can also be directly specified through the interface provided by the coupling problem manager.

[0053] The multi-level coupled scenario implementation layer encapsulates the main coupled scenario manager's pre-processing layer, execution layer, and post-processing layer. The main coupled scenario manager, as the top-level coupling manager, can be composed of multiple sub-coupled managers with a multi-level coupling relationship tree. The main coupled scenario manager execution layer executes the coupled scenarios sequentially according to the pre-assigned coupling manager numbers and the coupling model numbers in each sub-coupled manager, while simultaneously performing coupling data exchange between single / different coupling problem managers to complete the coupling data interaction.

[0054] The aforementioned multi-level coupling scenario implementation layer can further specify the coupling iteration control, coupling method, coupling time step, etc., through the input file provided by the coupling problem manager of each sub-level; thus flexibly and freely controlling the coupling strategy of each level.

[0055] Based on the aforementioned multi-level system architecture, the computational flow of a multi-level multi-physics coupled system for multiple program coupling proposed in this embodiment is as follows: First, at the bottom layer of the coupling infrastructure layer and the coupling model module encapsulation layer, the system calls the basic numerical library and mesh tools according to the computation request, loads and initializes the required underlying physics calculation programs (such as neutron, thermal, and other physics calculation programs), and transforms them into standardized "coupled model objects." At this point, the transformation from raw code to basic data units that can be called by the system is completed. Subsequently, the system injects the above-mentioned basic data units (coupled models) into the corresponding dedicated managers for containerized assembly according to physical properties, and calls the underlying mesh mapping tools to establish data interaction modules according to the correlation between physical fields. At this stage, the system completes the integration from single model data to local physical field data and establishes a bridge for data flow. Finally, at the top layer of the coupled multi-level scene implementation layer, the system further integrates the various local managers and interaction modules into a complete main coupled scene manager. After the main coupled scene manager takes over control, it drives the flow of underlying data between different levels: physical data generated by the underlying model calculations is processed by the coupled problem data interaction layer and then passed between managers; the calculation status of each sub-manager is fed back through the coupled problem manager encapsulation layer and converges level by level to the main coupled scene manager. The main coupled scene manager iteratively processes the data based on the convergence status, and finally integrates the data calculated at each level into a complete multiphysics solution, and outputs the final calculation result through the main post-processing interface.

[0056] This invention provides a multi-level, multi-physics coupling embodiment for multiple reactor coupling procedures, as illustrated in the flowchart of the specific embodiment. Figure 3 As shown.

[0057] The system is built and performs coupled computations step by step from the bottom to the top, based on a hierarchical architecture consisting of a coupled infrastructure layer, a coupled model module encapsulation layer, a coupled problem manager encapsulation layer, a coupled problem data interaction layer, and a coupled multi-level scenario implementation layer. The specific implementation process is detailed below:

[0058] First, at the system's bottom and sub-bottom layers, based on the numerical algebra library and grid processing tools provided by the coupled infrastructure layer, the coupled model module encapsulation layer is responsible for standardizing and encapsulating heterogeneous reactor physics calculation programs. Specifically, for the neutron transport calculation program, fuel rod thermal conductivity calculation program, thermal-hydraulic calculation program, and reactor loop system calculation program that require coupling, this embodiment strictly follows the "three-module" encapsulation paradigm of "program preprocessing - program core calculation - program postprocessing," encapsulating them into coupled models with unified calling interfaces. For example... Figure 3As shown, multiple neutron physics calculation models (such as Models 1 to 3), fuel rod thermal conductivity calculation models, thermal-hydraulic calculation models (such as Models 1 to 2), and reactor loop system calculation models are formed. These coupled models, through encapsulation, shield the implementation differences of the underlying code and enable unified instruction execution for initialization, single-step / multi-step calculations, and memory release operations, laying the foundation for upper-level scheduling.

[0059] Next, in the coupling problem manager encapsulation layer, this embodiment, based on the type and functional attributes of the physical fields, adopts an object-oriented derived inheritance paradigm to construct a dedicated coupling problem manager for containerized management of the aforementioned coupling models. Specifically, the system constructs a neutron physics problem manager, a fuel rod thermal conductivity problem manager, a thermal-hydraulic physics problem manager, and a core external system loop problem manager. Each manager integrates the corresponding type of coupling model as needed. For example, the neutron physics problem manager can integrate and sequentially number and manage multiple neutron physics calculation models of different precisions. It is worth noting that, considering the physical characteristics of multi-level reactor systems, this embodiment constructs a logically nested structure at this level: the three managers closely related to the reactor core—neutron, fuel rod, and thermal-hydraulic—are further summarized and assembled to construct a "reactor core coupling problem manager," thereby clearly distinguishing the two major levels of "reactor core" and "core external system loop" in the management architecture.

[0060] Building upon this foundation, the coupling problem data interaction layer constructs a "data bridge" connecting various coupling problem managers based on the multi-physics coupling scenarios encountered. Considering the differences in grid data structures between different physical models, this layer utilizes underlying grid data tools to construct mapping operators and develops and encapsulates multiple types of coupling data interaction modules. For the strong coupling requirements within the reactor core, a "neutron physics field-fuel rod thermal conductivity coupling data interaction module" and a "fuel rod thermal conductivity-thermal-hydraulic coupling data interaction module" are constructed to perform bidirectional transfer of power density, fuel temperature, heat flux density, and coolant parameters between different physical field grids. For system-level coupling requirements, an additional "reactor core-external system loop coupling data interaction module" is constructed to realize boundary condition interaction between the core outlet and the loop inlet.

[0061] Finally, the multi-level coupled scenario implementation layer, based on the preset reactor system coupling logic, assembles the components at each level to construct the main coupled scenario manager. This main coupled scenario manager internally forms a clear multi-level coupling relationship tree: its top level manages the "Reactor Core Coupling Problem Manager" and the "External System Loop Problem Manager," connecting them through a system-level interaction module; and the "Reactor Core Coupling Problem Manager" further coordinates the three sub-managers—neutron, thermal, and fuel rod—and their interaction modules.

[0062] Based on the system architecture described above, the multiphysics coupling calculation process in this embodiment is as follows: Under the unified scheduling of the main coupling scenario manager, global preprocessing is first executed, calling the initialization interfaces of each sub-model to complete parameter loading. Then, the main coupling scenario execution layer is entered, and the system starts a global time-stepping loop. Within each time step, the main coupling scenario manager first drives the reactor core coupling problem manager to perform internal iterative calculations. At this time, the neutron physics model calculates the power distribution and maps the power data to the fuel rod model through the interaction module; the fuel rod model calculates the temperature field accordingly and transmits the heat flow data to the thermal-hydraulic model; after calculating the flow field distribution, the thermal-hydraulic model feeds back the coolant density and temperature to the neutron model to update the macroscopic cross-section. When the residuals of each physics field inside the core meet the convergence requirements, the main coupling scenario manager triggers system-level interaction, transmitting the core outlet parameters to the external system loop problem manager for loop response calculation, and updating the loop feedback input parameters to the core. This process repeats until the simulation of all time steps is completed. After the calculation is finished, the main coupling scenario manager executes unified post-processing, exports the full-field data, and releases resources.

[0063] Figure 4 This paper demonstrates the power distribution results of multi-level multiphysics coupling calculations for reactors under zero-power and full-power conditions, as presented in this embodiment. Thanks to the precise control and efficient iteration of complex data interactions between physics fields through the aforementioned multi-level architecture, the system can accurately capture the power distribution variation characteristics caused by the thermal feedback mechanism under different operating conditions, verifying the effectiveness and computational accuracy of this invention in handling complex nuclear reactor multiphysics coupling problems.

[0064] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A multi-level, multi-physics coupling system for multiple reactor coupling procedures, characterized in that, include: Coupled model module encapsulation layer, coupled problem manager encapsulation layer, coupled problem data interaction layer, coupled multi-level scenario implementation layer, and coupled infrastructure layer; The coupling infrastructure layer includes preprocessing tools, numerical algebra libraries, grid data tools, and postprocessing tools required for coupling. Among them, the preprocessing tools are used for parameter configuration and computational domain initialization before multiphysics coupling calculations; the numerical algebra library is used for numerical solutions in multiphysics coupling; the grid data tools are used for low-level grid topology management and data interaction between different physics fields in multiphysics; and the postprocessing tools are used for the storage, transformation, and visualization of calculation results. The coupling model module encapsulation layer is used to unify and encapsulate multiple heterogeneous reactor physics calculation programs that need to be coupled into coupling modules; Based on the multiphysics coupling scenario, the coupling problem manager encapsulation layer further encapsulates and adds the coupling model already encapsulated in the coupling model module encapsulation layer into a dedicated coupling problem manager, including a neutron physics problem manager, a fuel rod heat conduction problem manager, a thermal-hydraulic physics problem manager, and a core external system loop problem manager; among them, the neutron physics problem manager, the fuel rod heat conduction problem manager, and the thermal-hydraulic physics problem manager are further summarized and assembled into a reactor core coupling problem manager; Based on the multi-physics coupling scenarios faced, the coupling problem data interaction layer develops, builds, and encapsulates coupling data interaction modules between coupling problem managers from the coupling problem manager encapsulation layer, including the neutron physics field-fuel rod thermal conduction coupling data interaction module, the fuel rod thermal conduction-thermal-hydraulic coupling data interaction module, and the reactor core-external system loop coupling data interaction module. Based on the multi-physics coupling scenario faced, the multi-level coupling implementation layer assembles and encapsulates the coupling problem manager from the coupling problem manager encapsulation layer and the coupling data interaction module from the coupling problem data interaction layer in multiple levels, and finally forms a multi-level coupling framework unified in the main coupling scenario manager. The main coupled scene manager includes a preprocessing layer, an execution layer, and a postprocessing layer. The preprocessing layer coordinates the initialization of each sub-manager, reads the parameters of each model, and allocates memory space. The execution layer is used for coupled iteration, coordinates the solution order of each manager, triggers the data interaction layer's transfer operations, and monitors the global coupled residuals to determine convergence. The postprocessing layer is used to release memory resources and output the calculation results after the coupled calculation is completed.

2. The multi-level multi-physics coupling system for multiple reactor coupling procedures as described in claim 1, characterized in that, The coupling model module encapsulation layer uses an object-oriented coupling encapsulation paradigm based on the three-module division of program preprocessing, program core calculation, and program postprocessing to encapsulate the heterogeneous reactor physics calculation programs that need to be coupled into a coupling model with a unified encapsulation paradigm. The heterogeneous reactor physics calculation programs include neutron transport calculation programs, fuel rod thermal conductivity calculation programs, thermal-hydraulic calculation programs, and reactor loop system calculation programs.

3. A multi-level multi-physics coupling system for multiple reactor coupling procedures as described in claim 2, characterized in that, The coupled model includes a model preprocessing submodule, a model core computation submodule, and a model postprocessing submodule.

4. A multi-level multi-physics coupling system for multiple reactor coupling procedures as described in claim 1, characterized in that, The Coupling Problem Manager encapsulation layer is used to represent single-physics problems, complex scene coupling problems, and multi-level sub-coupling problems. The coupling problem manager encapsulation layer assigns coupling model numbers to the added coupling models according to the order in which the coupling models are added after being encapsulated in the coupling model module encapsulation layer, or the coupling model number can be directly specified through the interface provided by the coupling manager.

5. A multi-level multi-physics coupling system for multiple reactor coupling procedures as described in claim 1, characterized in that, The coupled data interaction module in the coupled problem data interaction layer is also a public sub-function belonging to one or more coupled problem managers; The aforementioned coupling problem data interaction layer is used to complete the coupling data interaction from different coupling problem managers in the coupling problem manager encapsulation layer.

6. A multi-level multi-physics coupling system for multiple reactor coupling procedures as described in claim 5, characterized in that, The coupling problem data interaction layer specifies the coupling model from different coupling problem managers in the coupling problem manager encapsulation layer through the coupling model number, so as to realize the coupling data interaction of different coupling models in a single or different coupling problem managers.

7. A multi-level multi-physics coupling system for multiple reactor coupling procedures as described in claim 1, characterized in that, The multi-level coupling scenario implementation layer will add and encapsulate the lower-level coupling managers into the upper-level coupling problem manager based on the object-oriented derivation inheritance mechanism according to the multi-level coupling scenario. The multi-level coupling scenario implementation layer will assign coupling manager numbers to the added coupling managers according to the order in which the coupling managers are encapsulated in the coupling problem manager encapsulation layer, or directly specify the numbers through the interface provided by the coupling problem manager. The multi-level coupled scenario implementation layer is also used to encapsulate the main coupled scenario manager preprocessing layer, the main coupled scenario manager execution layer, and the main coupled scenario manager postprocessing layer. The main coupled scenario manager, as the top-level coupling manager, is composed of multiple sub-coupled managers under a multi-level coupling relationship tree. The main coupled scenario manager execution layer executes the coupled scenario sequentially according to the pre-compiled coupling manager number and the coupling model number in each sub-coupled manager, and simultaneously performs coupling data exchange between single or different coupling problem managers to complete the coupling data interaction.

8. A multi-level multi-physics coupling system for multiple reactor coupling procedures as described in claim 7, characterized in that, The multi-level scenario implementation layer further specifies the coupling iteration control, coupling method, and / or coupling time step through the input file provided by the coupling problem manager for each sub-level.

9. A multi-level multi-physics coupling system for multiple reactor coupling procedures as described in claim 7, characterized in that, In the multi-level coupled scene implementation layer, the main coupled scene manager preprocessing layer will complete the preprocessing of all sub-coupled managers in one go or selectively, and the main coupled scene manager postprocessing layer will complete the postprocessing of all sub-coupled managers in one go or selectively.

10. A multi-level multi-physics coupling method for multiple reactor coupling procedures, implemented based on the multi-level multi-physics coupling system according to any one of claims 1 to 9, characterized in that, include: First, in the coupled infrastructure layer and the coupled model module encapsulation layer, the corresponding basic numerical library and mesh tools are called according to the calculation request, the required underlying physical calculation program is loaded and initialized and transformed into a standardized coupled model, thereby completing the transformation from the original code to the basic data unit that can be called by the system. Subsequently, based on the physical properties, the standardized coupled model is injected into the corresponding proprietary manager for containerized assembly. Based on the correlation between physical fields, the underlying mesh mapping tool is called to establish a data interaction module, thereby completing the integration from single model data to local physical field data and establishing a bridge for data flow. Finally, at the top-level coupled multi-level scene implementation layer, the various local managers and interaction modules are further integrated upwards into a complete main coupled scene manager; After the main coupled scene manager takes over control, it drives the flow of underlying data between different levels: the physical data generated by the underlying model calculations is processed by the coupled problem data interaction layer and then passed between the various managers. The computation status of each sub-manager is fed back through the coupling problem manager encapsulation layer and converges to the main coupling scene manager level by level. The main coupling scene manager iteratively processes the data according to the convergence status, and finally integrates the data calculated at each level into a complete multiphysics solution, and outputs the final calculation result through the post-processing interface of the main coupling scene manager.

Images