Self-adaptive Monte Carlo region decomposition restarting method and system and medium

Abstract

The invention discloses a self-adaptive Monte Carlo region decomposition restart method and system and a medium. The method comprises the following steps: creating a restart database file at the end of each burnup step of a reactor core Monte Carlo burnup calculation program; creating data groups in the restart database file according to different data categories, and creating a one-dimensional data set of restart data for each data group; before the burnup step is interrupted, obtaining calculation data of each process in each region, and summarizing the data of each region to obtain multi-dimensional data of the reactor core; converting the multi-dimensional data of the reactor core into one-dimensional data, and writing the one-dimensional data into a restart database file; when calculation is restarted, a restarting database file corresponding to the burnup step is directly called, and calculation data decomposed in different areas are distributed to achieve self-adaptive restarting. According to the method and the device, a program interruption calculation function is realized, the problem of repeated calculation of the program is reduced, the method and the device can adapt to automatic decomposition of different numbers of regions, and restarting efficiency and correctness are improved.

Description

Technical Field

[0001] This invention relates to the field of nuclear reactor core programming technology, specifically to an adaptive Monte Carlo decomposition restart method, system, and medium. Background Technology

[0002] Spatial region decomposition is a parallel computing method that divides the reactor model into several spatial geometric regions and allocates processes (cores) according to the regions. In this way, each process only handles the simulation and calculation in one region, and the connection between regions is realized through inter-process communication, thereby achieving the decomposition of computational data and computational memory.

[0003] The reactor core Monte Carlo burnup calculation program (hereinafter referred to as the Monte Carlo program) simulates the entire core lifespan of the reactor core. During the calculation process, the operation may be interrupted due to various unpredictable problems such as the hardware platform or internal program issues. In order to save calculation time and avoid recalculating previously calculated results, it is necessary to flexibly save the information of the calculated burnup step and resume execution from the interrupted step, and the result of the resumed execution should be highly consistent with the original result.

[0004] Traditional Monte Carlo program restart functions typically write material nuclide information that changes with burnup into the original input file. When the program restarts, it reads fixed core geometry data and material nuclide data at the time of interruption from the original input file. This method enables the restart function.

[0005] Traditional methods achieve restart functionality by reconstructing the input file. Each restart input file contains copies of core geometry, burnup, and criticality information. However, when the core is large, the amount of material nuclide information written becomes excessive, reaching hundreds of millions of lines, thus limiting its applicability to large-scale core problems. Furthermore, traditional methods do not consider the impact of parameters such as random numbers, fission sources, and thermal feedback data on the program execution results. Therefore, when the program restarts and continues calculations, incomplete restart data leads to inconsistencies between the calculation results and those before the interruption.

[0006] When regional decomposition exists, traditional methods store data for each node and region separately. To restart computation, the number of upstream and downstream computing nodes and regions must be exactly the same. However, in practice, computing node resources may not meet the requirements of upstream computing nodes, potentially leading to computation failure. Therefore, traditional methods are inconvenient to use.

[0007] In view of the above, this application is hereby submitted. Summary of the Invention

[0008] The technical problem to be solved by the present application is that the restart method of the traditional Monte Carlo program is to directly write material information of each burn step into the original input file, and the function is realized by reconstructing the input file, but the calculation result after restart cannot be consistent with that before stopping due to incomplete stored data, and the efficiency of directly operating large amount of text files is low and the correctness is not high. The present application aims to provide a self-adaptive Monte Carlo region decomposition restart method, system and medium, which realizes the function of interrupting and continuing calculation by using HDF5 database as a restart database carrier, writes all data related to the calculation result (i.e. calculation data associated with the core geometry data of the reactor core Monte Carlo burnup calculation program) such as nuclide cross section data, material nuclide density, fission source, random number and other related data into the restart database file in HDF5 format in a one-dimensional continuous storage mode, realizes the function of interrupting and continuing calculation, reduces the problem of repeated calculation of the program, and can adapt to different number of region decomposition, thereby improving the restart efficiency and correctness.

[0009] The present application is realized by the following technical solutions: In a first aspect, the present application provides a self-adaptive Monte Carlo region decomposition restart method, which comprises: A restart database file is created at the end of each burn step of the reactor core Monte Carlo burnup calculation program; different data groups are created in the restart database file according to different data categories, and one-dimensional data sets of restart data are created for each data group; Before the interruption of the burn step, the calculation data of each process in each region is obtained, and the data of each region is summarized to obtain multi-dimensional data of the core; the calculation data is calculation data associated with the core geometry data of the reactor core Monte Carlo burnup calculation program; The multi-dimensional data of the core is converted into one-dimensional data, and the one-dimensional data is written into the restart database file; When restarting the calculation, the restart database file corresponding to the burn step is directly called to distribute the calculation data under different region decomposition to realize self-adaptive restart.

[0010] The technical design points of the application are: 1) using HDF5 as a restart database; 2) creating data groups in the HDF5 restart database according to different data categories, such as random number data group, material data group, nuclide cross-section data group, critical calculation data group, and thermal feedback data group; 3) adaptive domain decomposition, reconstructing the data related to Monte Carlo calculation, such as random parameter data, material data, and thermal feedback data (fuel temperature, moderator temperature, and moderator density) according to different region decomposition numbers and different process numbers, and realizing data mapping with the core geometry region; 4) during restart calculation, reading the restart database file corresponding to the burnup step, calling (3) to realize data mapping of the restart calculation, and then distributing the restart calculation data according to the calculation node to realize adaptive restart. The application does not store fixed core geometry data in the restart database file, greatly reducing the disadvantages of storing core geometry data in the original input file, and ensuring the consistency of the calculation results with those before interruption through the restart database.

[0011] Further, the restart database file is an HDF5 format database file.

[0012] Further, the data groups include random number data group, material nuclide data group, nuclide cross-section data group, critical calculation data group, and thermal feedback data group.

[0013] Further, one-dimensional data sets of restart data are created for each data group, including: a one-dimensional random number parameter data set of restart data is created below the random number data group; a one-dimensional nuclide data set of restart data is created below the material nuclide data group; a one-dimensional nuclide cross-section data set of restart data is created below the nuclide cross-section data group; a one-dimensional critical calculation data set of restart data is created below the critical calculation data group; a one-dimensional thermal feedback data set of restart data is created below the thermal feedback data group; the thermal feedback data includes fuel temperature, moderator temperature, and moderator density.

[0014] Further, the core multi-dimensional data is converted into one-dimensional data, specifically: the core multi-dimensional data is converted into one-dimensional data according to the grid arrangement rule.

[0015] Further, during restart calculation, the restart database file corresponding to the burnup step is directly called to distribute the calculation data under different region decompositions to realize adaptive restart, including: When the restart calculation is performed, if the restart is performed from the i-th burnup step, the core geometry data is read from the original input file by the reactor core Monte Carlo burnup calculation program, and the restart database file of the burnup step is called, wherein i is a positive integer; The corresponding calculation data is read from the restart database file of the burnup step, and the calculation data under different region decompositions is distributed to realize adaptive restart.

[0016] Further, the corresponding calculation data is read from the restart database file of the burnup step, and the calculation data under different region decompositions is distributed to realize adaptive restart, including: The restart database file is read, and configuration change judgment is performed; If the number of regions and processes of the restart is the same, the corresponding calculation data is directly read from the restart database file of the burnup step and the calculation data is loaded; If the number of regions and processes of the restart is different, the calculation data is redistributed, the data is remapped according to the newly created region division, and after remapping, the calculation data is distributed to the new processes.

[0017] In a second aspect, the application further provides an adaptive Monte Carlo region decomposition restart system, which comprises: A restart database creation unit is configured to create a restart database file at the end of each burnup step of the reactor core Monte Carlo burnup calculation program; the restart database file internally creates data groups according to different data categories, and creates a one-dimensional data set of restart data for each data group; A pre-interruption calculation data acquisition unit is configured to acquire the calculation data of each process in each region before the burnup step is interrupted, and to obtain core multi-dimensional data by summarizing the data of each region; the calculation data is calculation data associated with the core geometry data of the reactor core Monte Carlo burnup calculation program; A pre-interruption calculation data conversion unit is configured to convert the core multi-dimensional data into one-dimensional data, and write the one-dimensional data into the restart database file; An adaptive region decomposition unit is configured to directly call the restart database file of the corresponding burnup step to distribute the calculation data under different region decompositions and realize adaptive restart when the restart calculation is performed.

[0018] Further, the restart database file is an HDF5 format database file.

[0019] Further, the execution process of the adaptive region decomposition unit is as follows: When the restart calculation is performed, if the restart is performed from the i-th burnup step, the core geometry data is read from the original input file by the reactor core Monte Carlo burnup calculation program, and the restart database file of the burnup step is called, wherein i is a positive integer; read the corresponding calculation data from the restart database file of the burnup step, and distribute the calculation data under different region decompositions to realize adaptive restart, comprising: read the restart database file, and perform configuration change judgment; if the restart region and the process number configuration are the same, the corresponding calculation data is directly read from the restart database file of the burnup step, and the calculation data is loaded; if the restart region and the process number configuration are different, the calculation data is redistributed, the data is remapped according to the newly created region division, and after remapping, the calculation data is distributed to the new process.

[0020] In a third aspect, the present application further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor realizes the adaptive Monte Carlo region decomposition restart method described above when executing the computer program.

[0021] In a fourth aspect, the present application further provides a computer readable storage medium, which stores a computer program, and the computer program realizes the adaptive Monte Carlo region decomposition restart method described above when executed by a processor.

[0022] Compared with the prior art, the present application has the following advantages and beneficial effects: The present application uses HDF5 database as a restart database carrier, writes all data related to the calculation results (i.e. calculation data associated with the reactor core Monte Carlo burnup calculation program core geometry data) such as nuclear cross section data, material nuclide density, fission source, random number, etc. into the restart database file in HDF5 format in a one-dimensional continuous storage mode. When restarting, the program interruption and continuous calculation function is realized, the problem of repeated calculation of the program is reduced, and the restart efficiency and correctness are improved. BRIEF DESCRIPTION OF DRAWINGS

[0023] The accompanying drawings described herein are used to provide further understanding of the embodiments of the present application, constitute a part of the present application, and do not constitute a limitation of the embodiments of the present application. In the drawings: Figure 1 Flowchart of the adaptive Monte Carlo region decomposition restart method of the present application Figure 1 ; Figure 2 Flowchart of the adaptive Monte Carlo region decomposition restart method of the present application Figure 2 ; Figure 3 Specific structure diagram of the HDF5 restart database of the present application Figure 4 Fig. 1 is a schematic diagram of a data collection phase of the present application; Figure 5 Fig. 2 is a schematic diagram of a restart phase of the present application; Figure 6 Fig. 3 is a comparison diagram of different restart scenarios of the present application; Figure 7 Fig. 4 is a schematic diagram of data remapping of the present application; Figure 8 Fig. 5 is a structure block diagram of an adaptive Monte Carlo region decomposition restart system of the present application. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical scheme and advantages of the present application more clear and explicit, the present application is further described in detail below in combination with embodiments and drawings, the schematic embodiments of the present application and the description thereof are only used to explain the present application, and do not limit the present application.

[0025] First, basic principle The reactor core Monte Carlo burnup calculation program is a neutron simulation program based on probability theory, and the next behavior of particles, such as whether to collide with a nuclide, flight direction, flight distance, etc. are determined by random numbers, so the repeatability of the program calculation result is closely related to the random numbers.

[0026] The number of initial source particles, the initial parameters (energy, direction, initial weight) of particles affect the subsequent particle flight trajectory and parameter statistics, and the repeatability calculation of the calculation result is also necessary, so it will also be an important parameter for whether the restart is correct or not.

[0027] There is a close coupling relationship between the nuclear characteristics of the reactor core and the thermal-hydraulic characteristics, the core power distribution affects the fuel temperature distribution and the moderator temperature and density distribution, and these distributions affect the nuclear characteristics of the materials in the reactor, and further affect the neutron flux and power distribution in the core, so the thermal-hydraulic parameters will affect the calculation results of the program.

[0028] The spatial region decomposition method is a parallel computing method, which divides the reactor model into several spatial geometric regions, and distributes the processes (nuclei) according to the regions, so that each process only handles the simulation and calculation in one region, and the connection between regions is realized through inter-process communication, so as to realize the decomposition of the calculation data and the calculation memory.

[0029] The application of domain decomposition in Monte Carlo simulation of particle transport mainly includes two processes: "decomposition" and "communication". The "decomposition" is to divide the geometry region of the built reactor model, and to allocate different sub-regions to different processors, so that each processor simultaneously simulates the movement of particles in the region. The "communication" is to store the particle state and transfer the data between regions when the particles pass through the regions, so as to complete the particle transport process and realize the coupling calculation between regions.

[0030] In the existing core calculation, the geometric information, burnup lattice, critical information and other parameters of the core are fixed and will not change with the change of the burnup step, so the fixed information does not need to be saved every step when storing the restart information, resulting in space occupation.

[0031] Since the Monte Carlo burnup calculation program of the reactor core uses the direct simulation method to calculate the core parameters, the data volume is very large, and the use of ordinary text file reading and writing will cause a bottleneck in performance. HDF5 is a high-efficiency database for accessing, and can read large data in parallel.

[0032] For the case of domain decomposition, the data of different region materials, random numbers and fission source information need to be adaptively communicated and summarized for writing into the HDF5 file.

[0033] The traditional restart method of the Monte Carlo program is to directly write the material information of each burnup step into the original input file, and to realize the function by reconstructing the input file. Since the stored data is incomplete, the calculation result after restart cannot be consistent with the previous result, and directly operating large data volume text files is low in efficiency. Therefore, the present application solves the correctness and efficiency of restart by using the following methods: (1) HDF5 database is used as the restart database file.

[0034] (2) Groups are created in the restart database file according to the stored data categories.

[0035] (3) All data related to the calculation result, such as random numbers, fission sources, thermal parameters and material nuclides, are adaptively summarized and stored in the restart database file in one-dimensional form.

[0036] Second, main process The main process of the adaptive Monte Carlo domain decomposition restart method of the present application is as follows: 1) In the burnup step Step i , create an HDF5 format restart database file HRFile i . 2) In parallel multi-process environment, the main process collects the random number seed parameter ullRN_ISEED, the current random number position RN_POS and the total number of random numbers RandNum of each process by using MPI function; 3) The main process collects the fission source parameter FisSrc of each process; 4) The main process collects the thermal feedback data fuel temperature FuelTemp, moderator temperature ModTeff and moderator density ModDen; 5) The main process collects the burnup cell material nuclide data MatNuc; 6) In the restart database file HRFile in HDF5 format i The internal data is grouped according to different data categories, and one-dimensional data sets of random numbers, fission sources, thermal data and material nuclides are respectively created; 7) The collected core multidimensional data is converted into one-dimensional data according to the same arrangement sequence of the cell, and is written into the restart database file HRFile i 8) When restarting the calculation, if starting from the ith burnup step, after the program reads the core geometry information in the initial input file, the restart database file HRFile i of the burnup step is called to realize the update of random numbers, fission sources, thermal parameters and material nuclide data.

[0037] Compared with the traditional restart method based on the original input file, the main differences of the restart method based on the HDF5 library are as follows: 1) the restart database carrier is a separate HDF5 restart database, instead of the traditional construction of a new input file; 2) the influence of random numbers, fission sources and thermal feedback data on the calculation results is considered, and these data are stored in the restart database for restart calculation; 3) the data is grouped and the data set is one-dimensional in HDF5, and the access efficiency is high.

[0038] The application provides correct and efficient breakpoint continuation calculation function by using HDF5 database as the restart database file and storing all data related to the calculation results in the form of one-dimensional data set. The main processes include: (1) collecting all data affecting the calculation results in steps 2) to 5); (2) grouping the internal data of the restart database file according to the data categories and creating one-dimensional data set in step 6); (3) the fixed core geometry data is not stored in the restart library, when starting the calculation, the restart database file of the corresponding burnup step is directly called to distribute the calculation data under different region decomposition to realize adaptive restart.

[0039] Example 1 As Figure 1 ​As shown, the adaptive Monte Carlo region decomposition restart method of the present application comprises: A restart database file is created at the end of each burnup step of the reactor core Monte Carlo burnup calculation program; data groups are created inside the restart database file according to different data categories, and one-dimensional data sets of restart data are created for each data group; Before the burnup step is interrupted, the calculation data of each process is obtained in each region, and the data of each region is summarized to obtain the core multidimensional data; the calculation data is calculation data associated with the core geometry data of the reactor core Monte Carlo burnup calculation program; The core multidimensional data is converted into one-dimensional data, and the one-dimensional data is written into the restart database file; During restart calculation, the restart database file corresponding to the burnup step is directly called to distribute the calculation data under different region decompositions to realize adaptive restart.

[0040] In this embodiment, the restart database file is an HDF5 format database file.

[0041] In this embodiment, the data groups include random number data groups, material nuclide data groups, nuclide cross section data groups, criticality calculation data groups, and thermal feedback data groups.

[0042] In this embodiment, one-dimensional data sets of restart data are created for each data group, including: A one-dimensional random number parameter data set of restart data is created under the random number data group; A one-dimensional nuclide data set of restart data is created under the material nuclide data group; A one-dimensional nuclide cross section data set of restart data is created under the nuclide cross section data group; A one-dimensional criticality calculation data set of restart data is created under the criticality calculation data group; A one-dimensional thermal feedback data set of restart data is created under the thermal feedback data group; the thermal feedback data includes fuel temperature, moderator temperature, and moderator density.

[0043] In this embodiment, the core multidimensional data is converted into one-dimensional data, specifically: The core multidimensional data is converted into one-dimensional data according to the cell arrangement rule.

[0044] In this embodiment, during restart calculation, the restart database file corresponding to the burnup step is directly called to distribute the calculation data under different region decompositions to realize adaptive restart, including: When the restart calculation is performed, if it is started from the ith burnup step, the core geometry data is read from the original input file by the reactor core Monte Carlo burnup calculation program, and the restart database file of the burnup step is called; The corresponding calculation data is read from the restart database file of the burnup step, and the calculation data under different region decompositions is distributed to realize adaptive restart, including: The restart database file is read, and configuration change judgment is performed; If the number of regions and processes of the restart is the same, the corresponding calculation data is directly read from the restart database file of the burnup step and the calculation data is loaded; If the number of regions and processes of the restart is different, the calculation data is redistributed, the data is remapped according to the newly created region division, and after remapping, the calculation data is distributed to the new process.

[0045] As Figure 2 shown, the specific implementation is as follows: First step: in the burnup step Step i , the restart database file HRFile i in HDF5 format is created i , and the specific structure of HRFile i is shown in Figure 3 ; Second step: collect the random number parameter ullRNG, the fission source parameter FisSrc, the thermal feedback data (fuel temperature FuelTemp, moderator temperature ModTeff, and moderator density ModDen), and the burnup cell material nuclide data in each region, and then aggregate the data of each region; Third step: convert the collected core multidimensional data into one-dimensional storage format; Fourth step: group different data categories in the restart database file HRFile i , and create one-dimensional data sets of random numbers, fission sources, thermal data, and material nuclides; Fifth step: convert the collected core multidimensional data into one-dimensional data according to the cell arrangement rule, and write it into the restart database file HRFilei; Sixth step: when the restart calculation is performed, if it is started from the ith burnup step, the core geometry data is read from the original input file by the Monte Carlo program, and then the restart database file HRFilei of the burnup step is called, and the random number, fission source, thermal parameter, and material nuclide data of different processes in different regions are distributed, so as to realize the adaptive restart function; The sixth step specifically includes: S61, reading the restart database file, and performing configuration change judgment; S62, if the restart region and the process number configuration are the same, directly reading the corresponding calculation data from the restart database file of the burnup step and loading the calculation data; S63, if the restart region and the process number configuration are different, redistributing the calculation data, remapping the data according to the newly created region division, and distributing the calculation data to the new process after remapping.

[0046] The key of the adaptive restart of the present application is: A. Each burnup step saves complete state data (in the form of a one-dimensional array, containing sufficient information for reconstruction).

[0047] B. When restarting, the data can be remapped according to the new region division.

[0048] C. After remapping, the data is distributed to the new process.

[0049] Through the above method, the calculation data is first split according to different regions, and then the split calculation data is sent to each process in the region; the restart at any burnup step can be realized, and the change of parallel configuration (number of regions and number of processes) is supported, thereby improving the flexibility of calculation and resource utilization.

[0050] Among them, the data collection stage is mainly the second to fourth steps, see Figure 4 ; the restart stage is mainly the sixth step, see Figure 5 ; the comparison chart of different restart scenarios, see Figure 6 ; the data remapping diagram, see Figure 7 .

[0051] Embodiment 2 As shown in Figure 8 , the difference between this embodiment and embodiment 1 is that this embodiment provides an adaptive Monte Carlo region decomposition restart system, which corresponds to the adaptive Monte Carlo region decomposition restart method of embodiment 1; the system comprises: A restart database creation unit for creating a restart database file at the end of each burnup step of the reactor core Monte Carlo burnup calculation program; the restart database file internally creates data groups according to different data categories, and creates a one-dimensional data set of restart data for each data group; A pre-interruption calculation data acquisition unit for acquiring the calculation data of each process in each region before the interruption of the burnup step, and summarizing the data of each region to obtain multi-dimensional data of the core; the calculation data is the calculation data associated with the core geometry data of the reactor core Monte Carlo burnup calculation program; A pre-interruption calculation data conversion unit for converting the multi-dimensional data of the core into one-dimensional data and writing the one-dimensional data into the restart database file; An adaptive domain decomposition unit is used to directly call the restart database file of the corresponding burnup step to distribute the calculation data under different domain decompositions to realize adaptive restart when restart calculation is performed.

[0052] As a further implementation, the restart database file is an HDF5 format database file.

[0053] As a further implementation, the execution process of the adaptive domain decomposition unit is as follows: When restart calculation is performed, if restart is performed from the i-th burnup step, the core geometry data is read from the original input file by the reactor core Monte Carlo burnup calculation program, and the restart database file of the burnup step is called; The corresponding calculation data is read from the restart database file of the burnup step to distribute the calculation data under different domain decompositions to realize adaptive restart, including: The restart database file is read to perform configuration change judgment; If the restart domain and the process number configuration are the same, the corresponding calculation data is directly read from the restart database file of the burnup step and the calculation data is loaded; If the restart domain and the process number configuration are different, the calculation data is redistributed, the data is remapped according to the newly created domain division, and after remapping, the calculation data is distributed to the new processes.

[0054] The execution process of each unit can be performed according to the adaptive Monte Carlo domain decomposition restart method flow steps of embodiment 1, which will not be described again in this embodiment.

[0055] Meanwhile, the application further provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor realizes the adaptive Monte Carlo domain decomposition restart method when executing the computer program.

[0056] Meanwhile, the application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the adaptive Monte Carlo domain decomposition restart method.

[0057] Those skilled in the art should understand that the embodiments of the application can be provided as a method, a system, or a computer program product. Therefore, the application can be in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the application can be in the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program code.

[0058] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0059] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0060] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0061] The above description is only specific implementation of the present application, and is not intended to limit the protection scope of the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. An adaptive Monte Carlo region decomposition restart method, characterized in that, The method includes: At the end of each burnup step in the reactor core Monte Carlo burnup calculation program, a restart database file is created; the restart database file contains data groups created according to different data categories, and a one-dimensional dataset of restart data is created for each data group; Before the burnup step is interrupted, the calculation data of each process is acquired in each region, and the data of each region is summarized to obtain the multi-dimensional data of the reactor core; the calculation data is the calculation data associated with the core geometry data of the reactor core Monte Carlo burnup calculation program; The multi-dimensional data of the reactor core is converted into one-dimensional data, and the one-dimensional data is written into the restart database file; When restarting the calculation, the restart database file of the corresponding fuel consumption step is directly called to distribute the calculation data under different regional decomposition to achieve adaptive restart.

2. The adaptive Monte Carlo region decomposition restart method according to claim 1, characterized in that, The restart database file is an HDF5 format database file.

3. The adaptive Monte Carlo region decomposition restart method according to claim 1, characterized in that, The data set includes a random number data set, a material nuclide data set, a nuclide cross-section data set, a critical calculation data set, and a thermal feedback data set.

4. The adaptive Monte Carlo region decomposition restart method according to claim 3, characterized in that, Create a one-dimensional dataset of restart data for each data group, including: Create a one-dimensional random number parameter dataset for restarting data under the aforementioned random number data set; Create a one-dimensional nuclide dataset for restarting data under the aforementioned material nuclide data group; Create a one-dimensional nuclide cross-section dataset for the restart data under the aforementioned nuclide cross-section data group; Create a one-dimensional critical calculation dataset for restart data under the critical calculation data group; Create a one-dimensional thermal feedback dataset for restart data under the thermal feedback data group; the thermal feedback data includes fuel temperature, moderator temperature, and moderator density.

5. The adaptive Monte Carlo region decomposition restart method according to claim 1, characterized in that, The multi-dimensional data of the stack core is converted into one-dimensional data, specifically as follows: The multidimensional data of the core is converted into one-dimensional data according to the cell arrangement rules.

6. The adaptive Monte Carlo region decomposition restart method according to claim 1, characterized in that, When restarting the computation, the restart database file for the corresponding fuel consumption step is directly called to distribute the computation data under different regional decompositions to achieve adaptive restart, including: When restarting the calculation, if it starts from the i-th burnup step, the core geometry data is read from the original input file by the reactor core Monte Carlo burnup calculation program, and the restart database file for that burnup step is called. The corresponding computational data is read from the restart database file of the fuel consumption step, and the computational data under different regional decompositions is distributed to achieve adaptive restart.

7. The adaptive Monte Carlo region decomposition restart method according to claim 6, characterized in that, The corresponding computational data is read from the restart database file of this fuel consumption step, and the computational data under different regional decompositions is distributed to achieve adaptive restart, including: Read the restart database file and check for configuration changes; If the region and number of processes are configured to be restarted in the same way, the corresponding calculation data is directly read from the restart database file of that fuel consumption step and the calculation data is loaded. If the number of regions and processes configured for restarting are different, the computation data is redistributed, the data is remapped according to the newly created region division, and after remapping, the computation data is allocated to the new processes.

8. An adaptive Monte Carlo region decomposition and restart system, characterized in that, The system includes: The restart database creation unit is used to create a restart database file at the end of each burnup step in the reactor core Monte Carlo burnup calculation program; the restart database file creates data groups according to different data categories, and creates a one-dimensional dataset of restart data for each data group; The pre-interruption calculation data acquisition unit is used to acquire the calculation data of each process in each region before the burnup step is interrupted, and to summarize the data of each region to obtain the core multidimensional data; the calculation data is the calculation data associated with the core geometry data of the reactor core Monte Carlo burnup calculation program; The pre-interruption calculation data conversion unit is used to convert the multi-dimensional data of the core into one-dimensional data and write the one-dimensional data into the restart database file; The adaptive region decomposition unit is used to directly call the restart database file of the corresponding fuel consumption step when restarting the calculation, and distribute the calculation data under different region decompositions to achieve adaptive restart.

9. An adaptive Monte Carlo region decomposition restart system according to claim 8, characterized in that, The restart database file is an HDF5 format database file.

10. An adaptive Monte Carlo region decomposition restart system according to claim 8, characterized in that, The execution process of the adaptive region decomposition unit is as follows: When restarting the calculation, if it starts from the i-th burnup step, the core geometry data is read from the original input file by the reactor core Monte Carlo burnup calculation program, and the restart database file for that burnup step is called. The corresponding computational data is read from the restart database file of this fuel consumption step, and the computational data under different regional decompositions is distributed to achieve adaptive restart, including: Read the restart database file and check for configuration changes; If the region and number of processes are configured to be restarted in the same way, the corresponding calculation data is directly read from the restart database file of that fuel consumption step and the calculation data is loaded. If the number of regions and processes configured for restarting are different, the computation data is redistributed, the data is remapped according to the newly created region division, and after remapping, the computation data is allocated to the new processes.

11. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements an adaptive Monte Carlo region decomposition restart method as described in any one of claims 1 to 7.

12. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements an adaptive Monte Carlo region decomposition restart method as described in any one of claims 1 to 7.

Images