Low-group-energy group structure optimization calculation method for research reactor physical calculation

The small group energy group structure of the research reactor is optimized through the built-in genetic algorithm of DAKOTA, which solves the problem of excessive time-consuming core calculation of the research reactor, and achieves efficient calculation accuracy and efficiency, which is suitable for core layout and material replacement solutions. Optimized design.

CN119089986AActive Publication Date: 2024-12-06XI AN JIAOTONG UNIV
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202411211014.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-30
Publication Date
2024-12-06
Estimated Expiration
2044-08-30

AI Technical Summary

Technical Problem

In the prior art, when conducting large-scale calculation of research reactor cores, the precision of fine energy groups is high but takes too long, making it difficult to take into account both the calculation accuracy and efficiency, especially in the optimization design of core layout schemes and material replacement schemes. calculate.

Method used

The built-in genetic algorithm module of DAKOTA is used to optimize the structure of the energy group of the few groups. Through the Python interface program and the neutron computing program SARAX, an optimized energy group structure is generated, and the accuracy is controlled and the computing efficiency is improved.

Benefits of technology

On the basis of maintaining calculation accuracy, the efficiency of core calculation is significantly improved and the calculation amount is reduced. It is suitable for studying the optimization design of the core layout scheme and material replacement scheme.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119089986A_ABST
    Figure CN119089986A_ABST
Patent Text Reader

Abstract

A few-group-energy group structure optimization calculation method for research reactor physical calculation comprises the following steps: calling a built-in genetic algorithm module of an open source optimization toolbox DAKOTA, providing gene numbers to generate different few-group-energy group structures, and then providing the different few-group-energy group structures to a neutronics calculation program for calculation to obtain a reactor core physical calculation result. According to the method, target parameters are extracted from a physical calculation result of the reactor core to manufacture a fitness function, and an optimized low-group-energy group structure is obtained through iterative calculation between a gene number and a fitness function value, so that the purpose that calculation precision and calculation efficiency can be both considered for a large-scale reactor core calculation problem is achieved. According to the method disclosed by the invention, the small-group-energy group structure can be optimized through the genetic algorithm module built in the externally coupled optimization toolbox DAKOTA, the precision of the small-group-energy group structure is controlled at a relatively high level, the transportation calculation efficiency of a large-scale reactor core problem is improved, the calculation precision and the calculation efficiency are both considered, and the method is suitable for popularization and application. And the calculation amount is reduced for researching the optimization design of a reactor core arrangement scheme and a refueling scheme.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of nuclear reactor core design and nuclear reactor physical calculation technology, and in particular to a small group energy group structure optimization calculation method for research reactor physical calculation. Background Art

[0002] As the thermal neutron research reactor with the highest operating power in my country, HFETR is of great importance to the research on the irradiation performance of various reactor fuels and materials in my country. Due to the strong non-uniformity of the HFETR core layout, the complex core energy spectrum, and the heavy test tasks undertaken by HFETR, the simulation program needs to have flexible fuel management calculation functions. The advanced reactor neutronics calculation program NECP-SARAX, independently developed by the Nuclear Engineering Computational Physics Laboratory (NECP) of Xi'an Jiaotong University, has been verified by calculation on the 81-1S series simulated core loading of HFETR, indicating that the SARAX program has good calculation accuracy under the 32-group energy group structure.

[0003] Due to the need for optimizing the core layout and refueling schemes of the research reactor, a large number of full-core transport calculations are required. If the cross-sectional calculation of the fine energy group is used, the calculation accuracy is sufficient, but the calculation takes too long. The calculation efficiency of the small group structure is high, especially for large-scale core calculations. Controlling the accuracy of the small group energy group structure at a high level can greatly improve the efficiency of the core transport calculation. In order to take into account both calculation accuracy and calculation efficiency, it is necessary to optimize the small group energy group structure.

[0004] For multi-input, multi-objective, and constrained problems such as core calculations, genetic algorithms are a more suitable intelligent optimization algorithm. Genetic algorithms automatically search for the global optimal solution by simulating Darwin's biological evolution inheritance and natural selection process. The optimization method of probabilistic selection can adaptively adjust the search range and search direction. The optimization toolbox DAKOTA developed by Sandia National Laboratories in the United States has the advantages of being free, open source, and highly scalable. It includes optimization, sensitivity analysis and other functions. Through the Python interface program, the built-in genetic algorithm module of DAKOTA can be called for optimization calculations without changing the external program code.

[0005] In summary, when using the cross section of the fine energy group for large-scale core calculations, the calculation accuracy is sufficient but the calculation amount is large. When optimizing the core layout and refueling schemes of research reactors, a large number of core calculations are required, and the use of fine energy group calculations takes too long. Therefore, it is necessary to invent a small group energy group structure optimization calculation method for research reactor physics calculations, study the small group energy group structure with higher calculation accuracy, take into account both calculation accuracy and calculation efficiency, and reduce the calculation amount for the optimization design of the core layout and refueling schemes of research reactors. Summary of the invention

[0006] In order to solve the problems existing in the above-mentioned prior art, the purpose of the present invention is to provide a small group energy group structure optimization calculation method for research reactor physical calculation. The method of the present invention can optimize the small group energy group structure through the genetic algorithm module built in the externally coupled optimization toolbox DAKOTA, control the accuracy of the small group energy group structure at a high level, improve the transport calculation efficiency of large-scale core problems, take into account both calculation accuracy and calculation efficiency, and reduce the amount of calculation for the optimization design of research reactor core layout scheme and material replacement scheme.

[0007] In order to achieve the above objectives, the present invention is implemented by the following technical solutions:

[0008] A method for optimizing the structure of a small group of energy groups for research reactor physics calculations comprises the following steps:

[0009] Step 1: The genetic algorithm optimization module built into DAKOTA is coupled with the neutronics calculation program SARAX through the Python interface program. The gene number generated by DAKOTA, that is, a randomly generated natural number, is converted into an array through permutations and combinations. The interval width of the minority energy group is determined based on the array, so that the 1968 groups are divided into the corresponding minority energy group structure;

[0010] Step 2: Call the component calculation program TULIP of SARAX, read the small group energy group structure obtained in step 1 to perform component calculation, obtain the small group cross section, shape factor and other component small group parameters, and provide them to the subsequent core calculation program LAVENDER;

[0011] Step 3: Call the core calculation program LAVENDER of SARAX to perform core calculation based on the component group parameters obtained in step 2 to obtain the effective proliferation coefficient k of the thermal neutron research reactor eff , power distribution, flux distribution, burnup distribution, total isotope production and other core physics calculation results;

[0012] Step 4: Extract the effective multiplication factor k of the thermal neutron research reactor from the core physics calculation results obtained in step 3 eff , the total isotope yield is used as the target parameter to make the fitness function, which is fed back to DAKOTA for optimization calculation; the fitness function expression is shown in formula (1):

[0013]

[0014] Where:

[0015] Fitness—fitness function value;

[0016] w1, w2, w3—weights of the three target parameters;

[0017] k eff,calc,initial , k eff,calc,end —The effective proliferation coefficients at the beginning and end of core burnup are calculated based on the small group energy group structure;

[0018] k eff,fine,initial , k eff,fine,end —The effective multiplication coefficients at the beginning and end of core burnup are calculated based on the refined energy group structure;

[0019] production calc,end 、production fine,end —Total isotope production at the end of core burnup based on the few-group energy group structure and the fine energy group structure.

[0020] Step 5: DAKOTA performs genetic algorithm operations such as selection, crossover, and mutation according to the fitness function fed back in step 4, and then repeats the calculation process from step 1 to step 4 from the number of genes to the fitness function, iteratively generates a population with better genes, and according to the principle of survival of the fittest, genetic evolution from generation to generation produces better approximate solutions, approaches the global optimal solution, and finally converges to the global approximate optimal solution, that is, searches for the gene that minimizes the fitness function value and obtains the optimized small group energy group structure.

[0021] Compared with the existing technology, it has the following advantages:

[0022] 1. The method of the present invention uses a genetic algorithm to optimize the calculation of the energy group structure of a small group, and controls the accuracy of the energy group structure of a small group at a higher level. Compared with the existing fine energy group, the calculation efficiency can be improved under the premise of maintaining the approximate calculation accuracy, which helps to reduce the amount of calculation required for the optimization design of the core layout plan and the refueling plan.

[0023] 2. The method of the present invention uses the open source optimization toolbox DAKOTA, and its built-in genetic algorithm module can be called through the scalable Python interface program to perform optimization calculations. Compared with some optimization designs at home and abroad that require rewriting the optimization program or modifying the external program code, the method of the present invention has good scalability and is very practical in the engineering design of nuclear reactors. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 is a flow chart of the method of the present invention.

[0025] Figure 2 The method of the present invention is implemented in the neutronics calculation program SARAX, and the k eff , k at the end of burnup eff Flow chart of optimizing the calculation of the minority energy group structure with the total isotope yield at the end of burnup as the target parameters. DETAILED DESCRIPTION

[0026] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0027] The present invention is a method for optimizing the energy group structure of a small group for research reactor physics calculations. The method provides a genetic algorithm module built into the open source optimization toolbox DAKOTA, provides a gene number to generate different energy group structures of a small group, and then provides the gene number to a neutronics calculation program for calculation to obtain the core physics calculation results. The target parameters are extracted from the core physics calculation results to make a fitness function. The optimized energy group structure is obtained through iterative calculations between the gene number and the fitness function value, so as to achieve the purpose of balancing the calculation accuracy and efficiency of large-scale core calculation problems. The specific steps are as follows:

[0028] Step 1: The genetic algorithm optimization module built into DAKOTA is coupled with the neutronics calculation program SARAX through the Python interface program. The gene number generated by DAKOTA, that is, a randomly generated natural number, is converted into an array through permutations and combinations. The interval width of the minority energy group is determined based on the array, thereby dividing the 1968 groups into corresponding minority energy group structures.

[0029] Step 2: Call the component calculation program TULIP of SARAX, read the small group energy group structure obtained in step 1 to perform component calculation, obtain the small group cross section, shape factor and other component small group parameters, and provide them to the subsequent core calculation program LAVENDER.

[0030] Step 3: Call the core calculation program LAVENDER of SARAX to perform core calculation based on the component group parameters obtained in step 2 to obtain the effective proliferation coefficient k of the thermal neutron research reactor eff , power distribution, flux distribution, burnup distribution, and total isotope production are the results of core physics calculations.

[0031] Step 4: Extract the effective multiplication factor k of the thermal neutron research reactor from the core physics calculation results obtained in step 3 eff , and the total isotope yield as the target parameters to make the fitness function, which is fed back to DAKOTA for optimization calculation. For example, the optimization calculation of the 8-group minority energy group structure is carried out with the initial k eff , k at the end of burnup eff , the total isotope yield at the end of burnup is taken as the target parameter, and the result under 32 groups is taken as the reference value. The fitness function expression is shown in formula (1):

[0032]

[0033] Where:

[0034] fitness—fitness function value

[0035] w1, w2, w3—weights of the three target parameters;

[0036] k eff,8g,initial , k eff,8g,end —Effective proliferation coefficients at the beginning and end of core burnup under 8 groups;

[0037] k eff,32g,initial , k eff,32g,end —Effective proliferation coefficients at the beginning and end of core burnup under 32 groups;

[0038] production 8g,end 、production 32g,end —Total isotope production at the end of core burnup under 8 and 32 groups.

[0039] Step 5: DAKOTA performs genetic algorithm operations such as selection, crossover, and mutation according to the fitness function fed back in step 4, and then repeats the calculation process from step 1 to step 4 from the number of genes to the fitness function, iteratively generates a population with better genes, and according to the principle of survival of the fittest, genetic evolution from generation to generation produces better approximate solutions, approaches the global optimal solution, and finally converges to the global approximate optimal solution, that is, searches for the gene that minimizes the fitness function value and obtains the optimized small group energy group structure.

[0040] The neutronics calculation program used in step 1 uses a two-step method (component calculation first, then core calculation) to perform core calculations, such as NECP-SARAX developed by Xi'an Jiaotong University. A key step in simulating biological evolution in genetic algorithms is the decoding relationship between genes and design parameters. In this optimization problem, it is how to translate a gene number (i.e., a natural number) into different small group energy group structures. For example, the number of genes is converted into 8 small group energy group structures through the Python interface program, and the idea of ​​permutation and combination is adopted. In order to reduce the upper limit of the number of genes while ensuring a certain calculation accuracy, the fine 1968 groups are divided into 41 groups, each with 48 energy group intervals, and then 8 groups are selected from the 41 groups. At this time, the total number of energy group structures that can be selected is C(40,7)=18643560. These 18643560 minority energy group structures and their corresponding numbers are stored in an external file, so that the interface program can read the corresponding minority energy group structure from the external file when reading the gene number, and provide it to the subsequent neutronics calculation program for calculation. When the optimal gene number is searched through the genetic algorithm, it can also be converted into the optimal minority energy group structure.

[0041] Step 2 and step 3 are the processes in which the externally coupled neutronics calculation program performs core transport calculations based on the minority energy group structure obtained in step 1, and ultimately obtains core physics calculation results such as the effective multiplication coefficient keff, power distribution, flux distribution, burnup distribution, and total isotope production of the thermal neutron research reactor.

[0042] In step 4, the effective proliferation coefficient keff and the total isotope yield of the nuclear reactor are extracted from the core physics calculation results obtained in step 3 through the Python interface program as target parameters to make a fitness function, which is fed back to DAKOTA to convert the multi-objective optimization problem into a single-objective optimization problem.

[0043] After the method of the present invention was implemented in the neutronics calculation program SARAX developed by Xi'an Jiaotong University, the target parameter was selected as the k eff , k at the end of burnup eff , the total isotope production at the end of burnup, the specific flow chart is as follows Figure 2 shown.

[0044] After the method of the present invention is implemented in the neutronics calculation program SARAX, the 8-group small-group energy group structure optimization calculation is performed on the simulated core loading problem of the China Engineering Test Reactor HFETR 81-1S series. eff , k at the end of burnup eff , and the total isotope yield at the end of burnup are taken as target parameters, and the calculation results under different energy group structures are compared. The results under the 32-group energy group structure are taken as the reference solution, as shown in Table 1. The comparison of the few-group energy group structure before and after optimization is shown in Table 2.

[0045] Table 1 Comparison of optimized 8-group energy group daily standard parameters with 32-group reference solution

[0046]

[0047] Table 2 Comparison of 8-group energy group structures

[0048]

[0049] As can be seen from Table 1, compared with the original 8-group energy group structure, the optimized 8-group energy group structure has a closer calculation result to the 32-group energy group structure: the deviation of the effective multiplication coefficient at the beginning and end of burnup is reduced from the original -113pcm and -74pcm to -12pcm and -50pcm respectively, and the relative deviation of the total isotope yield at the end of burnup is changed from the original -3.4% to 0.828%. As can be seen from Table 2, compared with the original 8-group energy group structure, the overall energy of the optimized 8-group energy group structure is slightly reduced. This is because HFETR belongs to a thermal neutron research reactor, and the core energy spectrum is more inclined to the thermal energy zone. Therefore, the calculation accuracy of the optimized 8-group energy group structure is higher. The optimized 8-group energy group structure can be used to optimize the design of the core layout and refueling scheme of the HFETR 81-1S series simulation core loading problem, and improve the calculation efficiency.

Claims

1. A method for optimizing the structure of a small number of energy groups for research reactor physics calculations. Features: The following steps are involved: Step 1: The genetic algorithm optimization module built into DAKOTA is coupled with the neutronics calculation program SARAX through the Python interface program. The gene number generated by DAKOTA, that is, a randomly generated natural number, is converted into an array through permutations and combinations. The interval width of the minority energy group is determined based on the array, so that the 1968 groups are divided into the corresponding minority energy group structure; Step 2: Call the component calculation program TULIP of SARAX, read the small group energy group structure obtained in step 1 to perform component calculation, obtain the small group cross section, shape factor and other component small group parameters, and provide them to the subsequent core calculation program LAVENDER; Step 3: Call the core calculation program LAVENDER of SARAX to perform core calculation based on the component group parameters obtained in step 2 to obtain the effective proliferation coefficient k of the thermal neutron research reactor eff , power distribution, flux distribution, burnup distribution, total isotope production and other core physics calculation results; Step 4: Extract the effective multiplication factor k of the thermal neutron research reactor from the core physics calculation results obtained in step 3 eff , the total isotope yield is used as the target parameter to make the fitness function, which is fed back to DAKOTA for optimization calculation; the fitness function expression is shown in formula (1): Where: Fitness—fitness function value; w 1 、w 2 、w 3 —weights of the three objective parameters; k eff,calc,initial , k eff,calc,end —The effective proliferation coefficients at the beginning and end of core burnup are calculated based on the small group energy group structure; k eff,fine,initial , k eff,fine,end —The effective multiplication coefficients at the beginning and end of core burnup are calculated based on the refined energy group structure; production calc,end 、production fine,end —Total isotope production at the end of core burnup based on the few-group energy group structure and the fine energy group structure; Step 5: DAKOTA performs selection, crossover, and mutation genetic algorithm operations based on the fitness function fed back in step 4, and then repeats the calculation process from step 1 to step 4 from the number of genes to the fitness function, iteratively generates a population with better genes, and according to the principle of survival of the fittest, genetic evolution from generation to generation produces better approximate solutions, approaches the global optimal solution, and finally converges to the global approximate optimal solution, that is, searches for the gene that minimizes the fitness function value, and obtains the optimized small group energy group structure.

Citation Information

Patent Citations

  • Energy group structure optimization method suitable for pressurized water reactor Pin-by-pin calculation

    CN110826205A

  • Wide group shielding database energy group structure optimization method based on particle swarm algorithm

    CN114356894A

  • Improved subgrid method for processing environmental effect in pressurized water reactor core physical calculation

    CN116562035A

  • Method for generating stack small-group cross section energy group structure for intermediate energy spectrum neutronics analysis

    CN117150829A

  • Method for obtaining ratio of radiation shielding material based on genetic algorithm and computing device

    CN118335217A