Nuclear reactor design optimization method, device, computer equipment and medium

Abstract

The present application relates to a method, device, computer equipment, storage medium and computer program product for optimizing a nuclear reactor design scheme, wherein the method comprises: obtaining an overall machine learning model of a nuclear reactor; searching for a nuclear reactor design scheme using an optimization algorithm based on the overall machine learning model of the nuclear reactor to obtain a search result of a nuclear reactor design scheme; performing verification regression on the search result of the nuclear reactor design scheme; and generating a nuclear reactor design optimization scheme based on the verification regression result. In the whole process, a diverse set of nuclear reactor design schemes are generated through a machine learning model and an optimization algorithm is used to search for schemes and then perform verification regression, so that a nuclear reactor design optimization scheme can be efficiently generated while ensuring that the scheme is accurately constructed.

Description

Technical Field

[0001] The present application relates to the field of big data technology, and in particular to a method, device, computer equipment, storage medium and computer program product for optimizing a nuclear reactor design scheme. Background Art

[0002] The generation of nuclear reactor design plans is a prerequisite and an important link in the engineering construction and safe operation of nuclear reactors.

[0003] The design of nuclear reactors involves many disciplines, complex physical phenomena, and highly nonlinear problems. Nuclear reactor design will include multiple disciplines such as physics, shielding, fuel, thermal engineering, and safety. The preliminary nuclear reactor plan selects reasonable overall parameters according to the application scenario, and then physics, thermal engineering, safety analysis, and equipment mechanics carry out relevant designs in turn. The multi-disciplinary design iteration cycle in the nuclear reactor design process is long, inefficient, and not very accurate. Documents and design parameters need to be transferred between disciplines. This process requires a lot of human resources. It is often encountered that when downstream disciplines find that the design parameters are unreasonable, the entire plan will be redesigned and finalized from the upstream discipline. The reason is that the upstream discipline cannot directly consider the impact of the downstream discipline during the design and evaluation process. In the process of transmitting some calculation data, errors and errors occur in the processing and interface of professionals. All of the above contents lead to a long actual design cycle and low efficiency for nuclear reactors.

[0004] Therefore, there is an urgent need for an efficient nuclear reactor design optimization solution. Summary of the invention

[0005] Based on this, it is necessary to provide an efficient nuclear reactor design optimization method, device, computer equipment, computer readable storage medium and computer program product to address the above technical problems.

[0006] In a first aspect, the present application provides a method for optimizing a nuclear reactor design. The method comprises:

[0007] Get the overall machine learning model of the nuclear reactor;

[0008] According to the nuclear reactor overall machine learning model, an optimization algorithm is used to search for nuclear reactor design solutions to obtain nuclear reactor design solution search results;

[0009] Performing validation regression on the search results of the nuclear reactor design scheme;

[0010] Generate a nuclear reactor design optimization plan based on the validation regression results.

[0011] In one embodiment, obtaining the overall machine learning model of the nuclear reactor includes:

[0012] Obtaining nuclear reactor design parameters;

[0013] Performing an automatic process analysis of the nuclear reactor design according to the nuclear reactor design scheme parameters to obtain a nuclear reactor design process analysis result;

[0014] Generating a nuclear reactor design solution database according to the nuclear reactor design process analysis results;

[0015] An initial supervised learning model is trained based on the nuclear reactor design database to construct an overall nuclear reactor machine learning model.

[0016] In one embodiment, obtaining the nuclear reactor design parameters includes:

[0017] Obtain a table of nuclear reactor design parameters;

[0018] The design parameters in the nuclear reactor design parameter table are independently transformed to obtain nuclear reactor design scheme parameters.

[0019] In one embodiment, the nuclear reactor design scheme search is performed using an optimization algorithm based on the overall machine learning model of the nuclear reactor, and the search results of the nuclear reactor design schemes include:

[0020] Generate multiple initial design schemes for nuclear reactors based on the overall machine learning model of the nuclear reactor;

[0021] Determining a nuclear reactor design goal, a design method, and a design criterion, and establishing an optimization mathematical model based on the nuclear reactor design goal, the design method, and the design criterion, wherein the optimization mathematical model is used to determine optimization variables, an optimization objective function, and optimization constraints;

[0022] According to the optimization mathematical model, a composite optimization algorithm or a multi-objective optimization algorithm is used to solve the parameterized optimization problem of the entire nuclear reactor for the multiple nuclear reactor initial design schemes to obtain a search result of the nuclear reactor design scheme.

[0023] In one embodiment, performing validation regression on the search results of the nuclear reactor design solution includes:

[0024] Performing calculation analysis on the search results of the nuclear reactor design scheme using the nuclear reactor automatic process design to obtain calculation analysis results;

[0025] Iteratively correcting the nuclear reactor overall machine learning model according to the calculation and analysis results to obtain a corrected nuclear reactor overall machine learning model;

[0026] When the preset iterative correction stop condition is met, the latest corrected nuclear reactor overall machine learning model is output;

[0027] The nuclear reactor design scheme corresponding to the latest revised nuclear reactor overall machine learning model is verified.

[0028] In one embodiment, the above-mentioned nuclear reactor design optimization method further includes:

[0029] If the validation regression result characterization is passed, then the nuclear reactor design optimization scheme corresponding to the validation regression is obtained;

[0030] If the verification regression result characterization fails, the method returns to the step of searching for nuclear reactor design schemes based on the overall machine learning model of the nuclear reactor using an optimization algorithm to obtain the search results of the nuclear reactor design schemes.

[0031] In a second aspect, the present application also provides a nuclear reactor design optimization device. The device comprises:

[0032] A model acquisition module is used to obtain the overall machine learning model of the nuclear reactor;

[0033] A search module is used to search for nuclear reactor design solutions using an optimization algorithm based on the overall machine learning model of the nuclear reactor to obtain search results for the nuclear reactor design solutions;

[0034] A verification module, used for performing verification regression on the search results of the nuclear reactor design scheme;

[0035] The scheme generation module is used to generate a nuclear reactor design optimization scheme based on the verification regression results.

[0036] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0037] Get the overall machine learning model of the nuclear reactor;

[0038] According to the nuclear reactor overall machine learning model, an optimization algorithm is used to search for nuclear reactor design solutions to obtain nuclear reactor design solution search results;

[0039] Performing validation regression on the search results of the nuclear reactor design scheme;

[0040] Generate a nuclear reactor design optimization plan based on the validation regression results.

[0041] In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0042] Get the overall machine learning model of the nuclear reactor;

[0043] According to the nuclear reactor overall machine learning model, an optimization algorithm is used to search for nuclear reactor design solutions to obtain nuclear reactor design solution search results;

[0044] Performing validation regression on the search results of the nuclear reactor design scheme;

[0045] Generate a nuclear reactor design optimization plan based on the validation regression results.

[0046] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0047] Get the overall machine learning model of the nuclear reactor;

[0048] According to the nuclear reactor overall machine learning model, an optimization algorithm is used to search for nuclear reactor design solutions to obtain nuclear reactor design solution search results;

[0049] Performing validation regression on the search results of the nuclear reactor design scheme;

[0050] Generate a nuclear reactor design optimization plan based on the validation regression results.

[0051] The above-mentioned nuclear reactor design scheme optimization method, device, computer equipment, storage medium and computer program product obtain the overall machine learning model of the nuclear reactor; according to the overall machine learning model of the nuclear reactor, use the optimization algorithm to search for the nuclear reactor design scheme to obtain the search results of the nuclear reactor design scheme; perform verification regression on the search results of the nuclear reactor design scheme; and generate the nuclear reactor design optimization scheme according to the verification regression results. In the whole process, a variety of nuclear reactor design schemes are generated through the machine learning model and the optimization algorithm is used to search for the scheme and then perform verification regression, so that the nuclear reactor design optimization scheme can be efficiently generated under the premise of ensuring the accuracy of the scheme construction. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 An application environment diagram of a nuclear reactor design optimization method in one embodiment;

[0053] Figure 2 A schematic diagram of a process flow of a method for optimizing a nuclear reactor design scheme in one embodiment;

[0054] Figure 3 is a schematic diagram of a sub-process of S200 in one embodiment;

[0055] Figure 4is a schematic diagram of a sub-process of S400 in one embodiment;

[0056] Figure 5 It is a schematic diagram of the optimization algorithm flow of NSGA-II;

[0057] Figure 6 is a schematic diagram of a sub-process of S600 in an embodiment;

[0058] Figure 7 A schematic diagram of a regression verification process in an embodiment;

[0059] Figure 8 A structural block diagram of a nuclear reactor design optimization device in one embodiment;

[0060] Fig. 9 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0062] The nuclear reactor design optimization method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. The terminal 102 sends a nuclear reactor design optimization request to the server 104, and the server 104 obtains the overall machine learning model of the nuclear reactor; according to the overall machine learning model of the nuclear reactor, the optimization algorithm is used to search for the nuclear reactor design scheme to obtain the search results of the nuclear reactor design scheme; the search results of the nuclear reactor design scheme are verified and regressed; and the nuclear reactor design optimization scheme is generated according to the verification regression results. If not necessary, the server 104 can also feed back the generated nuclear reactor design optimization scheme to the terminal 102. Among them, the terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server 104 can be implemented with an independent server or a server cluster consisting of multiple servers.

[0063] In one embodiment, Figure 2 As shown, a method for optimizing the design of a nuclear reactor is provided. Figure 1 Taking the server 104 in the example as an example, the following steps are included:

[0064] S200: Obtain the overall machine learning model of the nuclear reactor.

[0065] The overall machine learning model of a nuclear reactor refers to a machine learning model for the entire nuclear reactor, which may specifically include a nuclear reactor physical model, a nuclear reactor thermal model, and a nuclear reactor safety analysis and assessment model. The overall machine learning model of a nuclear reactor can be obtained by temporary generation, or a pre-built model can be used directly. If a temporary construction method is used to obtain it, the design parameters of the nuclear reactor must be obtained first. The design parameters include seed physical parameters, thermal hydraulic parameters, fuel parameters, system parameters, and safety analysis result parameters. Based on these, the overall machine learning model of the nuclear reactor is trained using supervised learning.

[0066] S400: Based on the overall machine learning model of the nuclear reactor, an optimization algorithm is used to search for nuclear reactor design solutions to obtain nuclear reactor design solution search results.

[0067] According to the overall machine learning model of the nuclear reactor, various nuclear reactor design schemes can be generated. Among these design schemes, there are optimized design schemes. Here, an optimization algorithm is used to search for nuclear reactor design schemes to obtain the search results of nuclear reactor design schemes. Specifically, the optimization algorithm includes a composite multi-optimization method or a multi-objective optimization algorithm, through which the optimized design scheme is searched, and the searched results can be extracted.

[0068] S600: Perform verification regression on the search results of nuclear reactor design solutions.

[0069] There are optimized nuclear reactor design schemes in the search results of nuclear reactor design schemes, and these optimized nuclear reactor design schemes are verified to verify whether the design schemes are accurate. If the design schemes are accurate, the overall machine learning model of the nuclear reactor is corrected by regression according to the design schemes. If the verified design scheme is inaccurate, return to S400 to restart a new round of nuclear reactor design scheme search.

[0070] S800: Generate a nuclear reactor design optimization plan based on the validation regression results.

[0071] As mentioned above, if the verification regression result characterization passes, the overall machine learning model of the nuclear reactor is corrected according to the design scheme regression, and the final nuclear reactor design optimization scheme is generated according to the corrected overall machine learning model of the nuclear reactor; if the verification regression result characterization fails, return to S400 to continue to start a new nuclear reactor design scheme search.

[0072] The above-mentioned nuclear reactor design scheme optimization method obtains the overall machine learning model of the nuclear reactor; according to the overall machine learning model of the nuclear reactor, an optimization algorithm is used to search for nuclear reactor design schemes to obtain the search results of nuclear reactor design schemes; the search results of nuclear reactor design schemes are verified and regressed; and a nuclear reactor design optimization scheme is generated according to the verification and regression results. In the whole process, a variety of nuclear reactor design schemes are generated through machine learning models and optimization algorithms are used to search for schemes and then verification and regression are performed, so that the nuclear reactor design optimization scheme can be efficiently generated under the premise of ensuring the accuracy of the scheme construction.

[0073] like Figure 3 As shown, in one embodiment, S200 includes:

[0074] S220: Obtain nuclear reactor design parameters.

[0075] A nuclear reactor parameter table is established here. According to different types of data, it can be divided into: design input parameters and response parameters. Among them, the design input parameters include design parameters such as geometry, structure, material, and upstream professional calculation and analysis results. Furthermore, some initial parameters are not completely independent. If the input parameters are not independent, the feasible domain of the generated design scheme will become smaller. Therefore, preferably, a nuclear reactor design parameter table can be obtained; the design parameters in the nuclear reactor design parameter table are independently transformed to obtain the nuclear reactor design scheme parameters. That is, in the process of design input parameterization, the design parameters are independently transformed, which can not only reduce the dimension of the design parameters, but also increase the feasible domain of the design calculation. In addition, in order to better set the range of each variable, the design variables need to be preprocessed to reduce the correlation between the variables, such as the fuel element diameter, the inner diameter and the outer diameter of the fuel rod, which are transformed into the fuel rod diameter, the air gap diameter and the cladding thickness.

[0076] S240: Performing automatic process analysis of nuclear reactor design according to the nuclear reactor design scheme parameters to obtain nuclear reactor design process analysis results.

[0077] In the process of nuclear reactor design, it is necessary to consider core neutron physics calculation, core thermal hydraulics, structural design, and thermal system design. Therefore, in addition to objective functions such as volume, weight, and economic indicators, it is also necessary to consider its safety, such as reactivity keff, life, reactivity coefficient, fuel temperature, coolant flow rate, temperature, and other neutron physics design criteria, thermal hydraulic criteria, and structural design criteria. First, a parameterized design scheme is randomly generated, and the input cards of each design tool are automatically generated using the parameterized scheme; then, according to the traditional design ideas, the various design tools are linked using the external coupling method; and then they are automatically run. In particular, in this process, multi-threaded concurrent or CPU multi-core parallel methods can be used to accelerate the calculation and provide calculation efficiency. Finally, the design results are proposed.

[0078] S260: Generate a nuclear reactor design solution database based on the nuclear reactor design process analysis results.

[0079] An extraction tool is established to match the design scheme and the design analysis results, and finally a nuclear reactor design scheme evaluation database is formed. First, the parametric design scheme is extracted, and then the analysis results are extracted, including the target value and each limit value. The data contained in the actual database is shown in Table 1.

[0080] Table 1 shows an example of the data contained in the database

[0081]

[0082]

[0083] S280: Train the initial supervised learning model based on the nuclear reactor design database and build the overall machine learning model of the nuclear reactor.

[0084] This step uses machine learning to establish a nuclear reactor analysis and evaluation model. Machine learning models include supervised learning models such as generalized linear models, vector machine models, stochastic gradient descent models, Bayesian models, decision tree models, and random forest models. First, read the database produced by step S260, and then design variable preprocessing, including standardization, normalization, nonlinear transformation, category feature encoding, etc.; secondly, establish corresponding machine learning models, such as keff prediction model, life prediction model, maximum temperature prediction model; then, evaluate the model and perform uncertainty analysis to obtain the prediction error. Through multiple generations, a reasonable nuclear reactor analysis and evaluation model is finally saved. For example, design parameter 1, UO2 number 1, UN number 2, MOX number; design parameters can be obtained through Calculated, where a i is the jth parameter of the i-th group, is the normalized value of the jth parameter of the i-th group, is the maximum value of the i-th group of parameters, is the minimum value of the i-th group of parameters.

[0085] like Figure 4 As shown, in one embodiment, S400 includes:

[0086] S420: Generate multiple initial design plans for nuclear reactors based on the overall machine learning model of the nuclear reactor.

[0087] The overall machine learning model of a nuclear reactor is used to simulate the design parameters involved in the overall design process of a nuclear reactor, and multiple initial design schemes for nuclear reactors are generated by adjusting the design parameters.

[0088] S440: Determine the nuclear reactor design objectives, design methods and design criteria, and establish an optimization mathematical model based on the nuclear reactor design objectives, design methods and design criteria. The optimization mathematical model is used to determine optimization variables, determine optimization objective functions and determine optimization constraints.

[0089] The server responds to the design requirements of the designer and determines the optimization design problem, including (1) design goals; (2) design methods; and (3) design criteria. The design goal refers to the specific application scenarios of the nuclear reactor design, such as power supply for aerospace, ships, deep sea, polar regions, islands, etc. According to the design goal, the overall main parameters of the nuclear reactor, such as power and service life of the nuclear reactor, can be determined. The design method refers to the calculation and analysis method of the nuclear reactor as a whole, equipment or system. The analysis content includes physics, fuel, thermal engineering, system, safety analysis, etc. Physical analysis can be performed using deterministic software or Monte Carlo software for calculation and analysis. Thermal hydraulics can be performed using sub-channels, etc. Design criteria include neutron physics design criteria, thermal engineering design criteria, structural design criteria, etc. Among them, typical neutron physics design criteria include reactivity temperature coefficient, maximum controllable reactivity introduction rate, shutdown margin, burnup, etc. Thermal engineering design criteria include fuel temperature limit, thermal engineering limit of cladding, and coolant temperature limit. The core structure design criteria mainly meet the existing relevant specifications. For example, in order to optimize the economic efficiency of the nuclear reactor, uranium loading can be used as an optimization target.

[0090] Establishing a mathematical model includes (1) selecting decision variables; 2) determining the objective function; and 3) clarifying design requirements, setting constraints and input setting ranges. Decision variables, also known as optimization variables, refer to independent parameters that can be adjusted and selected in the target optimization design problem. The nuclear reactor design problem includes n decision variables, indicating that the nuclear design problem is an n-dimensional optimization problem. The types of decision variables include continuous and discrete types. These design variables are derived from the design parameters of the nuclear reactor, and each variable is relatively independent. The objective function refers to the design goal of the nuclear reactor, such as the best nuclear reactor layout, minimum volume, lightest volume, and most economical. The nuclear reactor optimization process is essentially the process of finding decision variables (design parameters) that can make the objective function reach the optimal value. The objective function is a scalar function of the decision variables, and the optimal value of the objective function is generally expressed in the form of a minimum (or maximum) value. Constraints are conditions that restrict the values ​​of decision variables. In the optimization process, constraints are also functions of decision variables and functions of nuclear reactor parameters. According to the nature of the constraints, they can be divided into performance constraints and side constraints. Performance constraints are mainly to meet performance indicators and design requirements, such as safety limits of nuclear reactors, design criteria, geometric limits, etc. Side constraints include upper and lower limits given to decision variables. The general model is as follows:

[0091] Decision variables or optimization variables are usually expressed as:

[0092] x=[x 1 ,x 2 ,…,x n ] (1)

[0093] The objective function is:

[0094] minf(x)=minf(x 1 ,x 2 ,…,x n ) (2)

[0095] The multi-objective function can be expressed as:

[0096] minF(x)=min[f 1 (x),f 2 (x),…,f m (x)] (3)

[0097] For multi-objective optimization problems, the ideal situation is to obtain each optimization sub-objective function f i(x) reaches the minimum value at the same time. However, in the actual calculation process, it is difficult to meet such a situation. In order to deal with conflicting design goals, the "preference solution" method and the "non-dominated" method can be solved. The "preference solution" method is to sort and combine each objective function by adopting the value judgment of each objective function. This type of method relies on the design experience and subjective opinions of the designer or individual. The preference solution method can be divided into: weight and method, dictionary order method, compromise method, etc. The non-dominated solution method minimizes the sub-objective function without weakening each sub-objective function. Non-dominated solution methods include vector evaluation genetic method, non-dominated solution sorting algorithm, distance algorithm and immune memory clone constraint algorithm. The restriction function can be expressed as,

[0098] g i (x)≤0,i=1,2…,q (4)

[0099] For example, F(m fuel )=m fuel +α 1 (max(G 1 ,0))+…+α n (max(G n ,0))

[0100] Among them, m fuel is the uranium loading, G 1 ……G n is the constraint condition, α 1 is the maximum coefficient of the constraint. Some of the constraints are shown in Table 2 below.

[0101] Table 2 shows the parameter values ​​corresponding to some restriction conditions.

[0102]

[0103] S460: According to the optimization mathematical model, a composite optimization algorithm or a multi-objective optimization algorithm is used to solve the parameterized optimization problem of the entire nuclear reactor for multiple initial design schemes of nuclear reactors, and a search result of the nuclear reactor design scheme is obtained.

[0104] The optimization algorithm includes a composite optimization algorithm or a multi-objective optimization algorithm. To explain the optimization calculation process in detail, the optimization algorithm of the target algorithm NSGA-II is used as an example to solve the full-stack parameterized optimization problem. Figure 5 As shown, the entire optimization algorithm includes the following steps:

[0105] 1. Confirm the solution setting parameters, objective function and constraints. The setting parameters for NSGA-II include at least the number of initial solutions, iteration number, crossover probability, mutation probability, crossover method, mutation method, fitness function options, etc. For example, the initial value of the individual crossover probability is set to 0.9, the mutation probability is 0.03, the population size ranges from 10 to 500, and the evolutionary number ranges from 50 to 1000. Then, set the population size and evolutionary number to the optimal values ​​obtained previously, the crossover probability ranges from 0.1 to 0.9, and the mutation probability ranges from 0.01 to 0.8. Now set the crossover probability to 0.9 and the mutation probability to 0.03.

[0106] 2. Generate N initialization schemes P0 in shorthand, then perform non-dominated sorting on them, initialize the rank values ​​of all individuals, and purify the algebra n;

[0107] 3. Randomly select individuals from Pn for genetic operation, then use normal distribution strategy for crossover operation, and use adaptive adjustment mutation method to make mutation;

[0108] 4. The previous step generates the offspring plan Xn, merge Pn and Xn to generate Qn and calculate the objective function value, perform a fast non-dominated sorting operation on Qn, calculate the crowding degree and crowding distance of the individuals in Qn, and select N individuals to form a new population Pn+1;

[0109] 5. When the calculation algebra reaches the maximum purification algebra and the termination condition is met, the loop ends, otherwise go to step 2.

[0110] 6. Obtain the Pareto optimal solution set.

[0111] Furthermore, after obtaining the search results of the nuclear reactor design scheme, the optimization scheme and results can be extracted. Specifically, the results can be saved by serialization operation. Common serialization operations include marshal, pickle, json, and shelve. Among them, marshal, pickle, and json can use dump and dumps commands to save data, and load and loads can be used to read the scheme. This method has fast loading speed and convenient reading, and can save and read the calculation content of each step in the optimization calculation process.

[0112] like Figure 6 As shown, in one embodiment, S600 includes:

[0113] S620: Performing calculation analysis on the nuclear reactor design solution search results using the nuclear reactor automatic process design to obtain calculation analysis results.

[0114] Here we first return to the automatic process design. The automatic process design of the nuclear reactor is used to perform calculation analysis on the multiple optimized nuclear reactor design schemes contained in the search results of the nuclear reactor design scheme to obtain the calculation analysis results. Specifically, the calculation analysis can be performed in a similar automatic process design method as in the above S240. This process also considers the core neutron physics calculation, core thermal hydraulic calculation, structural design, thermal system design, etc., and can obtain various indicators of the objective function, such as volume, weight, economic indicators, etc. As well as the various parameters in the analysis calculation, such as reactivity coefficient, fuel temperature, coolant flow rate, maximum coolant temperature, maximum fuel temperature, etc. This process can also use multi-threaded concurrency or multi-core parallel methods to accelerate the calculation.

[0115] S640: Iteratively modify the overall machine learning model of the nuclear reactor according to the calculation and analysis results to obtain a modified overall machine learning model of the nuclear reactor.

[0116] The overall machine learning model of the nuclear reactor is iteratively corrected according to the calculation and analysis results generated by S620. Specifically, the database corresponding to the overall machine learning model of the nuclear reactor is updated. On the basis of the database update, through the preprocessing of nuclear reactor design parameters, such as standardization, normalization, nonlinear transcoding and other technologies, machine learning prediction models are established, such as generalized linear models, random forest models, stochastic gradient descent models, etc., to evaluate and perform uncertainty quantification analysis on the models.

[0117] S660: When the preset iterative correction stop condition is met, the latest corrected nuclear reactor overall machine learning model is output.

[0118] Determine whether the output conditions are met. This step mainly determines whether to output the final design solution. Common judgment criteria include: the total iterations have reached the maximum number limit, and the design solution parameters have output the same calculation results in the last n times, and n is 20 times by default. If the judgment result is yes, proceed to the next step; if the judgment result is no, repeat the operation of step S640 until the judgment result is yes.

[0119] S680: Verify the nuclear reactor design corresponding to the latest revised nuclear reactor overall machine learning model.

[0120] The latest revised nuclear reactor overall machine learning model is verified again to verify the accuracy of the design scheme. Ensure that an accurate nuclear reactor design scheme can be obtained in the end. Furthermore, the optimization scheme is finally generated and the nuclear reactor engineering design scheme database is completed. The database includes nuclear reactor design schemes, various professional analysis and calculation results, including physical analysis results, thermal and hydraulic analysis results, as well as objective function values, constraint function values, corresponding machine learning model options and models, and optimization process content. It should be noted here that the scheme parameters and analysis results can be expanded to reduce repeated calculations. As the breadth and depth of the design expand, the corresponding models and optimization schemes will also be more accurate.

[0121] For example, the multi-layer machine learning agent model uses the gradient boosting regression tree model to calculate the effective proliferation factor k eff , using the extreme random forest model for thermal analysis and calculation and The specific regression verification process is as follows: Figure 7 During the optimization process, the database is updated as shown in Table 3 below.

[0122] Table 3 is the data table of the updated database during the optimization process

[0123]

[0124] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0125] Based on the same inventive concept, the embodiment of the present application also provides a nuclear reactor design scheme optimization device for implementing the nuclear reactor design scheme optimization method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more nuclear reactor design scheme optimization device embodiments provided below can refer to the limitations of the nuclear reactor design scheme optimization method above, and will not be repeated here.

[0126] like Figure 8As shown, the present application also provides a nuclear reactor design optimization device. The device includes:

[0127] A model acquisition module 200 is used to acquire an overall machine learning model of a nuclear reactor;

[0128] A search module 400 is used to search for nuclear reactor design solutions using an optimization algorithm based on the overall machine learning model of the nuclear reactor, and obtain search results for the nuclear reactor design solutions;

[0129] A verification module 600 is used to perform verification regression on the search results of the nuclear reactor design scheme;

[0130] The solution generation module 800 is used to generate a nuclear reactor design optimization solution according to the verification regression results.

[0131] The above-mentioned nuclear reactor design scheme optimization device obtains the overall machine learning model of the nuclear reactor; uses the optimization algorithm to search for nuclear reactor design schemes based on the overall machine learning model of the nuclear reactor to obtain the search results of the nuclear reactor design scheme; performs verification regression on the search results of the nuclear reactor design scheme; and generates a nuclear reactor design optimization scheme based on the verification regression results. In the whole process, a variety of nuclear reactor design schemes are generated through machine learning models and optimization algorithms are used to search for schemes and then perform verification regression, which can efficiently generate nuclear reactor design optimization schemes while ensuring that the schemes are accurately constructed.

[0132] In one of the embodiments, the model acquisition module 200 is also used to obtain nuclear reactor design scheme parameters; perform automatic process analysis of nuclear reactor design based on the nuclear reactor design scheme parameters to obtain nuclear reactor design process analysis results; generate a nuclear reactor design scheme database based on the nuclear reactor design process analysis results; train an initial supervised learning model based on the nuclear reactor design scheme database to construct an overall nuclear reactor machine learning model.

[0133] In one embodiment, the model acquisition module 200 is also used to obtain a nuclear reactor design parameter table; the design parameters in the nuclear reactor design parameter table are independently transformed to obtain nuclear reactor design scheme parameters.

[0134] In one embodiment, the search module 400 is also used to generate multiple initial design schemes for nuclear reactors based on the overall machine learning model of the nuclear reactor; determine the design objectives, design methods and design criteria of the nuclear reactor, and establish an optimization mathematical model based on the design objectives, design methods and design criteria of the nuclear reactor, and the optimization mathematical model is used to determine the optimization variables, determine the optimization objective function and determine the optimization constraints; according to the optimization mathematical model, a composite optimization algorithm or a multi-objective optimization algorithm is used to solve the parameterized optimization problem of the entire nuclear reactor for multiple initial design schemes of nuclear reactors to obtain the search results of nuclear reactor design schemes.

[0135] In one of the embodiments, the verification module 600 is also used to perform computational analysis on the search results of nuclear reactor design schemes using the nuclear reactor automatic process design to obtain computational analysis results; iteratively correct the nuclear reactor overall machine learning model according to the computational analysis results to obtain a corrected nuclear reactor overall machine learning model; when the preset iterative correction stop condition is met, the latest corrected nuclear reactor overall machine learning model is output; and the nuclear reactor design scheme corresponding to the latest corrected nuclear reactor overall machine learning model is verified.

[0136] In one of the embodiments, the verification module 600 is also used to control the scheme generation module 800 to obtain the nuclear reactor design optimization scheme corresponding to the verification regression when the verification regression result characterization passes; if the verification regression result characterization fails, the control search module 400 re-executes the operation of searching for nuclear reactor design schemes based on the overall machine learning model of the nuclear reactor and using an optimization algorithm to obtain the search results of the nuclear reactor design schemes.

[0137] Each module in the above-mentioned nuclear reactor design optimization device can be implemented in whole or in part by software, hardware and a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0138] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Fig. 9As shown. The computer device includes a processor, a memory and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store basic parameter data of nuclear reactor design and historical nuclear reactor design scheme data. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for optimizing a nuclear reactor design scheme is implemented.

[0139] Those skilled in the art will understand that Fig. 9 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0140] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0141] Get the overall machine learning model of the nuclear reactor;

[0142] According to the overall machine learning model of the nuclear reactor, an optimization algorithm is used to search for nuclear reactor design solutions, and the search results of nuclear reactor design solutions are obtained;

[0143] Perform validation regression on the search results of nuclear reactor design solutions;

[0144] Generate a nuclear reactor design optimization plan based on the validation regression results.

[0145] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0146] Acquire nuclear reactor design scheme parameters; perform automatic process analysis of nuclear reactor design based on the nuclear reactor design scheme parameters to obtain nuclear reactor design process analysis results; generate a nuclear reactor design scheme database based on the nuclear reactor design process analysis results; train an initial supervised learning model based on the nuclear reactor design scheme database to construct an overall nuclear reactor machine learning model.

[0147] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0148] Obtain a nuclear reactor design parameter table; independently transform the design parameters in the nuclear reactor design parameter table to obtain nuclear reactor design scheme parameters.

[0149] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0150] According to the overall machine learning model of the nuclear reactor, multiple initial design schemes of the nuclear reactor are generated; the design objectives, design methods and design criteria of the nuclear reactor are determined, and an optimization mathematical model is established according to the design objectives, design methods and design criteria of the nuclear reactor, and the optimization mathematical model is used to determine the optimization variables, determine the optimization objective function and determine the optimization constraints; according to the optimization mathematical model, a composite optimization algorithm or a multi-objective optimization algorithm is used to solve the parameterized optimization problem of the whole nuclear reactor for multiple initial design schemes of the nuclear reactor, and the search results of the nuclear reactor design scheme are obtained.

[0151] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0152] The nuclear reactor automatic process design is used to perform computational analysis on the search results of the nuclear reactor design scheme to obtain computational analysis results; the nuclear reactor overall machine learning model is iteratively corrected according to the computational analysis results to obtain a corrected nuclear reactor overall machine learning model; when the preset iterative correction stop condition is met, the latest corrected nuclear reactor overall machine learning model is output; and the nuclear reactor design scheme corresponding to the latest corrected nuclear reactor overall machine learning model is verified.

[0153] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0154] If the verification regression result characterization passes, then the nuclear reactor design optimization scheme corresponding to the verification regression is obtained; if the verification regression result characterization fails, then the step of searching for nuclear reactor design schemes based on the overall machine learning model of the nuclear reactor and using an optimization algorithm is returned to obtain the search results of the nuclear reactor design schemes.

[0155] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0156] Get the overall machine learning model of the nuclear reactor;

[0157] According to the overall machine learning model of the nuclear reactor, an optimization algorithm is used to search for nuclear reactor design solutions, and the search results of nuclear reactor design solutions are obtained;

[0158] Perform validation regression on the search results of nuclear reactor design solutions;

[0159] Generate a nuclear reactor design optimization plan based on the validation regression results.

[0160] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0161] Acquire nuclear reactor design scheme parameters; perform automatic process analysis of nuclear reactor design based on the nuclear reactor design scheme parameters to obtain nuclear reactor design process analysis results; generate a nuclear reactor design scheme database based on the nuclear reactor design process analysis results; train an initial supervised learning model based on the nuclear reactor design scheme database to construct an overall nuclear reactor machine learning model.

[0162] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0163] Obtain a nuclear reactor design parameter table; independently transform the design parameters in the nuclear reactor design parameter table to obtain nuclear reactor design scheme parameters.

[0164] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0165] According to the overall machine learning model of the nuclear reactor, multiple initial design schemes of the nuclear reactor are generated; the design objectives, design methods and design criteria of the nuclear reactor are determined, and an optimization mathematical model is established according to the design objectives, design methods and design criteria of the nuclear reactor, and the optimization mathematical model is used to determine the optimization variables, determine the optimization objective function and determine the optimization constraints; according to the optimization mathematical model, a composite optimization algorithm or a multi-objective optimization algorithm is used to solve the parameterized optimization problem of the whole nuclear reactor for multiple initial design schemes of the nuclear reactor, and the search results of the nuclear reactor design scheme are obtained.

[0166] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0167] The nuclear reactor automatic process design is used to perform computational analysis on the search results of the nuclear reactor design scheme to obtain computational analysis results; the nuclear reactor overall machine learning model is iteratively corrected according to the computational analysis results to obtain a corrected nuclear reactor overall machine learning model; when the preset iterative correction stop condition is met, the latest corrected nuclear reactor overall machine learning model is output; and the nuclear reactor design scheme corresponding to the latest corrected nuclear reactor overall machine learning model is verified.

[0168] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0169] If the verification regression result characterization passes, then the nuclear reactor design optimization scheme corresponding to the verification regression is obtained; if the verification regression result characterization fails, then the step of searching for nuclear reactor design schemes based on the overall machine learning model of the nuclear reactor and using an optimization algorithm is returned to obtain the search results of the nuclear reactor design schemes.

[0170] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0171] Get the overall machine learning model of the nuclear reactor;

[0172] According to the overall machine learning model of the nuclear reactor, an optimization algorithm is used to search for nuclear reactor design solutions, and the search results of nuclear reactor design solutions are obtained;

[0173] Perform validation regression on the search results of nuclear reactor design solutions;

[0174] Generate a nuclear reactor design optimization plan based on the validation regression results.

[0175] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0176] Acquire nuclear reactor design scheme parameters; perform automatic process analysis of nuclear reactor design based on the nuclear reactor design scheme parameters to obtain nuclear reactor design process analysis results; generate a nuclear reactor design scheme database based on the nuclear reactor design process analysis results; train an initial supervised learning model based on the nuclear reactor design scheme database to construct an overall nuclear reactor machine learning model.

[0177] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0178] Obtain a nuclear reactor design parameter table; independently transform the design parameters in the nuclear reactor design parameter table to obtain nuclear reactor design scheme parameters.

[0179] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0180] According to the overall machine learning model of the nuclear reactor, multiple initial design schemes of the nuclear reactor are generated; the design objectives, design methods and design criteria of the nuclear reactor are determined, and an optimization mathematical model is established according to the design objectives, design methods and design criteria of the nuclear reactor, and the optimization mathematical model is used to determine the optimization variables, determine the optimization objective function and determine the optimization constraints; according to the optimization mathematical model, a composite optimization algorithm or a multi-objective optimization algorithm is used to solve the parameterized optimization problem of the whole nuclear reactor for multiple initial design schemes of the nuclear reactor, and the search results of the nuclear reactor design scheme are obtained.

[0181] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0182] The nuclear reactor automatic process design is used to perform computational analysis on the search results of the nuclear reactor design scheme to obtain computational analysis results; the nuclear reactor overall machine learning model is iteratively corrected according to the computational analysis results to obtain a corrected nuclear reactor overall machine learning model; when the preset iterative correction stop condition is met, the latest corrected nuclear reactor overall machine learning model is output; and the nuclear reactor design scheme corresponding to the latest corrected nuclear reactor overall machine learning model is verified.

[0183] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0184] If the verification regression result characterization passes, then the nuclear reactor design optimization scheme corresponding to the verification regression is obtained; if the verification regression result characterization fails, then the step of searching for nuclear reactor design schemes based on the overall machine learning model of the nuclear reactor and using an optimization algorithm is returned to obtain the search results of the nuclear reactor design schemes.

[0185] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or authorized by all parties.

[0186] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0187] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0188] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the present application. It should be noted that, for a person of ordinary skill in the art, several modifications and substitutions can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for optimizing a nuclear reactor design, characterized in that: The method comprises: Get the overall machine learning model of the nuclear reactor; According to the nuclear reactor overall machine learning model, an optimization algorithm is used to search for nuclear reactor design solutions to obtain nuclear reactor design solution search results; Performing validation regression on the search results of the nuclear reactor design scheme; Generate a nuclear reactor design optimization plan based on the validation regression results; The method of searching for nuclear reactor design solutions by using an optimization algorithm according to the overall machine learning model of the nuclear reactor to obtain the search results of the nuclear reactor design solutions includes: generating a plurality of initial nuclear reactor design solutions according to the overall machine learning model of the nuclear reactor; determining the design objectives, design methods and design criteria of the nuclear reactor, and establishing an optimization mathematical model according to the design objectives, design methods and design criteria of the nuclear reactor, wherein the optimization mathematical model is used to determine optimization variables, determine optimization objective functions and determine optimization constraints; and solving the parameterized optimization problem of the whole nuclear reactor by using a composite optimization algorithm or a multi-objective optimization algorithm for the multiple initial nuclear reactor design solutions according to the optimization mathematical model to obtain the search results of the nuclear reactor design solutions; The verification regression of the nuclear reactor design scheme search results includes: performing computational analysis on the nuclear reactor design scheme search results using nuclear reactor automatic process design to obtain computational analysis results; iteratively correcting the nuclear reactor overall machine learning model according to the computational analysis results to obtain a corrected nuclear reactor overall machine learning model; when a preset iterative correction stop condition is met, outputting the latest corrected nuclear reactor overall machine learning model; and verifying the nuclear reactor design scheme corresponding to the latest corrected nuclear reactor overall machine learning model.

2. The method according to claim 1, characterized in that The obtaining of the overall machine learning model of the nuclear reactor comprises: Obtaining nuclear reactor design parameters; Performing an automatic process analysis of the nuclear reactor design according to the nuclear reactor design scheme parameters to obtain a nuclear reactor design process analysis result; Generating a nuclear reactor design solution database according to the nuclear reactor design process analysis results; An initial supervised learning model is trained based on the nuclear reactor design database to construct an overall nuclear reactor machine learning model.

3. The method according to claim 2, characterized in that The obtaining of nuclear reactor design parameters comprises: Obtain a table of nuclear reactor design parameters; The design parameters in the nuclear reactor design parameter table are independently transformed to obtain nuclear reactor design scheme parameters.

4. The method according to claim 1, characterized in that Also includes: If the validation regression result characterization is passed, then the nuclear reactor design optimization scheme corresponding to the validation regression is obtained; If the verification regression result characterization fails, the step of returning to the step of searching for nuclear reactor design schemes based on the overall machine learning model of the nuclear reactor using an optimization algorithm to obtain the search results of the nuclear reactor design schemes.

5. A nuclear reactor design optimization device, characterized in that: The device comprises: A model acquisition module is used to obtain the overall machine learning model of the nuclear reactor; A search module is used to search for nuclear reactor design solutions using an optimization algorithm based on the overall machine learning model of the nuclear reactor to obtain search results for the nuclear reactor design solutions; A verification module, used for performing verification regression on the search results of the nuclear reactor design scheme; A scheme generation module is used to generate a nuclear reactor design optimization scheme based on the verification regression results; The search module is also used to generate a plurality of initial design schemes for nuclear reactors according to the overall machine learning model of the nuclear reactor; determine the design objectives, design methods and design criteria of the nuclear reactor, and establish an optimization mathematical model according to the design objectives, design methods and design criteria of the nuclear reactor, wherein the optimization mathematical model is used to determine optimization variables, determine optimization objective functions and determine optimization constraints; according to the optimization mathematical model, a composite optimization algorithm or a multi-objective optimization algorithm is used to solve the parameterized optimization problem of the whole nuclear reactor for the plurality of initial design schemes of the nuclear reactor to obtain the search results of the nuclear reactor design scheme; The verification module is also used to perform computational analysis on the search results of the nuclear reactor design scheme using the nuclear reactor automatic process design to obtain computational analysis results; iteratively correct the nuclear reactor overall machine learning model according to the computational analysis results to obtain a corrected nuclear reactor overall machine learning model; when a preset iterative correction stop condition is met, output the latest corrected nuclear reactor overall machine learning model; and verify the nuclear reactor design scheme corresponding to the latest corrected nuclear reactor overall machine learning model.

6. The device according to claim 5, characterized in that The model acquisition module is also used to obtain the parameters of the nuclear reactor design scheme; perform automatic process analysis of the nuclear reactor design according to the parameters of the nuclear reactor design scheme to obtain the results of the nuclear reactor design process analysis; and generate a nuclear reactor design scheme database according to the results of the nuclear reactor design process analysis; The initial supervised learning model is trained based on the nuclear reactor design database, and the overall machine learning model of the nuclear reactor is constructed.

7. The device according to claim 6, characterized in that The model acquisition module is also used to obtain a nuclear reactor design parameter table; the design parameters in the nuclear reactor design parameter table are independently transformed to obtain nuclear reactor design scheme parameters.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images