Construction method and device of irradiation damage prediction model, terminal equipment and computer program product

Abstract

The invention is suitable for the technical field of safety assessment of reactor pressure vessels of nuclear power plants, and provides a construction method and device of an irradiation damage prediction model, terminal equipment and a computer program product, and the method comprises the steps: obtaining an irradiation feature data set of a reactor pressure vessel; according to the irradiation characteristic data set, constructing an irradiation damage prediction model of the reactor pressure vessel; wherein the irradiation damage prediction model is used for acquiring and outputting irradiation damage prediction data of the reactor pressure vessel to be predicted. According to the embodiment of the invention, when the irradiation damage prediction model is constructed, the essential reason causing the irradiation damage of the material of the reactor pressure vessel and the characteristics of the generated irradiation loss are fully considered, and when the irradiation damage prediction model is used for predicting the irradiation damage of the reactor pressure vessel, the prediction accuracy can be effectively improved; and the safety and reliability of the nuclear power station can be ensured.

Description

Technical Field

[0001] The present application belongs to the technical field of safety assessment of reactor pressure vessels in nuclear power plants, and in particular, relates to a method, device, terminal equipment and computer program product for constructing a radiation damage prediction model. Background Art

[0002] Reactor Pressure Vessel (RPV) is one of the most critical equipment in nuclear power plants, and it is essential to ensure the safe operation of nuclear power plants. During the service of the reactor pressure vessel, irradiation by high-energy neutrons and other factors will cause certain damage to the material of the reactor pressure vessel, thereby reducing the overall performance of the reactor pressure vessel. Therefore, it is necessary to predict and evaluate the reactor pressure vessel during the operation of the nuclear power unit.

[0003] The existing prediction methods or models for radiation damage to reactor pressure vessels fail to consider the essential causes of radiation damage to materials of reactor pressure vessels and the characteristics of the resulting radiation losses, resulting in low accuracy of the prediction results. Summary of the invention

[0004] In view of this, the embodiments of the present application provide a method, apparatus, terminal device and computer program product for constructing a radiation damage prediction model to solve the problem of low accuracy of prediction results in existing methods or models for predicting radiation damage to reactor pressure vessels.

[0005] A first aspect of an embodiment of the present application provides a method for constructing a radiation damage prediction model, the method comprising:

[0006] Obtaining the irradiation characteristic data set of the reactor pressure vessel;

[0007] constructing a radiation damage prediction model for the reactor pressure vessel according to the radiation characteristic data set;

[0008] The radiation damage prediction model is used to obtain and output radiation damage prediction data of the reactor pressure vessel to be predicted.

[0009] In one embodiment, the step of obtaining the radiation characteristic data set of the reactor pressure vessel includes:

[0010] Acquiring radiation defect simulation data and radiation defect experimental data of the reactor pressure vessel;

[0011] Preprocessing the irradiation defect simulation data and the irradiation defect experimental data to extract key characteristic parameters that have a strong correlation with the target performance of the material of the reactor pressure vessel;

[0012] Constructing the irradiation characteristic data set according to the key characteristic parameters;

[0013] Wherein, the target performance includes the hardness, strength and ductile-brittle transition temperature of the material of the reactor pressure vessel.

[0014] In one embodiment, the step of obtaining the radiation defect simulation data of the reactor pressure vessel includes:

[0015] Based on first principles, obtaining thermodynamic parameters and kinetic parameters related to defect formation of the reactor pressure vessel;

[0016] Based on the thermodynamic parameters and the kinetic parameters, a kinetic simulation analysis is performed on the defect evolution process and defect evolution morphology of the reactor pressure vessel under different irradiation conditions, and a correlation relationship between each chemical element under different irradiation conditions and the defect evolution process and defect evolution morphology of the reactor pressure vessel is established;

[0017] The radiation defect simulation data of the reactor pressure vessel is obtained based on the association relationship.

[0018] In one embodiment, the radiation characteristic data set includes k data subsets, where k is a positive integer;

[0019] The step of constructing the radiation damage prediction model for the reactor pressure vessel according to the radiation characteristic data set includes:

[0020] Using the k data subsets, multiple basic models are trained and predicted to obtain prediction results of each basic model;

[0021] According to the prediction results of each of the basic models, a target irradiation characteristic data set is obtained;

[0022] Screening the prediction results of the multiple basic models according to the evaluation index to obtain a target model;

[0023] The target model is trained using the target radiation feature data set, and the radiation damage prediction model is obtained after training.

[0024] In one embodiment, the basic model includes multiple ones of a random forest model, a linear regression model, a decision tree model, a support vector machine model, a K nearest neighbor model, an adaptive regression model, and a gradient boosting regression model.

[0025] In one embodiment, the k data subsets are used to train and predict multiple basic models to obtain prediction results of each basic model, including:

[0026] One of the k data subsets is used as a prediction set, and the remaining k-1 data subsets are used as training sets. Each basic model is trained on the training set by cross-validation, and prediction is performed using the prediction set, and after the prediction is completed, the prediction result of each basic model is obtained.

[0027] In one embodiment, the step of screening the prediction results of the plurality of base models according to the evaluation index to obtain the target model includes:

[0028] Obtaining the training score, prediction score, prediction root mean square error and prediction determination coefficient of each of the basic models;

[0029] According to the training score, prediction score, prediction root mean square error and prediction determination coefficient of each basic model, the basic model with the best performance is selected from the multiple basic models as the target model.

[0030] A second aspect of an embodiment of the present application provides a device for constructing a radiation damage prediction model, the device comprising:

[0031] A data acquisition module, used to acquire an irradiation characteristic data set of a reactor pressure vessel;

[0032] A model building module, used to build a radiation damage prediction model for the reactor pressure vessel according to the radiation characteristic data set;

[0033] The radiation damage prediction model is used to obtain and output radiation damage prediction data of the reactor pressure vessel to be predicted.

[0034] The third aspect of an embodiment of the present application provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the method for constructing a radiation damage prediction model as described in the first aspect of the embodiment of the present application are implemented.

[0035] The fourth aspect of the embodiments of the present application provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the steps of the method for constructing a radiation damage prediction model as described in the first aspect of the embodiments of the present application.

[0036] The first aspect of the embodiment of the present application provides a method for constructing an irradiation damage prediction model, which obtains an irradiation feature data set of a reactor pressure vessel; and constructs an irradiation damage prediction model of the reactor pressure vessel based on the irradiation feature data set; wherein the irradiation damage prediction model is used to obtain and output the irradiation damage prediction data of the reactor pressure vessel to be predicted. When constructing the irradiation damage prediction model, the essential causes of the radiation damage of the materials of the reactor pressure vessel and the characteristics of the irradiation loss generated are fully considered. When the irradiation damage of the reactor pressure vessel is predicted using the irradiation damage prediction model, the accuracy of the radiation damage prediction of the reactor pressure vessel can be effectively improved, which is conducive to ensuring the safety and reliability of the nuclear power plant.

[0037] It can be understood that the beneficial effects of the second to fourth aspects mentioned above can be found in the relevant description of the first aspect mentioned above, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0039] Figure 1 This is a first flow chart of a method for constructing a radiation damage prediction model provided in an embodiment of the present application;

[0040] Figure 2 This is a second flow chart of the method for constructing a radiation damage prediction model provided in an embodiment of the present application;

[0041] Figure 3 It is a schematic diagram of the evolution process of the Mn-Ni-Si-rich cluster provided in the embodiment of the present application;

[0042] Figure 4 This is a third flow chart of the method for constructing a radiation damage prediction model provided in an embodiment of the present application;

[0043] Figure 5 It is a structural schematic diagram of a device for constructing a radiation damage prediction model provided in an embodiment of the present application;

[0044] Figure 6 It is a schematic diagram of the structure of the terminal device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0045] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0046] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0047] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0048] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways. "Multiple" means "two" or "more than two".

[0049] The reactor pressure vessel is one of the most critical equipment in a nuclear power plant, and it is crucial to ensure the safe operation of the nuclear power plant. During the service of the reactor pressure vessel, irradiation by high-energy neutrons and other factors will cause certain damage to the material of the reactor pressure vessel, thereby reducing the overall performance of the reactor pressure vessel. Therefore, it is necessary to predict and evaluate the reactor pressure vessel during the operation of the nuclear power unit.

[0050] The existing prediction methods for radiation damage of reactor pressure vessels fail to consider the essential causes of radiation damage to the materials of reactor pressure vessels and the characteristics of the radiation loss, resulting in low accuracy of the prediction results. In addition, with the increase in the power of nuclear power plants, the operating environment of reactor pressure vessels has also changed, and a new prediction method is needed to adapt to the new operating environment.

[0051] Therefore, the embodiment of the present application provides a method for constructing an irradiation damage prediction model, by obtaining an irradiation feature data set of a reactor pressure vessel; constructing an irradiation damage prediction model of the reactor pressure vessel according to the irradiation feature data set; wherein the irradiation damage prediction model is used to obtain and output the irradiation damage prediction data of the reactor pressure vessel to be predicted. When constructing the irradiation damage prediction model, the essential causes of the radiation damage of the materials of the reactor pressure vessel and the characteristics of the irradiation loss generated are fully considered. When the irradiation damage of the reactor pressure vessel is predicted using the irradiation damage prediction model, the accuracy of the irradiation damage prediction of the reactor pressure vessel can be effectively improved, which is conducive to ensuring the safety and reliability of the nuclear power plant.

[0052] Embodiment 1

[0053] like Figure 1 As shown, the method for constructing the radiation damage prediction model provided in the embodiment of the present application includes the following steps S1 to S2:

[0054] Step S1, obtaining the radiation characteristic data set of the reactor pressure vessel, and proceeding to step S2.

[0055] In application, due to the challenges of neutron irradiation test, such as complex experimental equipment, strict safety requirements and long experimental cycle, it is difficult to carry out corresponding experiments to obtain corresponding test data, which in turn limits the problem of irradiation damage assessment of reactor pressure vessels. Therefore, in the embodiment of the present application, the simulation data is obtained by machine simulation of neutron irradiation experiment, and the corresponding irradiation characteristic data set is obtained by combining the experimental data disclosed in the existing literature, and the corresponding irradiation damage prediction model is constructed based on the irradiation characteristic data set, thereby realizing the irradiation damage prediction of the reactor pressure vessel.

[0056] In one embodiment, Figure 2 As shown, step S1 specifically includes the following steps S11 to S13:

[0057] Step S11, obtaining radiation defect simulation data and radiation defect experimental data of the reactor pressure vessel, and proceeding to step S12.

[0058] In application, the experimental data of radiation defects of reactor pressure vessels can be obtained from some publicly published documents such as academic databases, professional journals, and experimental reports published by various institutions. Specific radiation defect experimental data include but are not limited to data generated by neutron irradiation experiments, data generated by proton irradiation experiments, data generated by ion irradiation experiments, etc. Correspondingly, the data generated in different experiments include but are not limited to the chemical composition of the materials of the reactor pressure vessels used for the experiments, irradiation experimental conditions (for example, neutron fluence and neutron fluence rate in neutron irradiation experiments, proton fluence in proton irradiation experiments, and ion fluence in ion irradiation experiments), data on the evolution law of defects in the materials of the reactor pressure vessels caused by radiation, and data related to mechanical properties such as hardness, strength, and ductile-brittle transition temperature of the materials of the reactor pressure vessels after the irradiation experiment.

[0059] In applications, defects caused by irradiation in the material of the reactor pressure vessel include but are not limited to dislocation loops and solute clusters. The material of the reactor pressure vessel includes a variety of chemical components (the proportion of each chemical component is set according to certain standards), and there are also a variety of solute clusters generated under irradiation, specifically including but not limited to Mn-Ni-Si rich clusters, Cu rich clusters, P rich clusters, CN rich clusters, O rich clusters, and H rich clusters.

[0060] In applications, dislocation loops are defects formed by dislocation lines closing into a ring structure inside a crystal material. These loops can be full dislocations or partial dislocations. Each dislocation loop has a fixed Burgers vector, which indicates the type and size of the dislocation. The direction and size of the Burgers vector determine the nature of the dislocation. In a nuclear reactor environment, irradiation-induced dislocation loops are an important component of irradiation damage. With the increase of irradiation dose, the number and size of dislocation loops will gradually increase, resulting in the degradation of the mechanical properties of the material of the reactor pressure vessel.

[0061] In applications, the Mn-Ni-Si-rich clusters in the solute clusters are mainly composed of manganese (Mn), nickel (Ni) and silicon (Si) atoms. Mn atoms, Ni atoms and Si atoms accelerate diffusion and aggregation under high temperature and irradiation (neutron irradiation, proton irradiation, ion irradiation), thus forming Mn-Ni-Si-rich clusters; the formation of Mn-Ni-Si-rich clusters will affect the mechanical properties of the materials of the reactor pressure vessel, especially its toughness and radiation resistance, and increase the risk of brittle fracture.

[0062] In applications, Cu-rich clusters are mainly composed of copper (Cu) atoms and may contain a small amount of other elements such as nickel (Ni) and silicon (Si). Cu atoms have a high diffusion capacity at high temperatures. Under the action of irradiation (neutron irradiation, proton irradiation, ion irradiation), Cu atoms will accelerate diffusion and aggregation to form Cu-rich clusters. The formation of Cu-rich clusters will lead to hardening and embrittlement of the material of the reactor pressure vessel and reduce its toughness, especially under high-dose irradiation conditions.

[0063] It is understandable that the formation mechanism of P-rich clusters, CN-rich clusters, O-rich clusters and H-rich clusters is similar to the formation mechanism of the above-mentioned Mn-Ni-Si-rich clusters or Cu-rich clusters, and will not be repeated here. The formation of any solute cluster will not only affect the microstructure of the material of the reactor pressure vessel, but may also cause changes in material properties, such as a decrease in toughness and radiation resistance. Therefore, in order to ensure the safety, reliability and economy of nuclear power plants, the radiation defect characteristics of the reactor pressure vessel must be fully considered to obtain more accurate radiation damage prediction data for the reactor pressure vessel.

[0064] In one embodiment, when step S11 acquires the radiation defect simulation data of the reactor pressure vessel, it specifically includes the following steps S111 to S113:

[0065] Step S111, based on the first principles, obtaining thermodynamic parameters and kinetic parameters related to defect formation of the reactor pressure vessel;

[0066] Step S112: Based on thermodynamic parameters and kinetic parameters, a kinetic simulation analysis is performed on the defect evolution process and defect evolution morphology of the reactor pressure vessel under different irradiation conditions, and a correlation relationship between each chemical element under different irradiation conditions and the defect evolution process and defect evolution morphology of the reactor pressure vessel is established;

[0067] Step S113: acquiring radiation defect simulation data of the reactor pressure vessel based on the association relationship.

[0068] In application, the first principle is a method based on quantum mechanics, which is used to directly calculate the properties of materials based on the most basic physical laws without relying on empirical data. The following takes the Mn-Ni-Si rich cluster defects generated by the material of the reactor pressure vessel under neutron irradiation as an example to illustrate steps S111 to S113.

[0069] In the application, when obtaining the thermodynamic parameters and kinetic parameters related to the formation of Mn-rich Ni-Si clusters based on the first principles, some existing experimental and theoretical studies can be consulted to obtain known information about the formation of Mn-rich Ni-Si clusters, including but not limited to phase diagrams, diffusion coefficients, binding energy, etc.; then determine which thermodynamic parameters (such as free energy) and kinetic parameters (such as diffusion coefficients, reaction rate constants) are the main factors affecting cluster formation, and then calculate the initial thermodynamic parameters and kinetic parameters based on the first principles.

[0070] In applications, the main stages in which chemical elements affect the evolution of Mn-Ni-Si-rich clusters are point defect formation, point defect diffusion, cluster nucleation, and cluster growth. The typical microstructural evolution processes of these four stages are as follows: Figure 3 As shown, the black circles represent Fe atoms and the gray circles represent solute atoms.

[0071] In step S112, a kinetic Monte Carlo simulation calculation method can be used to perform a kinetic simulation analysis on the evolution process of the Mn-Ni-Si-rich clusters of the reactor pressure vessel, which specifically includes the following three analysis processes: first, analyzing the alloy composition containing only Fe, Ni, Mn, and Si elements, and the composition is recorded as Fe-Mn-Ni-Si precipitation phase-0; second, on the basis of Fe-Mn-Ni-Si precipitation phase-0, in order to investigate the influence of the content of chemical elements in the cluster composition, a composition is designed by adjusting the Mn content, which is recorded as Fe-Mn-Ni-Si precipitation phase-1; third, In order to investigate the influence of the content change of non-constituent chemical elements in the clusters, the results of thermodynamic and kinetic parameter analysis were used (wherein, the thermodynamic analysis includes: using the method of calculating the phase diagram combined with the corresponding thermodynamic database to predict the phase equilibrium of the Fe-Mn-Ni-Si-M system under different M contents; then evaluating the change in the formation free energy of each phase after the addition of M; and finally finding the concentration range of M that makes the precipitation phase most stable, as well as the concentration threshold that may lead to the appearance of undesirable precipitation phases. The kinetic analysis includes: based on the results of thermodynamic analysis, constructing a kinetic Monte Carlo (KMC) model to simulate the influence of the change in the content of the chemical element M on the evolution of the precipitation phase; analyzing how the presence of the chemical element M changes the kinetic behavior of the precipitation phase under different temperature and time conditions.), a non-main chemical element M with a greater effect on RPV steel was selected, and the content of the chemical element M was adjusted on the basis of the Fe-Mn-Ni-Si precipitation phase-0 to determine a composition, recorded as the Fe-Mn-Ni-Si precipitation phase M, and so on, and finally the correlation between each chemical element and the evolution process of the Mn-Ni-Si precipitation phase cluster was constructed.

[0072] In addition, we can also refer to the average number of displacements per atom (Displacement Per Atom, DPA), simulate neutron irradiation by continuously introducing vacancies and self-interstitial pairs in the matrix, set appropriate injection rate and injection time, and use the Lattice Kinetic Monte Carlo (LKMC) model to simulate and analyze the Fe-Mn-Ni-Si precipitation phase M to obtain information such as the size, density and volume fraction of the Mn-Ni-Si precipitation phase clusters, and then investigate the effects of different chemical elements on the size, density and volume fraction of the Mn-Ni-Si precipitation phase clusters, and compare them with the aforementioned laws of the generation, evolution and distribution of rich Mn-Ni-Si clusters determined by different chemical elements to verify the correctness of the LKMC simulation analysis. Finally, based on the simulation analysis results, the correlation between each chemical element and the evolutionary morphology of the Mn-Ni-Si clusters under different irradiation conditions is constructed.

[0073] Finally, based on the relationship between the chemical elements and the evolution process and evolution morphology of the Mn-Ni-Si clusters under different irradiation conditions, the defect simulation data of the corresponding rich Mn-Ni-Si clusters can be obtained. It can be understood that the process of obtaining defect simulation data of other solute clusters is similar to the process of obtaining defect simulation data of rich Mn-Ni-Si clusters mentioned above, which will not be repeated here.

[0074] In the application, defect simulation data of dislocation loops can also be obtained according to needs. Since there is no strong correlation between dislocation loops and the composition changes of chemical elements, it is possible to establish a correlation between different irradiation conditions (such as different irradiation doses, irradiation temperatures, etc.) and the density changes of dislocation loops for specific component materials based on the results of dynamic simulation analysis. The specific dynamic simulation analysis process can be found in the above-mentioned related content, which will not be repeated here.

[0075] Step S12, preprocessing the irradiation defect simulation data and the irradiation defect experimental data, extracting key characteristic parameters that have a strong correlation with the target performance of the material of the reactor pressure vessel, and proceeding to step S13;

[0076] Step S13, constructing an irradiation characteristic data set according to key characteristic parameters;

[0077] Among them, the target performance includes the hardness, strength and ductile-brittle transition temperature of the reactor pressure vessel material.

[0078] In the application, after obtaining the irradiation defect simulation data and the irradiation defect experimental data, the data is also cleaned to remove missing values, abnormal values ​​and duplicate recorded values ​​to ensure the integrity and consistency of the data; determine the target performance (in the embodiment of the present application, the target performance includes the hardness, strength and ductile-brittle transition temperature of the material of the reactor pressure vessel), list all characteristic parameters that may affect the target performance, including the chemical composition of the material, defect characteristics (such as defect characteristics of dislocation rings, solute clusters, etc.), irradiation conditions (for example, temperature, time, particle type (including neutrons, protons, ions, etc.), injection), etc.; calculate the Pearson correlation coefficient between all characteristic parameters and each target performance, and the range of the Pearson correlation coefficient can be [-1,1], where 1 represents a complete positive correlation, -1 represents a complete negative correlation, and 0 represents no correlation; according to needs, a correlation coefficient threshold is set for each target performance (for example, it can be 0.5 or 0.6 or 0.7 or 0.8 or 0.9, etc., not limited here), delete the characteristic parameters whose absolute value of the correlation coefficient is lower than the threshold, and retain the characteristic parameters whose absolute value of the correlation coefficient is not lower than the threshold as the key characteristic parameters. Finally, an irradiation characteristic dataset is constructed based on the retained key characteristic parameters that have a strong correlation with the performance of each target.

[0079] In applications, according to actual needs, the radiation feature data set can also be randomly divided into k data subsets of equal size, where k is a positive integer, and each subset should maintain the consistency of data distribution as much as possible for training the radiation damage prediction model.

[0080] In the embodiment of the present application, k can be 100. It can be understood that, depending on the size of the data set and experimental requirements, in other embodiments, k can be any positive integer, which is not limited here.

[0081] The embodiment of the present application fully considers the essential causes of radiation damage to materials of the reactor pressure vessel and the characteristics of radiation defects, combines a variety of radiation defect simulation data and radiation defect experimental data for the construction of a radiation damage prediction model, which is conducive to improving the accuracy of the prediction model.

[0082] Step S2: constructing a radiation damage prediction model for the reactor pressure vessel according to the radiation characteristic data set.

[0083] In an embodiment of the present application, a radiation damage prediction model for a reactor pressure vessel can be constructed based on the Stacking integration strategy. Stacking can provide higher accuracy than a single model by integrating the predictions of multiple models and improve the generalization ability of the model. The specific construction process is as follows.

[0084] In one embodiment, the irradiation characteristic data set includes k data subsets, where k is a positive integer;

[0085] like Figure 4 As shown, step S2 specifically includes the following steps S21 to S24:

[0086] Step S21, using k data subsets, train and predict multiple basic models to obtain the prediction results of each basic model, and then go to step S22.

[0087] In the application, when constructing a radiation damage prediction model, multiple algorithm models commonly used in machine learning can be selected as basic models for training and testing. For example, the basic models can include any of a random forest model, a linear regression model, a decision tree model, a support vector machine model, a K nearest neighbor model, an adaptive regression model, and a gradient boosting regression model.

[0088] In applications, different models have their own advantages and disadvantages, so they can be applied to different types of prediction tasks. For example, the Random Forest model can reduce the risk of overfitting and can handle high-dimensional data, but it is not sensitive to missing values ​​and is suitable for scenarios with large data sets and more features. The Linear Regression model has high computational efficiency, but has poor modeling effect on nonlinear relationships and is suitable for scenarios when the relationship between the target variable and the feature is roughly linear. The Decision Tree model is prone to overfitting and is sensitive to small changes in data, so it is suitable for application scenarios with strong explanatory power. The Support Vector Machine model (SVM) is suitable for high-dimensional space and can effectively handle small sample data, but it has high computational complexity for large-scale data sets and is suitable for scenarios with high feature dimensions and moderate sample numbers. The K-Nearest Neighbors model (KNN) has poor performance on high-dimensional data and is sensitive to outliers, but requires a lot of memory and computing resources and is suitable for scenarios with relatively uniform data distribution and small feature space. The Adaptive Regression Model (ARM) can dynamically adjust model parameters to improve the flexibility and adaptability of the model, but the model complexity is high and requires more computing resources. It is suitable for scenarios where the data has time series characteristics or nonlinear changes. The Gradient Boosting Regression Model (GBR) has high prediction accuracy and can handle complex nonlinear relationships, but the training time is long. It is suitable for scenarios with large data sets and more features. Therefore, when selecting a basic model, it is necessary to select an appropriate model based on the task requirements and the characteristics of the data set.

[0089] In one embodiment, step S21 specifically includes:

[0090] One of the k data subsets is used as the prediction set, and the remaining k-1 data subsets are used as the training set. Each basic model is trained on the training set through cross-validation, and the prediction set is used for prediction. After the prediction is completed, the prediction results of each basic model are obtained.

[0091] In the application, any one of the k data subsets can be selected as the prediction set, and the remaining k-1 data subsets can be selected as the training set, so that each basic model can be trained and predicted on the training set through k-fold cross-validation.

[0092] In the application, when training each basic model, each of the k data subsets can be used as a prediction set in turn. For each (fold) data subset (from the 1st (fold) to the (kth (fold)), the (fold) data subset is used as a prediction set (also called a validation set), and the remaining (k-1) data subsets are used as training sets. In each iteration, the (k-1) (fold) data subsets are used to train the model, and the remaining 1 (fold) data subset is used to verify the performance of the model. The above process is repeated (k) times, and a different data subset is selected as the prediction set each time. Finally, the results of the (k) verifications can be averaged to obtain a more stable and reliable verification data. Through multiple training and testing, each basic model can be better generalized to unseen data, reducing the risk of overfitting; and the data subsets have the opportunity to be used as prediction sets, ensuring that the performance of each basic model on different data subsets can be verified; compared with simple training / testing division, k-fold cross-validation provides more stable and reliable performance verification.

[0093] Step S22: Obtain a target irradiation feature data set according to the prediction results of each basic model, and proceed to step S23.

[0094] In the application, the prediction result can also be called the verification result. Each basic model can get a corresponding verification result in each iteration. It can be understood that the more iterations, the more verification results are obtained, and each verification result has a corresponding data subset. Finally, according to the prediction results of each basic model, the target irradiation feature data set obtained includes the prediction results corresponding to each basic model.

[0095] Step S23: Filter the prediction results of multiple basic models according to the evaluation index to obtain the target model, and then proceed to step S24.

[0096] In applications, different basic models may obtain different prediction results for the same prediction set. Therefore, multiple different evaluation indicators can be used to evaluate the prediction performance of each basic model, so as to screen out the basic model with the best prediction performance.

[0097] In one embodiment, step S23 specifically includes:

[0098] Get the training score, prediction score, prediction root mean square error, and prediction determination coefficient of each base model;

[0099] According to the training score, prediction score, prediction root mean square error and prediction determination coefficient of each basic model, the basic model with the best performance is selected from multiple basic models as the target model.

[0100] In the application, in order to evaluate the prediction performance of each basic model, the basic model with the best prediction performance can be screened out by obtaining any multiple evaluation indicators including the training score, prediction score, prediction root mean square error and prediction determination coefficient of each basic model and combining the values ​​of each indicator.

[0101] Assume that y1, y2, y3, ... y n is the true value, is the predicted value, for y i The average value of , where i and n are both positive integers, 1≤i≤n (in this application, the value of n can be equal to the number of data subsets, for example, both are set to 100), then the prediction determination coefficient R 2 The formula can be expressed as:

[0102]

[0103] The formula for the prediction root mean square error RMSE can be expressed as:

[0104]

[0105] In the application, the training score and prediction score of each basic model can be directly obtained through the model, which will not be repeated here. It can be understood that each iteration training of each basic model can obtain a corresponding training score and prediction score. After the training is completed, all training scores can be averaged to obtain the final training score, and all prediction scores can be averaged to obtain the final prediction score, so as to screen the basic model with the best performance according to the final training score, prediction score, prediction root mean square error and prediction determination coefficient of each model.

[0106] In this embodiment, the specific evaluation indicators of each basic model are shown in Table 1 below:

[0107] Table 1

[0108] Algorithm Name Training score (%) Prediction score (%) Prediction root mean square error Prediction coefficient of determination Random Forest 96.85 96.85 0.00347 0.9685 Linear Regression 90.23 90.23 0.01078 0.9023 Decision Tree Regression 99.96 99.96 0.000044 0.9996 Support Vector Machine 1.44 1.44 0.1087 0.0144 K nearest neighbors 7.73 7.73 0.101769 0.0773 Adaptive Regression 96.13 96.13 0.004263 0.9613 Gradient Boosting Regression 99.99 99.99 1.532894 0.9999

[0109] In this example, a decision tree regression model can be selected as the target model to construct a radiation damage prediction model.

[0110] Step S24: training the target model using the target radiation feature data set, and obtaining a radiation damage prediction model after training.

[0111] In the application, according to the characteristics of the selected target model, the target irradiation feature data set can be first divided into different training sets, prediction sets and verification sets, and can be divided according to the division ratios of 70%: 25%: 5%; 75%: 20%: 5%; 80%: 15%: 5%; 85%: 10%: 5%; 90%: 5%: 5%; respectively (it can be understood that it can also be divided according to other division ratios, which are only examples and not limited here); then, based on different division ratios, the target model is trained, predicted and verified, and the optimal division ratio is selected from different division ratios through training as the data division ratio of the final irradiation damage prediction model. At the same time, the parameters of the target model can also be tuned, and the model parameters corresponding to the optimal prediction results are the optimal parameters of the target model. Finally, based on the optimal data division ratio and the optimal parameters, the final irradiation damage prediction model is determined.

[0112] In the application, after the target model is screened out through the above steps, in order to further improve the prediction accuracy of the final radiation damage prediction model, the target radiation feature data set can also be expanded by adding some new data (new defect simulation data and / or new defect experimental data), and then the data set is divided according to different division ratios to finally obtain the radiation damage prediction model.

[0113] In one embodiment, the radiation damage prediction model is used to obtain and output radiation damage prediction data of a reactor pressure vessel to be predicted.

[0114] In the application, based on the radiation damage prediction model, the relevant data of the reactor pressure vessel to be predicted can be input into the radiation damage prediction model to obtain the radiation damage prediction data output by the radiation damage prediction model, thereby realizing the radiation damage prediction of the reactor pressure vessel to be predicted. Among them, the relevant data input to the model include but are not limited to the composition and initial mechanical properties of the material of the reactor pressure vessel, irradiation conditions (for example, irradiation temperature, cumulative irradiation dose) and operating conditions (for example, working temperature, working pressure) and other relevant data. Correspondingly, the radiation damage prediction data can also be output to the user end for the staff to view and analyze.

[0115] In the application, the user terminal may include but is not limited to mobile phones, tablet computers, wearable devices, augmented reality (AR) / virtual reality (VR) devices, laptops, ultra-mobile personal computers (UMPC), netbooks, personal digital assistants (PDA) and other devices.

[0116] In the embodiments of the present application, when constructing a radiation damage prediction model, the essential causes of radiation damage to the materials of the reactor pressure vessel and the characteristics of the resulting radiation losses are fully considered, and the constructed radiation damage prediction model can be applicable to the new operating environment of the reactor pressure vessel, and can more accurately predict the radiation damage of the reactor pressure vessel, which helps to identify potential safety risks, take preventive measures in advance, and ensure the safe operation of the equipment.

[0117] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0118] Embodiment 2

[0119] The embodiment of the present application also provides a device for constructing a radiation damage prediction model, which is used to execute the method steps in the above-mentioned method for constructing a radiation damage prediction model. The device can be a virtual appliance in a terminal device, which is run by a processor of the terminal device, or it can be the terminal device itself.

[0120] like Figure 5 As shown, the device 100 for constructing a radiation damage prediction model provided in an embodiment of the present application includes a data acquisition module 101 and a model construction module 102.

[0121] The data acquisition module 101 is used to acquire the radiation characteristic data set of the reactor pressure vessel;

[0122] A model building module 102 is used to build a radiation damage prediction model for a reactor pressure vessel according to the radiation characteristic data set;

[0123] The radiation damage prediction model is used to obtain and output the radiation damage prediction data of the reactor pressure vessel to be predicted.

[0124] In one embodiment, the data acquisition module 101 is further configured to:

[0125] Obtain radiation defect simulation data and radiation defect experimental data of reactor pressure vessels;

[0126] Preprocess the irradiation defect simulation data and irradiation defect experimental data to extract key characteristic parameters that have a strong correlation with the target performance of the reactor pressure vessel material;

[0127] Construct an irradiation characteristic data set based on key characteristic parameters;

[0128] Among them, the target performance includes the hardness, strength and ductile-brittle transition temperature of the reactor pressure vessel material.

[0129] In one embodiment, the data acquisition module 101 is further configured to:

[0130] Based on first principles, obtain the thermodynamic and kinetic parameters related to the defect formation of the reactor pressure vessel;

[0131] Based on thermodynamic parameters and kinetic parameters, the defect evolution process and defect evolution morphology of the reactor pressure vessel under different irradiation conditions are subjected to kinetic simulation analysis, and the correlation between each chemical element under different irradiation conditions and the defect evolution process and defect evolution morphology of the reactor pressure vessel is established;

[0132] The irradiation defect simulation data of the reactor pressure vessel is obtained based on the association relationship.

[0133] In one embodiment, the irradiation characteristic data set includes k data subsets, where k is a positive integer;

[0134] The model building module 102 is further specifically used for:

[0135] Using k data subsets, multiple basic models are trained and predicted to obtain the prediction results of each basic model;

[0136] According to the prediction results of each basic model, the target irradiation characteristic data set is obtained;

[0137] The prediction results of multiple basic models are screened according to the evaluation indicators to obtain the target model;

[0138] The target model is trained using the target radiation feature data set, and a radiation damage prediction model is obtained after training.

[0139] In one embodiment, the basic model includes multiple ones of a random forest model, a linear regression model, a decision tree model, a support vector machine model, a K nearest neighbor model, an adaptive regression model, and a gradient boosting regression model.

[0140] In one embodiment, the model building module 102 is further configured to:

[0141] One of the k data subsets is used as the prediction set, and the remaining k-1 data subsets are used as the training set. Each basic model is trained on the training set through cross-validation, and the prediction set is used for prediction. After the prediction is completed, the prediction results of each basic model are obtained.

[0142] In one embodiment, the model building module 102 is further configured to:

[0143] Get the training score, prediction score, prediction root mean square error, and prediction determination coefficient of each base model;

[0144] According to the training score, prediction score, prediction root mean square error and prediction determination coefficient of each basic model, the basic model with the best performance is selected from multiple basic models as the target model.

[0145] In application, each unit in the above device may be a software program module, or may be implemented by different logic circuits integrated in a processor or independent physical components connected to a processor, or may be implemented by multiple distributed processors.

[0146] Embodiment 3

[0147] like Figure 6 As shown, the embodiment of the present application further provides a terminal device 200, including: at least one processor 201 ( Figure 6 Only one processor is shown in the figure), memory 202, and a computer program 203 stored in the memory 202 and executable on at least one processor 201. When the processor 201 executes the computer program 203, the steps in the above-mentioned method embodiments are implemented.

[0148] In the application, the terminal device includes but is not limited to the processor, memory, Figure 6 It is only an example of a terminal device and does not constitute a limitation on the terminal device. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components, such as human-computer interaction devices, input and output devices, network access devices, etc. The network access device may include a communication module for the terminal device to communicate with the user terminal.

[0149] In applications, the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. For example, the processor may be a timing controller (TCON). A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.

[0150] In applications, the memory may be an internal storage unit of a terminal device in some embodiments, for example, a hard disk or memory of the terminal device. The memory may also be an external storage device of the terminal device in other embodiments, for example, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc., equipped on the terminal device. The memory may also include both an internal storage unit of the terminal device and an external storage device. The memory is used to store operating systems, applications, boot loaders, data, and other programs, such as program codes of computer programs. The memory may also be used to temporarily store data that has been output or is to be output.

[0151] In the application, the communication module can be set as any device that can directly or indirectly perform long-distance wired or wireless communication with the user terminal according to actual needs. For example, the communication module can provide communication solutions including wireless local area networks (WLAN) (such as Wi-Fi networks), Bluetooth, Zigbee, mobile communication networks, global navigation satellite systems (GNSS), frequency modulation (FM), near field communication technology (NFC), infrared technology (IR), etc., which are applied to network devices. The communication module may include an antenna, and the antenna may have only one array element or an antenna array including multiple array elements. The communication module can receive electromagnetic waves through the antenna, frequency modulate and filter the electromagnetic wave signals, and send the processed signals to the processor. The communication module can also receive the signal to be sent from the processor, frequency modulate and amplify it, and convert it into electromagnetic waves for radiation through the antenna.

[0152] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / modules are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0153] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example for illustration. In actual applications, the above-mentioned function allocation can be completed by different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. The functional modules in the embodiment can be integrated into a processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The above-mentioned integrated modules can be implemented in the form of hardware or in the form of software functional modules. In addition, the specific names of the functional modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0154] The embodiment of the present application further provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments can be implemented.

[0155] An embodiment of the present application provides a computer program product. When the computer program product runs on a terminal device, the terminal device can implement the steps in the above-mentioned various method embodiments.

[0156] If the integrated module is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include at least: any entity or device capable of carrying the computer program code to the terminal device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electric carrier signal, a telecommunication signal, and a software distribution medium. For example, a USB flash drive, a mobile hard disk, a disk or an optical disk.

[0157] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0158] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0159] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0160] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0161] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for constructing a radiation damage prediction model, characterized in that: The method comprises: Obtaining the irradiation characteristic data set of the reactor pressure vessel; constructing a radiation damage prediction model for the reactor pressure vessel according to the radiation characteristic data set; The radiation damage prediction model is used to obtain and output radiation damage prediction data of the reactor pressure vessel to be predicted.

2. The method for constructing a radiation damage prediction model according to claim 1, characterized in that: The step of obtaining the radiation characteristic data set of the reactor pressure vessel comprises: Acquiring radiation defect simulation data and radiation defect experimental data of the reactor pressure vessel; Preprocessing the irradiation defect simulation data and the irradiation defect experimental data to extract key characteristic parameters that have a strong correlation with the target performance of the material of the reactor pressure vessel; Constructing the irradiation characteristic data set according to the key characteristic parameters; Wherein, the target performance includes the hardness, strength and ductile-brittle transition temperature of the material of the reactor pressure vessel.

3. The method for constructing a radiation damage prediction model according to claim 2, characterized in that: The obtaining of the radiation defect simulation data and the radiation defect experimental data of the reactor pressure vessel includes: Based on first principles, obtaining thermodynamic parameters and kinetic parameters related to defect formation of the reactor pressure vessel; Based on the thermodynamic parameters and the kinetic parameters, a kinetic simulation analysis is performed on the defect evolution process and defect evolution morphology of the reactor pressure vessel under different irradiation conditions, and a correlation relationship between each chemical element under different irradiation conditions and the defect evolution process and defect evolution morphology of the reactor pressure vessel is established; The radiation defect simulation data of the reactor pressure vessel is obtained based on the association relationship.

4. The method for constructing a radiation damage prediction model according to any one of claims 1 to 3, characterized in that: The radiation characteristic data set includes k data subsets, where k is a positive integer; The step of constructing the radiation damage prediction model for the reactor pressure vessel according to the radiation characteristic data set includes: Using the k data subsets, multiple basic models are trained and predicted to obtain prediction results of each basic model; According to the prediction results of each of the basic models, a target irradiation characteristic data set is obtained; Screening the prediction results of the multiple basic models according to the evaluation index to obtain a target model; The target model is trained using the target radiation feature data set, and the radiation damage prediction model is obtained after training.

5. The method for constructing a radiation damage prediction model according to claim 4, characterized in that: The basic models include multiple ones of a random forest model, a linear regression model, a decision tree model, a support vector machine model, a K nearest neighbor model, an adaptive regression model and a gradient boosting regression model.

6. The method for constructing a radiation damage prediction model according to claim 5, characterized in that: The method of using k data subsets to train and predict multiple basic models to obtain prediction results of each basic model includes: One of the k data subsets is used as a prediction set, and the remaining k-1 data subsets are used as training sets. Each basic model is trained on the training set by cross-validation, and prediction is performed using the prediction set, and after the prediction is completed, the prediction result of each basic model is obtained.

7. The method for constructing a radiation damage prediction model according to any one of claims 5 to 6, characterized in that: The step of screening the prediction results of the plurality of basic models according to the evaluation index to obtain the target model includes: Obtaining the training score, prediction score, prediction root mean square error and prediction determination coefficient of each of the basic models; According to the training score, prediction score, prediction root mean square error and prediction determination coefficient of each basic model, the basic model with the best performance is selected from the multiple basic models as the target model.

8. A device for constructing a radiation damage prediction model, characterized in that: Applied to the radiation damage prediction scenario of a reactor pressure vessel, the device comprises: A data acquisition module, used to acquire an irradiation characteristic data set of a reactor pressure vessel; A model building module, used to build a radiation damage prediction model for the reactor pressure vessel according to the radiation characteristic data set; The radiation damage prediction model is used to obtain and output radiation damage prediction data of the reactor pressure vessel to be predicted.

9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method for constructing the radiation damage prediction model as described in any one of claims 1 to 7 are implemented.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the method for constructing a radiation damage prediction model as described in any one of claims 1 to 7 are implemented.

Images