Nuclear power material life evaluation method and electronic equipment

Abstract

The invention provides a nuclear power material life evaluation method and electronic equipment, and relates to the technical field of nuclear power materials, and the method comprises the steps: obtaining first test data of a sample nuclear power material; obtaining feature data corresponding to the input features from the first test data; analyzing distribution characteristics of the input characteristics based on the characteristic data; preliminarily screening the input features according to the distribution feature analysis result; carrying out correlation analysis on the life characteristics in the first test data and the input characteristics after preliminary screening; screening target input features from the input features after preliminary screening according to a correlation analysis result, and constructing a training sample based on the target input features and life features corresponding to the target input features; training the life prediction model according to the training sample; and inputting second test data of the to-be-evaluated nuclear power material into the trained life prediction model to obtain a life evaluation result of the to-be-evaluated nuclear power material.

Description

Technical Field

[0001] This application relates to the field of nuclear power materials technology, and in particular to a method for assessing the lifespan of nuclear power materials and an electronic device. Background Technology

[0002] A pressurized water reactor (PWR) is a reactor that uses pressurized, unboiled light water (i.e., ordinary water) as both a moderator and coolant. As the mainstream reactor type in commercial nuclear power, PWRs generally have a design life of 60 years. Further extending reactor life has become a focus of related research. Meanwhile, the operating temperatures of next-generation advanced reactors are showing an upward trend, placing higher demands on the long-term reliability of materials. Materials of key components such as the reactor pressure vessel, heat exchanger piping, and control rod drive mechanisms are subjected to high temperatures, high pressures, and alternating loads for extended periods during service. Under these extreme conditions, materials may experience creep damage, fatigue failure, or complex failure modes resulting from the interaction of both. Therefore, accurately assessing the lifespan of nuclear power materials under creep-fatigue interaction is crucial for ensuring the safety of nuclear power systems, developing life-extending strategies, and selecting materials. Summary of the Invention

[0003] In view of this, this application provides a method and electronic device for assessing the lifespan of nuclear power materials.

[0004] In a first aspect, this application provides a method for assessing the lifespan of nuclear power materials, comprising: acquiring first test data of a sample nuclear power material, the first test data including first pure creep test data, first pure fatigue test data, first creep fatigue test data with peak retention, and first creep fatigue test data under different test conditions; acquiring feature data corresponding to input features from the first test data; analyzing the distribution characteristics of the input features based on the feature data to obtain distribution feature analysis results; performing preliminary screening of the input features based on the distribution feature analysis results; and performing correlation analysis between the lifespan features in the first test data and the preliminary screened input features.

[0005] Based on the correlation analysis results, target input features are selected from the initial screening input features, and training samples are constructed based on the target input features and the corresponding lifetime features. The lifetime prediction model is trained according to the training samples. The second test data of the nuclear power material to be evaluated is input into the trained lifetime prediction model to obtain the lifetime assessment result of the nuclear power material to be evaluated. The second test data includes the second pure creep test data, the second pure fatigue test data, the second creep fatigue test data with peak retention, and the second creep fatigue test data with sequential loading of the nuclear power material to be evaluated under the set test conditions.

[0006] Secondly, this application provides an electronic device, comprising:

[0007] At least one processor; and

[0008] At least one memory storing instructions that, when executed individually or jointly by the at least one processor, cause the electronic device to perform the method as described in the first aspect.

[0009] The proposed life assessment method for nuclear power materials incorporates various data from sample nuclear power materials under different experimental conditions during model training: firstly, pure creep test data, firstly, pure fatigue test data, firstly, creep fatigue test data with peak hold, and firstly, sequentially loaded creep fatigue test data. This comprehensively considers creep and fatigue damage mechanisms, enabling accurate assessment of the creep fatigue life of nuclear power materials. Secondly, it combines the correlation between distribution characteristics, input characteristics, and life characteristics for coupled analysis, retaining target input characteristics whose correlation and distribution states both meet the requirements. This optimization helps ensure the generalization ability and practical usability of the subsequent life prediction model. Furthermore, the method of first screening input features based on distribution characteristics and then performing correlation analysis based on the screening results has the following advantages: because distribution characteristics are relatively convenient and efficient to calculate, redundant input features can be effectively eliminated during the initial screening, reducing the workload for subsequent correlation analysis. Attached Figure Description

[0010] The accompanying drawings are included to provide a further understanding of this application; they are incorporated into and constitute a part of this application. The drawings illustrate embodiments of this application and, together with this specification, serve to explain the principles of this application. In the drawings:

[0011] Figure 1 This is a schematic flowchart of a life assessment method for nuclear power materials provided in an embodiment of this application;

[0012] Figure 2 This is a matrix diagram showing the relationship between lifetime features and input features;

[0013] Figure 3 This is a scatter plot corresponding to PCA analysis;

[0014] Figure 4 This is a flowchart illustrating another method for assessing the lifespan of nuclear power materials provided in this application embodiment;

[0015] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0016] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are merely some examples or embodiments of this application. For those skilled in the art, these drawings can be applied to other similar scenarios without creative effort. Unless obvious from the context or otherwise specified, the same reference numerals in the drawings represent the same structures or operations.

[0017] As indicated in this application, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" do not specifically refer to the singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0018] Furthermore, this application uses specific terms to describe embodiments of the application. For example, "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic related to at least one embodiment of the application. Therefore, it should be emphasized and noted that "an embodiment," "one embodiment," or "an alternative embodiment" mentioned twice or more in different locations in this specification do not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of the application can be appropriately combined.

[0019] Unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps described in these embodiments do not limit the scope of this application. It should also be understood that, for ease of description, the dimensions of the parts shown in the drawings are not drawn to actual scale. Techniques, methods, and devices known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and devices should be considered part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values. It should be noted that similar reference numerals and letters in the following drawings denote similar items; therefore, once an item is defined in one drawing, it need not be further discussed in subsequent drawings.

[0020] Furthermore, although the terminology used in this application is selected from commonly known and used terms, some terms mentioned in this application's specification may have been chosen by the applicant according to his or her judgment, and their detailed meanings are explained in the relevant sections of the description herein. Moreover, this application is to be understood not only by the actual terms used, but also by the meaning implied by each term.

[0021] This application uses flowcharts to illustrate the operations performed by an apparatus or device according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously. Furthermore, other operations may be added to these processes, or one or more steps may be removed from these processes.

[0022] Nuclear power materials are widely used in critical components such as reactor pressure vessels, heat exchanger piping, and control rod drive mechanisms. During service, these materials are typically subjected to extreme conditions such as high temperature, high pressure, and alternating loads. Long-term operation may lead to creep damage, fatigue failure, or complex failure modes resulting from the interaction of both. Therefore, accurately assessing the lifespan of nuclear power materials under creep-fatigue interaction is crucial for ensuring the safety of nuclear power systems, formulating life extension strategies, and selecting appropriate materials.

[0023] However, the typical test commonly used in the industry for creep fatigue life assessment is the strain control test with peak and valley values. This test introduces relatively small creep damage and cannot reflect the impact of behaviors such as precipitation phase growth on creep performance. As a result, the creep fatigue interaction diagram calculated by the existing scheme is mainly located in the fatigue damage-dominant region and cannot fully cover the multiple mechanisms of creep fatigue interaction. Therefore, the existing creep fatigue test data cannot accurately and comprehensively reflect the life loss process of nuclear power materials.

[0024] To alleviate at least one of the above problems, see Figure 1 This application proposes a method for assessing the lifespan of nuclear power materials, including:

[0025] S10: Obtain the first test data of the sample nuclear power material. The first test data includes the first pure creep test data, the first pure fatigue test data, the first creep fatigue test data with peak retention, and the first creep fatigue test data under sequential loading of the sample nuclear power material under different test conditions.

[0026] Currently, the industry commonly uses strain-controlled creep fatigue interaction tests with peak and valley retention to obtain creep fatigue test data. However, this type of test introduces relatively small creep damage (usually fatigue damage is dominant, for example, fatigue damage accounts for 90% and creep damage only accounts for 10%), which cannot reflect the influence of precipitation phase growth and other behaviors on creep performance. This does not match the actual service damage process of nuclear power materials. Consequently, the creep fatigue test data obtained through this test cannot accurately and comprehensively reflect the life loss process of nuclear power materials.

[0027] To alleviate the above problems, this application proposes a method for obtaining creep fatigue test data. In one embodiment, the creep fatigue test data is the first creep fatigue test data of the above-mentioned sequential loading. The acquisition process includes: applying a preset number of cyclic fatigue tests to the sample nuclear power material under a first preset service temperature zone and a first preset stress condition to form pre-fatigue damage; subjecting the pre-fatigue damaged sample nuclear power material to constant stress creep loading until the sample nuclear power material fractures or continues until a first preset termination time; and outputting the creep life of the sample nuclear power material under the first preset service temperature zone, the first preset stress condition, and the pre-fatigue damage.

[0028] Unlike traditional creep-fatigue interaction tests where fatigue damage dominates, this approach introduces more creep damage throughout the test. Constant stress creep loading is applied to pre-fatigue-damaged nuclear power material samples, maintaining a relatively high stress level and a longer creep time. This allows precipitates (such as carbides and intermetallic compounds) to gradually grow or aggregate within the sample material, leading to structural changes and accelerating the creep damage process. Introducing more creep damage during the test ensures that the final creep fatigue test data obtained from the sequential loading process better reflects the interaction between creep and fatigue during actual service, impacting the lifespan of the sample material and improving data accuracy.

[0029] In one embodiment, the process of acquiring the first pure creep test data includes: subjecting the sample nuclear power material to constant stress creep loading under a second preset service temperature zone and a second preset stress condition until the sample nuclear power material fractures or continues until a second preset termination time, and outputting the fracture time of the sample nuclear power material under the second preset service temperature zone and the second preset stress condition.

[0030] In one embodiment, the process of acquiring the first pure fatigue test data includes: applying cyclic strain to the sample nuclear power material under a third preset service temperature range and fatigue loading conditions until the sample nuclear power material fractures or meets a third preset termination condition, and outputting the fatigue life of the sample nuclear power material under the third preset service temperature range and fatigue loading conditions.

[0031] In one embodiment, the process of acquiring the first creep fatigue test data with peak hold includes: applying cyclic strain for a preset holding time to the sample nuclear power material under the conditions of a fourth preset service temperature zone and a preset strain amplitude, until the sample nuclear power material fractures or meets the fourth preset termination condition, and outputting the number of fracture cycles of the sample nuclear power material under the conditions of the fourth preset service temperature zone, preset strain amplitude, and preset holding time.

[0032] The following uses N (N>1) austenitic stainless steel nuclear grade 316 materials (hereinafter referred to as 316 materials) from the same batch with the sample nuclear power material as the standard size as an example to illustrate the process of obtaining the first creep fatigue test data, the first pure creep test data, the first pure fatigue test data, and the first creep fatigue test data with peak retention in the above sequential loading.

[0033] The process of generating the first pure creep test data includes: setting key service temperature zones such as 650℃ and 700℃ (i.e., the aforementioned second preset service temperature zone); selecting different levels of initial stress as the second preset stress condition based on existing literature and previous creep strength assessment results; and conducting at least three replicate samples for each temperature-stress combination test. Constant stress creep loading is applied to the 316 material under different temperatures and stress conditions until the 316 material fractures or continues until the second preset termination time. The creep fracture time of the 316 material under different temperature and stress conditions is output (i.e., the fracture time of the aforementioned sample nuclear power material under the second preset service temperature zone and the second preset stress condition), etc.

[0034] The process of acquiring the first pure fatigue test data includes: setting key service temperature ranges such as 650℃ and 700℃ (i.e., the aforementioned third preset service temperature range); based on the previous fatigue performance evaluation results of the 316 material, selecting multiple strain amplitudes as fatigue loading conditions under strain control mode, with strain amplitudes of 0.18%, 0.2%, 0.4%, 0.6%, 0.8%, 1.0%, and 1.2%. Under the third preset service temperature range and fatigue loading conditions, cyclic strain is applied to the 316 material until the 316 material fractures or meets the third preset termination condition, and the stress-strain hysteresis curves and fatigue life (i.e., the fatigue life of the aforementioned sample nuclear power material under the third preset service temperature range and fatigue loading conditions) of the 316 material at different temperatures and strain amplitudes are output.

[0035] The process of acquiring the first creep fatigue test data with peak hold includes: setting the temperature to key service temperature ranges such as 650℃ and 700℃ (i.e., the aforementioned fourth preset service temperature range); selecting a typical 0.8% strain amplitude in the preset strain amplitude conditions; setting the holding time at the peak strain, with the holding time (i.e., the aforementioned preset holding time) ranging from 0.110 minutes to 1060 minutes. Under the fourth preset service temperature range and preset strain amplitude conditions, cyclic strains with holding times ranging from 0.110 minutes to 1060 minutes are applied to the 316 material until the 316 material fractures or meets the fourth preset termination condition. The number of fracture cycles of the 316 material under the fourth preset service temperature range, preset strain amplitude conditions, and preset holding time is output, etc.

[0036] The process of acquiring the first creep fatigue test data under sequential loading includes: setting the temperature to key service temperature ranges such as 650℃ and 700℃ (i.e., the aforementioned first preset service temperature range); selecting the peak fatigue stress in the first preset stress condition to be the same as the stress at a creep fracture time of 1000h; conducting fatigue tests on 316 material for a preset number of cycles (100, 1000, 3000, 10000, 30000, and 100000 cycles, etc.) to form different pre-fatigue damages; subjecting 316 material with different pre-fatigue damages to constant stress creep loading until the 316 material fractures; and outputting the creep life under different temperatures, the first preset stress condition, and different fatigue pre-damages (i.e., the creep life of the aforementioned sample nuclear power material under the first preset service temperature range, the first preset stress condition, and pre-fatigue damages), etc.

[0037] It should be noted that in the above examples, the first, second, third, and fourth preset service temperature zones are all set to the same value, such as 650℃ and 700℃. However, in other examples, the first, second, third, and fourth preset service temperature zones may be different or partially the same, and this application does not impose specific limitations on this.

[0038] In one embodiment, after obtaining the first test data of the sample nuclear power materials, some preprocessing can be performed on the first test data, such as statistical consistency testing (e.g., standard deviation, skewness), outlier removal, etc., and the preprocessed first test data can be used to perform subsequent steps S11 to S16.

[0039] S11: Obtain feature data corresponding to the input features from the first experimental data.

[0040] S12: Analyze the distribution characteristics of input features based on feature data.

[0041] In one embodiment, input features may include temperature, maximum / minimum strain, loading rate, holding time, number of cycles, etc. These preset features are pre-defined, and the specific input features can be configured according to requirements.

[0042] The feature data corresponding to the aforementioned input features can be understood as the specific numerical value of each input feature, such as the input feature "temperature" corresponding to a certain number of degrees Celsius (°C). In one embodiment, feature data corresponding to the input features can be obtained from the first experimental data, and statistical analysis can be performed on these feature data to determine the distribution characteristics of the input features. The distribution characteristics may include the mean, data range, and dispersion, etc., and the dispersion can be obtained by calculating the standard deviation.

[0043] For example, assuming the input features include temperature, strain, loading rate, holding time, and number of cycles (Nf), statistical analysis of the feature data corresponding to each input feature yields the distribution feature analysis results shown in Table 1. Where: the range is the difference between the maximum and minimum values, reflecting the range of data fluctuation; the mean is the average value, representing the central tendency; the standard deviation is an indicator of dispersion, with a larger standard deviation indicating greater dispersion; and the coefficient of variation (CV) is the standard deviation divided by the mean.

[0044] Table 1

[0045] Feature name Range mean Standard deviation Coefficient of variation (CV) Temperature (°C) 50 676.25 25.39 0.038 strain(%) 0.88 0.50 0.259 0.518 Loading speed 0 0.001 0 0 Hold time (min) 100 19.44 35.10 1.81 Number of iterations (Nf) 340405 41245 84145 2.04

[0046] S13: Perform initial screening of input features based on the distribution feature analysis results.

[0047] In one embodiment, the distribution feature analysis results include statistical indicators such as mean, standard deviation, and range. Step S13, which involves initial screening of input features based on the distribution feature analysis results, includes: quantifying the distribution pattern of each input feature using statistical indicators such as mean, standard deviation, range, and coefficient of variation; identifying input features with abnormal distributions (hereinafter referred to as abnormal distribution features); and then, based on the number of abnormal distribution features, filtering out abnormal distribution features with a quantity > L (L is, for example, 2), and retaining abnormal distribution features with a quantity ≤ L. Abnormal distribution features may include input features with abnormally narrow distributions (e.g., coefficient of variation ≤ lower limit of coefficient of variation) and extremely outlier input features (e.g., determined according to a preset distribution rule). The lower limit of the coefficient of variation can be, for example, 0.1~0.2. For example, for an input feature: CV < 0.1~0.2: The feature has very small fluctuations and the data is extremely concentrated. Such features are usually considered redundant or lack discriminative power and can generally be considered for removal for machine learning and regression modeling. CV > 0.2: This feature is generally considered to have a certain degree of volatility and discriminative power and can be included in subsequent analysis. CV > 0.5: Significant fluctuations, clearly distinguishing different samples, worthy of attention. However, extreme outliers causing abnormally high CVs should be noted. In such cases, manual assistance can be used to screen input features. If the CV is much greater than 1 (e.g., greater than 2 or 3), it is also necessary to determine whether it is due to an abnormal distribution based on the reasonableness of the data distribution itself. In this case, manual assistance can be used to screen input features.

[0048] For example, as shown in Table 1, assuming the lower limit of the coefficient of variation is 0.1 and L is 2, the coefficient of variation of the input feature "temperature" in Table 1 is less than the lower limit of the coefficient of variation, so it can be identified as an abnormal distribution feature. However, there are only two input features "temperature" in the initial dataset, and the amount of data is small. Therefore, the input feature "temperature" is retained, and the determination will be made later by combining correlation analysis.

[0049] S14: Perform correlation analysis between the lifetime characteristics in the first experimental data and the input characteristics after initial screening.

[0050] In one embodiment, step S14 includes: obtaining lifetime data corresponding to the lifetime characteristics from the first experimental data, and calculating the correlation coefficient between the characteristic data of the input characteristics after initial screening and the lifetime data according to the following calculation formula. :

[0051] .

[0052] Where Cov is the covariance, and X is the variance. i Y is the input feature, Var is the lifetime feature, and Var is the variance.

[0053] The lifetime data corresponding to the aforementioned lifetime characteristics can be understood as the specific numerical values ​​of each lifetime characteristic. In one embodiment, lifetime characteristics may refer to features related to the lifetime of nuclear power materials. The specific lifetime characteristics can be preset, and lifetime characteristics may include, for example, fatigue life, cycle count, creep life, fracture time, etc.

[0054] S15: Based on the correlation analysis results, target input features are selected from the initially screened input features, and training samples are constructed based on the target input features and their corresponding lifetime features. There are multiple training samples, each consisting of a lifetime feature and multiple corresponding target input features.

[0055] In one embodiment, the correlation analysis results include the aforementioned correlation coefficients. Step S15 includes: determining the input features whose correlation coefficients meet the threshold condition (e.g., the correlation coefficients need to be greater than the set threshold) as target input features.

[0056] Alternatively, in another embodiment, the correlation analysis result is a correlation matrix representing lifetime characteristics versus input characteristics, such as... Figure 2 As shown, there are 13 features, including stress, left deformation, right deformation, average deformation, temperature rise, preheating time, holding temperature, holding time, high temperature, medium temperature, low temperature, displacement, and creep fracture life. Creep fracture life is the life feature, and the rest are input features. To identify the input features most strongly correlated with the life features and reduce redundancy, analysis was performed. Figure 2It can be observed that the correlation coefficient between the five features (high temperature, low temperature, medium temperature, heat preservation time, and temperature rise) is 1, and the correlation coefficient between heat preservation time and creep fracture life is -0.38, which is stronger than the other four features. Therefore, it is selected as the key feature, and the other four features are considered redundant and removed from the dataset. Similarly, left deformation, right deformation, and average deformation are deformation features with strong correlation (e.g., the difference in their correlation coefficients is less than 0.2). Average deformation (whose correlation coefficient with creep fracture life is greater than other features) is selected as the main feature, thereby removing the other two redundant features. Following this logic, stress, heat preservation temperature, preheating time, average deformation, and displacement are ultimately selected as the target input features.

[0057] By implementing steps S13-S15, a coupled analysis is performed combining distribution characteristics (such as dispersion), the correlation between input features and lifetime features. Features with abnormally narrow distributions or extreme outliers are automatically filtered out, retaining target input features that are both highly correlated with lifetime features and have good distribution representativeness. This optimization ensures the generalization ability and practical usability of the subsequent lifetime prediction model. Furthermore, by first screening input features based on distribution characteristics and then performing correlation analysis based on the screening results, the following advantages are available: because distribution characteristics are relatively easy and efficient to calculate, redundant input features can be effectively eliminated during the initial screening, reducing the workload for subsequent correlation analysis.

[0058] S16: Train the life prediction model using training samples, and input the second test data of the nuclear power material to be evaluated into the trained life prediction model to obtain the life assessment results of the nuclear power material to be evaluated. The second test data includes second pure creep test data, second pure fatigue test data, second creep fatigue test data with peak hold, and second creep fatigue test data under set test conditions. This life prediction model can be, for example, a neural network (ANN) model.

[0059] In this embodiment, the first experimental data obtained (first pure creep test data, first pure fatigue test data, first creep fatigue test data with peak hold, and first creep fatigue test data under different test conditions for sample nuclear power materials) can comprehensively cover multiple mechanisms of creep-fatigue interaction: fatigue damage-dominated, creep damage-dominated, and creep-fatigue interaction. It can reflect the bilinear characteristics in the creep-fatigue interaction diagram, thus solving the problem that the creep-fatigue interaction diagram calculated by existing schemes is mainly located in the fatigue damage-dominated region. This provides a comprehensive reflection of the life loss process of nuclear power materials. Furthermore, the aforementioned first creep fatigue test data under sequential loading can better reflect the interaction between creep and fatigue during actual service. Therefore, training a life prediction model based on such first experimental data can make the final life assessment result output by the life prediction model more consistent with the actual service damage process of nuclear power materials, resulting in higher accuracy.

[0060] The aforementioned lifetime prediction model is a neural network model, such as an ANN. Currently, to ensure the accuracy of the output results of a neural network model, a large number of high-fidelity training samples are usually required to perform machine learning on the neural network model. However, in the actual industrial scenarios of nuclear power materials, the available creep fatigue interaction data (such as the first creep fatigue test data with peak hold and the first creep fatigue test data with sequential loading) are very limited, making it difficult for the neural network model to produce effective output results with a small sample size.

[0061] To alleviate the above problems, step S15 above constructs training samples based on the target input features and the lifetime features corresponding to the target input features. This includes: constructing an original dataset based on the target input features and the lifetime features corresponding to the target input features; using an affinity propagation clustering algorithm to decompose the original dataset into at least one data cluster; constructing and training a generative adversarial network model (which includes a generator and a discriminator) for each data cluster; merging the data generated by the final generators corresponding to each data cluster into the generated dataset; and finally, using the generator corresponding to the convergence state of the adversarial network model training. Furthermore, training samples are constructed based on the generated dataset and the original dataset to expand the original dataset, thereby achieving accurate assessment of the lifetime of nuclear power materials under limited creep fatigue interaction test data.

[0062] Generative Adversarial Network (GAN) models, for example, include a generator G and a discriminator D. The generator G aims to learn to generate samples G(z) similar to real data x~pdata(x) from random noise z ~ pz(z), while the discriminator D aims to distinguish between real data x and generated data G(z). The related architecture design of GAN models includes:

[0063] 1. Generator G: Input dimension = 16-dimensional noise + condition (key statistics); 2–4 fully connected layers; hidden layer width increases linearly with the number of cluster samples; LeakyReLU is used for activation; output layer has no / linear activation.

[0064] 2. Discriminator D: 2–4 fully connected layers; spectral normalization; activation of LeakyReLU; addition of Dropout 0.1 to prevent overfitting.

[0065] 3. The loss functions for both the generator G and the discriminator D are the loss functions of currently known GAN models.

[0066] 4. Learning rate: 1e-4 to 3e-4 (Adam optimizer, β1=0.5, β2=0.9), batch size 16.

[0067] 5. Training strategy: Discriminator: Generator = 5:1 update, to accelerate discriminator convergence.

[0068] First, the Affinity Propagation (AP) clustering algorithm is used to perform unsupervised clustering on the original dataset. Unlike algorithms such as K-Means, which require pre-specifying the number of clusters (k), AP clustering can autonomously determine the optimal number of clusters and cluster centers based on the similarity between data points. The purpose of this step is to group data samples with similar intrinsic distribution characteristics into the same subset (cluster). In this way, the complex multimodal dataset is decomposed into several simpler, more uniformly distributed subsets, with each subset corresponding to a data cluster.

[0069] Furthermore, GANs are trained independently for each cluster: For each data cluster generated in AP clustering, a dedicated GAN model is independently built and trained. The architecture of each GAN (including the number of network layers, neurons, activation function, optimizer, and learning rate of the generator G and discriminator D) is configured according to the data characteristics of the corresponding cluster. Each GAN only needs to learn a simpler data distribution, thus reducing the training difficulty and significantly improving the quality and realism of the generated data.

[0070] The generator loss and discriminator loss will oscillate wildly in the early stages of training, but as training progresses, they should gradually enter a dynamic equilibrium and stabilize within a certain numerical range (equivalent to the adversarial network model reaching convergence). This indicates that the generator and discriminator have reached Nash Equilibrium, meaning that the generator's generative ability and the discriminator's discriminative ability mutually constrain and enhance each other, and the model has not experienced problems such as mode collapse or gradient vanishing. The data included in the above-mentioned generated dataset are the generator output data corresponding to the convergence state of the adversarial network model.

[0071] In one embodiment, constructing training samples based on the generated dataset and the original dataset includes: performing data distribution analysis and feature correlation analysis on the generated dataset and the original dataset; based on the results of the data distribution analysis and feature correlation analysis, if it is determined that the generated dataset matches the original dataset in terms of feature correlation and data distribution, then the generated dataset is determined as the target generated dataset, and training samples are constructed according to the target generated dataset and the original dataset. In this way, the target generated dataset used to construct the training samples can be more consistent with the original dataset in terms of statistical characteristics, thus ensuring the quality and reliability of the expanded training samples.

[0072] One feasible approach is data distribution analysis, which includes histogram analysis and principal component analysis (PCA). Histogram analysis is primarily used to verify the consistency of marginal distributions between the generated and original datasets, while PCA is primarily used to verify the consistency of global structure between the generated and original datasets. For example, the histogram analysis process and decision-making process are as follows:

[0073] Analyze each input feature in the generated dataset and the original dataset, and plot histograms for each input feature in the original dataset and the generated dataset. The histograms visually reflect the numerical distribution characteristics of a single feature (such as shape, peak position, and data range). Compare the shape (e.g., whether they are all normal or uniform), peak position (e.g., if the "holding time" of the input feature in the original dataset is 0.8, the input feature in the generated dataset should also be around 0.8, with a fluctuation of 0.1 allowed), and data range (the interval between the minimum and maximum values). If the distribution characteristics such as shape, peak position, and data range are similar (e.g., similarity greater than 90%), the dataset is considered to have passed; otherwise, it is considered to have failed. Generate the histogram distribution analysis results representing whether the dataset has passed or failed.

[0074] For example, the analysis and judgment process of principal component analysis is as follows:

[0075] Generate a centralized data matrix from the generated dataset or the original dataset. (where N is the number of samples and D is the feature dimension), its covariance matrix Defined as:

[0076] .

[0077] Performing eigenvalue decomposition on the covariance matrix ∑ yields the eigenvectors W = [w1, w2, w3]. D ] and eigenvalues ​​Λ=diag[λ1, λ2, λ D ], where each w i (representing the i-th eigenvector) is the "principal component direction", corresponding to the direction with the largest variance, each λ i (Representing the i-th eigenvalue) reflects the "importance" of the corresponding principal component, i.e., the magnitude of its variance.

[0078] The projection matrix is ​​constructed by selecting the first K (K=2 or 3, for easy 2D / 3D visualization) eigenvectors with the largest eigenvalues. Generate the dataset and each data point x in the original dataset. i K-dimensional projection y i Represented as:

[0079] .

[0080] Projecting the original dataset and the generated dataset into y i The original and generated datasets are plotted as K-dimensional scatter plots. The shapes (whether the clustering patterns of the points are consistent, e.g., if the original dataset shows an "elliptical distribution," the generated dataset should also have similar elliptical patterns), ranges (whether the coverage areas of the points on the coordinate axes overlap, e.g., if the original dataset's range on the PCA1 axis is [-3, 3], the generated dataset should also be within this range), and densities (whether the density distribution of the points is consistent, e.g., if the original dataset is dense in a certain area, the generated dataset should also be dense in that area). The criteria for judgment are that the shapes, ranges, and densities of the scatter plots of the original and generated datasets should highly overlap, without any deviation of the generated data from the original data area or the formation of new artificial clusters. The two scatter plots can be manually checked to see if they meet the judgment criteria. If they do, the dataset passes the judgment; otherwise, it fails, and the principal component analysis results indicating whether the judgment passes are output. Alternatively, image analysis techniques can be used to compare the shapes, ranges, and densities of the two scatter plots. If they are similar, the dataset passes the judgment; otherwise, it fails, and the principal component analysis results indicating whether the judgment passes are output.

[0081] For example, the above scatter plot can be, for instance, as shown in the example... Figure 3 As shown, Figure 3Hollow circles correspond to the generated dataset, and solid circles correspond to the original dataset. As can be seen from the figure, the generated dataset and the original dataset are almost completely overlapping in the figure. The shape, range and density of the scatter plot are basically the same, indicating that the difference between the two is small.

[0082] As a feasible approach, the correlation analysis results between the above features may include any one or more of the following: generating correlation coefficient matrices corresponding to the dataset and the original dataset respectively, and heatmaps of the correlation coefficient matrices. The correlation analysis and determination process between features is as follows:

[0083] Calculate the Pearson correlation coefficient matrix between all pairs of features in both the generated and original datasets. For two variables X and Y, the Pearson correlation coefficient is... for:

[0084]

[0085] Where N is the number of samples. and These are the means of X and Y, respectively. The criterion is that the first correlation coefficient matrix of the generated dataset should maintain a high degree of similarity to the second correlation coefficient matrix of the original dataset in both structure and value. Quantitatively, the mean absolute error (MAE) of the difference between corresponding elements of the first and second correlation coefficient matrices can be calculated. If this value is lower than a preset threshold (e.g., 0.2), it indicates that the generated dataset successfully reproduces the linear relationship structure between variables in the original dataset, and is thus deemed acceptable; otherwise, it is deemed unacceptable, and the feature correlation analysis results representing whether the judgment is acceptable are output.

[0086] Alternatively, intuitively, heatmaps corresponding to the first and second correlation coefficient matrices can be generated. If the color patterns and textures of the two heatmaps are similar (e.g., the similarity reaches 90%), the test is passed; otherwise, it is not passed, and the correlation analysis results between features representing whether the test is passed are output.

[0087] In this embodiment, the data distribution analysis results include histogram distribution analysis results and principal component analysis results. On one hand, when both histogram distribution analysis and principal component analysis results pass the characterization judgment, it is determined that the data distribution of the generated dataset matches that of the original dataset. This ensures that the internal patterns (marginal distribution) of single features and the correlation patterns (global structure) between features in the subsequently selected target generated dataset are consistent with the original dataset. If only the single feature distribution is similar, but the relationships between features are chaotic (e.g., the generated data are isolated clusters), it will still lead to inaccurate samples. On the other hand, when the correlation analysis results between features pass the characterization judgment, it is determined that the feature correlation of the generated dataset matches that of the original dataset. Further, combining the correlation analysis results between features and the data distribution analysis results, when both the feature correlation and data distribution of the generated dataset match those of the original dataset, the generated dataset is determined as the target generated dataset for subsequent training sample construction. By comprehensively considering the distribution characteristics and feature correlations between the generated dataset and the original dataset, the quality and reliability of the target generated dataset used for subsequent training sample construction are ensured. See also... Figure 4 This application also proposes another method for lifetime assessment of nuclear power materials, wherein the sample nuclear power materials include a first sample nuclear power material and a second sample nuclear power material. This method is applicable in... Figure 1 Based on the previous embodiment, the following steps S17 to S21 are added, and before performing step S10, the method further includes:

[0088] Step S17: Obtain nuclear power materials of the same size and batch as multiple initial sample nuclear power materials, and perform heat treatment on multiple initial sample nuclear power materials to eliminate residual stress (such as stress introduced by welding and processing) and microstructure deviations (such as uneven grain size and phase composition differences).

[0089] Step S18: Perform X-ray residual stress analysis (e.g., XRD). The microstructure and residual stress of the nuclear power materials in each initial sample after heat treatment were quantitatively characterized by electron backscatter diffraction (EBSD) and electron backscatter diffraction (EBSD) analysis.

[0090] Step S19: Obtain microstructure characteristic factors related to the heat treatment state from the quantitative characterization, and analyze the dispersion of these microstructure characteristic factors. These microstructure characteristic factors include residual stress factors (such as the mean and standard deviation of surface residual stress) and microstructure factors (such as grain size, dislocation density, phase ratio, etc.). Methods for analyzing the dispersion of microstructure characteristic factors include, for example, analysis of variance (ANOVA) and F-test, mainly used to compare the differences in microstructure characteristic factors among the initial nuclear power materials samples.

[0091] Step S20: Based on the dispersion analysis results, screen the first sample nuclear power material and the second sample nuclear power material from multiple initial sample nuclear power materials. The first sample nuclear power material is the initial sample nuclear power material whose dispersion meets the first condition, and the second sample nuclear power material is the initial sample nuclear power material whose dispersion meets the second condition.

[0092] Both the first and second conditions can be preset based on experimental data or theoretical analysis and can be adjusted as needed later. For example, the first condition can be that the fluctuation is less than 30%, and the second condition can be that the fluctuation is equal to 30%. Assuming multiple initial sample nuclear power materials include initial sample nuclear power material 01, initial sample nuclear power material 02, and initial sample nuclear power material 03, according to the dispersion analysis results, if the tissue characteristic factor of initial sample nuclear power material 01 is found to have a fluctuation exceeding 30% compared to the other two initial sample nuclear power materials, it can be removed and not used as a sample nuclear power material for subsequent step S10. If the tissue characteristic factor of initial sample nuclear power material 02 is found to have a fluctuation less than 30% compared to the other two initial sample nuclear power materials (i.e., meeting the first condition), it is retained as the first sample nuclear power material and used as one of the sample nuclear power materials for subsequent step S10. If the microstructure characteristic factor of the initial sample nuclear power material 03 is found to fluctuate by 30% compared to the other two initial sample nuclear power materials (i.e., meeting the second condition), it is not directly eliminated, but retained as the second sample nuclear power material, serving as one of the sample nuclear power materials for subsequent step S10. Its corresponding microstructure characteristic factor is associated with the identifier (such as material number) of the initial sample nuclear power material 03 and stored in the sample material library. When subsequent step S10 obtains the first test data of the initial sample nuclear power material 03, it retrieves the corresponding microstructure characteristic factor of the initial sample nuclear power material 03 from the sample database, and corrects the initial lifetime characteristics in the first test data of the initial sample nuclear power material 03 using this microstructure characteristic factor. Subsequent steps are then performed based on the corrected lifetime characteristics. This approach eliminates microstructure deviations in sample nuclear power materials from different batches and with different heat treatment histories, establishing an equivalent screening mechanism based on heat treatment status.

[0093] Step S21: Obtain the tissue characteristic factor corresponding to the second sample nuclear power material, and associate the tissue characteristic factor corresponding to the second sample nuclear power material with the second sample nuclear power material and store it in the sample material library.

[0094] In one embodiment, the lifetime characteristics in step S14 above include sub-lifetime characteristics corresponding to the second sample nuclear power material. The sub-lifetime characteristics are values ​​corrected based on the microstructure characteristics factor corresponding to the second sample nuclear power material stored in the sample material library, which eliminates the microstructure deviation of sample nuclear power materials under different batches and different heat treatment histories.

[0095] Specifically, the first experimental data mentioned above includes initial lifetime characteristics. Assuming that the sample nuclear power material in step S10 includes a second sample nuclear power material, after obtaining the first experimental data of the second sample nuclear power material by executing step S10, the microstructure characteristic factor corresponding to the second sample nuclear power material can be obtained from the sample database. The initial lifetime data corresponding to the second sample nuclear power material can be corrected by the microstructure characteristic factor. The subsequent steps all use the corrected lifetime characteristics. The microstructure characteristic factor characterizing the heat treatment state of the material can be filtered out during the entire lifetime assessment process. Lifetime equivalence prediction can be achieved between nuclear power materials of different batches and different heat treatment histories, thereby improving the stability and generalization ability of the subsequent lifetime prediction model.

[0096] As a feasible approach, correcting the initial lifetime data using tissue characteristic factors includes: measuring tissue characteristic factors and lifetime for several batches of sample materials, establishing the relationship between lifetime deviation and tissue characteristic factors, and obtaining a correction function F(S) to characterize the decay rate of lifetime due to tissue characteristic factors (0 < F ≤ 1), which is also the correction factor. This correction factor can be used to correct the lifetime.

[0097] By executing steps S17 to S21, it is possible to effectively ensure that the microstructure and stress deviations of each sample nuclear power material are within a reasonable range before executing subsequent steps S10 to S16. Furthermore, in addition to directly eliminating abnormal initial sample nuclear power materials (such as those with fluctuations exceeding 30%), lifetime equivalence prediction can be achieved between the same material from different batches and with different heat treatment histories by filtering microstructural characteristic factors that characterize the material's heat treatment state in lifetime prediction. This improves the stability and generalization ability of subsequent model predictions.

[0098] This application also provides a chip, including a circuit system configured to perform the life assessment method for nuclear power materials described above. The chip includes a Field Programmable Gate Array (FPGA) chip, a Complex Programmable Logic Device (CPLD) chip, and an Application Specific Integrated Circuit (ASIC) chip, etc.

[0099] Figure 5This is a simplified block diagram of an electronic device 500 suitable for implementing embodiments of this application. For example, the life assessment method for nuclear power materials described above can be implemented by the electronic device 500. As shown, the electronic device 500 includes one or more processors 510, one or more memories 520 coupled to the processors 510, and one or more communication modules 540 coupled to the processors 510.

[0100] Communication module 540 is used for bidirectional communication. Communication module 540 has at least one antenna to facilitate communication. The communication interface can represent any interface necessary for communication with other network elements.

[0101] Processor 510 can be of any type suitable for a local technology network, and as a non-limiting example, can include one or more of the following: general-purpose computer, special-purpose computer, microprocessor, digital signal processor (DSP), and processor based on a multi-core processor architecture. Electronic device 500 can have multiple processors, such as application-specific integrated circuit (ASIC) chips, which are timely driven to a clock that synchronizes with the main processor.

[0102] Memory 520 may include one or more non-volatile memories and one or more volatile memories. Examples of non-volatile memories include, but are not limited to, read-only memory (ROM), electrically programmable read-only memory (EPROM), flash memory, hard disk, optical disc (CD), digital video disc (DVD), and other magnetic and / or optical storage. Examples of volatile memories include, but are not limited to, random access memory (RAM) and other volatile memories that do not persist during power-off periods.

[0103] Computer program 530 includes computer-executable instructions that are executed by the associated processor 510. Program 530 may be stored in ROM 524. Processor 510 may perform any appropriate actions and processes by loading program 530 into RAM 522.

[0104] The embodiments of this application can be implemented by program 530, enabling electronic device 500 to execute the reference. Figure 1 or Figure 4 Any process disclosed in the discussion. Embodiments of this application may also be implemented by hardware or by a combination of software and hardware.

[0105] In some embodiments, program 530 may be tangibly contained in a computer-readable medium, which may be contained in an electronic device 500 (e.g., memory 520) or other storage device accessible to the electronic device 500. The electronic device 500 may load program 530 from the computer-readable medium into RAM 522 for execution. The computer-readable medium may include any type of tangible non-volatile memory, such as ROM, EPROM, flash memory, hard disk, CD, DVD, etc. Program 530 is stored on the computer-readable medium.

[0106] Generally, the various embodiments of this application can be implemented in hardware or dedicated circuitry, software, logic, or any combination thereof. Some aspects may be implemented in hardware, while others may be implemented in firmware or software, which may be executed by a controller, microprocessor, or other electronic device. Although various aspects of the embodiments of this application are shown and described as block diagrams, flowcharts, or other graphical representations, it should be understood that, as non-limiting examples, the blocks, devices, systems, techniques, or methods described herein may be implemented in hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware or controllers or other electronic devices, or some combination thereof.

[0107] This application also provides at least one computer program product tangibly stored on a non-transitory computer-readable storage medium. The computer program product includes computer-executable instructions, such as instructions included in a program module, which execute in a device on a target real or virtual processor to perform the aforementioned references. Figure 1 or Figure 4 The method described herein. Typically, a program module includes routines, programs, libraries, objects, classes, components, data structures, etc., that perform specific tasks or implement specific abstract data types. In various embodiments, the functionality of a program module can be combined or separated among program modules as needed. The machine-executable instructions used in the program module can execute on a local or distributed device. In a distributed device, the program module can reside on both local and remote storage media.

[0108] The program code used to perform the methods of this application can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the program code is executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The program code can be executed entirely on a machine, partially on a machine, partially on a remote machine, partially on a remote machine, or entirely on a remote machine or server as a standalone software package.

[0109] In the context of this application, computer program code or related data may be carried by any suitable carrier to enable a device, apparatus, or processor to perform the various processes and operations described above. Examples of carriers include signals, computer-readable media, etc.

[0110] Computer-readable media can be computer-readable signal media or computer-readable storage media. Computer-readable media can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or apparatuses, or any suitable combination thereof. More specific examples of computer-readable storage media include electrical connections having one or more wires, portable computer floppy disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable optical disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0111] Furthermore, although the operations are described in a specific order, this should not be construed as requiring that these operations be performed in the specific order or sequence shown, or that all of the operations shown be performed to obtain the desired result. In some cases, multitasking and parallel processing may be advantageous. Similarly, while several specific implementation details are included in the foregoing discussion, these details should not be construed as limiting the scope of this application, but rather as descriptions of features specific to particular embodiments. Certain features described in the context of a single embodiment may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments.

[0112] Although this application has been described in language specific to structural features and / or methodological behavior, it should be understood that the application as defined in the appended claims is not necessarily limited to the specific features or behaviors described above. Rather, the specific features and actions described above are disclosed as exemplary forms for implementing the claims.

Claims

1. A method for life assessment of a nuclear power material, characterized by, The method comprises the following steps: obtaining first test data of a sample nuclear power material, wherein the first test data comprises first pure creep test data, first pure fatigue test data, first creep-fatigue test data with peak holding and first sequential loading creep-fatigue test data of the sample nuclear power material under different test conditions; obtaining feature data corresponding to input features from the first test data; analyzing distribution characteristics of the input features based on the feature data to obtain a distribution characteristic analysis result; preliminarily screening the input features according to the distribution characteristic analysis result; performing correlation analysis on life characteristics in the first test data and the preliminarily screened input features; selecting target input features from the preliminarily screened input features according to a correlation analysis result, and constructing a training sample based on the target input features and life characteristics corresponding to the target input features; training a life prediction model according to the training sample; inputting second test data of a nuclear power material to be evaluated into the trained life prediction model to obtain a life evaluation result of the nuclear power material to be evaluated, wherein the second test data comprises second pure creep test data, second pure fatigue test data, second creep-fatigue test data with peak holding and second sequential loading creep-fatigue test data of the nuclear power material to be evaluated under a set test condition.

2. The method of claim 1, wherein, The process of obtaining the first sequential loading creep-fatigue test data comprises the following steps: applying a preset number of cycle fatigue tests to the sample nuclear power material under a first preset service temperature zone and a first preset stress condition to form a pre-fatigue damage; performing constant stress creep loading on the sample nuclear power material with the pre-fatigue damage until the sample nuclear power material is fractured or until a first preset termination time, and outputting a creep life of the sample nuclear power material under the first preset service temperature zone, the first preset stress condition and the pre-fatigue damage.

3. The method of claim 1, wherein, The process of constructing a training sample based on the target input features and life characteristics corresponding to the target input features comprises the following steps: constructing an original data set based on the target input features and life characteristics corresponding to the target input features; decomposing the original data set into at least one data cluster by using an affinity propagation clustering algorithm; constructing and training a generative adversarial network model corresponding to each data cluster respectively, wherein the generative adversarial network model comprises a generator and a discriminator; merging data generated by final generators corresponding to each data cluster into a generated data set, wherein the final generator is a generator corresponding to the adversarial network model when the adversarial network model is trained to reach a convergence state; constructing a training sample based on the generated data set and the original data set.

4. The method of claim 3, wherein, The process of constructing a training sample based on the generated data set and the original data set comprises the following steps: performing data distribution analysis and feature correlation analysis on the generated data set and the original data set to obtain a data distribution analysis result and a feature correlation analysis result; determining the generated data set as a target generated data set based on the data distribution analysis result and the feature correlation analysis result if it is determined that the feature correlation and data distribution of the generated data set and the original data set are matched. The training sample is constructed according to the target generated data set and the original data set.

5. The method of claim 4, wherein, The data distribution analysis result includes a histogram distribution analysis result and a principal component analysis result of the generated data set and the original data set.

6. The method of claim 4, wherein, The inter-feature correlation analysis result is determined based on respective correlation coefficient matrices of the generated data set and the original data set.

7. The method of claim 1, wherein, The sample nuclear power material includes a first sample nuclear power material and a second sample nuclear power material, and the method further includes: Obtaining nuclear power materials of the same size and batch as a plurality of initial sample nuclear power materials; Performing heat treatment on the plurality of initial sample nuclear power materials to eliminate residual stress and organizational deviation; Quantitatively characterizing the microstructure and residual stress of each initial sample nuclear power material after heat treatment by X-ray residual stress analysis and electron backscatter diffraction analysis; Obtaining a heat treatment state related organizational characteristic factor from the quantitative characterization, and analyzing the dispersion of the organizational characteristic factor; Based on the dispersion analysis result, the first sample nuclear power material and the second sample nuclear power material are selected from the plurality of initial sample nuclear power materials, the first sample nuclear power material is an initial sample nuclear power material satisfying a first condition in terms of dispersion, and the second sample nuclear power material is an initial sample nuclear power material satisfying a second condition in terms of dispersion; Obtaining the organizational characteristic factor corresponding to the second sample nuclear power material, and storing the organizational characteristic factor corresponding to the second sample nuclear power material in association with the identification of the second sample nuclear power material in a sample material library.

8. The method of claim 7, wherein, The life characteristic includes a sub-life characteristic corresponding to the second sample nuclear power material, and the sub-life characteristic is a numerical value corrected based on the organizational characteristic factor corresponding to the second sample nuclear power material stored in the sample material library.

9. The method of claim 1, wherein, The first pure creep test data acquisition process includes: Under a second preset service temperature zone and a second preset stress condition, the sample nuclear power material is subjected to constant stress creep loading until the sample nuclear power material is fractured or continues to a second preset termination time; The fracture time of the sample nuclear power material under the second preset service temperature zone and the second preset stress condition is output.

10. The method of claim 1, wherein, The first pure fatigue test data acquisition process includes: Under a third preset service temperature zone and fatigue loading conditions, the sample nuclear power material is subjected to cyclic strain until the sample nuclear power material is fractured or satisfies a third preset termination condition, and the fatigue life of the sample nuclear power material under the third preset service temperature zone and the fatigue loading conditions is output.

11. An electronic device, comprising: Comprise: At least one processor; And At least one memory having instructions stored thereon, which, when executed by the at least one processor alone or collectively, cause the electronic device to perform the method of any one of claims 1-10.

Images