Rubber aging duration prediction model training method and device and computer equipment

Abstract

The application relates to a rubber aging time prediction model training method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: obtaining a target training set for nuclear power rubber materials; the target training set comprises a plurality of training samples and a label aging time corresponding to each training sample, and each training sample comprises an aging index of a material sample belonging to a target nuclear power rubber material type and aging environment information including a radiation dose and an aging temperature corresponding to the aging index; the training sample is input into an initial rubber aging time prediction model to obtain a predicted aging time corresponding to the training sample; a sample error between the predicted aging time and the label aging time corresponding to the same training sample is obtained to obtain a target loss; the model parameters are adjusted based on the target loss until a convergence condition is met, and a target rubber aging time prediction model corresponding to the target nuclear power rubber material type is obtained. The method can improve the rubber aging time prediction efficiency.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus, computer equipment, storage medium and computer program product for training a rubber aging time prediction model. Background Technology

[0002] Rubber materials, due to their outstanding elasticity, have been widely used in many fields and have played a significant role. However, the performance of rubber materials or products deteriorates due to aging during storage, processing, or use. A current concern is how long rubber materials or products can be stored or used to ensure timely replacement with new materials or products, maintain the normal operation of equipment or components, and prevent accidents.

[0003] In the traditional method, staff members periodically collect current status data of nuclear power rubber materials or products during the production process and predict the current aging time of nuclear power rubber materials or products based on past work experience. However, this method suffers from low efficiency in predicting the aging time of rubber. Summary of the Invention

[0004] Therefore, it is necessary to provide a rubber aging time prediction model training method, apparatus, computer equipment, computer-readable storage medium, and computer program product that can improve the prediction efficiency of rubber aging time in response to the above-mentioned technical problems.

[0005] This application provides a method for training a rubber aging time prediction model. The method includes:

[0006] Obtain a target training set for nuclear power rubber materials; the target training set includes multiple training samples and the label aging time corresponding to each training sample. The training samples include the aging index of material samples belonging to the target nuclear power rubber material type and the aging environment information pairs corresponding to the aging index; the aging environment information pairs include radiation dose and aging temperature.

[0007] Each training sample is input into the initial rubber aging time prediction model to obtain the predicted aging time corresponding to each training sample.

[0008] The target loss is obtained based on the sample error between the predicted aging time and the label aging time corresponding to the same training sample.

[0009] The model parameters of the initial rubber aging time prediction model are adjusted based on the target loss until the convergence condition is met, thus obtaining the target rubber aging time prediction model corresponding to the target nuclear power rubber material type.

[0010] This application also provides a training device for a rubber aging time prediction model. The device includes:

[0011] The training set acquisition module is used to acquire a target training set for nuclear power rubber materials. The target training set includes multiple training samples and the label aging time corresponding to each training sample. The training samples include the aging index of material samples belonging to the target nuclear power rubber material type and the aging environment information pairs corresponding to the aging index. The aging environment information pairs include radiation dose and aging temperature.

[0012] The aging time determination module is used to input each training sample into the initial rubber aging time prediction model to obtain the predicted aging time corresponding to each training sample.

[0013] The target loss determination module is used to obtain the target loss based on the sample error between the predicted aging time and the label aging time corresponding to the same training sample.

[0014] The target model determination module is used to adjust the model parameters of the initial rubber aging time prediction model based on the target loss until the convergence condition is met, so as to obtain the target rubber aging time prediction model corresponding to the target nuclear power rubber material type.

[0015] A computer device includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described rubber aging time prediction model training method.

[0016] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described rubber aging time prediction model training method.

[0017] A computer program product includes a computer program that, when executed by a processor, implements the steps of the above-described rubber aging time prediction model training method.

[0018] The aforementioned rubber aging time prediction model training method, apparatus, computer equipment, storage medium, and computer program product acquire a target training set for nuclear power rubber materials. The target training set includes multiple training samples and corresponding labeled aging times for each training sample. Each training sample includes an aging index of the material sample belonging to the target nuclear power rubber material type and a pair of aging environment information corresponding to the aging index. The aging environment information pair includes radiation dose and aging temperature. Each training sample is input into the initial rubber aging time prediction model to obtain the predicted aging time for each training sample. Based on the sample error between the predicted aging time and the labeled aging time for the same training sample, a target loss is obtained. The model parameters of the initial rubber aging time prediction model are adjusted based on the target loss until the convergence condition is met, resulting in the target rubber aging time prediction model corresponding to the target nuclear power rubber material type. Thus, the aging index of the material sample, the radiation dose corresponding to the aging index, and the aging temperature are used as training samples, and the aging time required for the material sample to reach the aging index under the aforementioned radiation dose and aging temperature is used as the sample label, resulting in the target training set. The initial rubber aging time prediction model is trained based on the target training set to obtain the target rubber aging time prediction model corresponding to the target nuclear power rubber material type. By inputting the aging environment information pair and aging index corresponding to the material belonging to the target nuclear power rubber material type into the target rubber aging time prediction model, the aging time required for the material to reach the aging index under the radiation dose and aging temperature indicated by the aging environment information pair can be directly obtained, which can effectively improve the rubber aging time prediction efficiency. Attached Figure Description

[0019] Figure 1 This is a diagram illustrating the application environment of a rubber aging time prediction model training method in one embodiment.

[0020] Figure 2 This is a flowchart illustrating the training method for a rubber aging time prediction model in one embodiment.

[0021] Figure 3 This is a flowchart illustrating the training method for a rubber aging time prediction model in another embodiment;

[0022] Figure 4 This is a flowchart illustrating the process of obtaining a target training set for nuclear power rubber materials in another embodiment;

[0023] Figure 5 This is a schematic diagram of the structure of a rubber aging time prediction model in one embodiment;

[0024] Figure 6 This is a schematic diagram showing the mean square error corresponding to different numbers of hidden layer nodes in one embodiment;

[0025] Figure 7 This is a schematic diagram showing the mean squared error corresponding to different test set proportions in one embodiment;

[0026] Figure 8 This is a schematic diagram illustrating the prediction accuracy of a target lifespan prediction model in one embodiment;

[0027] Figure 9 This is a schematic diagram illustrating the percentage of prediction error of a target lifespan prediction model in one embodiment;

[0028] Figure 10 This is a structural block diagram of a rubber aging time prediction model training device in one embodiment;

[0029] Figure 11 This is a structural block diagram of a rubber aging time prediction model training device in another embodiment;

[0030] Figure 12 This is an internal structural diagram of a computer device in one embodiment;

[0031] Figure 13 This is a diagram of the internal structure of a computer device in another embodiment. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0033] The rubber aging time prediction model training method provided in this application embodiment can be applied to, for example... Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or located on the cloud or other network servers. Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can be smart TVs, smart in-vehicle devices, etc. Portable wearable devices can be smartwatches, smart bracelets, head-mounted devices, etc. Server 104 can be implemented using a standalone server or a server cluster consisting of multiple servers. Terminal 102 and server 104 can be directly or indirectly connected via wired or wireless communication, which is not limited herein.

[0034] Both the terminal and the server can be used independently to execute the rubber aging time prediction model training method provided in the embodiments of this application.

[0035] For example, the terminal acquires a target training set for nuclear power rubber materials. The target training set includes multiple training samples and their corresponding label aging times. Each training sample includes an aging index for a material sample belonging to the target nuclear power rubber material type and corresponding aging environment information pairs, including radiation dose and aging temperature. The terminal inputs each training sample into an initial rubber aging time prediction model to obtain the predicted aging time for each training sample. Based on the sample error between the predicted aging time and the label aging time for the same training sample, the terminal obtains the target loss. The terminal adjusts the model parameters of the initial rubber aging time prediction model based on the target loss until the convergence condition is met, thus obtaining the target rubber aging time prediction model corresponding to the target nuclear power rubber material type.

[0036] The terminal and server can also work together to execute the rubber aging time prediction model training method provided in the embodiments of this application.

[0037] For example, the terminal sends a training request to the server for a rubber aging time prediction model, carrying a training set identifier. The server obtains a target training set for nuclear power rubber materials based on the training set identifier. The target training set includes multiple training samples and their corresponding labeled aging times. Each training sample includes an aging index for a material sample belonging to the target nuclear power rubber material type and a pair of aging environment information corresponding to that index. The aging environment information pair includes radiation dose and aging temperature. The server inputs each training sample into the initial rubber aging time prediction model to obtain the predicted aging time for each training sample. The server calculates the target loss based on the sample error between the predicted aging time and the labeled aging time for the same training sample. The server adjusts the model parameters of the initial rubber aging time prediction model based on the target loss until the convergence condition is met, obtaining the target rubber aging time prediction model corresponding to the target nuclear power rubber material type. The server can then send the target rubber aging time prediction model to the terminal, which uses this model to predict the aging time of materials belonging to the target nuclear power rubber material type.

[0038] In one embodiment, such as Figure 2 As shown, a method for training a rubber aging time prediction model is provided. The method is illustrated using a computer device as an example. The computer device can be a terminal or a server, and the model can be executed independently by the terminal or server, or through interaction between the terminal and server. The rubber aging time prediction model training method includes the following steps:

[0039] Step S202: Obtain the target training set for nuclear power rubber materials; the target training set includes multiple training samples and the label aging time corresponding to each training sample. The training samples include the aging index of material samples belonging to the target nuclear power rubber material type and the aging environment information pair corresponding to the aging index; the aging environment information pair includes radiation dose and aging temperature.

[0040] Nuclear power rubber materials refer to rubber materials used in nuclear power equipment. The target training set includes multiple training samples belonging to the target nuclear power rubber material type and the corresponding label aging time for each training sample, used to train a rubber aging time prediction model for the target nuclear power rubber material type. Nuclear power rubber material type refers to the type used to distinguish different nuclear power rubber materials, which can be classified based on at least one of the raw materials, models, or applications of the nuclear power rubber material. The target nuclear power rubber material type refers to the type of nuclear power rubber material whose aging time prediction model needs to be trained to predict its aging time among all nuclear power rubber material types. Material samples refer to the materials used for aging tests. Aging tests on material samples yield the aging index corresponding to the material sample after a specified aging time under the radiation dose and aging temperature indicated by the specified aging environment information, i.e., the training sample and its corresponding label aging time. Label aging time refers to the actual time required for the material sample to reach the specified aging index under the specified aging environment information, obtained from aging tests. Aging time refers to the time required from when a brand new material is placed in a specified aging environment until the material reaches a specified aging index.

[0041] The aging index is a numerical value used to characterize the current degree of aging of a material sample, and it is positively correlated with the degree of aging. In practice, the aging index can be calculated based on at least one of the following indicators: compressive strength retention, tensile strength, and elongation at break. An aging environment information pair refers to a set of information used to indicate environmental characteristics. Radiation dose indicates the dose of radiation the material receives in the environment. Aging temperature refers to the temperature of the environment in which the material is located.

[0042] For example, the computer device acquires the target training set corresponding to the target nuclear power rubber material type. In actual implementation, multiple training samples can be randomly collected from all training samples corresponding to the target nuclear power rubber material type, and the target training set is obtained based on each collected training sample and the label aging time corresponding to each training sample.

[0043] Step S204: Input each training sample into the initial rubber aging time prediction model to obtain the predicted aging time corresponding to each training sample.

[0044] The rubber aging time prediction model is used to predict the time required for a material belonging to the target nuclear power rubber material type to reach a specified aging index under a given aging environment information pair. The model's input data includes the aging index corresponding to the material and the aging environment information pair corresponding to the aging index. The model's output data is the time required for the material to reach the aging index under the environment indicated by the aging environment information pair. The initial rubber aging time prediction model refers to the rubber aging time prediction model that has not yet completed training. The predicted aging time corresponding to the training samples refers to the predicted aging time output by the initial rubber aging time prediction model when the training samples are input.

[0045] For example, the computer device inputs the aging index, radiation dose, and aging temperature contained in the same training sample into the initial rubber aging time prediction model, and outputs the predicted aging time corresponding to each training sample in the target training set from the initial rubber aging time prediction model.

[0046] Step S206: Based on the sample error between the predicted aging time and the label aging time corresponding to the same training sample, the target loss is obtained.

[0047] Here, sample error refers to the difference between the predicted aging time and the labeled aging time for the same training sample. Target loss refers to the model loss value calculated based on the sample error corresponding to each training sample in the target training set.

[0048] For example, the computer device calculates the sample error for each training sample based on the predicted aging time and the label aging time corresponding to the same training sample. In actual implementation, the mean squared error between the predicted aging time and the label aging time corresponding to the same training sample can be directly used as the sample error. Then, the sample errors corresponding to each training sample in the target training set are fused to obtain the target loss for the target training set. Specifically, the mean squared error of the sample errors corresponding to each training sample in the target training set can be used as the target loss.

[0049] Step S208: Adjust the model parameters of the initial rubber aging time prediction model based on the target loss until the convergence condition is met, and obtain the target rubber aging time prediction model corresponding to the target nuclear power rubber material type.

[0050] The convergence condition refers to the preset conditions for judging whether the model has converged. For example, the convergence condition can be that the model has been trained for more than a preset number of rounds, or that the model error is less than a preset error.

[0051] For example, the computer device updates the model parameters in the initial aging time based on the target loss to obtain an intermediate rubber aging time prediction model. Using this intermediate rubber aging time prediction model as the initial rubber aging time prediction model, the step of obtaining the target training set for nuclear power rubber materials is returned. Each training sample in the target training set is input into the initial aging time prediction to obtain the target loss corresponding to the target training set. Based on the target loss, the model parameters of the initial rubber aging time prediction model are further adjusted until the convergence condition is met, thus obtaining the target rubber aging time prediction model.

[0052] In one embodiment, such as Figure 3 As shown, the computer device first initializes the model parameters such as weights and thresholds in the rubber aging time prediction model to obtain the initial rubber aging time prediction model. Then, a training sample set is collected from all training samples, and the training samples in the training sample set are input into the initial rubber aging time prediction model to obtain the output of the output layer node in the model, i.e., the predicted aging time corresponding to the training sample. Based on the label aging time and predicted aging time corresponding to the same training sample, the sample error corresponding to each training sample is calculated. Based on the sample error corresponding to each training sample, the mean squared error corresponding to the training sample set is obtained. The model parameters in the initial rubber aging time prediction model are adjusted based on the mean squared error. Simultaneously, the training samples are improved based on the sample error corresponding to the training sample set. Specifically, training samples with sample errors less than a preset threshold can be filtered from all training samples. The process continues to collect training sample sets from the remaining training samples, returning to the step of inputting training samples, until all training samples are trained and the number of learning iterations is updated. If the current model loss (mean squared error) corresponding to the rubber aging time prediction model is greater than or equal to a set value, or the number of learning iterations is less than or equal to a set value, the process returns to the step of inputting training samples. The target rubber aging time prediction model is obtained when the mean square error of the current rubber aging time prediction model is less than the set value or the number of learning iterations is greater than the set value.

[0053] In the aforementioned method for training the rubber aging time prediction model, a target training set for nuclear power rubber materials is obtained. This target training set includes multiple training samples and corresponding label aging times for each sample. The training samples include the aging index of the material sample belonging to the target nuclear power rubber material type and the corresponding aging environment information pair, which includes radiation dose and aging temperature. Each training sample is input into the initial rubber aging time prediction model to obtain the predicted aging time for each training sample. The target loss is obtained based on the sample error between the predicted aging time and the label aging time for the same training sample. The model parameters of the initial rubber aging time prediction model are adjusted based on the target loss until the convergence condition is met, resulting in the target rubber aging time prediction model corresponding to the target nuclear power rubber material type. Thus, the aging index of the material sample, the radiation dose corresponding to the aging index, and the aging temperature are used as training samples, and the aging time required for the material sample to reach the aging index under the aforementioned radiation dose and aging temperature is used as the sample label, resulting in the target training set. The initial rubber aging time prediction model is trained based on the target training set to obtain the target rubber aging time prediction model corresponding to the target. By inputting the aging environment information pair and aging index corresponding to the target nuclear power rubber material type into the target rubber aging time prediction model, the aging time required for the material to reach the aging index under the radiation dose and aging temperature indicated by the aging environment information pair can be directly obtained, which can effectively improve the rubber aging time prediction efficiency.

[0054] In one embodiment, such as Figure 4 As shown, the target training set for nuclear power rubber materials is obtained, including:

[0055] Step S402: Obtain the aging index information corresponding to the material sample belonging to the target nuclear power rubber material type under multiple aging environment information pairs; the aging environment information pairs include radiation dose and aging temperature, and the aging index information includes preset aging time and aging index under preset aging time.

[0056] Step S404: Combine the aging index and the corresponding aging environment information to obtain training samples, and use the preset aging time corresponding to the aging index as the label aging time corresponding to the training samples.

[0057] Step S406: Based on each training sample and the corresponding label aging time, the initial training set is obtained.

[0058] Step S408: Determine the target training set from the initial training set.

[0059] The initial training set refers to the training set containing all training samples belonging to the target nuclear power rubber material type and the label aging time corresponding to each training sample. Each round of model training requires collecting multiple training samples from the initial training set to obtain the target training set.

[0060] For example, multiple material samples belonging to the target nuclear power rubber material type are placed in experimental environments corresponding to different aging environment information pairs for aging experiments. Multiple preset aging times are set. When the preset aging time is reached, the aging index corresponding to each material sample is tested. Based on the preset aging time and the aging index corresponding to the preset aging time, the corresponding aging index information is obtained. For example, the preset aging time can be set to 200h, 400h, and 600h. For each material sample, the aging index corresponding to the material sample is calculated every 200h, which can obtain multiple aging index information corresponding to each material sample under the same aging environment information pair. Computer equipment obtains the aging index information corresponding to the material samples belonging to the target nuclear power rubber material type under multiple aging environment information pairs. The aging index in the aging environment information pair and the aging index information corresponding to the material samples under the aging environment information pair are combined to obtain training samples. The preset aging time contained in the aging index information is used as the sample label corresponding to the training sample, i.e., the label aging time. Based on each training sample and the label aging time corresponding to each training sample, an initial training set is obtained. Multiple training samples are randomly collected from the initial training set, and the target training set is obtained based on each collected training sample and the corresponding label aging time.

[0061] In the above embodiments, by acquiring the aging index information corresponding to material samples belonging to the target nuclear power rubber material type under multiple aging environment information pairs, and based on the aging environment information and the corresponding aging index information, multiple training samples and the label aging time corresponding to each training sample can be quickly combined to obtain an initial training set. The rubber aging time prediction model trained based on the initial training set can predict the aging time required for the material to reach a specified aging index under the radiation dose and aging temperature indicated by the aging environment information pair, thus effectively improving the rubber aging time prediction efficiency.

[0062] In one embodiment, the target loss is obtained based on the sample error between the predicted aging time and the label aging time corresponding to the same training sample, including:

[0063] Based on the sample error between the predicted aging time and the labeled aging time corresponding to the training samples, each training sample is divided into valid samples and invalid samples; based on the sample error corresponding to each valid sample, the target loss is obtained; invalid samples are filtered in the initial training set to obtain an updated sample set; the model parameters of the initial rubber aging time prediction model are adjusted based on the target loss until the convergence condition is met, resulting in the target rubber aging time prediction model corresponding to the target nuclear power rubber material type. This includes: adjusting the model parameters of the initial rubber aging time prediction model based on the target loss to obtain an intermediate rubber aging time prediction model; using the intermediate rubber aging time prediction model as the initial rubber aging time prediction model; using the updated sample set as the initial sample set; and returning to the step of determining the target training set from the initial training set until the convergence condition is met, resulting in the target rubber aging time prediction model corresponding to the target nuclear power rubber material type.

[0064] Valid samples refer to those that promote model convergence and improve training efficiency during model training. Invalid samples refer to those that have a poor effect on promoting model convergence during model training.

[0065] For example, the computer device compares the sample error corresponding to the training samples with a preset threshold, identifying training samples with sample errors less than the preset threshold as invalid samples and training samples with sample errors greater than or equal to the preset threshold as valid samples. The sample errors corresponding to each valid sample are then fused to obtain the target loss corresponding to the target training set. Invalid samples in the initial sample set are filtered out, and an updated sample set is obtained based on the remaining training samples in the initial sample set. The model parameters of the initial rubber aging time prediction model are updated based on the target loss to obtain an intermediate rubber aging time prediction model. This intermediate rubber aging time prediction model is used as the initial rubber aging time prediction model, and the updated sample set is used as the initial sample set. The process of determining the target training set from the initial training set is repeated until the convergence condition is met, resulting in the target rubber aging time prediction model corresponding to the target nuclear power rubber material type.

[0066] In the above embodiments, during the model training process, timely filtering of invalid samples in the initial sample set whose sample error is less than a preset threshold can significantly reduce model training time and computational consumption, thereby improving model training efficiency.

[0067] In one embodiment, the method for training a rubber aging time prediction model further includes:

[0068] The number of input layer nodes is determined based on the amount of information contained in the training samples; multiple candidate hidden layer nodes are determined based on the number of input layer nodes; candidate rubber aging time prediction models are constructed corresponding to each candidate hidden layer node number; each candidate rubber aging time prediction model is trained based on the candidate training set, and the prediction accuracy of each trained candidate rubber aging time prediction model is tested based on the test set; the candidate hidden layer node number corresponding to the candidate rubber aging time prediction model with the highest prediction accuracy is taken as the target hidden layer node number; an initial rubber aging time prediction model is constructed based on the target hidden layer node number.

[0069] The information quantity refers to the number of input data corresponding to the rubber aging time prediction model, indicating the number of attributes in the sample. For example, when the model input data includes three attributes: radiation dose, aging temperature, and aging index, the information quantity is 3. The number of input layer nodes refers to the number of nodes contained in the input layer of the rubber aging time prediction model, indicating the amount of information contained in the sample. The number of hidden layer nodes refers to the number of hidden layer nodes contained in the model's hidden layer. The number of candidate hidden layer nodes refers to the number of multiple candidate hidden layer nodes. For example, such as... Figure 5 As shown, the rubber aging time prediction model includes an input layer, a hidden layer, and an output layer. When the model input data includes attribute data of three aspects: radiation dose, temperature, and compression set retention rate, the corresponding number of input layer nodes is 3 and the number of hidden layer nodes is 3.

[0070] The candidate training set refers to the training set used to train the candidate rubber aging time prediction model, including multiple training samples belonging to the target nuclear power rubber material type and training sets of labeled aging times for each training sample. The test set refers to the sample set used to test the prediction accuracy of the trained candidate rubber aging time prediction model. The target hidden layer node number refers to the optimal number of hidden layer nodes determined among the candidate hidden layer node numbers for constructing the initial rubber aging time prediction model.

[0071] For example, the computer device uses the amount of information contained in the training samples as the number of input layer nodes. The number of input layer nodes is expanded to obtain the maximum value of the number of hidden layer nodes. Based on the minimum and maximum values ​​corresponding to the number of hidden layer nodes, multiple candidate hidden layer node numbers are determined. For each candidate hidden layer node number, a corresponding candidate rubber aging time prediction model is constructed. A candidate training set corresponding to the target nuclear power rubber material type is obtained. In actual implementation, multiple training samples can be randomly collected from all training samples corresponding to the target nuclear power rubber material type. A candidate training set is obtained based on each collected training sample and its corresponding label aging time. Each candidate rubber aging time prediction model is trained based on the candidate training set. Then, each test sample in the test set is input into the trained candidate rubber aging time prediction model. The predicted aging time of each test sample output by the trained candidate rubber aging time prediction model, based on the label aging time and predicted aging time corresponding to the same test sample, yields the sample error for each test sample. Based on the test errors corresponding to each test sample, the prediction accuracy of the candidate rubber aging time prediction model is obtained. After obtaining the prediction accuracy of each candidate rubber aging time prediction model, the number of candidate hidden layer nodes corresponding to the candidate rubber aging time prediction model with the highest prediction accuracy is taken as the target number of hidden layer nodes. An initial rubber aging time prediction model is constructed based on the target number of hidden layer nodes. The initial rubber aging time prediction model is trained based on the initial training set until the model converges, thus obtaining the target rubber aging time prediction model.

[0072] In one embodiment, the maximum number of hidden layer nodes can be calculated using the following formula:

[0073] M=2N+1

[0074] Where M is the maximum number of hidden layer nodes and N is the number of input layer nodes.

[0075] In the above embodiments, multiple candidate hidden layer node numbers are determined based on the number of input layer nodes. Candidate rubber aging time prediction models are constructed corresponding to each candidate hidden layer node number. Based on the candidate training and test sets, the prediction accuracy corresponding to each candidate hidden layer node number is determined, thereby determining the optimal number of hidden layer nodes, i.e., the target number of hidden layer nodes. Constructing the initial rubber aging time prediction model based on the target number of hidden layer nodes can improve the prediction accuracy of the rubber aging time prediction model.

[0076] In one embodiment, the method for training a rubber aging time prediction model further includes:

[0077] Obtain the current aging environment information pair corresponding to the material to be predicted, which belongs to the target nuclear power rubber material type; input the current aging environment information pair and the target aging index into the target rubber aging time prediction model to obtain the predicted aging time corresponding to the material to be predicted; the target aging index is used to indicate the material's abnormality; the predicted aging time corresponding to the material to be predicted is used as the service life corresponding to the material to be predicted.

[0078] In this context, "material to be predicted" refers to materials whose aging duration needs to be predicted. "Current aging environment information" refers to the aging environment information corresponding to the material to be predicted, which can be the current or future aging environment information of the material. "Target aging index" refers to the aging index corresponding to the material to be predicted. The target aging index can characterize the critical value of the aging degree of the material to be predicted. If the current aging index of the material to be predicted is less than or equal to the target aging index, it indicates that the material to be predicted is currently normal and can function normally. If the current aging index of the material to be predicted is greater than the target aging index, it indicates that the aging degree of the material to be predicted has exceeded the critical value, meaning that the material to be predicted is abnormal and needs to be replaced promptly.

[0079] For example, a computer device acquires the aging temperature and radiation dose corresponding to the environment in which the material to be predicted, belonging to the target nuclear power rubber material type, is located, obtains the current aging environment information pair, and acquires the target aging index corresponding to the material to be predicted. The current aging information and target aging index corresponding to the material to be predicted are input into the target rubber aging time prediction model to obtain the predicted aging time corresponding to the material to be predicted. The predicted aging time is used as the service life of the material to be predicted under the current aging environment information pair.

[0080] In the above embodiments, the current aging environment information and target aging index of the material to be predicted are input into the target rubber aging time prediction model to obtain the corresponding predicted aging time, which is then used as the service life of the material to be predicted. This achieves accurate prediction of the service life of materials belonging to the target category under different aging environment information pairs.

[0081] In one embodiment, the target nuclear power rubber material type is a nuclear power rubber O-ring sealing material, the aging index is the compression set retention rate, and the method further includes:

[0082] Obtain the initial height of the current nuclear power plant rubber O-ring sealing material; obtain the compression amount obtained by compressing the current nuclear power plant rubber O-ring sealing material, and obtain the decompression amount obtained by decompressing the current nuclear power plant rubber O-ring sealing material; based on the height difference between the initial height and the decompression amount, obtain the deformation amount of the current nuclear power plant rubber O-ring sealing material; based on the ratio between the deformation amount and the compression amount, determine the compression deformation retention rate of the current nuclear power plant rubber O-ring sealing material.

[0083] Among them, rubber O-ring sealing materials for nuclear power plants refer to O-ring rubber sealing materials used in nuclear power equipment. These materials can be O-ring sealing materials made of specified materials, such as EPDM rubber. Compression deformation retention rate characterizes the amount of permanent deformation that occurs after a material is compressed to a certain height at a certain temperature and maintained for a certain period. The compression deformation retention rate is positively correlated with the degree of material aging and negatively correlated with the material's resilience.

[0084] The term "nuclear power plant rubber O-ring sealing material" refers to the nuclear power plant rubber O-ring sealing material for which compression set retention rate needs to be calculated. Initial height refers to the initial height of the current nuclear power plant rubber O-ring sealing material before compression. Current aging environment information refers to the aging environment information corresponding to the current environment in which the nuclear power plant rubber O-ring sealing material is located. Compression amount refers to the height of the current nuclear power plant rubber O-ring sealing material after compression. Decompression amount refers to the height of the current nuclear power plant rubber O-ring sealing material after it has recovered to its free state after compression. Deformation is the difference between the initial height and the decompression amount.

[0085] For example, the computer device acquires the initial height of the current nuclear power plant rubber O-ring sealing material, as well as the corresponding compression and decompression amounts. The height difference between the initial height and the decompression amount is used as the deformation. The compression deformation retention rate of the current nuclear power plant rubber O-ring sealing material is determined based on the ratio between the deformation and the compression amount. For example, the difference between a preset value and the above ratio can be used as the compression deformation retention rate.

[0086] In one embodiment, the compression set retention rate can be calculated using the following formula:

[0087]

[0088]

[0089] Where P is the compressive strength retention rate and C is the compressive strength. The initial height, For decompression height, For the limiter height, For compression amount, For deformable variables.

[0090] In the above embodiments, based on the initial height and decompression amount corresponding to the current nuclear power material, the deformation of the rubber O-ring sealing material for the current nuclear power is determined. The compression deformation retention rate determined based on the ratio between the deformation and the compression amount can accurately characterize the aging degree of the material. Using the compression deformation retention rate as an aging index can improve the accuracy of aging time prediction.

[0091] In a specific embodiment, the rubber aging time prediction model training method proposed in this application can be applied to predict the service life of rubber O-ring sealing materials used in nuclear power plants. The rubber aging time prediction model training method includes the following steps:

[0092] 1. Obtaining Samples

[0093] The sealing material was subjected to radiation of different doses and accelerated aging experiments at different temperatures. The compressive permanent deformation retention rate at different time points was calculated, resulting in multiple sets of experimental data. The computer system used radiation dose, aging temperature, and compressive permanent deformation retention rate as input data and aging time as output data. The experimental data were normalized to obtain training samples and their corresponding labels. In the actual implementation, the sealing material was subjected to gamma-ray radiation with cumulative absorbed doses of 30 kGy, 300 kGy, and 1000 kGy, followed by accelerated aging experiments at 70℃, 85℃, and 100℃. The compressive permanent deformation retention rate at different time points was calculated, resulting in 200 sets of experimental data. Some of the original data are shown in Table 1.

[0094] Table 1

[0095] serial number Radiation dose (kGy) Temperature (°C) Compression deformation retention rate Aging time (h) 1 1000 70 0.71 132 2 1000 70 0.70 153 3 1000 70 0.69 176 4 1000 70 0.68 210 5 1000 70 0.67 270 6 1000 70 0.66 494 7 1000 70 0.65 813

[0096] 2. Construct a lifespan prediction model

[0097] The computer equipment uses a lifespan prediction model to determine the approximate range of hidden layer nodes based on the number of input layer nodes. An "trial and error" method is then employed to determine the optimal number of hidden layer nodes. Specifically, the neural network is run 30 times for each of different hidden layer node numbers, with a target of 50 iterations. The minimum mean squared error (MSE) is used as the neural network's fitting accuracy. Ultimately, it is determined that 3 hidden layer nodes provide the best fitting accuracy. The MSE values ​​corresponding to different hidden layer node numbers are shown below. Figure 6As shown, the computer device then constructs a lifetime prediction model based on the optimal number of hidden layer nodes. Specifically, a lifetime prediction model based on the Levenberg-Marquardt algorithm is established, containing 3 input nodes, 1 output node, and 3 hidden layer nodes. The lifetime prediction model is a backpropagation neural network model, and the activation functions for the hidden and output layers are the tansig function (hyperbolic tangent sigmoid transfer function) and the purelin function (linear transfer function), respectively.

[0098] 3. Model Training

[0099] The computer equipment trains an initial lifespan prediction model based on training samples until the model converges, resulting in a target lifespan prediction model. In the actual model training process, such as... Figure 7 As shown, the mean square error of the service life prediction model is relatively small when the test set accounts for 10% to 20% and 60% of the total test set. However, if the test set accounts for too high a proportion, the training set will be too small, which may lead to insufficient data. Therefore, 20% was selected. 160 groups were used as the training set, and 40 groups were used as the test set. The target service life prediction model was used to predict the time required to achieve different compression set retention rates under working conditions of 1000 KGy radiation dose and 70℃. Some prediction results are shown in Table 2.

[0100] Table 2

[0101] serial number Radiation dose (kGy) Temperature (°C) Compression deformation retention rate Aging time (h) 1 1000 70 0.71 131.7 2 1000 70 0.7 150.4 3 1000 70 0.69 191.5 4 1000 70 0.68 223.23 5 1000 70 0.67 265.8 6 1000 70 0.66 467.8 7 1000 70 0.65 765

[0102] In actual implementation, the comparison results between the predicted values ​​and the actual values ​​of the target lifespan prediction model are as follows: Figure 8 As shown in the figure, the target life prediction model has high prediction accuracy. The percentage of prediction error of the target life prediction model is as follows: Figure 9 As shown, the prediction error percentage remains within 9%, indicating that the target service life prediction model has high accuracy and can be used to predict the service life of rubber O-ring sealing materials for nuclear power plants. Figure 8 , 9 In this context, C represents the compressive deformation rate.

[0103] In the above embodiments, compared with the prior art, the method for predicting the service life of rubber O-ring sealing materials for nuclear power plants can describe complex nonlinear relationships. Moreover, the method is simple, has high prediction accuracy, strong fault tolerance, is highly flexible, has no limit on the number of modeling relationships, and does not require explicit definition of specific relationships. For complex environments such as high temperatures and radiation, the nonlinear fitting results are good. By training the service life prediction model, accurate prediction of the service life of rubber O-ring sealing materials for nuclear power plants can be achieved.

[0104] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0105] Based on the same inventive concept, this application also provides a rubber aging time prediction model training device for implementing the rubber aging time prediction model training method described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the rubber aging time prediction model training device provided below can be found in the limitations of the rubber aging time prediction model training method described above, and will not be repeated here.

[0106] In one embodiment, such as Figure 10 As shown, a training device for a rubber aging time prediction model is provided, comprising: a training set acquisition module 1002, an aging time determination module 1004, a target loss determination module 1006, and a target model determination module 1008, wherein:

[0107] The training set acquisition module 1002 is used to acquire a target training set for nuclear power rubber materials. The target training set includes multiple training samples and the label aging time corresponding to each training sample. The training samples include the aging index of material samples belonging to the target nuclear power rubber material type and the aging environment information pairs corresponding to the aging index. The aging environment information pairs include radiation dose and aging temperature.

[0108] The aging time determination module 1004 is used to input each training sample into the initial rubber aging time prediction model to obtain the predicted aging time corresponding to each training sample.

[0109] The target loss determination module 1006 is used to obtain the target loss based on the sample error between the predicted aging time and the label aging time corresponding to the same training sample.

[0110] The target model determination module 1008 is used to adjust the model parameters of the initial rubber aging time prediction model based on the target loss until the convergence condition is met, so as to obtain the target rubber aging time prediction model corresponding to the target nuclear power rubber material type.

[0111] In one embodiment, the training set acquisition module 1002 is further configured to:

[0112] Acquire aging index information for material samples belonging to the target nuclear power rubber material type under multiple aging environment information pairs; the aging environment information pairs include radiation dose and aging temperature, and the aging index information includes preset aging time and aging index at the preset aging time; combine the aging index and the corresponding aging environment information pair to obtain training samples, and use the preset aging time corresponding to the aging index as the label aging time corresponding to the training sample; obtain the initial training set based on each training sample and the corresponding label aging time; determine the target training set from the initial training set.

[0113] In one embodiment, the target loss determination module 1006 is further configured to:

[0114] Based on the sample error between the predicted aging time and the labeled aging time corresponding to the training samples, each training sample is divided into valid samples and invalid samples; based on the sample error corresponding to each valid sample, the target loss is obtained; invalid samples are filtered in the initial training set to obtain an updated sample set; the model parameters of the initial rubber aging time prediction model are adjusted based on the target loss until the convergence condition is met, resulting in the target rubber aging time prediction model corresponding to the target nuclear power rubber material type. This includes: adjusting the model parameters of the initial rubber aging time prediction model based on the target loss to obtain an intermediate rubber aging time prediction model; using the intermediate rubber aging time prediction model as the initial rubber aging time prediction model; using the updated sample set as the initial sample set; and returning to the step of determining the target training set from the initial training set until the convergence condition is met, resulting in the target rubber aging time prediction model corresponding to the target nuclear power rubber material type.

[0115] In one embodiment, such as Figure 11 As shown, the rubber aging time prediction model training device also includes:

[0116] The hidden layer node number determination module 1102 is used to determine the number of input layer nodes based on the amount of information contained in the training samples; determine multiple candidate hidden layer node numbers based on the number of input layer nodes; construct candidate rubber aging time prediction models corresponding to each candidate hidden layer node number; train each candidate rubber aging time prediction model based on the candidate training set; test the prediction accuracy of each candidate rubber aging time prediction model after training based on the test set; take the candidate hidden layer node number corresponding to the candidate rubber aging time prediction model with the highest prediction accuracy as the target hidden layer node number; and construct an initial rubber aging time prediction model based on the target hidden layer node number.

[0117] The service life determination module 1104 is used to obtain the current aging environment information pair corresponding to the material to be predicted, which belongs to the target nuclear power rubber material type; input the current aging environment information pair and the target aging index into the target rubber aging time prediction model to obtain the predicted aging time corresponding to the material to be predicted; the target aging index is used to indicate that the material is abnormal; the predicted aging time corresponding to the material to be predicted is used as the service life corresponding to the material to be predicted.

[0118] The compression deformation retention rate determination module 1106 is used to obtain the initial height corresponding to the current nuclear power rubber O-ring sealing material; obtain the compression amount obtained by compressing the current nuclear power rubber O-ring sealing material; obtain the decompression amount obtained by decompressing the current nuclear power rubber O-ring sealing material; obtain the deformation amount corresponding to the current nuclear power rubber O-ring sealing material based on the height difference between the initial height and the decompression amount; and determine the compression deformation retention rate corresponding to the current nuclear power rubber O-ring sealing material based on the ratio between the deformation amount and the compression amount.

[0119] The aforementioned rubber aging time prediction model training device uses the aging index of material samples, the radiation dose corresponding to the aging index, and the aging temperature as training samples, and the aging time required for the material samples to reach the aging index under the aforementioned radiation dose and aging temperature as sample labels to obtain a target training set. Based on the target training set, the initial rubber aging time prediction model is trained to obtain a target rubber aging time prediction model corresponding to the target nuclear power rubber material type. By inputting the aging environment information pair and aging index corresponding to the material belonging to the target nuclear power rubber material type into the target rubber aging time prediction model, the aging time required for the material to reach the aging index under the radiation dose and aging temperature indicated by the aging environment information pair can be directly obtained, effectively improving the rubber aging time prediction efficiency.

[0120] Each module in the aforementioned rubber aging time prediction model training device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0121] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 12 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores target training sets, aging environment information pairs, and other data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network. When the computer program is executed by the processor, it implements a training method for a rubber aging time prediction model.

[0122] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 13As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a training method for a rubber aging time prediction model. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0123] Those skilled in the art will understand that Figure 12 , 13 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0124] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0125] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0126] In one embodiment, a computer program product or computer program is provided, the computer product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the steps in the above-described method embodiments.

[0127] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0128] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0129] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0130] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for training a rubber aging time prediction model, characterized in that, The method includes: Obtain a target training set for nuclear power rubber materials; the target training set includes multiple training samples and the label aging time corresponding to each training sample; the training samples include the aging index of material samples belonging to the target nuclear power rubber material type and the aging environment information pair corresponding to the aging index; the aging environment information pair includes radiation dose and aging temperature; The target nuclear power rubber material type is nuclear power rubber O-ring sealing material, and the aging index is the compression deformation retention rate; Obtain the initial height corresponding to the current rubber O-ring sealing material used in nuclear power plants; Obtain the compression amount obtained by compressing the current nuclear power plant rubber O-ring sealing material, and obtain the decompression amount obtained by decompressing the current nuclear power plant rubber O-ring sealing material; Based on the height difference between the initial height and the decompression amount, the deformation of the current nuclear power plant rubber O-ring sealing material is obtained; Based on the ratio between the deformation and the compression amount, the compression deformation retention rate corresponding to the current nuclear power rubber O-ring sealing material is determined; The acquisition of the target training set for nuclear power rubber materials includes: Acquire aging index information for material samples belonging to the target nuclear power rubber material type under multiple aging environment information pairs; the aging environment information pairs include radiation dose and aging temperature, and the aging index information includes preset aging time and aging index under preset aging time. The aging index and the corresponding aging environment information are combined to obtain training samples, and the preset aging time corresponding to the aging index is used as the label aging time corresponding to the training sample. The initial training set is obtained based on each training sample and its corresponding label aging time; Determine the target training set from the initial training set; Each training sample is input into the initial rubber aging time prediction model to obtain the predicted aging time corresponding to each training sample. Based on the sample error between the predicted aging time and the label aging time corresponding to the training sample, each training sample is divided into valid samples and invalid samples. The target loss is obtained based on the sample error corresponding to each valid sample; Invalid samples are filtered out from the initial training set to obtain an updated sample set; Based on the target loss, adjust the model parameters of the initial rubber aging time prediction model to obtain an intermediate rubber aging time prediction model. Use the intermediate rubber aging time prediction model as the initial rubber aging time prediction model and the updated sample set as the initial sample set. Return to the step of determining the target training set from the initial training set until the convergence condition is met to obtain the target rubber aging time prediction model corresponding to the target nuclear power rubber material type.

2. The method according to claim 1, characterized in that, The method further includes: The number of input layer nodes is determined based on the amount of information contained in the training samples. The number of candidate hidden layer nodes is determined based on the number of input layer nodes; Construct a prediction model for the aging time of candidate rubber corresponding to the number of nodes in each candidate hidden layer; Based on the candidate training set, train each candidate rubber aging time prediction model respectively, and test the prediction accuracy of each candidate rubber aging time prediction model after training based on the test set. The number of candidate hidden layer nodes corresponding to the candidate rubber aging time prediction model with the highest prediction accuracy is taken as the target number of hidden layer nodes. The initial rubber aging time prediction model is constructed based on the target number of hidden layer nodes.

3. The method according to claim 1, characterized in that, The method further includes: Obtain the current aging environment information pair corresponding to the material to be predicted that belongs to the target nuclear power rubber material type; The current aging environment information and the target aging index are input into the target rubber aging time prediction model to obtain the predicted aging time of the material to be predicted; the target aging index is used to indicate that the material is abnormal. The predicted aging time of the material to be predicted is taken as the service life of the material to be predicted.

4. A training device for a rubber aging time prediction model, characterized in that, The device includes: The training set acquisition module is used to acquire a target training set for nuclear power rubber materials. The target training set includes multiple training samples and the label aging time corresponding to each training sample. The training samples include the aging index of material samples belonging to the target nuclear power rubber material type and the aging environment information pair corresponding to the aging index. The aging environment information pair includes radiation dose and aging temperature. The training set acquisition module is also used for: The target nuclear power rubber material type is nuclear power rubber O-ring sealing material, and the aging index is the compression deformation retention rate; Obtain the initial height corresponding to the current rubber O-ring sealing material used in nuclear power plants; Obtain the compression amount obtained by compressing the current nuclear power plant rubber O-ring sealing material, and obtain the decompression amount obtained by decompressing the current nuclear power plant rubber O-ring sealing material; Based on the height difference between the initial height and the decompression amount, the deformation of the current nuclear power plant rubber O-ring sealing material is obtained; Based on the ratio between the deformation and the compression amount, the compression deformation retention rate corresponding to the current nuclear power rubber O-ring sealing material is determined; The acquisition of the target training set for nuclear power rubber materials includes: Acquire aging index information for material samples belonging to the target nuclear power rubber material type under multiple aging environment information pairs; the aging environment information pairs include radiation dose and aging temperature, and the aging index information includes preset aging time and aging index under preset aging time. The aging index and the corresponding aging environment information are combined to obtain training samples, and the preset aging time corresponding to the aging index is used as the label aging time corresponding to the training sample. The initial training set is obtained based on each training sample and its corresponding label aging time; Determine the target training set from the initial training set; The aging time determination module is used to input each training sample into the initial rubber aging time prediction model to obtain the predicted aging time corresponding to each training sample. The target loss determination module is used to divide each training sample into valid samples and invalid samples based on the sample error between the predicted aging time and the labeled aging time corresponding to the training samples; obtain the target loss based on the sample error corresponding to each valid sample; filter invalid samples in the initial training set to obtain an updated sample set; adjust the model parameters of the initial rubber aging time prediction model based on the target loss to obtain an intermediate rubber aging time prediction model; use the intermediate rubber aging time prediction model as the initial rubber aging time prediction model; use the updated sample set as the initial sample set; return to the step of determining the target training set from the initial training set; continue until the convergence condition is met to obtain the target rubber aging time prediction model corresponding to the target nuclear power rubber material type.

5. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 3.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 3.

7. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 3.

Images