Method for predicting the service life of a rubber core

Abstract

The embodiment of the application provides a method for predicting the service life of a rubber core. The method comprises: obtaining measurement information of a rubber core to be predicted in a current environment; wherein the measurement information comprises current temperature information, stress information, and strain information; determining strain energy information of the rubber core to be predicted according to the stress information and the strain information; and determining service life information of the rubber core to be predicted based on a preset service life prediction model according to the temperature information and the strain energy information; wherein the preset service life prediction model is a pre-trained neural network model used to output the service life information. The method can improve the accuracy of the prediction of the service life of the rubber core.

Description

Technical Field

[0001] This application relates to the fields of oil and gas wells and blowout preventers, and in particular to a method for predicting the service life of a rubber core. Background Technology

[0002] Blowout preventers (BOPs) are important equipment used in oil and gas well drilling to control blowouts and prevent blowout accidents. The rubber core of the BOP is a key component that directly affects the sealing performance and service life of the BOP.

[0003] However, existing methods for predicting the service life of rubber cores are rather crude, generally relying on manual judgment based on experience. This cannot accurately reflect the fatigue damage of the rubber cores under actual working conditions, resulting in a lack of scientific basis for maintenance and replacement work, and increasing the risk of accidents and maintenance costs.

[0004] Therefore, there is a problem with the inaccurate prediction of the service life of the blowout preventer core. Summary of the Invention

[0005] This application provides a method for predicting the service life of a blowout preventer core, which can improve the accuracy of predicting the service life of the blowout preventer core.

[0006] In a first aspect, embodiments of this application provide a method for predicting the service life of a rubber core, comprising:

[0007] Obtain measurement information of the rubber core to be predicted under the current environment; the measurement information includes the current temperature information, stress information, and strain information;

[0008] Based on stress and strain information, determine the strain energy information of the core to be predicted;

[0009] Based on temperature and strain energy information, and using a pre-set life prediction model, the lifespan information of the rubber core to be predicted is determined. The pre-set life prediction model is a pre-trained neural network model used to output lifespan information.

[0010] In one possible implementation, based on temperature information and strain energy information, and using a preset life prediction model, the lifespan information of the rubber core to be predicted is determined, including:

[0011] Temperature and strain energy information are input into a preset lifetime prediction model;

[0012] By using the network layer in the preset lifetime prediction model, the temperature information and strain energy information are processed into feature vectors respectively to obtain the feature vector data corresponding to the temperature information and the feature vector data corresponding to the strain energy information.

[0013] Based on the eigenvector data corresponding to temperature information and the eigenvector data corresponding to strain energy information, the service life information of the rubber core to be predicted is determined.

[0014] In one possible implementation, the service life information of the rubber core to be predicted is determined based on the feature vector data corresponding to the temperature information and the feature vector data corresponding to the strain energy information, including:

[0015] The feature vector data corresponding to temperature information and the feature vector data corresponding to strain energy information are fused to obtain fused vector data.

[0016] Based on the fused vector data, the lifespan information of the rubber core to be predicted is determined.

[0017] In one possible implementation, the service life information of the rubber core to be predicted is determined based on the feature vector data corresponding to the temperature information and the feature vector data corresponding to the strain energy information, including:

[0018] Based on the feature vector data corresponding to the temperature information, the first service life of the rubber core to be predicted is determined; whereby the first service life characterizes the service life information of the rubber core under the temperature information.

[0019] Based on the feature vector data corresponding to the strain energy information, the second service life of the rubber core to be predicted is determined; wherein, the second service life characterizes the service life information of the rubber core under the strain energy information.

[0020] Based on the first service life and the second service life, the service life information of the rubber core to be predicted is determined.

[0021] In one possible implementation, the above method further includes:

[0022] Obtain a pre-collected training dataset; wherein the training dataset includes training temperature information, training strain energy information, a first label, and a second label. The first label represents the actual service life of the training rubber core under the training temperature information, and the second label represents the actual service life of the training rubber core under the training strain energy information.

[0023] Based on the dataset to be trained, the preset initial model is trained to obtain the trained life prediction model.

[0024] In one possible implementation, obtaining a pre-collected training dataset includes:

[0025] Obtain the first dataset and the second dataset; wherein, the first dataset includes the temperature information to be trained, the stress information to be trained, and the number of experimental cycles, and the second dataset includes the temperature information to be trained, the stress information to be trained, and the strain information to be trained, and the number of experimental cycles represents the number of times the training core is subjected to the preset experiment under the stress information to be trained.

[0026] Based on the first dataset, determine the first label corresponding to the temperature information to be trained;

[0027] Based on the second dataset, determine the strain energy information and second label to be trained.

[0028] In one possible implementation, determining a first label corresponding to the temperature information to be trained based on a first dataset includes:

[0029] Based on the temperature information to be trained in the first dataset, the stress information to be trained, and the number of experimental cycles, a first curve corresponding to the temperature information to be trained in the first dataset is determined; wherein, the first curve represents the relationship between the stress information to be trained and the number of experimental cycles.

[0030] The first curve is processed by Weibull distribution to obtain the first label corresponding to the temperature information to be trained in the first dataset.

[0031] In one possible implementation, based on a second dataset, the strain energy information to be trained and a second label are determined, including:

[0032] Based on the temperature information to be trained in the second dataset, the stress information to be trained and the strain information to be trained corresponding to the temperature information to be trained in the second dataset are determined.

[0033] Based on the temperature information to be trained in the second dataset, corresponding to the stress information and strain information to be trained, a second curve corresponding to the temperature information to be trained in the second dataset is determined; wherein, the second curve represents the relationship between the stress information and strain information to be trained.

[0034] Integrating the second curve yields the second label corresponding to the strain energy information to be trained.

[0035] In one possible implementation, a pre-defined initial model is trained based on the dataset to be trained to obtain a trained lifespan prediction model, including:

[0036] The target service life is determined based on the actual service life represented by the first label and the actual service life represented by the second label;

[0037] The temperature information and strain energy information to be trained are input into the preset initial model to obtain the predicted service life.

[0038] The Pearson correlation coefficient between the target service life and the predicted service life is determined and given as the target coefficient.

[0039] If the target coefficients meet the preset training completion conditions, then the trained life prediction model is obtained.

[0040] In one possible implementation, a pre-defined initial model is trained based on the dataset to be trained to obtain a trained lifespan prediction model, including:

[0041] The temperature information and strain energy information to be trained are input into the preset initial model to obtain the first predicted lifetime corresponding to the temperature information to be trained and the second predicted lifetime corresponding to the strain energy information to be trained.

[0042] The Pearson correlation coefficient between the actual service life represented by the first label and the first predicted service life is determined as the first coefficient, and the Pearson correlation coefficient between the actual service life represented by the second label and the second predicted service life is determined as the second coefficient;

[0043] If both the first and second coefficients meet the preset training completion conditions, then the trained life prediction model is obtained.

[0044] Secondly, embodiments of this application provide a device for predicting the service life of a rubber core, comprising:

[0045] The acquisition module is used to acquire measurement information of the rubber core to be predicted under the current environment; the measurement information includes the current temperature information, stress information, and strain information.

[0046] The module determines the strain energy information of the core to be predicted based on stress and strain information.

[0047] The prediction module is used to determine the service life information of the rubber core to be predicted based on temperature information and strain energy information and a preset service life prediction model; wherein, the preset service life prediction model is a pre-trained neural network model used to output service life information.

[0048] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;

[0049] The memory stores the instructions that the computer executes;

[0050] The processor executes computer execution instructions stored in memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0051] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0052] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0053] The method for predicting the service life of a rubber core provided in this application considers the temperature and strain energy information of the environment in which the rubber core is located. Based on a preset service life prediction model, it determines the predicted service life information of the rubber core. The strain energy information can be obtained from stress and strain information. Furthermore, the pre-trained service life prediction model ensures the reliability of the service life prediction, improves the efficiency of service life prediction, and saves manpower and time. The method for predicting the service life of a rubber core provided in this application improves the accuracy of the service life prediction. Attached Figure Description

[0054] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0055] Figure 1 This is a schematic diagram of the adhesive core provided in this application;

[0056] Figure 2 A flowchart illustrating a method for predicting the service life of a rubber core provided in this application;

[0057] Figure 3 A flowchart illustrating another method for predicting the service life of a rubber core provided in this application;

[0058] Figure 4 A flowchart illustrating another method for predicting the service life of a rubber core provided in this application;

[0059] Figure 5 A schematic diagram of the temperature-lifetime curve provided for this application;

[0060] Figure 6 A flowchart illustrating another method for predicting the service life of a rubber core provided in this application;

[0061] Figure 7 A flowchart illustrating a method for predicting the service life of a rubber core, as provided in this application.

[0062] Figure 8A schematic diagram of the structure of the device for predicting the service life of the rubber core provided in this application;

[0063] Figure 9 A schematic diagram of the structure of the electronic device provided in this application.

[0064] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0065] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0066] First, let me explain the terms used in this application:

[0067] A blowout preventer is a critical safety device used in oil and gas drilling operations. Its main function is to prevent blowout accidents, that is, to control the sudden outflow of high-pressure fluids (such as oil, gas, water, etc.) from the well to the surface during the drilling process, so as to prevent danger and loss.

[0068] The rubber core is a key sealing component widely used in various mechanical and industrial equipment. In blowout preventers (BOPs) used in oil and gas drilling operations, the BOP core is typically made of high-strength rubber or other elastic materials, possessing excellent elasticity, pressure resistance, and abrasion resistance.

[0069] Stress refers to the internal force exerted on a product per unit area. The unit of stress is Pascal (Pa), which can also be expressed in Newtons per square millimeter (N / mm²). 2 Stress, expressed as megapascal (MPa), describes the internal mechanical state of a product material under the action of external forces.

[0070] Strain refers to the degree of deformation of a material under the action of external force. Strain is a dimensionless quantity, that is, it has no unit. Strain describes the deformation of a material.

[0071] Service life refers to the length of time a product, equipment, or system can maintain its expected function and performance under normal operating conditions. In this application, service life may also be referred to as lifespan.

[0072] A neural network model is a computational model that simulates the structure and function of biological neural networks. It is widely used in the fields of artificial intelligence and machine learning. It consists of a large number of artificial neurons (also known as nodes or units), which are interconnected through connections (also known as weights) to form a complex network structure.

[0073] The network layer of a neural network model is the basic building block of the model. Each layer consists of several neurons (nodes). Common network layers may include, but are not limited to, the input layer, the hidden layer, and the output layer.

[0074] Figure 1 A schematic diagram of the adhesive core provided in this application is shown below. Figure 1 As shown, the rubber core is a conical rubber core, and the material of the rubber core can be high-temperature silicone rubber. In an optional embodiment, the material of the rubber core can also be an acid and alkali resistant polymer material, which is not limited here.

[0075] In actual operation, for blowout preventers, the top cover-core assembly mainly consists of a conical rubber core and a metal support, with surface contact formed between the conical rubber core and the metal support.

[0076] In one possible implementation, the blowout preventer of this application can be an annular blowout preventer.

[0077] In one possible implementation, the blowout preventer's core includes a heat-conducting pipe, which can be used to ensure uniform temperature of the core.

[0078] In one possible implementation, the blowout preventer core is equipped with at least a temperature sensor, a stress sensor, and a strain sensor to acquire temperature, stress, and strain information of the environment in which the core is located.

[0079] As we know, blowout preventers (BOPs) are a key safety device used in oil and gas drilling operations to prevent blowout accidents. The rubber core, as an important sealing component of the BOP, ensures the normal operation of the BOP. If the lifespan of the rubber core is predicted and it is replaced regularly, blowout accidents will be effectively prevented, and the safety of workers and equipment will be protected.

[0080] However, in the existing technology, the prediction of the service life of the rubber core is usually based on manual judgment based on experience, which lacks a certain degree of reliability and accuracy.

[0081] This application provides a method, apparatus, device, and medium for predicting the service life of a rubber core. By considering the temperature and strain energy information of the environment in which the rubber core is located, and based on a preset service life prediction model, the predicted service life information of the rubber core is determined. The strain energy information can be obtained through stress and strain information, and the service life prediction model is obtained using preset experimental data, ensuring the reliability of the service life prediction. The method for predicting the service life of a rubber core provided in this application improves the accuracy of the service life prediction.

[0082] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0083] Figure 2 This application provides a flowchart illustrating a method for predicting the service life of a rubber core. The execution entity of this method can be a host, server, or other device, such as... Figure 2 As shown, the method includes:

[0084] S201. Obtain measurement information of the rubber core to be predicted under the current environment; wherein, the measurement information includes the current temperature information, stress information, and strain information.

[0085] Among them, the rubber core to be predicted can refer to the rubber core whose service life is to be predicted, and the current environment can refer to the environment in which the rubber core is installed in the wellhead blowout preventer.

[0086] Temperature information refers to the temperature data of the rubber core in the current environment. Specifically, the unit of temperature data can be ℃, and the temperature data can be obtained in real time through devices such as temperature sensors or thermocouples.

[0087] Stress information can refer to the force or pressure data acting on the rubber core. Specifically, the unit of pressure data can be MPa. Stress is usually measured by devices such as stress sensors or strain gauges.

[0088] Strain information refers to the deformation data of the rubber core under stress; specifically, it characterizes the degree of deformation of the rubber core under stress. Strain information generally has no units and can be obtained through strain gauges or other strain measurement devices.

[0089] For example, at a specific moment, the measurement information of the rubber core to be predicted under the current environment may include: temperature information of 100℃, stress information of 300MPa, and strain information of 0.003.

[0090] S202. Based on the stress and strain information, determine the strain energy information of the rubber core to be predicted.

[0091] Among them, strain energy information can refer to the strain energy data of the rubber core. Strain energy can refer to the energy stored in the material during deformation. For elastic materials, strain energy can be calculated through the relationship between stress and strain.

[0092] The strain energy information of the rubber core to be predicted can be obtained from the stress and strain information of the rubber core. The strain energy information can be expressed as strain energy.

[0093] In one possible implementation, the strain energy information of the core to be predicted can be obtained by integration, that is, by integrating the stress-strain curve. Numerical integration methods such as the trapezoidal method or Simpson's method can be used to calculate the area between the stress-strain curve and the horizontal axis. This area is the strain energy density. Multiplying the strain energy density by the area or volume of strain borne by the product gives the strain energy of the product.

[0094] The stress-strain curve can be obtained by fitting a pre-collected stress-strain dataset, with the horizontal axis representing strain and the vertical axis representing stress. The data can be collected from historical stress and strain sensor data.

[0095] In one possible implementation, the strain energy information of the rubber core to be predicted can also be calculated using the expression of the Yeoh constitutive model, which satisfies:

[0096]

[0097] Where U is the strain energy density, multiplying the strain energy density by the stress-bearing area or volume of the core to be predicted yields the strain energy of the core to be predicted. 10 C 20 and C 30 All of these are material constants for the rubber core to be predicted, where C 10 In the Yeoh constitutive model, C is used to ensure that the initial shear modulus of the core material is positive. 20 Used to reflect the softness of rubber materials under moderate deformation, C 30 Used to describe the hardening effect of rubber materials under large deformation.

[0098] It should be noted that the method for determining the strain energy information of the rubber core to be predicted in this embodiment is not limited, as long as the strain energy information of the rubber core to be predicted can be obtained.

[0099] S203. Based on the temperature information and strain energy information, determine the service life information of the rubber core to be predicted based on the preset service life prediction model; wherein, the preset service life prediction model is a pre-trained neural network model used to output service life information.

[0100] Among them, the neural network model is a type of machine learning model that mimics the workings of neurons in the human brain. Through training with a large amount of data, it can learn from input data and predict output results.

[0101] The preset lifespan prediction model is a neural network model pre-trained by the staff, that is, a preset prediction model for the lifespan of the rubber core.

[0102] It is understandable that by inputting temperature and strain energy information into a preset life prediction model, the life prediction model can output the lifespan information of the rubber core.

[0103] Service life information can be understood as the time data of the rubber core when it reaches the failure boundary.

[0104] In an optional implementation, the method in S203 described above may include:

[0105] Temperature and strain energy information are input into a preset life prediction model. The network layer in the preset life prediction model performs feature vector processing on the temperature and strain energy information respectively to obtain the feature vector data corresponding to the temperature information and the feature vector data corresponding to the strain energy information. Based on the feature vector data corresponding to the temperature information and the feature vector data corresponding to the strain energy information, the lifespan information of the rubber core to be predicted is determined.

[0106] Among them, the lifespan information of the rubber core to be predicted is the lifespan data of the rubber core to be predicted, which is usually in time units.

[0107] In the fields of machine learning and deep learning, neural networks typically consist of multiple layers, each containing several neurons. These layers can be categorized into three main types: input layers, hidden layers, and output layers. Specifically, the input layer typically receives the raw input data, the hidden layers process and extract features from the data, and the output layer generates the final prediction result.

[0108] In this application, the network layer may include at least one of the above-mentioned input layer, hidden layer, and output layer. In the prediction model of the service life of the rubber core, the network layer extracts and comprehensively analyzes the features of temperature information and strain energy information, and finally outputs the prediction result of the service life of the rubber core.

[0109] Specifically, the network layer performs feature vector processing on the temperature information to obtain the feature vector data corresponding to the temperature information.

[0110] For example, temperature information is input into the input layer, which then passes the temperature information to the hidden layer for further processing. The hidden layer performs a nonlinear transformation on the temperature information, converting it into a feature vector, which is the feature vector data corresponding to the temperature information.

[0111] In this context, nonlinear transformation can be understood as obtaining it using a nonlinear function. The nonlinear function can include, but is not limited to, the ReLU function, the Sigmoid function, and the Tanh function.

[0112] Specifically, the network layer also performs feature vector processing on the strain energy information to obtain the feature vector data corresponding to the strain energy information.

[0113] For example, strain energy information is input into the input layer, and the input layer passes the temperature information to the hidden layer for further processing. The hidden layer performs a nonlinear transformation on the strain energy information and converts it into a feature vector, which is the feature vector data corresponding to the temperature information.

[0114] The advantage of this setup is that it considers the impact of both temperature and strain energy on the lifespan of the rubber core, which helps improve the reliability and accuracy of the prediction method for the lifespan of the rubber core.

[0115] In an optional implementation, determining the lifespan information of the rubber core to be predicted based on the feature vector data corresponding to the temperature information and the feature vector data corresponding to the strain energy information may include:

[0116] The feature vector data corresponding to temperature information and the feature vector data corresponding to strain energy information are fused to obtain fused vector data; based on the fused vector data, the service life information of the rubber core to be predicted is determined.

[0117] The fusion process refers to combining feature vector data from different sources to comprehensively consider the impact of multiple factors on the service life of the rubber core. In this application, the fusion process can be achieved by merging the feature vector data corresponding to temperature information and the feature vector data corresponding to strain energy information to obtain merged fused vector data.

[0118] In one possible implementation, the fusion process may include, but is not limited to, simple splicing, weighted averaging, and fusion layer processing.

[0119] Simple concatenation can be achieved by directly concatenating the temperature feature vector and the strain energy feature vector into a longer vector; weighted averaging can be performed by weighting the feature vectors according to the importance of the temperature and strain energy information; fusion layer processing can be achieved by using fusion layers in neural networks (such as fully connected layers, convolutional layers, etc.) to automatically learn how to optimally fuse these feature vectors. The implementation method of fusion processing is not limited here.

[0120] The beneficial effect of the above settings is that the fusion process can comprehensively consider the impact of various environmental factors on the service life of the rubber core, thereby providing more accurate prediction results.

[0121] In an optional implementation, the process of determining the lifespan information of the rubber core to be predicted based on the feature vector data corresponding to the temperature information and the feature vector data corresponding to the strain energy information may further include:

[0122] Based on the feature vector data corresponding to the temperature information, the first service life of the rubber core to be predicted is determined; wherein, the first service life characterizes the service life information of the rubber core under the temperature information; based on the feature vector data corresponding to the strain energy information, the second service life of the rubber core to be predicted is determined; wherein, the second service life characterizes the service life information of the rubber core under the strain energy information; based on the first service life and the second service life, the service life information of the rubber core to be predicted is determined.

[0123] The first service life can refer to the first service life value obtained by processing the feature vector data corresponding to the temperature information, and the second service life can refer to the second service life value obtained by processing the feature vector data corresponding to the strain energy information.

[0124] In one possible implementation, the life prediction model may include two sub-models: the first sub-model is used to determine the first lifespan of the rubber core to be predicted based on the feature vector data corresponding to the temperature information; the second sub-model is used to determine the second lifespan of the rubber core to be predicted based on the feature vector data corresponding to the strain energy information.

[0125] For example, the life prediction model inputs the feature vector data corresponding to the temperature information into the first sub-model and outputs the first lifespan; the life prediction model inputs the feature vector data corresponding to the strain energy information into the second sub-model and outputs the second lifespan.

[0126] In one possible implementation, the method for determining the lifespan information of the rubber core to be predicted based on the first lifespan and the second lifespan may include, but is not limited to, the weighted average method, the minimum value method, and the maximum value method.

[0127] The weighted average method can be used to calculate a weighted average value as the comprehensive service life by assigning different weights to the first service life and the second service life. This comprehensive service life is the service life information of the rubber core to be predicted. For example, the weight ratio of the first service life and the second service life is 6:4. The minimum value method can be used to take the minimum value between the first service life and the second service life as the comprehensive service life. This comprehensive service life is the service life information of the rubber core to be predicted. The minimum value method is suitable for cases where the rubber core fails when it reaches its limit under any condition. The maximum value method can be used to take the maximum value between the first service life and the second service life as the comprehensive service life. This comprehensive service life is the service life information of the rubber core to be predicted. The maximum value method is suitable for cases where the rubber core fails when it reaches its limit under all conditions.

[0128] In real-world scenarios, staff can choose one of the following methods—weighted average, minimum, or maximum—based on the environment in which the rubber core is located. This method determines the lifespan information of the rubber core to be predicted based on the first and second lifespans. The implementation method of this method is not limited here.

[0129] The advantage of this setup is that by separately considering the effects of temperature and strain energy on the lifespan of the rubber core, the actual lifespan of the core can be predicted more accurately. Each sub-model focuses on processing a specific type of information, thereby improving the accuracy and reliability of the prediction. The final comprehensive lifespan information provides a more complete prediction result, taking into account all relevant factors.

[0130] It should be noted that temperature and strain energy information are not the only factors that determine the service life of the rubber core to be predicted. If other information exists, it can also affect the service life of the rubber core and is also applicable to the above-mentioned method for predicting the service life of the rubber core.

[0131] This application provides a method for predicting the service life of a rubber core. By combining temperature, stress, and strain information under the current environment, a neural network model is used to predict the service life of the rubber core, thereby improving the accuracy and reliability of the prediction.

[0132] Figure 3 A flowchart illustrating another method for predicting the service life of a rubber core provided in this application is shown. The execution subject of this method can be a host, server, or other device, such as... Figure 3 As shown, the method includes:

[0133] S301. Obtain the pre-collected training dataset; wherein, the training dataset includes the temperature information to be trained, the strain energy information to be trained, the first label, and the second label, the first label representing the actual service life of the rubber core under the temperature information to be trained, and the second label representing the actual service life of the rubber core under the strain energy information to be trained.

[0134] The pre-collected training dataset can be sourced from experiments, historical records, or sensor monitoring data. The training dataset includes temperature information to be trained, a first label corresponding to the temperature information, strain energy information to be trained, and a second label corresponding to the strain energy information. This training dataset is crucial for obtaining the lifespan prediction model.

[0135] In machine learning and data science, a label can refer to the target output or result of each sample in a dataset, or it can be the value that the model needs to predict. Labels can be divided into classification labels and regression labels. In this application, both the first label and the second label are regression labels, meaning that both the first label and the second label are numerical values ​​representing the lifespan of the training core.

[0136] Actual service life refers to the length of time a rubber core can function normally or maintain its function under specific environmental conditions. The actual service life of a rubber core under training temperature information refers to the length of time a rubber core can function normally or maintain its function under specific temperature conditions; the actual service life of a rubber core under training strain energy information refers to the length of time a rubber core can function normally or maintain its function under specific strain energy conditions.

[0137] In this application, the first label indicates the actual service life of the rubber core under specific temperature information. The second label indicates the actual service life of the rubber core under specific strain energy information.

[0138] For example, the actual service life under the temperature information to be trained can be expressed as follows: when the temperature information to be trained is 23℃, 70℃, 80℃, 90℃, and 100℃, the actual service life under the temperature information to be trained is 4.92 months, 5.17 months, 4.89 months, 3.96 months, and 3.24 months, respectively.

[0139] For example, the actual service life under the strain energy information to be trained can be expressed as follows: when the strain energy information to be trained is 0.5 J / m 2 0.7J / m 2 0.9J / m 2 1.1J / m 2 At that time, the actual service life under the strain energy information to be trained was 6.21 months, 5.29 months, 4.92 months, and 4.45 months.

[0140] S302. Based on the dataset to be trained, train the preset initial model to obtain the trained life prediction model.

[0141] The preset initial model can be understood as model initialization, that is, initializing the model parameters according to the preset initial model structure. For example, a linear regression model is initialized, which can be expressed as: service life = a*temperature + b*strain energy + c, where a, b, and c are the parameters of the preset initial model, and this linear regression model is the preset initial model.

[0142] For example, the temperature information to be trained, the strain energy information to be trained, the first label, and the second label in the training dataset are the aforementioned {23℃, 70℃, 80℃, 90℃, 100℃} and {0.5J / m 2 0.7J / m 2 0.9J / m 2 1.1J / m 2 {4.92 months, 5.17 months, 4.89 months, 3.96 months, 3.24 months}, {6.21 months, 5.29 months, 4.92 months, 4.45 months}, take the temperature information and strain energy information of the dataset to be trained as the input of the preset initial model, and take the first label and the second label as the corresponding output of the input of the preset initial model to train the preset initial model.

[0143] During training, parameters a, b, and c can be adjusted using methods such as gradient descent to minimize the error between the model's predicted lifetime value and the actual label values ​​(first label and second label). After model training, parameters a, b, and c are adaptively adjusted to A, B, and C, resulting in the trained lifetime prediction model. The structure of the trained lifetime prediction model can be expressed as: Lifetime = A * Temperature + B * Strain Energy + C, where A, B, and C are the parameters of the trained lifetime prediction model.

[0144] If temperature information and / or strain energy information are input into the trained life prediction model, the predicted life value can be obtained, thereby enabling the prediction of the life of the rubber core.

[0145] The alternative method for predicting the service life of the rubber core provided in this application comprehensively considers temperature and strain energy information and uses actual collected data for training, which significantly improves the accuracy and reliability of the service life prediction model.

[0146] Figure 4 This application provides a flowchart illustrating another method for predicting the service life of a rubber core. The execution entity of this method can be a host, server, or other device, such as... Figure 4 As shown, the method includes:

[0147] S401. Obtain the first dataset and the second dataset; wherein, the first dataset includes the temperature information to be trained, the stress information to be trained, and the number of experimental cycles, and the second dataset includes the temperature information to be trained, the stress information to be trained, and the strain information to be trained, and the number of experimental cycles represents the number of times the training core is subjected to the preset experiment under the stress information to be trained.

[0148] The first dataset and the second dataset are used to determine the first label and the second label, respectively. The meaning of the first label and the second label here is the same as that of S301 above, and will not be repeated here.

[0149] In the first dataset, the number of experimental cycles can refer to the number of preset experiments that the core or core material of a product can withstand under specific stress conditions. For example, a preset experiment can be understood as applying a specific stress to the core at fixed time intervals, and when the core just reaches the point of failure (i.e. unusable), the number of cycles in which the specific stress is applied to the core is the number of experimental cycles.

[0150] S402. Based on the first dataset, determine the first label corresponding to the temperature information to be trained.

[0151] The first dataset includes the temperature information to be trained, the stress information to be trained, and the number of experimental cycles. It can be understood that, given the temperature information to be trained, the stress information to be trained, and the number of experimental cycles, the first label corresponding to the temperature information to be trained can be calculated.

[0152] In an optional implementation, determining the first label corresponding to the temperature information to be trained based on the first dataset may include:

[0153] Based on the temperature information to be trained in the first dataset, the stress information to be trained, and the number of experimental cycles, the first curve corresponding to the temperature information to be trained in the first dataset is determined; wherein, the first curve represents the relationship between the stress information to be trained and the number of experimental cycles; the first curve is processed by Weibull distribution to obtain the first label corresponding to the temperature information to be trained in the first dataset.

[0154] The first curve can be an SN (stress-cycle number) curve, and there will be a corresponding first curve for each temperature information.

[0155] For example, at a temperature of 23°C, in the first curve, the stress is used as the horizontal axis of the first curve and the number of cycles is used as the vertical axis of the first curve. The training stress information and the number of experimental cycles corresponding to a temperature of 23°C are plotted on the coordinate graph, and the points on the coordinate axes are smoothly connected to obtain the first curve.

[0156] Among them, Weibull distribution processing can refer to using the statistical distribution method of Weibull distribution to process and obtain the actual service life corresponding to the temperature information to be trained in the first dataset.

[0157] In the Weibull distribution method, the lifetime value can be calculated by determining only the shape and scale parameters of the Weibull distribution. In some possible embodiments, the lifetime value calculated by the Weibull distribution method can also be called the fatigue time value.

[0158] The shape parameter of the Weibull distribution describes its shape. If the shape parameter is less than 1, it indicates that the product failure rate decreases over time; if the shape parameter is equal to 1, it indicates that the product failure rate remains constant over time; and if the shape parameter is greater than 1, it indicates that the product failure rate increases over time. The failure rate refers to the instantaneous probability of product failure at a given moment.

[0159] The scaling parameter of the Weibull distribution describes the scale of the distribution. Often called the characteristic lifetime or scaling factor, it represents a characteristic value of time or cycle number within the Weibull distribution. Specifically, the scaling parameter determines the horizontal position of the distribution. A larger scaling parameter indicates a longer product lifetime, while a smaller scaling parameter indicates a shorter product lifetime.

[0160] In this embodiment, the product described above can be understood as the rubber core of a blowout preventer.

[0161] In one alternative approach, determining the shape and scale parameters of the Weibull distribution may include:

[0162] The SN curves are converted to obtain the failure time dataset; based on the failure time dataset, the shape and scale parameters of the Weibull distribution are calculated.

[0163] The conversion process can refer to selecting p points from the SN curve at equal intervals to obtain the experimental cycle count dataset, where the size of the cycle count dataset is p; and multiplying all elements in the cycle count dataset by the time of a single cycle of the experiment to obtain the failure time dataset.

[0164] For example, if 5 points are selected from the SN curve at the same interval, the number of cycles dataset is {1000 times, 1600 times, 1800 times, 1900 times, 1950 times}, and the single cycle time is 2 seconds, then the failure time dataset is {2000 seconds, 3200 seconds, 3600 seconds, 3800 seconds, 3900 seconds}.

[0165] Based on the failure time dataset, the shape parameters of the Weibull distribution are calculated. The failure time dataset can then be linearly fitted using the least squares method to obtain quantile points. These quantile points are then linearly fitted using the least squares method to obtain a straight line. The slope of this line is the shape parameter of the Weibull distribution. Here, a quantile point refers to a point in the statistical distribution that corresponds to a specific cumulative probability value.

[0166] In one possible implementation, software (such as MATLAB) can be used to perform the above method of calculating the shape parameters of the Weibull distribution based on the failure time dataset.

[0167] Based on the failure time dataset, the scaling parameter of the Weibull distribution is calculated, which satisfies the formula:

[0168]

[0169] Among them, K t Let C be the scale parameter of the Weibull distribution. t Let p be the shape parameter of the Weibull distribution, and p be the failure time dataset.

[0170] Therefore, given the shape and scale parameters of the Weibull distribution, the lifetime value can be calculated, and the lifetime value satisfies the formula:

[0171] n1 = K t ·Gamma(1+1 / C t ),

[0172] Where n1 is the lifetime value, K t Let C be the scale parameter of the Weibull distribution. t Let be the shape parameter of the Weibull distribution.

[0173] For example, when the temperature is 23℃, there is a corresponding SN curve. Based on the SN curve, using the least squares method in linear regression, the shape parameter of the Weibull distribution is obtained as 48.29. Then, the scale parameter of the Weibull distribution is calculated as 4.98 using the above formula. Substituting the shape parameter and scale parameter of the Weibull distribution into the formula for the lifetime value, the lifetime value is obtained as 4.92. This lifetime value is the first label corresponding to the temperature information of 23℃.

[0174] The advantage of using the Weibull distribution to obtain the first label is that it can adapt to the failure modes of different types of materials by adjusting the shape and scale parameters. By fitting experimental data and calculating the parameters of the Weibull distribution, the lifespan of materials or products can be predicted more accurately.

[0175] S403. Based on the second dataset, determine the strain energy information and second label to be trained.

[0176] The second dataset includes temperature information, stress information, and strain information to be trained. It can be understood that, given the temperature information, stress information, and strain information to be trained, the strain energy information and the second label corresponding to the strain energy information to be trained can be calculated.

[0177] In one alternative implementation, determining the strain energy information and the second label to be trained based on the second dataset may include:

[0178] Based on the training stress and strain information corresponding to the training temperature information in the second dataset, the training strain energy information corresponding to the training temperature information in the second dataset is determined; based on the training stress and strain information corresponding to the training temperature information in the second dataset, the second curve corresponding to the training temperature information in the second dataset is determined; wherein, the second curve represents the relationship between the training stress information and the training strain information; the second curve is integrated to obtain the second label corresponding to the training strain energy information.

[0179] In this context, it can be understood that for a given temperature information, there exists at least one stress information and one strain information to be trained, where stress information and strain information have a corresponding relationship, with one stress value corresponding to one strain value.

[0180] In one possible implementation, the strain energy satisfies the formula:

[0181]

[0182] Where W is the strain energy, δ is the stress, and Δl is the change in strain.

[0183] The horizontal axis of the second curve represents strain information, and the vertical axis of the second curve represents stress information.

[0184] For example, at a temperature of 23°C, there are stress and strain information of {50MPa, 100MPa, 150MPa, 200MPa, 200MPa} and {0.001, 0.002, 0.003, 0.004, 0.005}, respectively. If the horizontal axis is taken as strain and the vertical axis is taken as stress, a strain-stress curve is plotted. This curve is the second curve corresponding to the temperature information.

[0185] In one possible implementation, the second curve is integrated using numerical integration methods such as the trapezoidal method or Simpson's method. The area between the stress-strain curve and the horizontal axis is calculated, and this area is the strain energy density. Multiplying the strain energy density by the area or volume of strain borne by the product yields the strain energy of the product.

[0186] The second label is the lifetime value corresponding to the strain energy information to be trained, and this lifetime value satisfies the formula:

[0187] n2=C(ΔU) m *t,

[0188] Where n2 is the lifetime value, i.e. the second label, C and m are the material constants of the product. Typically, C and m can be obtained experimentally or estimated based on the relevant properties of the material, or they can be obtained from public websites. ΔU is the change in strain energy density, and t is the time interval between a single experiment in the stress-strain dataset. This time interval is the time interval between two adjacent stress values ​​at the same temperature in the second dataset.

[0189] It can be seen that each temperature value corresponds to the stress information and strain information to be trained. A strain energy can be calculated, and each strain energy corresponds to a lifetime value. It can be concluded that each strain energy information corresponds to a second label value.

[0190] The beneficial effect of this setting is that, for methods using elastic materials, the influence of strain energy on service life is considered, ensuring the reliability of the method for predicting the service life of the rubber core.

[0191] S404. Based on the dataset to be trained, train the preset initial model to obtain the trained life prediction model.

[0192] For example, this step can refer to step S302 above, and will not be repeated here.

[0193] The present application provides another method for predicting the service life of rubber cores, which comprehensively considers temperature and strain energy information and uses actual collected data for training, thus significantly improving the accuracy and reliability of the service life prediction model.

[0194] To illustrate the above more intuitively Figure 4 The association between the temperature information to be trained and the first label corresponding to the temperature information to be trained. Figure 5 A schematic diagram of the temperature-lifetime curve provided in this application, such as Figure 5 As shown, the horizontal axis of the curve represents temperature, and the vertical axis represents the life value (also known as fatigue life time value).

[0195] When the temperature is 23℃, the strain energy can be assumed to be 0.5 J / m. 2 1.1J / m 2 The calculated lifespan values ​​are 6.21 months and 4.45 months, respectively.

[0196] When the temperature is 70℃, the strain energy can be assumed to be 0.5 J / m. 2 1.1J / m 2 The calculated lifespan values ​​are 6.00 months and 4.40 months, respectively.

[0197] When the temperature is 80℃, the strain energy can be assumed to be 0.5 J / m. 2 1.1J / m 2 The calculated lifespan values ​​were 5.63 months and 4.04 months, respectively.

[0198] When the temperature is 90℃, the strain energy can be assumed to be 0.5 J / m. 2 1.1J / m 2 The calculated lifespan values ​​are 4.55 months and 2.99 months, respectively.

[0199] When the temperature is 100℃, the strain energy can be assumed to be 0.5 J / m. 2 1.1J / m 2 The calculated lifespan values ​​were 3.86 months and 8.96 months, respectively.

[0200] Group 1 represents the temperature information of 23℃, 70℃, 80℃, 90℃, and 100℃ under Group 1 when the strain energy is 0.5. The corresponding lifetime values ​​are smoothly connected on the coordinates to form a curve.

[0201] Group 2 represents the temperature information of 23℃, 70℃, 80℃, 90℃, and 100℃ under Group 2 when the strain energy is 1.1. The corresponding lifetime values ​​are smoothly connected on the coordinates to form a curve.

[0202] from Figure 5 It can be seen that the fatigue life of rubber decreases significantly under higher temperatures and higher strain energies. This again vividly demonstrates the impact of temperature and strain energy on the service life of the rubber core or rubber core material.

[0203] Figure 6A flowchart illustrating another method for predicting the service life of a rubber core provided in this application is shown. The execution subject of this method can be a host, server, or other device, such as... Figure 6 As shown, the method includes:

[0204] S601. Obtain the pre-collected training dataset; wherein, the training dataset includes the temperature information to be trained, the strain energy information to be trained, the first label, and the second label, the first label representing the actual service life of the rubber core under the temperature information to be trained, and the second label representing the actual service life of the rubber core under the strain energy information to be trained.

[0205] For example, this step can refer to step S301 above, and will not be repeated here.

[0206] S602. Determine the target service life based on the actual service life represented by the first label and the actual service life represented by the second label.

[0207] The actual service life represented by the first label can be multiple, and the actual service life represented by the second label can also be multiple. The target service life can be obtained by selecting a portion of the data from multiple first labels and multiple second labels.

[0208] In one possible implementation, a set is formed by multiple first labels and multiple second labels, and a portion of the data in this set is randomly selected as the target lifetime.

[0209] For example, taking the data mentioned in S302 above as an example, the temperature information to be trained, the strain energy information to be trained, the first label, and the second label in the training dataset are {23℃, 70℃, 80℃, 90℃, 100℃}, {0.5J / m 2 0.7J / m 2 0.9J / m 2 1.1J / m 2 {4.92 months, 5.17 months, 4.89 months, 3.96 months, 3.24 months}, {6.21 months, 5.29 months, 4.92 months, 4.45 months}, since the temperature information to be trained corresponds to the first label and the strain energy information to be trained corresponds to the second label, the target lifespan corresponds to the corresponding temperature information or strain energy information to be trained.

[0210] For example, the temperature information or strain energy information to be trained is randomly selected as follows: 23℃, 90℃, and 0.7J / m. 2 The corresponding target lifespans are 4.92 months, 3.96 months, and 5.29 months, respectively.

[0211] S603. Input the temperature information and strain energy information to be trained into the preset initial model to obtain the predicted service life.

[0212] In this context, it can be understood that the temperature information and strain energy information to be trained are used as inputs to a preset initial model to obtain the output results.

[0213] The preset initial model can be either an initialized life prediction model or a trained life prediction model. Only when the preset initial model meets the training completion conditions preset by the staff can the trained life prediction model be obtained.

[0214] For example, taking the data mentioned above regarding the target lifespan as an example, and taking the temperature information and strain energy information to be trained corresponding to the target lifespan as an example, in this embodiment, the temperature information and / or strain energy information to be trained are 23℃, 90℃, and 0.7J / m. 2 The temperature information and strain energy information to be trained are input into the preset initial model to obtain the predicted service life. For example, the predicted service life is 4.23 months, 3.79 months and 5.35 months respectively.

[0215] It should be noted that, in order to describe the content of the embodiment simply and clearly, this embodiment only provides a small amount of temperature information and / or strain energy information to be trained. In actual application scenarios, the amount of temperature information and / or strain energy information to be trained can be much greater.

[0216] S604. Determine the Pearson correlation coefficient between the target service life and the predicted service life, which is the target coefficient.

[0217] The target service life is obtained by combining the actual service life represented by the first label and the actual service life represented by the second label, while the predicted service life is the output result obtained by inputting the temperature information and / or strain energy information to be trained into the preset initial model.

[0218] The Pearson correlation coefficient is a statistical indicator used to measure the degree of linear correlation between two variables. The Pearson correlation coefficient ranges from -1 to 1. The larger the Pearson correlation coefficient, the stronger the correlation between the two variables.

[0219] The commonly used formula for the Pearson correlation coefficient can be expressed as:

[0220]

[0221] Where ρ is the Pearson correlation coefficient, X and Y are the two variables used to calculate the Pearson correlation coefficient, cov(X,Y) represents the covariance between X and Y, and σXσY represents the product of the standard deviations of X and Y.

[0222] For example, X can be the target service life, and Y can be the predicted service life.

[0223] For example, the target lifetimes are 4.92 months, 3.96 months, and 5.29 months as described above, and the predicted lifetimes are 4.23 months, 3.79 months, and 5.35 months as described above. The Pearson correlation coefficient is calculated to be 0.88 using the formula for calculating the Pearson correlation coefficient, and this coefficient is the target coefficient.

[0224] S605. If the target coefficients meet the preset training completion conditions, the trained life prediction model is obtained.

[0225] Among them, the preset training completion conditions can be the training completion conditions set in advance by the staff.

[0226] For example, the preset training completion condition can be a target coefficient greater than or equal to 0.5. When the target coefficient is greater than 0.5, the preset initial model at this time is the life prediction model that has been trained.

[0227] In one possible implementation, if the target coefficient does not meet the preset training completion conditions, the method described above for obtaining the pre-collected training dataset and determining the Pearson correlation coefficient between the target lifespan and the predicted lifespan as the target coefficient can be re-executed until the target coefficient meets the preset training completion conditions.

[0228] This application provides another method for predicting the service life of a rubber core. By combining temperature information and strain energy information, a machine learning model is used to predict the service life of the rubber core. Furthermore, by setting training completion conditions, a trained service life prediction model is obtained, thereby improving the accuracy and reliability of the prediction.

[0229] Figure 7 The flowchart illustrating a method for predicting the service life of a rubber core provided in this application shows that the execution entity of this method can be a host, server, or other device, such as... Figure 7 As shown, the method includes:

[0230] S701. Obtain the pre-collected training dataset; wherein, the training dataset includes the temperature information to be trained, the strain energy information to be trained, the first label, and the second label, the first label representing the actual service life of the rubber core under the temperature information to be trained, and the second label representing the actual service life of the rubber core under the strain energy information to be trained.

[0231] For example, this step can refer to step S301 above, and will not be repeated here.

[0232] S702. Input the temperature information to be trained and the strain energy information to be trained into the preset initial model to obtain the first predicted lifetime corresponding to the temperature information to be trained and the second predicted lifetime corresponding to the strain energy information to be trained.

[0233] The preset initial model can be either an initialized life prediction model or a trained life prediction model. Only when the preset initial model meets the training completion conditions preset by the staff can the trained life prediction model be obtained.

[0234] For example, if the temperature information to be trained is 23°C and 90°C as mentioned above, inputting 23°C and 90°C into the preset initial model yields a first predicted lifetime of 4.23 months and 3.79 months, respectively. If the strain energy information to be trained is 0.7 J / m² as mentioned above... 2 0.9J / m 2 The second predicted lifespan was 5.35 months and 5.07 months.

[0235] S703. Determine the Pearson correlation coefficient between the actual service life represented by the first label and the first predicted service life as the first coefficient, and determine the Pearson correlation coefficient between the actual service life represented by the second label and the second predicted service life as the second coefficient.

[0236] The first coefficient is the Pearson correlation coefficient between the actual service life represented by the first label and the first predicted service life, and the second coefficient is the Pearson correlation coefficient between the actual service life represented by the second label and the second predicted service life.

[0237] For example, the formula for calculating the Pearson correlation coefficient can be referred to step S604 above, and will not be repeated here.

[0238] For example, if the actual service life represented by the first label is 4.92 months and 3.96 months corresponding to the temperature information of 23°C and 90°C, and the first predicted service life is 4.23 months and 3.79 months, then the first coefficient is 1;

[0239] The actual service life represented by the second label is 0.7 J / m, based on the strain energy information mentioned above. 2 0.9J / m 2 The corresponding 5.29 months and 4.92 months, the second predicted lifespan is 5.35 months and 5.07 months as mentioned above, then the second coefficient is 1.

[0240] It should be noted that, in order to describe the content of the embodiment simply and clearly, this embodiment only provides a small amount of temperature information and / or strain energy information to be trained. In actual application scenarios, the amount of temperature information and / or strain energy information to be trained can be much greater.

[0241] S704. If it is determined that both the first coefficient and the second coefficient meet the preset training completion conditions, then the trained life prediction model is obtained.

[0242] Among them, the preset training completion conditions can be the training completion conditions set in advance by the staff.

[0243] For example, the preset training completion conditions can be that the first coefficient is greater than or equal to 0.5 and the second coefficient is greater than or equal to 0.6. When the first coefficient is greater than or equal to 0.5 and the second coefficient is greater than or equal to 0.6, the preset initial model at this time is the life prediction model that has been trained.

[0244] In one possible implementation, if at least one of the first coefficient and the second coefficient fails to meet the preset training completion condition, the above-described method of obtaining the pre-collected training dataset and determining the Pearson correlation coefficient between the actual lifespan represented by the first label and the first predicted lifespan as the first coefficient, and determining the Pearson correlation coefficient between the actual lifespan represented by the second label and the second predicted lifespan as the second coefficient, is repeated until it is determined that both the first coefficient and the second coefficient meet the preset training completion condition.

[0245] The present application provides a method for predicting the service life of a rubber core. By combining temperature information and strain energy information, a machine learning model is used to predict the service life of the rubber core. Furthermore, by setting training completion conditions, a trained service life prediction model is obtained, thereby improving the accuracy and reliability of the prediction.

[0246] Figure 8 A schematic diagram of the structure of the device for predicting the service life of the rubber core provided in this application is shown below. Figure 4 As shown, the lifespan prediction device 80 for the adhesive core provided in this embodiment includes: an acquisition module 801, a determination module 802, and a prediction module 803;

[0247] The acquisition module 801 is used to acquire measurement information of the rubber core to be predicted under the current environment; the measurement information includes the current temperature information, stress information, and strain information.

[0248] The determination module 802 determines the strain energy information of the rubber core to be predicted based on the stress and strain information.

[0249] The prediction module 803 is used to determine the service life information of the rubber core to be predicted based on temperature information and strain energy information and a preset service life prediction model; wherein, the preset service life prediction model is a pre-trained neural network model used to output service life information.

[0250] The prediction module 803 is also used to acquire a pre-collected training dataset; wherein the training dataset includes training temperature information, training strain energy information, a first label, and a second label, the first label representing the actual service life of the training rubber core under the training temperature information, and the second label representing the actual service life of the training rubber core under the training strain energy information.

[0251] Based on the dataset to be trained, the preset initial model is trained to obtain the trained life prediction model.

[0252] In one possible implementation, the prediction module 803 is also used to input temperature information and strain energy information into a preset lifetime prediction model;

[0253] By using the network layer in the preset lifetime prediction model, the temperature information and strain energy information are processed into feature vectors respectively to obtain the feature vector data corresponding to the temperature information and the feature vector data corresponding to the strain energy information.

[0254] Based on the eigenvector data corresponding to temperature information and the eigenvector data corresponding to strain energy information, the service life information of the rubber core to be predicted is determined.

[0255] In one possible implementation, the prediction module 803 is further used to perform fusion processing on the feature vector data corresponding to the temperature information and the feature vector data corresponding to the strain energy information to obtain fused vector data.

[0256] Based on the fused vector data, the lifespan information of the rubber core to be predicted is determined.

[0257] In one possible implementation, the prediction module 803 is further configured to determine the first service life of the rubber core to be predicted based on the feature vector data corresponding to the temperature information; wherein the first service life characterizes the service life information of the rubber core under the temperature information.

[0258] Based on the feature vector data corresponding to the strain energy information, the second service life of the rubber core to be predicted is determined; wherein, the second service life characterizes the service life information of the rubber core under the strain energy information.

[0259] Based on the first service life and the second service life, the service life information of the rubber core to be predicted is determined.

[0260] In one possible implementation, the prediction module 803 is further used to acquire a first dataset and a second dataset; wherein the first dataset includes temperature information to be trained, stress information to be trained, and number of experimental cycles, and the second dataset includes temperature information to be trained, stress information to be trained, and strain information to be trained, and the number of experimental cycles represents the number of times the training core is subjected to a preset experiment under the stress information to be trained.

[0261] Based on the first dataset, determine the first label corresponding to the temperature information to be trained;

[0262] Based on the second dataset, determine the strain energy information and second label to be trained.

[0263] In one possible implementation, the prediction module 803 is further configured to determine a first curve corresponding to the temperature information to be trained in the first dataset based on the stress information to be trained corresponding to the temperature information to be trained in the first dataset and the number of experimental cycles; wherein the first curve represents the relationship between the stress information to be trained and the number of experimental cycles.

[0264] The first curve is processed by Weibull distribution to obtain the first label corresponding to the temperature information to be trained in the first dataset.

[0265] In one possible implementation, the prediction module 803 is further configured to determine the strain energy information to be trained corresponding to the temperature information to be trained in the second dataset based on the stress information to be trained and the strain information to be trained corresponding to the temperature information to be trained in the second dataset.

[0266] Based on the temperature information to be trained in the second dataset, corresponding to the stress information and strain information to be trained, a second curve corresponding to the temperature information to be trained in the second dataset is determined; wherein, the second curve represents the relationship between the stress information and strain information to be trained.

[0267] Integrating the second curve yields the second label corresponding to the strain energy information to be trained.

[0268] In one possible implementation, the prediction module 803 is further configured to determine the target service life based on the actual service life represented by the first label and the actual service life represented by the second label.

[0269] The temperature information and strain energy information to be trained are input into the preset initial model to obtain the predicted service life.

[0270] The Pearson correlation coefficient between the target service life and the predicted service life is determined and given as the target coefficient.

[0271] If the target coefficients meet the preset training completion conditions, then the trained life prediction model is obtained.

[0272] In one possible implementation, the prediction module 803 is further configured to input the temperature information to be trained and the strain energy information to be trained into a preset initial model to obtain a first predicted lifetime corresponding to the temperature information to be trained and a second predicted lifetime corresponding to the strain energy information to be trained.

[0273] The Pearson correlation coefficient between the actual service life represented by the first label and the first predicted service life is determined as the first coefficient, and the Pearson correlation coefficient between the actual service life represented by the second label and the second predicted service life is determined as the second coefficient;

[0274] If both the first and second coefficients meet the preset training completion conditions, then the trained life prediction model is obtained.

[0275] The device for predicting the service life of the rubber core provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0276] Figure 9 A schematic diagram of the structure of the electronic device provided in this application, such as... Figure 9 As shown, the electronic device 90 provided in this embodiment includes at least one processor 901 and a memory 902. Optionally, the device 90 further includes a communication component 903. The processor 901, memory 902, and communication component 903 are connected via a bus 904.

[0277] In a specific implementation, at least one processor 901 executes computer execution instructions stored in memory 902, causing at least one processor 901 to perform the above-described method.

[0278] The specific implementation process of processor 901 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0279] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0280] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0281] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0282] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0283] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0284] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0285] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0286] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0287] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0288] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0289] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0290] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0291] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for predicting the service life of a rubber core, characterized in that, include: Acquire measurement information of the rubber core to be predicted under the current environment; wherein, the measurement information includes current temperature information, stress information, and strain information; wherein, the strain information refers to the deformation data of the rubber core under stress, characterizing the degree of deformation of the rubber core under stress; Based on the stress information and the strain information, the strain energy information of the core to be predicted is determined; wherein, the strain energy information refers to the energy stored in the core during deformation. Based on the temperature information and the strain energy information, and using a preset life prediction model, the lifespan information of the rubber core to be predicted is determined; wherein, the preset life prediction model is a pre-trained neural network model used to output lifespan information. The lifetime prediction model is obtained based on the following method: Obtain a first dataset and a second dataset; wherein, the first dataset includes multiple temperature information to be trained, as well as stress information to be trained and experimental cycle number corresponding to each temperature information to be trained; the second dataset includes multiple temperature information to be trained, as well as stress information to be trained and strain information to be trained corresponding to each temperature information to be trained; the number of experimental cycles represents the number of times the training core is subjected to a preset experiment under the stress information to be trained. Based on the temperature information to be trained in the first dataset, the stress information to be trained, and the number of experimental cycles, a first curve corresponding to the temperature information to be trained in the first dataset is determined; wherein, the first curve represents the relationship between the stress information to be trained and the number of experimental cycles. The first curve is converted to obtain a failure time dataset; based on the failure time dataset, the shape parameter and scale parameter of the Weibull distribution are calculated, and based on the shape parameter and scale parameter, the first label corresponding to each temperature information to be trained in the first dataset is obtained; the first label represents the actual service life of the rubber core to be trained under the temperature information to be trained. Based on the stress information and strain information to be trained corresponding to the temperature information to be trained in the second dataset, determine the strain energy information to be trained corresponding to the temperature information to be trained in the second dataset. Based on the training stress information and training strain information corresponding to the training temperature information in the second dataset, a second curve corresponding to the training temperature information in the second dataset is determined; wherein, the second curve represents the relationship between the training stress information and the training strain information. The second curve is integrated to obtain the second label corresponding to the strain energy information to be trained; The second label characterizes the actual service life of the rubber core under the strain energy information to be trained; Based on the first dataset, the second dataset, the first label, and the second label, a preset initial model is trained, and the model parameters are adjusted to obtain the trained life prediction model.

2. The method according to claim 1, characterized in that, Based on the temperature information and the strain energy information, and using a preset life prediction model, the lifespan information of the rubber core to be predicted is determined, including: The temperature information and the strain energy information are input into a preset lifetime prediction model; The temperature information and the strain energy information are processed by feature vectors through the network layer in the preset lifetime prediction model to obtain feature vector data corresponding to the temperature information and feature vector data corresponding to the strain energy information. The service life information of the rubber core to be predicted is determined based on the feature vector data corresponding to the temperature information and the feature vector data corresponding to the strain energy information.

3. The method according to claim 2, characterized in that, Based on the feature vector data corresponding to the temperature information and the feature vector data corresponding to the strain energy information, the service life information of the rubber core to be predicted is determined, including: The feature vector data corresponding to the temperature information and the feature vector data corresponding to the strain energy information are fused to obtain fused vector data. Based on the fused vector data, the lifespan information of the core to be predicted is determined.

4. The method according to claim 2, characterized in that, Based on the feature vector data corresponding to the temperature information and the feature vector data corresponding to the strain energy information, the service life information of the rubber core to be predicted is determined, including: Based on the feature vector data corresponding to the temperature information, the first service life of the rubber core to be predicted is determined; wherein, the first service life characterizes the service life information of the rubber core under the temperature information. The second service life of the rubber core to be predicted is determined based on the feature vector data corresponding to the strain energy information; wherein, the second service life characterizes the service life information of the rubber core under the strain energy information. Based on the first service life and the second service life, the service life information of the rubber core to be predicted is determined.

5. The method according to claim 1, characterized in that, Based on the first dataset, the second dataset, the first label, and the second label, a preset initial model is trained to obtain the trained lifespan prediction model, including: The target service life is determined based on the actual service life represented by the first label and the actual service life represented by the second label; The temperature information and strain energy information to be trained are input into a preset initial model to obtain the predicted service life. The Pearson correlation coefficient between the target service life and the predicted service life is determined as the target coefficient. If the target coefficients meet the preset training completion conditions, then the trained lifetime prediction model is obtained.

6. The method according to claim 1, characterized in that, Based on the first dataset, the second dataset, the first label, and the second label, a preset initial model is trained to obtain the trained lifespan prediction model, including: The temperature information to be trained and the strain energy information to be trained are input into a preset initial model to obtain a first predicted lifetime corresponding to the temperature information to be trained and a second predicted lifetime corresponding to the strain energy information to be trained. The Pearson correlation coefficient between the actual service life represented by the first tag and the first predicted service life is determined as the first coefficient, and the Pearson correlation coefficient between the actual service life represented by the second tag and the second predicted service life is determined as the second coefficient; If it is determined that both the first coefficient and the second coefficient meet the preset training completion conditions, then the trained life prediction model is obtained.

Images