Neural network algorithm-based steel service life prediction method and device, equipment and medium

Abstract

The invention provides a steel service life prediction method, device, equipment and medium based on a neural network algorithm, and the method comprises the steps: obtaining a material parameter and an operation parameter of a pipeline steel to obtain a stress parameter and a test temperature of the steel, taking the stress parameter and the test temperature as input nodes, taking the fracture time as an output node, and calculating the service life of the steel. Constructing a neural network model, training the neural network model to obtain a creep fatigue prediction model, obtaining the fracture time of the pipeline steel based on the creep fatigue prediction model, generating a life prediction curve of the pipeline steel according to the fracture time, and predicting the life of the target pipeline steel through the life prediction curve; according to the method, training can be carried out by utilizing limited experimental data and simulation data through the neural network model, and a high-precision creep fatigue prediction model can be constructed in a short time, so that the residual life of the material is rapidly evaluated, and the working efficiency is greatly improved.

Description

Technical Field

[0001] The present application relates to the technical field of materials science, and particularly relates to a method, device, equipment and medium for predicting the service life of steel based on a neural network algorithm. Background Art

[0002] With the continuous improvement of the global requirements for energy efficiency and environmental protection, the technology of thermal power generating units is also constantly advancing. As a key technology for improving the power generation efficiency of boilers, reducing the use of primary energy, and reducing environmental pollution, ultra-supercritical units have gradually become an important part of modern thermal power plants. Ultra-supercritical units significantly improve the thermal efficiency by increasing the pressure and temperature of steam, and reduce the emissions of carbon dioxide and other pollutants.

[0003] However, the working environment of ultra-supercritical units is extremely harsh. Its main pipelines are in a high-temperature and high-pressure state for a long time, which poses extremely high requirements on the performance of materials. In particular, common materials such as 9Cr series steel are prone to problems such as material aging and damage under such extreme conditions, such as creep fatigue, corrosion, and crack propagation. These problems seriously threaten the safe operation and service life of the units.

[0004] Currently, the L-M parameter method is a relatively reliable life prediction method, which is widely used to evaluate the remaining life of materials under specific working conditions. The L-M parameter method can accurately predict the remaining life of materials under any working stress and temperature conditions by establishing a relationship between the temperature, stress and time of the materials. This method relies on experimental data to determine key parameters and obtains the corresponding prediction model through fitting. Although the L-M parameter method has high accuracy in theory, the application of the L-M parameter method requires a large amount of experimental data support, and these data usually need to be obtained through long-term creep fatigue tests in a laboratory environment. This not only consumes a large amount of time and resources, but also limits its rapid application in practical engineering. Summary of the Invention

[0005] In view of the above-mentioned disadvantages of the prior art, the present invention provides a method, device, equipment and medium for predicting the service life of steel based on a neural network algorithm to solve the above technical problems.

[0006] A method for predicting the service life of steel based on a neural network algorithm provided by the present invention, the method comprising: obtaining material parameters of pipeline steel and operating parameters of the unit where the pipeline steel operates; determining stress parameters of the pipeline steel based on the material parameters and the operating parameters, and obtaining the test temperature of the pipeline steel based on the operating parameters; using the stress parameters and the test temperature as input nodes of a neural network, and using the fracture time of the pipeline steel as an output node, constructing an initial neural network model, and training the initial neural network model to obtain a creep-fatigue prediction model; obtaining the fracture time of the pipeline steel based on the creep-fatigue prediction model, and generating a service life prediction curve of the pipeline steel according to the fracture time, so as to predict the service life of the target pipeline steel based on the service life prediction curve.

[0007] In an embodiment of the present invention, determining the stress parameters of the pipeline steel based on the material parameters and the operating parameters includes: obtaining the ambient temperature; determining the elastic modulus and the coefficient of linear expansion of the pipeline steel based on the material parameters, and obtaining the pipeline parameters and the operating temperature based on the operating parameters; calculating the pipeline parameters and the operating temperature to obtain the internal pressure stress of the pipeline steel; calculating the difference between the operating temperature and the ambient temperature to obtain a temperature difference, and calculating the temperature difference, the elastic modulus, and the coefficient of linear expansion to obtain the thermal stress of the pipeline steel; obtaining the total stress of the pipeline steel according to the internal pressure stress and the thermal stress, so as to generate the stress parameters of the pipeline steel.

[0008] In an embodiment of the present invention, constructing an initial neural network model includes: identifying the input nodes and the output nodes to obtain the number of input layer nodes and the number of output layer nodes; calculating the number of input layer nodes and the number of output layer nodes to determine the number of hidden layer nodes of the initial neural network model; setting a first transfer function based on the number of input layer nodes and the number of hidden layer nodes, and setting a second transfer function based on the number of output layer nodes and the number of hidden layer nodes; transmitting test data from the input layer to the hidden layer based on the first transfer function, triggering the hidden layer to perform feature transformation on the test data, and transmitting the data after feature transformation to the output layer based on the second transfer function to generate a predicted value; calculating the error between the predicted value and the test data, and feeding back the error to the hidden layer to update the first transfer function and the second transfer function until the error reaches a preset error range; generating an initial neural network model based on the updated first transfer function and second transfer function.

[0009] In an embodiment of the present invention, training the initial neural network model includes: obtaining a historical data set composed of multiple stress parameters and test temperatures based on historical data, and annotating the historical data set according to the fracture time, where there is a one-to-one correspondence between the historical data set and the fracture time; dividing the annotated historical data set into a training data set and a validation data set; training the initial neural network model based on the training data set to learn the initial mapping relationship between the input historical data set and the fracture time; validating the initial mapping relationship based on the validation data set, and optimizing the initial mapping relationship based on the validation result to obtain an optimized neural network model, and determining the optimized neural network model as the creep fatigue prediction model.

[0010] In an embodiment of the present invention, generating a life prediction model for the pipeline steel according to the fracture time includes: obtaining fracture time data at multiple different temperatures to generate test data; generating multiple sets of prediction data based on the optimized neural network model, where the stress parameters of the test data and the prediction data are the same, and the test temperatures are different; combining the test data and the prediction data to obtain a target data set, where the target data set includes stress parameters, test temperatures, fracture times, and the mapping relationship among the three; calculating the data sets in the target data set to obtain the mean value of the material constants of the pipeline steel and the material constants at different test temperatures; fitting a life prediction curve based on the mean value of the material constants and / or the material constants at different test temperatures, where the life prediction curve is used to map the fracture time under the stress-temperature relationship.

[0011] In an embodiment of the present invention, predicting the life of the target pipeline steel based on the life prediction curve includes: obtaining the current stress parameter and the current test temperature of the target pipeline steel; retrieving in the life prediction curve to obtain the target fracture time corresponding to the current stress parameter and the current test temperature, so as to obtain the remaining life of the target pipeline steel based on the target fracture time.

[0012] In an embodiment of the present invention, after predicting the life of the target pipeline steel, it further includes: if the remaining life of the target pipeline steel is less than the preset expected operation duration, updating the target pipeline steel; if the remaining life of the target pipeline steel is greater than or equal to the preset expected operation duration, maintaining the target pipeline steel, where the maintenance includes detecting defects of the pipeline steel to obtain defect parameters of the pipeline steel; querying the corresponding corrosion assessment result according to the defect parameters in the preset corrosion damage assessment database; retrieving in the preset pipeline maintenance plan library to obtain a pipeline maintenance plan matching the corrosion assessment result.

[0013] This application provides a steel lifespan prediction device based on a neural network algorithm. The device includes: a data acquisition module for obtaining the material parameters of pipeline steel and the operating parameters of the unit where the pipeline steel operates; an input parameter calculation module for determining the stress parameters of the pipeline steel based on the material parameters and the operating parameters, and obtaining the test temperature of the pipeline steel based on the operating parameters; a creep-fatigue prediction model training module for constructing an initial neural network model with the stress parameters and the test temperature as the input nodes of the neural network and the fracture time of the pipeline steel as the output node, and training the initial neural network model to obtain a creep-fatigue prediction model; a steel lifespan prediction module for obtaining the fracture time of the pipeline steel based on the creep-fatigue prediction model and generating a lifespan prediction curve of the pipeline steel according to the fracture time, so as to predict the lifespan of the target pipeline steel based on the lifespan prediction curve.

[0014] This application provides an electronic device, including a processor, a memory, and a communication bus; the communication bus is used to connect the processor and the memory; the processor is used to execute the computer program stored in the memory to implement the steel lifespan prediction method based on the neural network algorithm as described above.

[0015] This application provides a computer-readable storage medium, on which a computer program is stored, and the computer program is used to make a computer execute the steel lifespan prediction method based on the neural network algorithm as described above.

[0016] Advantages of the present invention: In the steel lifespan prediction method based on the neural network algorithm of the present invention, by combining material parameters and operating parameters and using the neural network algorithm to construct a creep-fatigue prediction model, complex non-linear relationships can be better captured, thus significantly improving the accuracy of fracture time prediction; moreover, based on the characteristics of the neural network model, only a small amount of experimental data is required to train an efficient prediction model, which not only saves experimental time and costs, but also can relatively accurately evaluate the remaining lifespan of the material in a short time, greatly improving work efficiency; in addition, the lifespan prediction curve generated based on the prediction model provides an intuitive reference for users, enabling users to clearly understand the change trend of the service life of pipeline steel under different stress and temperature conditions through the lifespan prediction curve, facilitating the formulation of reasonable maintenance and overhaul plans, and enhancing the scientificity and operability of decision-making.

[0017] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit this application. Description of the Drawings

[0018] The accompanying drawings here are incorporated into the specification and form a part of this specification, showing embodiments consistent with this application, and are used together with the specification to explain the principles of this application. Obviously, the accompanying drawings in the following description are only some embodiments of this application, and those of ordinary skill in the art can obtain other accompanying drawings based on these drawings without creative efforts. In the accompanying drawings:

[0019] Figure 1 It is a schematic diagram of the implementation environment of the steel life prediction method based on the neural network algorithm shown in an exemplary embodiment of this application;

[0020] Figure 2 It is a flowchart of the steel life prediction method based on the neural network algorithm shown in an exemplary embodiment of this application;

[0021] Figure 3 It is a schematic diagram of the whole process of the steel life prediction method based on the neural network algorithm shown in an exemplary embodiment of this application;

[0022] Figure 4 It is a schematic diagram of the prediction result of the creep fatigue prediction model shown in an exemplary embodiment of this application;

[0023] Figure 5 It is a block diagram of the steel life prediction device based on the neural network algorithm shown in an exemplary embodiment of this application;

[0024] Figure 6 It shows a schematic diagram of the structure of the computer system of the electronic device suitable for implementing the embodiments of this application. Detailed Embodiments

[0025] The following will describe the embodiments of the present invention with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for explaining the present invention, rather than for limiting the protection scope of the present invention.

[0026] It should be noted that the diagrams provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner. Therefore, only the components related to the present invention are shown in the diagrams, rather than being drawn according to the number, shape, and size of the components in actual implementation. The types, quantities, and proportions of the components in actual implementation can be arbitrarily changed, and the component layout type may also be more complex.

[0027] In the following description, numerous details are explored to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details. In other embodiments, well-known structures and devices are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present invention difficult to understand.

[0028] First of all, it should be noted that the BP (Back Propagation) neural network is a multi-layer feedforward neural network trained according to the error backpropagation algorithm.

[0029] The 9Cr series of steels refers to a series of steels with a chromium content of approximately 9%. Such steels are widely used in different industrial fields due to their specific chemical compositions and performance characteristics.

[0030] Ultra-supercritical (USC) units refer to a highly efficient and low-emission coal-fired power generation technology. It achieves higher thermal efficiency and reduces the emissions of carbon dioxide and other pollutants by increasing the steam parameters (temperature and pressure).

[0031] The L-M parameter method, namely the Levenberg-Marquardt algorithm (abbreviated as the L-M algorithm), is an optimization algorithm for solving nonlinear least-squares problems. It combines the advantages of the gradient descent method and the Gauss-Newton method and is applicable to the problem of fitting model parameters to a data set.

[0032] SEM (Scanning Electron Microscope) is a technique that uses a focused fine beam of high-energy electrons to scan the surface of a sample to generate an image. When the electron beam strikes the sample, various signals such as secondary electrons and backscattered electrons are generated, and these signals can be converted into an image after being captured by a detector.

[0033] EDS (Energy Dispersive X-ray Spectroscopy, also known as EDX), usually used as an attachment to the SEM, can perform qualitative and quantitative analysis of the elemental composition of a sample. When the electron beam excites the sample, characteristic X-rays are emitted. Different elements emit X-rays with specific energies, and by analyzing these energies, it is possible to determine which elements are present in the sample and their relative abundances.

[0034] Figure 1 It is a schematic diagram of the implementation environment of a steel life prediction method based on a neural network algorithm shown in an exemplary embodiment of the present application.

[0035] Such as Figure 1As shown in the figure, the implementation environment of the steel life prediction method based on the neural network algorithm includes a data acquisition module 101 and a computer device 102. Among them, the data acquisition module 101 is responsible for collecting various data required for constructing and training the neural network model, and its relevant data includes material parameters, operating parameters, defect detection data, and corrosion behavior data. The computer device 102 is responsible for processing and analyzing the data collected by the data acquisition module 101, and using the neural network algorithm to construct and train the creep fatigue prediction model. Specifically, it includes: performing preprocessing operations such as data cleaning and normalization on the collected data to ensure the quality and consistency of the data; using the neural network algorithm, according to the design of the input layer (related factors such as stress state, test temperature, etc.), hidden layer, and output layer (fracture time), writing Matlab code to implement the construction and training of the neural network model; using the trained neural network model to predict the fracture time of 9Cr series steel, and comparing it with the actual value to continuously correct and optimize the model; evaluating the prediction effect through the fitting equation and adjusting coefficients, and generating an intuitive life prediction curve for users to refer to.

[0036] Through the detailed description of the above implementation environment, it can be understood that the steel life prediction method based on the neural network algorithm proposed by the present invention not only has efficient data acquisition and processing capabilities, but also can quickly and accurately predict the remaining life of steel using advanced neural network algorithms, providing strong technical support for ensuring the safe operation of ultra-supercritical units.

[0037] Figure 2 It is a flowchart of the steel life prediction method based on the neural network algorithm shown in an exemplary embodiment of the present application.

[0038] As Figure 2 shown, in an exemplary embodiment, the steel life prediction method based on the neural network algorithm at least includes steps S210 to S240, which are introduced in detail as follows:

[0039] Step S210, obtain the material parameters of the pipeline steel and the operating parameters of the unit where the pipeline steel is located.

[0040] In an embodiment of the present application, taking the target pipeline steel as the 9Cr series steel used in the main pipeline of the ultra-supercritical unit as an example. Obtain the detailed information of the 9Cr series steel used in the main pipeline of the ultra-supercritical unit from the power plant, specifically including material parameters such as the chemical composition, microstructure, and mechanical properties of the steel, and operating parameters such as the actual operating temperature, pressure, operating time, and fluid flow rate of the unit.

[0041] In a specific embodiment of the present application, first, a material sample of 9Cr series steel is collected from an ultra-supercritical unit, and its chemical composition (such as the proportions of elements like chromium and molybdenum), microstructure (grain size, phase composition), and mechanical properties (hardness, tensile strength, etc.) are obtained through laboratory analysis. The material parameters are obtained, including but not limited to the following:

[0042] Chemical composition: Cr 9%, Mo 1%

[0043] Microstructure: The average grain size is 50 μm

[0044] Mechanical properties: The hardness is 200 HB, and the tensile strength is 600 MPa

[0045] In addition, the key parameters during the operation of the unit are recorded, including temperature, pressure, flow rate, etc., and the operation parameters are obtained, including but not limited to the following:

[0046] Actual operating temperature: 600 °C

[0047] Pressure: 30 MPa

[0048] Flow rate: 5 m / s

[0049] Operating time: Accumulated 1000 hours.

[0050] It can be understood that through the above specific embodiment, the material parameters of the main pipeline of 9Cr series steel and the operation parameters of the ultra-supercritical unit can be obtained comprehensively and accurately, providing a solid data foundation for the subsequent construction of the neural network model and life prediction.

[0051] Step S220, determine the stress parameters of the pipeline steel based on the material parameters and operation parameters, and obtain the test temperature of the pipeline steel based on the operation parameters.

[0052] In an embodiment of the present application, to determine the stress parameters of the pipeline steel from the material parameters and operation parameters, it includes: obtaining the ambient temperature; determining the elastic modulus and coefficient of linear expansion of the pipeline steel based on the material parameters, and obtaining the pipeline parameters and operating temperature based on the operation parameters; calculating the pipeline parameters and operating temperature to obtain the internal pressure stress of the pipeline steel; calculating the difference between the operating temperature and the ambient temperature to obtain the temperature difference, and calculating the temperature difference, elastic modulus, and coefficient of linear expansion to obtain the thermal stress of the pipeline steel; obtaining the total stress of the pipeline steel according to the internal pressure stress and thermal stress to generate the stress parameters of the pipeline steel. It can be understood that the aforementioned pipeline parameters at least include the pipeline internal pressure, pipeline outer diameter, and pipeline wall thickness.

[0053] In one embodiment of the present application, the material parameters and operating parameters of the pipeline steel are obtained respectively, the internal pressure stress is calculated based on the operating parameters, the thermal stress is calculated based on the combination of the material parameters and the operating parameters, and then the total stress of the pipeline steel is calculated based on the internal pressure stress and the thermal stress, and the corresponding stress parameters are generated according to the total stress.

[0054] In a specific embodiment of the present application, taking 9Cr steel as the target steel as an example, referring to the material handbook of 9Cr steel, assuming its elastic modulus E = 200GPaE = 200GPa, and the linear expansion coefficient α = 12×10 -6 / ℃. Assuming the ambient temperature is 20℃, and the operating parameters are as follows:

[0055] Internal pressure p = 30MPap = 30MPa

[0056] Operating temperature T = 600℃T = 600℃

[0057] Pipeline outer diameter D = 500mmD = 500mm

[0058] Wall thickness t = 20mmt = 20mm.

[0059] First, use the thin-walled container theory formula to calculate the circumferential stress and axial stress, specifically as follows:

[0060]

[0061] Among them, σ h represents the circumferential stress, σ a represents the axial stress, p represents the internal pressure, D represents the pipeline outer diameter, and t represents the pipeline wall thickness.

[0062] Then, the uneven thermal expansion caused by the temperature gradient will generate thermal stress. The thermal stress can be estimated by the following formula:

[0063] σ th = EαΔT Equation (2)

[0064] Among them, σ th represents the thermal stress, E represents the elastic modulus, α represents the linear expansion coefficient, and ΔT represents the temperature difference between the ambient temperature and the operating temperature.

[0065] Finally, based on the circumferential stress, axial stress, and thermal stress obtained above, calculate the total stress of the pipeline steel, and its calculation formula is as follows:

[0066]

[0067] Among them, σ total represents the total stress, σ h represents the circumferential stress, σ adenotes the axial stress, σ th denotes the thermal stress.

[0068] It should be noted that after calculating the circumferential stress, axial stress, thermal stress, and total stress according to the above embodiments, appropriate parameters can be selected from them as the stress parameters required for model construction based on the actual needs of model construction. It can be a combination of multiple stresses or a single total stress. This application does not limit the specific data categories and data quantities of its stress parameters.

[0069] Step S230: Use the stress parameter and test temperature as the input nodes of the neural network, and use the fracture time of the pipeline steel as the output node to construct an initial neural network model and train the initial neural network model to obtain a creep fatigue prediction model.

[0070] In an embodiment of the present application, constructing the initial neural network model includes: identifying the input nodes and output nodes to obtain the number of input layer nodes and the number of output layer nodes; calculating the number of input layer nodes and the number of output layer nodes to determine the number of hidden layer nodes of the initial neural network model; setting a first transfer function based on the number of input layer nodes and the number of hidden layer nodes, and setting a second transfer function based on the number of output layer nodes and the number of hidden layer nodes; transmitting the test data from the input layer to the hidden layer based on the first transfer function, triggering the hidden layer to perform feature transformation on the test data, and transmitting the data after feature transformation to the output layer based on the second transfer function to generate a predicted value; calculating the error between the predicted value and the test data, and feeding back the error to the hidden layer to update the first transfer function and the second transfer function until the error reaches the preset error range; generating the initial neural network model based on the updated first transfer function and second transfer function.

[0071] In an embodiment of the present application, three parameters, namely circumferential stress, axial stress, and thermal stress, are used as stress parameters, one parameter represents the test temperature, and one parameter represents the fracture time. Therefore, the number of input nodes of the model is set to 4, and the number of output nodes is set to 1. Then, based on the determined number of input nodes and the number of output nodes, the number of hidden layer nodes is calculated as follows:

[0072]

[0073] where L represents the number of hidden layer nodes, a represents the number of input nodes, b represents the number of output nodes, and c is an adjustment constant.

[0074] In view of the fact that in related applications of fine-tuning model complexity, the value of the adjustment constant usually ranges from 0 to 5. Therefore, based on Equation (4), the number of hidden layer nodes in this embodiment should be between 2.5 and 7.5. Also, since the number of hidden layer nodes must be an integer, in this embodiment, the number of hidden layer nodes is taken as 6 for illustrative purposes.

[0075] Then, the input test data enters the network through the input layer. The data of the input layer is passed to the hidden layer through the action of the weight matrix W1 and the bias b1 (i.e., the first transfer function), and the input is non-linearly transformed through the activation function (such as the Sigmoid function) of the hidden layer; then, the output of the hidden layer is passed to the output layer through another weight matrix W2 and bias b2 (i.e., the second transfer function), and the output layer also has an activation function (for regression problems, a linear function can be directly used).

[0076] Then, calculate the error of the output layer and propagate it back to the hidden layer through the chain rule, and then update the weight matrices W1, W2 and the biases b1, b2 to minimize the error function (such as the mean square error MSE).

[0077] In a specific embodiment, Equation (5) shown below is used as the transfer function of the BP neural network model, and the continuous perceptron learning rule (Delta) is used to improve the accuracy of the model, so that the error function of Equation (6) reaches the minimum value. Specifically as follows:

[0078]

[0079] Δw ij =η(y i -o j )o i Equation (6)

[0080] Among them, Δw ij is the weight value, y i is the expected output value of i, η is the learning efficiency, o i and o j are the activation values of neurons i and j.

[0081] Finally, use programming languages such as Matlab to write the function transfer code between the input layer, hidden layer and output layer, establish the implementation program of the model, input the relevant parameters into the neural network nodes, so as to construct the BP neural network creep fatigue prediction topological structure, and train the constructed BP neural network model through the learning function of Matlab to meet the accuracy requirements. In addition, the trained BP neural network model can be used to predict the fracture time of 9Cr series steel, and compare it with the actual value, and continuously correct and optimize the neural network model so that it can accurately predict the fracture time of 9Cr series steel.

[0082] In one embodiment of the present application, training the initial neural network model includes: obtaining a historical data set composed of multiple stress parameters and test temperatures based on historical data, and annotating the historical data set according to the fracture time, where there is a one-to-one correspondence between the historical data set and the fracture time; dividing the annotated historical data set into a training data set and a validation data set; training the initial neural network model based on the training data set to learn the initial mapping relationship between the input historical data set and the fracture time; validating the initial mapping relationship based on the validation data set, and optimizing the initial mapping relationship based on the validation result to obtain an optimized neural network model, and determining the optimized neural network model as the creep fatigue prediction model.

[0083] In a specific embodiment of the present application, the collected material-related data is used as the training set, and multiple iterative trainings are performed through a learning function until the model reaches a predetermined accuracy (for example, the mean square error is less than 0.00001). And during the training process, the weights and biases are continuously adjusted to minimize the error function to ensure that the model can accurately predict the fracture time. In addition, the trained BP neural network model can be further used to predict the creep fatigue fracture time of the material, and the prediction result is compared with the actual value to further correct and optimize the model so that it can more accurately predict the fracture time of the material.

[0084] Step S240, obtaining the fracture time of the pipeline steel based on the creep fatigue prediction model, and generating a life prediction curve of the pipeline steel according to the fracture time, so as to predict the life of the target pipeline steel based on the life prediction curve.

[0085] In one embodiment of the present application, generating a life prediction model of pipeline steel according to the fracture time includes: obtaining fracture time data at multiple different temperatures to generate test data; generating multiple groups of prediction data based on the optimized neural network model, where the stress parameters of the test data and the prediction data are the same and the test temperatures are different; merging the test data and the prediction data to obtain a target data set, where the target data set includes stress parameters, test temperatures, fracture times, and the mapping relationships among the three; calculating the data groups in the target data set to obtain the mean value of the material constants of the pipeline steel and the material constants at different test temperatures; fitting a life prediction curve based on the mean value of the material constants and / or the material constants at different test temperatures, where the life prediction curve is used to map the fracture time under the stress-temperature relationship.

[0086] In one embodiment of the present application, first, fracture time data of 9Cr series steel under various stress conditions and different test temperatures are obtained through experiments. For example, 5 different stress parameters (σ1, σ2,..., σ5) are selected, and 3 experiments are conducted under each stress parameter, with different test temperatures (T1, T2,..., Tn) set for each experiment.

[0087] Record the fracture time (tr) under each experimental condition to obtain a set of initial test data as shown in Table 1.

[0088] Table 1

[0089] Stress parameter (MPa) Test temperature (K) Fracture time (h) σ1 T1 tr1 ... ... ... σ5 Tn trn

[0090] Using the trained and optimized BP neural network model, input the same stress parameters as the test data, but select different test temperature ranges. Assume that the neural network model has been fully trained and can accurately predict the fracture time at different temperatures. Generate an additional prediction data set as shown in Table 2:

[0091] Table 2

[0092]

[0093]

[0094] Merge the data obtained above to form a target data set containing stress parameters, test temperature, and fracture time, as shown in Table 3:

[0095] Table 3

[0096] Stress parameter (MPa) Test temperature (K) Fracture time / Predicted fracture time (h) σ1 T1 tr1 ... ... ... σ5 Tn trn σ1 T'1 tr'1 ... ... ... σ5 T'm tr'm

[0097] Then, using the fracture time in the target data set and the corresponding test temperature, calculate the material constant C at different test temperatures. The material constant C can be determined by fitting the relationship between the fracture time and the test temperature. Its calculation method can adopt the L-M parameter method or other suitable methods. The present application does not impose any restrictions on its specific calculation method. Finally, obtain the material constant values at different temperatures and calculate their average value.

[0098] Finally, based on the average value of the material constants obtained above and the material constants at different test temperatures, fit one or more life prediction curves. These curves will map the fracture time under the stress-temperature relationship and provide a basis for subsequent life prediction. The adjustment coefficient R 2 and the adjustable coefficient Adj.R 2 can be used to evaluate the fitting effect and ensure the accuracy of the fitting result.

[0099] In a specific embodiment of the present application, assume that the target data set obtained based on testing and prediction is shown in Table 4:

[0100] Table 4

[0101]

[0102]

[0103] Using the Levenberg-Marquardt (L-M) parameter method or other suitable methods, calculate the material constant C for each group of stress-temperature combinations, and calculate its average value. Assume that the material constant values obtained by fitting are: T1: C1 = 0.001, T2: C2 = 0.0012, T3: C3 = 0.0015, T4: C4 = 0.002, T5: C5 = 0.003, then the calculated average value is 0.00174.

[0104] Finally, based on the average value of the material constant and the material constants at different test temperatures, fit to obtain the life prediction curve. Assume that the following relationship is fitted using linear regression or non-linear regression methods:

[0105]

[0106] where, t f represents the fracture time, σ represents the stress, represents the average value of the material constant, Q represents the activation energy, T represents the absolute temperature, and R represents the gas constant.

[0107] By adjusting the coefficients n and the activation energy Q, make the fitted curve as close as possible to the experimental data points, so as to obtain the final life prediction curve. This step can be implemented by software such as Matlab, and the present application does not impose any specific restrictions on its implementation method.

[0108] In an embodiment of the present application, predicting the life of the target pipeline steel based on the life prediction curve includes: obtaining the current stress parameter and the current test temperature of the target pipeline steel; retrieving in the life prediction curve to obtain the target fracture time corresponding to the current stress parameter and the current test temperature, so as to obtain the remaining life of the target pipeline steel based on the target fracture time.

[0109] In one embodiment of the present application, after predicting the lifespan of the target pipeline steel, the following steps are further included: If the remaining lifespan of the target pipeline steel is less than the preset expected operation duration, the target pipeline steel is updated; if the remaining lifespan of the target pipeline steel is greater than or equal to the preset expected operation duration, maintenance is performed on the target pipeline steel. The maintenance includes detecting defects in the pipeline steel to obtain defect parameters of the pipeline steel; querying the corresponding corrosion assessment result according to the defect parameters in the preset corrosion damage assessment database; and retrieving in the preset pipeline maintenance plan library to obtain a pipeline maintenance plan that matches the corrosion assessment result.

[0110] In one embodiment of the present application, if it is calculated that the remaining lifespan of the target pipeline steel is less than the expected operation duration, then this section of pipeline steel is replaced, a replacement plan is arranged, and it is ensured that the newly installed pipeline steel meets the design specifications and safety standards.

[0111] In another embodiment of the present application, if it is calculated that the remaining lifespan of the target pipeline steel is greater than the expected operation duration, then technologies such as ultrasonic testing, industrial CT, and phased array testing are further used to comprehensively detect the pipeline steel to obtain information such as the location, depth, height, quantity, and shape of the defects, and according to the defect parameters (such as crack location, depth, length, etc.), the corresponding corrosion assessment result is searched in the preset corrosion damage assessment database. Suppose the detection result shows that there is a small crack, its location is in the middle of the pipeline, the depth is 0.5 mm, and the length is 2 mm; the database shows that this type of crack belongs to minor corrosion damage and will not immediately affect the safety of the pipeline, but regular monitoring and repair are required. Therefore, in the preset pipeline maintenance plan library, a maintenance plan that matches the corrosion assessment result (minor corrosion damage) is retrieved.

[0112] In a specific embodiment, the found maintenance plan includes: grinding the crack area, then filling and strengthening it with high-strength repair materials, and it is recommended to conduct a recheck every 6 months. Then, according to the maintenance plan, the found crack area is ground to remove the surface oxide layer and impurities; the crack is filled with high-strength repair materials, and it is ensured that the repaired area is smooth and flat to avoid stress concentration; finally, a subsequent regular inspection plan is arranged to ensure that the pipeline steel maintains a good state during future operation.

[0113] It can be understood that through the solutions proposed in the above embodiments, it can be seen how to decide whether to update or maintain based on the remaining life of the target pipeline steel. If the remaining life is not sufficient to meet the expected operation duration, it needs to be updated in a timely manner; otherwise, the service life of the pipeline steel can be extended through detailed defect detection, corrosion assessment, and matching maintenance plans to ensure the safe and stable operation of the equipment. This systematic management method not only improves the reliability and safety of the equipment but also can effectively reduce maintenance costs and improve the overall operation efficiency.

[0114] Figure 3 It is a schematic diagram of the whole process of the steel life prediction method based on the neural network algorithm shown in an exemplary embodiment of the present application.

[0115] As Figure 3 shown, in an embodiment of the present application, the whole process of the steel life prediction method based on the neural network algorithm includes: collecting relevant data, at least including material data, operation data, and defect data; performing T92 material detection and analysis based on the collected data, and respectively constructing a BP neural network model, conducting creep-fatigue tests, and identifying damage modes based on the analysis results; during the process of constructing the BP neural network model, it is necessary to continuously improve the accuracy of the model until it meets the preset requirements. Finally, a creep-fatigue simulation test is performed based on the trained model to obtain creep-fatigue prediction data, and it is combined with the creep-fatigue data at the same stress and different temperatures obtained from the creep-fatigue test to obtain a comprehensive data set; then, the L-M parameter method is used to fit the data in the comprehensive data set to solve for the material constant C value, and further, a T92 material remaining life prediction equation is generated based on the obtained material constant C value to predict the remaining life of the target steel. If the prediction result shows that the remaining life is less than the expected operation duration, the target steel is modified or replaced; if the prediction result shows that the remaining life is greater than or equal to the expected operation duration, the damage assessment result generates future maintenance suggestions for the steel, and the damage assessment result is obtained by querying in a preset corrosion damage database based on the previously identified damage mode.

[0116] Figure 4 It is a schematic diagram of the prediction result of the creep-fatigue prediction model shown in an exemplary embodiment of the present application.

[0117] In a specific embodiment of the present application, first, the material parameters of the 9Cr series steel of the main pipeline of the ultra-supercritical unit and the operating parameters of the unit during its service are collected, including the chemical composition, microstructure, mechanical properties of the material, the actual operating temperature, pressure, operating time, fluid flow rate, etc. of the unit; and a comprehensive inspection is carried out on the internal micro-defects of the 9Cr series steel main pipeline to obtain information such as the location, depth, height, quantity, and shape of the defects. Then, by using analysis means such as SEM and EDS, the corrosion behavior of the 9Cr series steel in a high-temperature steam environment is studied, the composition and structure of the corrosion products are analyzed, and a corrosion damage assessment database based on the 9Cr series steel is established. After that, it is executed in 3 parallel logics:

[0118] First, the stress state of the 9Cr series steel and factors related to the test temperature are used as the input layer, the number of input layer nodes is 2, the output layer is the fracture time, and the number of output layer nodes is 1. The number of hidden nodes is obtained by using Equation 1, and the range of hidden nodes is [3, 8]; different hidden nodes are used to repeatedly optimize the model to improve the prediction accuracy. Finally, a BP neural network with a structure of 2-5-1 is constructed; and Equation (5) is used as the transfer function of the BP neural network model, and the continuous perceptron learning rule (Delta) is used to improve the accuracy of the model, so that the error function of Equation (6) reaches the minimum value; furthermore, Matlab is used to write the function transfer code between the input layer, hidden layer, and output layer, establish the implementation program of the model, input the relevant parameters into the neural network nodes, and thus construct the creep-fatigue prediction topological structure of the BP neural network. In addition, a total of 500 pieces of data of the 9Cr series steel are used as training data and input into the creep-fatigue prediction neural network model, and the learning function "trainbr" is used for training, the number of training times is 5000 times, and the accuracy is 0.00001; and the trained BP neural network model is used to predict the creep-fatigue fracture time of the 9Cr series steel, and it is compared with the actual value, and the neural network model is continuously corrected and optimized to enable it to accurately predict the fracture time of the 9Cr series steel. The results of the trained model are as Figure 4 shown.

[0119] Second, a creep-fatigue experiment is carried out to obtain the fracture times of 3 groups of 9Cr series steel under the same stress and different temperatures. In addition, to obtain more accurate C values and fitting effects, the creep-fatigue prediction model of the 9Cr series steel is used to additionally predict the fracture times of 7 groups under the same stress and different temperatures. The experimental data and the predicted data total 10 groups (5 data in each group). The 10 groups of data with the same stress and different temperatures are fitted to obtain the C values in 10 L-M parameter methods, and the average value is calculated. Subsequently, the fitting equations of P(σ)-σ are calculated in the case of 11 C values, and the adjustment coefficient R2 and the adjustable coefficient Adj.R2 are used to evaluate the fitting effect. The closer the value is to 1, the better the fitting effect. Finally, the life prediction equation of the 9Cr series steel is obtained.

[0120] Thirdly, compare the service data of the existing 9Cr series steel with the corrosion damage assessment database to obtain the aging and damage types of the 9Cr series steel. Input the relevant parameters into the life prediction equation of the 9Cr series steel to obtain the remaining life of the 9Cr series steel. According to the prediction results, put forward targeted maintenance and repair suggestions to ultimately ensure the safe operation of the unit.

[0121] It should be noted that the steel life prediction method based on the neural network algorithm proposed in this application has the following advantages compared with the traditional life detection method: First, by combining multiple detection means and analysis methods, it can accurately evaluate the aging and damage of the 9Cr series steel. Using the BP neural network to construct a remaining creep prediction model, the remaining life prediction equation after fitting and optimization can accurately predict the remaining life of the 9Cr series steel in the service unit, and the accuracy of its prediction results is higher; Second, this method is easy to operate. Only a small amount of relevant data needs to be provided to use the creep-fatigue prediction model to provide a large number of prediction results for the fitting of the remaining life, greatly reducing the time cost and labor consumption, making the method highly practical; Third, by predicting the material life in advance, it can effectively avoid accidents caused by material aging and failure, improve the reliability and safety of the 9Cr series steel pipelines in service in the power station, and reduce the downtime and maintenance costs.

[0122] Figure 5 It is a block diagram of a steel life prediction device based on a neural network algorithm shown in an exemplary embodiment of this application. This device can be applied to Figure 1 the implementation environment shown. This device can also be applicable to other exemplary implementation environments and be specifically configured in other devices. This embodiment does not limit the implementation environment applicable to this device.

[0123] As Figure 5 shown, this exemplary steel life prediction device based on a neural network algorithm includes:

[0124] Among them, a data acquisition module 510 is used to obtain the material parameters of the pipeline steel and the operation parameters of the unit where the pipeline steel is working; an input parameter calculation module 520 is used to determine the stress parameters of the pipeline steel based on the material parameters and operation parameters, and obtain the test temperature of the pipeline steel based on the operation parameters; a creep-fatigue prediction model training module 530 is used to construct an initial neural network model with the stress parameters and test temperature as the input nodes of the neural network and the fracture time of the pipeline steel as the output node, and train the initial neural network model to obtain a creep-fatigue prediction model; a steel life prediction module 540 is used to obtain the fracture time of the pipeline steel based on the creep-fatigue prediction model, and generate a life prediction curve of the pipeline steel according to the fracture time to predict the life of the target pipeline steel based on the life prediction curve.

[0125] It should be noted that the steel life prediction device based on the neural network algorithm provided in the above embodiments and the steel life prediction method based on the neural network algorithm provided in the above embodiments belong to the same concept. The specific ways in which each module and unit perform operations have been described in detail in the method embodiments, and will not be repeated here. In practical applications, the steel life prediction device based on the neural network algorithm provided in the above embodiments can, according to needs, allocate the above functions to different functional modules, that is, divide the internal structure of the device into different functional modules to complete all or part of the functions described above. This is not limited here either.

[0126] An embodiment of the present application also provides an electronic device, including: one or more processors; a storage device for storing one or more programs, which, when executed by the one or more processors, cause the electronic device to implement the steel life prediction method based on the neural network algorithm provided in each of the above embodiments.

[0127] Figure 6 The structural schematic diagram of a computer system of an electronic device suitable for implementing the embodiments of the present application is shown. It should be noted that Figure 6 The computer system 600 of the electronic device shown is only an example and should not impose any limitations on the functions and usage scope of the embodiments of the present application.

[0128] As Figure 6 shown, the computer system 600 includes a central processing unit (CPU) 601, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 602 or the program loaded from the storage section 608 into the random access memory (RAM) 603, such as executing the method described in the above embodiments. In the RAM 603, various programs and data required for system operation are also stored. The CPU 601, ROM 602, and RAM 603 are connected to each other through a bus 604. The input / output (I / O) interface 605 is also connected to the bus 604.

[0129] The following components are connected to the I / O interface 605: an input section 606 including a keyboard, a mouse, etc.; an output section 607 including, for example, a cathode ray tube (CRT), a liquid crystal display (LCD), etc. and a speaker, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as required. A removable medium 611 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc. is mounted on the drive 610 as required so that a computer program read therefrom is installed into the storage section 608 as required.

[0130] Specifically, according to an embodiment of the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, an embodiment of the present application includes a computer program product including a computer program carried on a computer-readable medium, the computer program including a computer program for performing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 609, and / or installed from the removable medium 611. When the computer program is executed by a central processing unit (CPU) 601, various functions defined in the system of the present application are executed.

[0131] It should be noted that the computer-readable medium shown in the embodiments of the present application can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable computer program. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium can be transmitted using any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0132] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present application. Among them, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the above module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram or flowchart, and the combination of blocks in the block diagram or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0133] The units involved in the embodiments described in this application can be implemented in software or in hardware, and the described units can also be provided in a processor. Among them, the names of these units do not, in some cases, constitute a limitation on the units themselves.

[0134] On the other hand, this application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor of a computer, the computer is enabled to execute the steel life prediction method based on a neural network algorithm as described above. The computer-readable storage medium can be included in the electronic device described in the above embodiments, or can exist alone without being assembled into the electronic device.

[0135] On the other hand, this application also provides a computer program product or a computer program. The computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the steel life prediction method based on a neural network algorithm provided in the above various embodiments.

[0136] The above embodiments are only used to exemplarily illustrate the principles and effects of the present invention, rather than to limit the present invention. Any person familiar with this technology can modify or change the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or changes completed by those with ordinary knowledge in the technical field without departing from the spirit and technical ideas disclosed by the present invention should still be covered by the claims of the present invention.

Claims

1. A steel lifespan prediction method based on a neural network algorithm, characterized in that, The method includes: Obtaining the material parameters of the pipeline steel and the operating parameters of the unit where the pipeline steel operates; Determining the stress parameters of the pipeline steel based on the material parameters and the operating parameters, and obtaining the test temperature of the pipeline steel based on the operating parameters; Taking the stress parameters and the test temperature as the input nodes of a neural network, and taking the fracture time of the pipeline steel as the output node, constructing an initial neural network model, and training the initial neural network model to obtain a creep-fatigue prediction model; Obtaining the fracture time of the pipeline steel based on the creep-fatigue prediction model, and generating a life prediction curve for the pipeline steel according to the fracture time, so as to predict the life of the target pipeline steel based on the life prediction curve.

2. The method for predicting the service life of steel based on the neural network algorithm according to claim 1, wherein Determining the stress parameters of the pipeline steel based on the material parameters and the operating parameters includes: Obtaining the ambient temperature; Determining the elastic modulus and the coefficient of linear expansion of the pipeline steel based on the material parameters, and obtaining the pipeline parameters and the operating temperature based on the operating parameters; Calculating the pipeline parameters and the operating temperature to obtain the internal pressure stress of the pipeline steel; Calculating the difference between the operating temperature and the ambient temperature to obtain the temperature difference, and calculating the temperature difference, the elastic modulus, and the coefficient of linear expansion to obtain the thermal stress of the pipeline steel; Obtaining the total stress of the pipeline steel according to the internal pressure stress and the thermal stress to generate the stress parameters of the pipeline steel.

3. The method for predicting the service life of steel based on the neural network algorithm according to claim 1, wherein, Constructing an initial neural network model includes: Identifying the input nodes and the output nodes to obtain the number of input layer nodes and the number of output layer nodes; Calculating the number of input layer nodes and the number of output layer nodes to determine the number of hidden layer nodes of the initial neural network model; Setting a first transfer function based on the number of input layer nodes and the number of hidden layer nodes, and setting a second transfer function based on the number of output layer nodes and the number of hidden layer nodes; Transmitting the test data from the input layer to the hidden layer based on the first transfer function, triggering the hidden layer to perform feature transformation on the test data, and transmitting the data after feature transformation to the output layer based on the second transfer function to generate a predicted value; Calculating the error between the predicted value and the test data, and feeding back the error to the hidden layer to update the first transfer function and the second transfer function until the error reaches a preset error range; Generating an initial neural network model based on the updated first transfer function and second transfer function.

4. The steel lifespan prediction method based on the neural network algorithm according to claim 3, wherein, Training the initial neural network model includes: Obtaining a historical data set composed of multiple stress parameters and test temperatures based on historical data, and annotating the historical data set according to the fracture time, and there is a one-to-one correspondence between the historical data set and the fracture time; Dividing the annotated historical data set into a training data set and a validation data set; Training the initial neural network model based on the training data set to learn the initial mapping relationship between the input historical data set and the fracture time; Verify the initial mapping relationship based on the verification data set, optimize the initial mapping relationship based on the verification result to obtain an optimized neural network model, and determine the optimized neural network model as the creep fatigue prediction model.

5. The method for predicting the service life of steel based on the neural network algorithm according to claim 1, wherein, Generate a life prediction model for the pipeline steel according to the fracture time, including: Obtain fracture time data at multiple different temperatures to generate test data; Generate multiple groups of prediction data based on the optimized neural network model, where the stress parameters of the test data and the prediction data are the same, and the test temperatures are different; Merge the test data and the prediction data to obtain a target data set, where the target data set includes stress parameters, test temperatures, fracture times, and the mapping relationships among the three; Calculate the data groups in the target data set to obtain the mean value of the material constants of the pipeline steel and the material constants at different test temperatures; Fit a life prediction curve based on the mean value of the material constants and / or the material constants at different test temperatures, where the life prediction curve is used to map the fracture time under the stress-temperature relationship.

6. The method for predicting the service life of steel based on the neural network algorithm according to claim 5, wherein Predict the life of the target pipeline steel based on the life prediction curve, including: Obtain the current stress parameter and the current test temperature of the target pipeline steel; Retrieve in the life prediction curve to obtain the target fracture time corresponding to the current stress parameter and the current test temperature, and obtain the remaining life of the target pipeline steel based on the target fracture time.

7. The method for predicting the service life of steel based on the neural network algorithm according to any one of claims 1-6, characterized in that, After predicting the life of the target pipeline steel, it further includes: If the remaining life of the target pipeline steel is less than the preset expected operation duration, update the target pipeline steel; If the remaining life of the target pipeline steel is greater than or equal to the preset expected operation duration, maintain the target pipeline steel, and the maintenance includes: Detect the defects of the pipeline steel to obtain the defect parameters of the pipeline steel; In the preset corrosion damage assessment database, query the corresponding corrosion assessment result according to the defect parameters; Retrieve in the preset pipeline maintenance plan library to obtain a pipeline maintenance plan matching the corrosion assessment result.

8. A device for predicting the service life of steel based on a neural network algorithm, characterized in that, The device includes: A data acquisition module for obtaining the material parameters of the pipeline steel and the operation parameters of the unit where the pipeline steel works; An input parameter calculation module for determining the stress parameter of the pipeline steel based on the material parameter and the operation parameter, and obtaining the test temperature of the pipeline steel based on the operation parameter; A creep fatigue prediction model training module for constructing an initial neural network model with the stress parameter and the test temperature as the input nodes of the neural network and the fracture time of the pipeline steel as the output node, and training the initial neural network model to obtain a creep fatigue prediction model; A steel life prediction module for obtaining the fracture time of the pipeline steel based on the creep fatigue prediction model, generating a life prediction curve for the pipeline steel according to the fracture time, and predicting the life of the target pipeline steel based on the life prediction curve.

9. An electronic device, characterized in that, It includes a processor, a memory, and a communication bus; the communication bus is used to connect the processor and the memory; the processor is used to execute the computer program stored in the memory to implement the steel lifespan prediction method based on the neural network algorithm as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, A computer program is stored thereon, and the computer program is used to cause a computer to execute the steel lifespan prediction method based on the neural network algorithm as described in any one of claims 1-7.

Images