A method and device for predicting crack propagation rate and evaluating failure of a hydrogen transmission pipeline

Abstract

The application discloses a hydrogen transmission pipeline crack propagation rate prediction and failure evaluation method and device. The method comprises the following steps: obtaining material attribute parameters and pipeline geometric parameters, obtaining a hydrogen transmission pipeline model library comprising a prefabricated crack through modeling and parameterization configuration; obtaining mechanical response data and hydrogen concentration distribution data through hydrogen-force coupling numerical simulation, correcting an initial fatigue crack propagation model based on the hydrogen concentration distribution data, converting the mechanical response data to obtain crack propagation rate data and constructing a model training data set; taking working condition parameters as input features and crack propagation rate as output labels, training to obtain an intelligent prediction model; finally, inputting real-time working condition parameters of a pipeline to be evaluated to obtain a prediction result, mapping through a failure evaluation graph and comparing with a safety boundary, and outputting a pipeline failure risk level and a safety state evaluation result. The application can replace high-pressure hydrogen environment entity test, give consideration to evaluation accuracy and prediction efficiency, and provide reliable support for hydrogen-doped gas transmission pipeline safety operation and maintenance.

Description

Technical Field

[0001] This application relates to the field of oil and gas pipeline safety engineering technology, and in particular to a method and apparatus for predicting crack propagation rate and assessing failure in hydrogen pipelines. Background Technology

[0002] Utilizing existing natural gas pipeline networks for hydrogen blending and transportation has become the mainstream method for achieving low-cost, long-distance hydrogen transport. However, the introduction of hydrogen can trigger hydrogen embrittlement in pipeline steel, significantly reducing the material's fracture toughness, accelerating crack propagation, and easily leading to pipeline leaks, explosions, and other safety accidents, seriously threatening the long-term stable operation of hydrogen transportation networks. To ensure the safe operation of hydrogen transportation pipelines, it is necessary to accurately predict the crack propagation rate and conduct scientific and efficient assessments of pipeline failure risks.

[0003] Current methods for assessing crack propagation and failure in hydrogen pipelines typically rely on physical fatigue tests under high-pressure hydrogen conditions combined with traditional empirical formulas in fracture mechanics. These methods usually require obtaining material property parameters through slow strain rate tensile testing and autoclave fatigue testing, followed by calculation of crack propagation rate and risk assessment using empirical formulas. However, these physical tests place extremely high demands on equipment, pose safety hazards due to hydrogen leaks and explosions, and are time-consuming and costly, making it difficult to cover various combinations of hydrogen doping ratios, pressure fluctuations, and crack sizes encountered in actual pipeline operation. Furthermore, traditional methods do not fully consider the strong coupling effect between hydrogen diffusion and the stress field, resulting in insufficient accuracy in crack propagation prediction. Conventional finite element simulation methods are also time-consuming and cannot meet the real-time online assessment needs of the massive number of defects along the entire long-distance pipeline. Ultimately, this makes it difficult to guarantee the accuracy and timeliness of hydrogen pipeline failure risk assessment.

[0004] Therefore, how to avoid the safety risks of high-pressure hydrogen physical testing, while achieving rapid and accurate prediction of crack propagation rate in hydrogen pipelines and efficiently completing the quantitative assessment of pipeline failure risk and safety status, is a technical problem that urgently needs to be solved. Summary of the Invention

[0005] In view of this, the method and apparatus for predicting the crack propagation rate and assessing the failure of a hydrogen transportation pipeline provided in this application can avoid the safety risks of high-pressure hydrogen physical testing, while achieving rapid and accurate prediction of the crack propagation rate of the hydrogen transportation pipeline, and efficiently completing the quantitative assessment of pipeline failure risk and safety status. The method and apparatus for predicting the crack propagation rate and assessing the failure of a hydrogen transportation pipeline provided in this application are implemented as follows: This application provides a method for predicting the crack propagation rate and assessing the failure of a hydrogen transportation pipeline, including: Obtain material property parameters and pipeline geometric parameters, and perform modeling and configuration processing on the material property parameters and pipeline geometric parameters to obtain a hydrogen transportation pipeline model library including pre-cracked ones; Based on the hydrogen pipeline model library, hydrogen-force coupling numerical simulation was performed on the hydrogen pipeline to obtain mechanical response data and hydrogen concentration distribution data. Based on the hydrogen concentration distribution data, the initial fatigue crack propagation model is modified to address the hydrogen embrittlement effect, resulting in the target modified crack propagation model. Based on the target modified crack propagation model, the mechanical response data is transformed to obtain crack propagation rate data. The crack propagation rate data is then paired and integrated to obtain the model training dataset. Using operating condition parameters as input features and the crack propagation rate as output label, the model training dataset is subjected to model training and accuracy verification processing to obtain an intelligent prediction model. The operating condition parameters include at least one of crack size, pipeline internal pressure, and hydrogen partial pressure. The real-time operating parameters of the hydrogen pipeline to be evaluated are obtained, and the real-time operating parameters are input into the intelligent prediction model for prediction processing to obtain the crack propagation rate prediction result. The crack propagation rate prediction results and the real-time operating parameters are subjected to failure assessment map mapping and safety boundary comparison processing to obtain the pipeline failure risk level and safety status assessment results.

[0006] In some embodiments, the process of training and verifying the model training dataset using operating condition parameters as input features and crack propagation rate as output label to obtain an intelligent prediction model includes: The model training dataset is preprocessed to obtain a standardized training dataset; The standardized training dataset is shuffled and split to obtain training and test subsets. An initial training model is constructed, and the training subset is input into the initial training model for updating and optimization to obtain an initial prediction model. The initial prediction model is subjected to accuracy verification based on the test subset to obtain an intelligent prediction model.

[0007] In some embodiments, the process of mapping the crack propagation rate prediction results and the real-time operating parameters to a failure assessment map and comparing them with a safety boundary to obtain the pipeline failure risk level and safety status assessment results includes: A two-dimensional evaluation coordinate system composed of fracture ratio and load ratio is constructed to obtain the failure evaluation benchmark diagram; Based on the hydrogen partial pressure data in the real-time operating parameters of the hydrogen pipeline to be evaluated, the fracture toughness value is dynamically corrected to obtain the fracture toughness correction value. Based on the fracture toughness correction value, the failure assessment benchmark map is adjusted to obtain a dedicated failure assessment map. Based on the crack propagation rate prediction results and real-time operating parameters, the fracture ratio and load ratio corresponding to the current crack state are calculated and processed to obtain the coordinate point data to be evaluated. The coordinate point data to be evaluated is mapped onto the dedicated failure evaluation map, and the relative position of the coordinate point data to be evaluated and the safety boundary envelope curve is compared to obtain the position comparison result. Based on the location comparison results, failure mode determination and safety risk level classification are performed to obtain the pipeline failure risk level and safety status assessment results.

[0008] In some embodiments, the mechanical response data is transformed based on the target modified crack propagation model to obtain crack propagation rate data, and the crack propagation rate data is paired and integrated to obtain a model training dataset, including: Based on the target modified crack propagation model, the mechanical response data is transformed and calculated to obtain crack propagation rate data; The operating parameters and the crack propagation rate data are bound together to obtain multiple sets of structured data units; Outlier removal and format standardization are performed on the multiple sets of structured data units to obtain a regularized dataset; The regularized dataset is matched and bound to obtain the model training dataset.

[0009] In some embodiments, the step of obtaining material property parameters and pipeline geometric parameters, and modeling and configuring the material property parameters and pipeline geometric parameters to obtain a hydrogen transport pipeline model library including pre-existing cracks includes: The material property parameters and the pipe geometric parameters are collected and standardized to obtain the basic parameters; A three-dimensional solid model is constructed based on the aforementioned basic parameters. The three-dimensional solid model is then subjected to fine mesh division to obtain an initial pipeline simulation model. The value range of the operating condition parameters is configured to obtain parameterized control rules; Based on the parametric control rules, the initial pipeline simulation model is adapted and updated to obtain a hydrogen transportation pipeline model library.

[0010] In some embodiments, the step of correcting the initial fatigue crack propagation model for hydrogen embrittlement based on the hydrogen concentration distribution data to obtain the target corrected crack propagation model includes: The initial fatigue crack propagation model is analyzed to obtain the initial model baseline framework; The hydrogen concentration distribution data is fitted to obtain the hydrogen embrittlement influencing factor; Based on the hydrogen embrittlement influencing factor, the initial model baseline framework is modified to obtain the crack propagation model; The crack propagation model is subjected to adaptability verification and parameter optimization to obtain the target modified crack propagation model.

[0011] In some embodiments, the crack size-to-crack depth ratio ranges from 0.1 to 0.7, the internal pressure of the pipeline ranges from 4 MPa to 8 MPa, and the hydrogen partial pressure ranges from 0 MPa to 1.0 MPa.

[0012] This application provides an embodiment of a device for predicting the crack propagation rate and assessing the failure of a hydrogen transportation pipeline, comprising: The acquisition module is used to acquire material property parameters and pipeline geometric parameters, and to perform modeling and configuration processing on the material property parameters and pipeline geometric parameters to obtain a hydrogen transport pipeline model library including pre-cracked structures. The processing module is used to perform hydrogen-force coupling numerical simulation processing on the hydrogen pipeline based on the hydrogen pipeline model library to obtain mechanical response data and hydrogen concentration distribution data. The processing module is also used to perform hydrogen embrittlement correction processing on the initial fatigue crack propagation model based on the hydrogen concentration distribution data to obtain the target corrected crack propagation model. The processing module is further configured to transform the mechanical response data based on the target modified crack propagation model to obtain crack propagation rate data, and to perform pairing and integration processing on the crack propagation rate data to obtain a model training dataset. The training module is used to train and verify the model on the model training dataset by taking the operating condition parameters as input features and the crack propagation rate as the output label, so as to obtain an intelligent prediction model. The operating condition parameters include at least one of crack size, pipeline internal pressure and hydrogen partial pressure. The acquisition module is also used to acquire real-time operating parameters of the hydrogen pipeline to be evaluated, input the real-time operating parameters into the intelligent prediction model for prediction processing, and obtain the crack propagation rate prediction result. The processing module is also used to perform failure assessment map mapping and safety boundary comparison processing on the crack propagation rate prediction results and the real-time operating parameters to obtain the pipeline failure risk level and safety status assessment results.

[0013] The computer device provided in this application includes a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the program, it implements the method described in this application.

[0014] The computer-readable storage medium provided in this application embodiment stores a computer program thereon, which, when executed by a processor, implements the method provided in this application embodiment.

[0015] This application provides a method and apparatus for predicting crack propagation rate and assessing failure in hydrogen transportation pipelines. It acquires material property parameters and pipeline geometric parameters, and through modeling and parameterization, obtains a hydrogen transportation pipeline model library including pre-existing cracks. Mechanical response data and hydrogen concentration distribution data are obtained through hydrogen-force coupled numerical simulation. Based on the hydrogen concentration distribution data, the initial fatigue crack propagation model is corrected, and the mechanical response data is transformed into crack propagation rate data to construct a model training dataset. Using operating parameters as input features and crack propagation rate as the output label, an intelligent prediction model is trained. Finally, the real-time operating parameters of the pipeline to be assessed are input to obtain the prediction results. After failure assessment map mapping and safety boundary comparison, the pipeline failure risk level and safety status assessment results are output. This approach can replace high-pressure hydrogen environment physical testing, balancing assessment accuracy and prediction efficiency, providing reliable support for the safe operation and maintenance of hydrogen-blended gas transportation pipelines, and solving the technical problems mentioned in the background art. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A schematic diagram illustrating the implementation process of a method for predicting the crack propagation rate and assessing the failure of a hydrogen transportation pipeline, provided in an embodiment of this application. Figure 2 A schematic diagram illustrating the implementation process of acquiring an intelligent prediction model, provided in an embodiment of this application; Figure 3 This is a schematic diagram of a device for predicting the crack propagation rate and assessing the failure of a hydrogen transportation pipeline, provided in an embodiment of this application. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0019] The following description of some technologies involved in the embodiments of this application is provided to aid understanding and should be considered merely exemplary. Therefore, those skilled in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this application. Similarly, for clarity and brevity, some descriptions of well-known functions and structures are omitted in the following description.

[0020] Figure 1 This is a schematic diagram illustrating the implementation flow of a method for predicting the crack propagation rate and assessing the failure of a hydrogen transportation pipeline, as provided in an embodiment of this application, including steps 101 to 107. Wherein, Figure 1 This is merely one execution order shown in the embodiments of this application and does not represent the only execution order for a method of predicting crack propagation rate and assessing failure in hydrogen transportation pipelines. Where the final result can be achieved, the following execution order may be used: Figure 1 The steps shown can be performed in parallel or in reverse order.

[0021] Step 101: Obtain material property parameters and pipeline geometric parameters, and perform modeling and configuration processing on the material property parameters and pipeline geometric parameters to obtain a hydrogen transport pipeline model library including pre-cracked ones.

[0022] In this embodiment, the obtained pipeline steel material properties include basic mechanical properties such as elastic modulus, Poisson's ratio, yield strength, and fracture toughness. The pipeline geometric parameters include structural dimensions such as outer diameter, wall thickness, and pipe length. These parameters are imported into general-purpose finite element analysis software to construct a three-dimensional solid model of the pipeline containing a semi-elliptical pre-fabricated crack on the inner wall. The crack tip region is then finely meshed to ensure the accuracy of the simulation calculation. Simultaneously, the crack depth ratio, pipeline internal pressure amplitude, and hydrogen partial pressure are set as variable input parameters, and the value ranges of each parameter are configured. The crack depth ratio covers the full gradient from 0.1 to 0.7, the pipeline internal pressure amplitude covers the fluctuation range of 4 MPa to 8 MPa, and the hydrogen partial pressure covers the full hydrogen doping ratio range of 0 MPa to 1.0 MPa. Based on the configuration of the above variable input parameters, the parameterization adaptation of the three-dimensional solid model is completed, enabling automatic updates of the model corresponding to the working conditions by adjusting a single parameter. Finally, a parameterized model library of hydrogen transportation pipelines that can be called in batches is obtained.

[0023] Step 102: Perform hydrogen-force coupling numerical simulation on the hydrogen pipeline based on the hydrogen pipeline model library to obtain mechanical response data and hydrogen concentration distribution data.

[0024] In this embodiment, a sequentially coupled or fully coupled numerical algorithm is used to simulate the diffusion process of hydrogen atoms within the pipe wall and their enrichment process at the crack tip. First, mechanical simulation calculations are performed on the three-dimensional pipe model under different operating conditions to obtain stress field distribution data for the entire pipe and the crack tip. Then, based on the hydrostatic stress gradient of the current stress field, the hydrogen atom concentration distribution within the pipe wall is updated to obtain hydrogen concentration field distribution data for the entire pipe wall. After completing batch simulation calculations for all operating conditions, the simulation results are automatically post-processed. The maximum and minimum stress intensity factors at the crack front edge are extracted in each calculation increment step to calculate the effective stress intensity factor range data. Simultaneously, the average hydrogen concentration data within the plastic zone at the crack tip is extracted, ultimately obtaining the mechanical response data and hydrogen concentration distribution data corresponding to the crack tip under different operating conditions.

[0025] Step 103: Based on the hydrogen concentration distribution data, the initial fatigue crack propagation model is modified to correct for hydrogen embrittlement, thus obtaining the target modified crack propagation model.

[0026] In this embodiment, the initial fatigue crack propagation model adopts the classic Paris fatigue crack propagation model commonly used in oil and gas pipeline engineering. First, the basic structure and inherent material constants of the classic Paris fatigue crack propagation model are analyzed to determine the coefficients affected by hydrogen embrittlement as the coefficients to be corrected. The basic calculation logic and fixed structure unaffected by the hydrogen environment are then removed to obtain the baseline framework of the initial model. Next, based on the aforementioned crack tip hydrogen concentration distribution data, the correlation between hydrogen embrittlement and crack tip hydrogen concentration is fitted to obtain the hydrogen embrittlement influence factor that dynamically changes with the crack tip hydrogen concentration. Using this hydrogen embrittlement influence factor, the coefficients to be corrected in the baseline framework of the initial model are dynamically adapted and corrected, so that the model coefficients can be adjusted synchronously with the change in crack tip hydrogen concentration, resulting in a preliminarily corrected crack propagation model. The corrected model expression is as follows: ,in, The fatigue crack propagation rate is expressed in meters per cycle (m / cycle), representing the increase in crack length for a single load cycle under alternating internal pressure loading. The first material constant in the Paris formula is a dynamic material coefficient related to the hydrogen concentration at the crack tip. Its value changes dynamically with the hydrogen concentration at the crack tip and is used to quantify the accelerating effect of hydrogen embrittlement on crack propagation. The effective stress intensity factor range at the crack tip, in MPa. m¹ / ² refers to the difference between the maximum and minimum values ​​of the crack tip stress intensity factor within one load cycle, and is a core mechanical parameter driving fatigue crack propagation. The second material constant in the Paris formula is a dynamic material index related to the hydrogen concentration at the crack tip. Its value changes dynamically with the hydrogen concentration at the crack tip, characterizing the sensitivity of the crack propagation rate to the range of stress intensity factors under hydrogen embrittlement conditions.

[0027] Finally, the adaptability verification and parameter optimization of the initially modified crack propagation model were carried out to ensure that the model can accurately reflect the fatigue crack propagation law of pipeline steel under hydrogen transportation environment, and finally obtain the target modified crack propagation model adapted to hydrogen transportation environment.

[0028] Step 104: Based on the target modified crack propagation model, the mechanical response data is transformed to obtain crack propagation rate data. The crack propagation rate data is then paired and integrated to obtain the model training dataset.

[0029] In this embodiment, using the aforementioned target-corrected crack propagation model, the effective stress intensity factor range at the crack tip in the mechanical response data corresponding to each set of working conditions is substituted into the corrected Paris formula. Combined with the dynamic material parameters corresponding to the hydrogen concentration at the crack tip under that working condition, the fatigue crack propagation rate corresponding to the pipe crack under that working condition is calculated and transformed. The working condition parameters corresponding to each set of working conditions, including crack size, pipe internal pressure, and hydrogen partial pressure, are bound one-to-one with the crack propagation rate data calculated under that working condition to form multiple sets of structured data units. Outlier removal processing is performed on all structured data units to remove invalid data caused by simulation calculation deviations. The remaining data is then format-standardized to obtain a regularized dataset. Finally, the input features and output labels are matched and bound on the regularized dataset to obtain a model training dataset suitable for subsequent model training.

[0030] Step 105: Using the working condition parameters as input features and the crack propagation rate as the output label, perform model training and accuracy verification on the model training dataset to obtain the intelligent prediction model.

[0031] In this embodiment, the model training dataset is first standardized by removing dimensions to eliminate the interference of dimensional differences between different features on model training, resulting in a standardized training dataset. The standardized training dataset is then randomly shuffled to eliminate the influence of data order on model training. The shuffled dataset is then split into a training subset and a test subset according to a preset ratio. The training subset is used for iterative model training, and the test subset is used for model accuracy verification. A multi-layer backpropagation neural network is constructed as a machine learning regression model. The input feature dimension, output label dimension, and number of network layers are configured, where the input features are working condition parameters and the output label is the crack propagation rate. The training subset is input into the constructed initial training model, and the model's weight parameters are repeatedly updated and tuned using an iterative optimization algorithm to obtain an initial prediction model. The test subset is then input into the initial prediction model to verify the model's prediction accuracy and generalization ability. Upon successful verification, a final intelligent prediction model suitable for predicting crack propagation rate is obtained.

[0032] Step 106: Obtain the real-time operating parameters of the hydrogen pipeline to be evaluated, input the real-time operating parameters into the intelligent prediction model for prediction processing, and obtain the crack propagation rate prediction result.

[0033] In this embodiment, real-time operating parameters of the hydrogen pipeline to be evaluated are obtained, including the current crack size, internal pressure amplitude, and hydrogen partial pressure in the transport medium. The obtained real-time operating parameters are subjected to standardized preprocessing consistent with the training data, and then the preprocessed real-time operating parameters are input into the trained intelligent prediction model. The model can output the prediction result of the pipeline crack propagation rate under the real-time operating condition in milliseconds without performing complex finite element simulation calculations, and at the same time output the critical crack size data of the pipeline under the operating condition.

[0034] Step 107: Perform failure assessment map mapping and safety boundary comparison on the crack propagation rate prediction results and real-time operating parameters to obtain the pipeline failure risk level and safety status assessment results.

[0035] In this embodiment, a two-dimensional evaluation coordinate system composed of fracture ratio and load ratio is first constructed, and the corresponding safety boundary envelope curve is plotted to obtain a failure evaluation benchmark diagram for hydrogen pipelines. The fracture ratio characterizes the utilization degree of the material's fracture toughness, and the load ratio characterizes the degree of plastic instability of the pipeline. Based on the hydrogen partial pressure data in the real-time operating parameters of the hydrogen pipeline to be evaluated, the fracture toughness value of the pipeline steel is dynamically corrected to quantify the decrease in material toughness caused by hydrogen embrittlement, thus obtaining a material fracture toughness correction value adapted to the current hydrogen environment. Based on this correction value, the failure evaluation benchmark diagram is then used to evaluate the fracture toughness of the pipeline steel. The safety boundary envelope curve is parameter-adjusted to obtain a dedicated failure assessment map adapted to the current operating condition. Based on the crack propagation rate prediction results and real-time operating parameters, the fracture ratio and load ratio corresponding to the current crack state of the pipeline are calculated to determine the coordinate points to be assessed. The coordinate points to be assessed are mapped onto the dedicated failure assessment map, and the relative positions of the coordinate points and the safety boundary envelope curve are compared to obtain the position comparison results. Finally, based on the position comparison results, the pipeline failure mode is determined and the safety risk level is classified, and the pipeline failure risk level and safety status assessment results are output.

[0036] If the coordinate point is located inside the safety boundary envelope curve, it is determined to be in a safe state, and a safe operation prompt is output; if the coordinate point is close to or beyond the safety boundary envelope curve and the fracture ratio is high, it is determined to be at risk of hydrogen-induced embrittlement fracture, and a depressurization operation suggestion is output; if the coordinate point is close to or beyond the safety boundary envelope curve and the load ratio is high, it is determined to be at risk of plastic collapse, and a shutdown and maintenance suggestion is output.

[0037] This application's embodiments construct a complete closed-loop methodology for predicting crack propagation rates and assessing failures in hydrogen transportation pipelines. By replacing physical fatigue tests under high-pressure hydrogen conditions with numerical simulation, it completely avoids the safety risks of hydrogen leakage and explosion associated with physical tests. This significantly shortens the testing cycle and reduces the economic cost of pipeline integrity assessment. It can cover a full range of hydrogen doping ratios, pressure fluctuations, and crack size combinations in actual hydrogen transportation pipeline operation, solving the problem that traditional methods struggle to cover multiple operating scenarios. It fully considers the strong coupling effect of hydrogen diffusion and stress field under hydrogen transportation conditions, and corrects for hydrogen embrittlement effects in the fatigue crack propagation model using hydrogen concentration distribution data. This ensures the physical consistency and engineering accuracy of crack propagation rate calculations, addressing the core pain points of traditional models that are not adapted to the hydrogen transportation environment and suffer from large prediction deviations due to the lack of quantification of hydrogen embrittlement effects. By integrating numerical simulation and machine learning technologies, the trained intelligent prediction model can achieve millisecond-level rapid prediction of crack propagation rate without repeatedly running complex finite element simulations. Combined with failure assessment diagrams, it completes the quantitative assessment of pipeline failure risk, balancing prediction accuracy and assessment efficiency. It can adapt to the real-time online assessment needs of massive defect points along the entire long-distance pipeline, providing reliable technical support and data basis for the integrity management and safe operation and maintenance of hydrogen-blended natural gas pipelines.

[0038] In the above Figure 1 Based on the above, this application embodiment also provides a schematic diagram of the implementation process for obtaining an intelligent prediction model. For example... Figure 2 As shown, steps 201 to 204 are included: Step 201: Preprocess the model training dataset to obtain a standardized training dataset.

[0039] In this embodiment, the input features and output labels in the model training dataset are first sorted out. The input features are three types of working condition parameters: crack size, pipeline internal pressure, and hydrogen partial pressure. The output labels are the crack propagation rate data under the corresponding working condition. The sorted full dataset is then subjected to dimensionless standardization processing to map the values ​​of all parameters to a uniform numerical range, eliminating the interference caused by differences in dimensions and numerical magnitudes of different parameters on model training. Simultaneously, outlier data in the dataset is removed to eliminate invalid data caused by simulation calculation deviations, avoiding outliers from affecting the stability of model training. Finally, a standardized training dataset with a uniform format and even distribution is obtained.

[0040] Step 202: The standardized training dataset is shuffled and split to obtain training subsets and test subsets.

[0041] In this embodiment, the standardized training dataset is randomly shuffled to disrupt its original order arranged according to the gradient of operating conditions. This avoids the regularity of the data arrangement interfering with the model training effect and ensures the model's generalization ability. After the shuffling, the dataset is split according to a preset fixed ratio. 80% of the dataset is divided into a training subset for subsequent iterative training and parameter tuning of the model, and the remaining 20% ​​is divided into a test subset for subsequent verification of the model's prediction accuracy and generalization ability. After the split, the operating condition coverage and data distribution characteristics of the training and test subsets are verified to ensure that both sets of data fully cover the full range of crack size, pipeline internal pressure, and hydrogen partial pressure operating conditions, and that the data distribution characteristics are consistent. This avoids overfitting or underfitting problems caused by uneven data distribution, and finally obtains training and test subsets that meet the model training requirements.

[0042] Step 203: Construct an initial training model by inputting a subset of training data into the initial training model for updating and optimization, thereby obtaining the initial prediction model.

[0043] In this embodiment, a backpropagation neural network with three hidden layers is constructed as the initial training model. The number of nodes in the input layer is configured according to the number of input working condition parameters, and the number of nodes in the output layer is configured according to the output crack propagation rate data. At the same time, the basic training parameters such as the number of hidden layer nodes, the initial learning rate, the maximum number of iterations, and the loss function are configured to complete the initial training model. The aforementioned training subset is input into the initial training model. The working condition parameters are used as the model input, and the crack propagation rate of the corresponding working condition is used as the expected output of the model. Through the gradient descent iterative optimization algorithm, the weight parameters inside the model are repeatedly updated and adjusted to continuously reduce the deviation between the model prediction result and the actual crack propagation rate obtained by simulation calculation, until the model's loss function value drops to a preset stable range and no longer decreases significantly. The iterative training and parameter tuning of the model are completed, and finally, the initial prediction model that has been preliminarily trained is obtained.

[0044] Step 204: Perform accuracy verification processing on the initial prediction model based on the test subset to obtain the intelligent prediction model.

[0045] In this embodiment, the aforementioned test subset is input into the trained initial prediction model. The model automatically outputs the corresponding crack propagation rate prediction value based on the operating condition parameters in the test subset. The predicted value is compared with the corresponding actual crack propagation rate value in the test subset one by one to calculate the overall prediction deviation of the model, thus verifying the model's prediction accuracy and generalization ability across all operating conditions. If the overall prediction deviation of the model is within a preset allowable range and has stable prediction capabilities for all operating conditions, including low hydrogen doping, high hydrogen doping, shallow cracks, and deep cracks, the model is deemed to have passed verification, and the initial prediction model is determined as the final intelligent prediction model for crack propagation rate of hydrogen pipelines that can be used for engineering applications. If the model's prediction deviation exceeds a preset range, or the prediction effect for some operating conditions is not up to standard, the model training stage is returned to. The network structure and training parameters of the model are adjusted and optimized, and the model training and accuracy verification are carried out again until the model passes all verification requirements.

[0046] This application's embodiments eliminate interference from differences in the dimensions and numerical magnitudes of parameters under different operating conditions on model training through standardized preprocessing of the training dataset. Abnormal and invalid data are removed, ensuring the stability and uniformity of the training data and avoiding model training bias caused by data defects. Through comprehensive control of the entire process, including random shuffling and standardized splitting of the dataset, iterative model optimization, and multi-dimensional accuracy verification, overfitting and underfitting problems are effectively avoided. This significantly improves the generalization ability and prediction accuracy of the intelligent prediction model across all operating conditions, ensuring the model's prediction stability under different hydrogen doping ratios and operating conditions, and further enhancing the efficiency and accuracy of crack propagation rate prediction.

[0047] In some embodiments, the crack propagation rate prediction results and real-time operating parameters are processed by failure assessment map mapping and safety boundary comparison to obtain the pipeline failure risk level and safety status assessment results, including: constructing a two-dimensional assessment coordinate system composed of fracture ratio and load ratio to obtain a failure assessment benchmark map.

[0048] Specifically, the fracture ratio, which characterizes the degree of utilization of material fracture toughness, is used as the vertical axis of a two-dimensional coordinate system, and the load ratio, which characterizes the degree of plastic instability of the pipeline, is used as the horizontal axis of the two-dimensional coordinate system to complete the construction of a two-dimensional coordinate system for hydrogen pipeline failure assessment. Combining the structural integrity assessment specifications commonly used in oil and gas pipeline engineering, a safety boundary envelope curve is drawn within the coordinate system to distinguish between the safe region and the failure region, and the limit range of the safety boundary envelope curve is clarified, ultimately forming a failure assessment benchmark map that can be used for hydrogen pipeline failure risk assessment.

[0049] Furthermore, based on the hydrogen partial pressure data in the real-time operating parameters of the hydrogen pipeline to be evaluated, the fracture toughness value is dynamically corrected to obtain the fracture toughness correction value.

[0050] Specifically, the initial fracture toughness benchmark value of the pipeline steel used in the hydrogen pipeline to be evaluated is first obtained in a hydrogen-free environment. Then, combined with the hydrogen partial pressure data in the real-time operating parameters, the hydrogen embrittlement performance degradation law of the pipeline steel under this hydrogen partial pressure environment is matched, and the reduction of material fracture toughness caused by hydrogen embrittlement effect is quantified. The initial fracture toughness benchmark value is dynamically corrected according to this degradation range to eliminate the deviation of the influence of hydrogen embrittlement effect on material toughness, and finally, the fracture toughness correction value of the pipeline steel adapted to the current hydrogen partial pressure environment is obtained.

[0051] Furthermore, the failure assessment baseline map is adjusted based on the fracture toughness correction value to obtain a dedicated failure assessment map.

[0052] Specifically, the relevant parameters of the vertical axis of the safety boundary envelope curve in the failure assessment benchmark are adjusted synchronously to correct the safety boundary offset caused by hydrogen embrittlement, so that the adjusted safety boundary envelope curve is fully adapted to the current hydrogen partial pressure environment of the pipeline under assessment. After the parameter adjustment is completed, the correspondence between the coordinate system and the safety boundary envelope curve is verified to ensure the accuracy of the assessment benchmark, and finally a specific failure assessment chart that is only applicable to the current operating condition under assessment is obtained.

[0053] Furthermore, based on the crack propagation rate prediction results and real-time operating parameters, the fracture ratio and load ratio corresponding to the current crack state are calculated and processed to obtain the coordinate point data to be evaluated.

[0054] Specifically, by combining the real-time operating parameters of the pipeline to be evaluated, including the pipeline operating internal pressure, crack size, pipeline wall thickness, material yield strength, and the crack tip stress state corresponding to the crack propagation rate prediction results, the fracture ratio and load ratio corresponding to the current crack state are calculated respectively. The calculated load ratio is used as the abscissa parameter and the fracture ratio is used as the ordinate parameter to form the coordinate point data of the pipeline to be evaluated in a two-dimensional coordinate system.

[0055] Furthermore, the coordinate point data to be evaluated is mapped onto a dedicated failure assessment map, and the relative positions of the coordinate point data to be evaluated and the safety boundary envelope curve are compared to obtain the position comparison results.

[0056] Specifically, the coordinate points to be evaluated obtained above are precisely projected onto the two-dimensional coordinate system of the dedicated failure assessment map to clarify the specific location of the coordinate points within the coordinate system; the relative positional relationship between the coordinate points and the safety boundary envelope curve is compared to confirm whether the coordinate points are located inside the safety boundary envelope curve, near the curve edge, or beyond the curve edge; at the same time, it is confirmed whether the coordinate points are closer to the vertical coordinate or the horizontal coordinate within the coordinate system, thus forming a complete position comparison result.

[0057] Furthermore, based on the location comparison results, failure mode determination and safety risk level classification are performed to obtain the pipeline failure risk level and safety status assessment results.

[0058] Specifically, if the location comparison results show that the coordinate point is located inside the safety boundary envelope curve, the pipeline is determined to be in a safe operating state with a low risk level, and a green safe operation prompt is output. If the location comparison results show that the coordinate point is close to the safety boundary envelope curve and is closer to the vertical axis, the pipeline is determined to have a high risk of hydrogen-induced embrittlement fracture with a medium-high risk level, and a yellow warning and pressure reduction operation suggestion are output. If the location comparison results show that the coordinate point is close to or exceeds the safety boundary envelope curve and is closer to the horizontal axis, the pipeline is determined to have a high risk of plastic collapse with an extremely high risk level, and a red alarm and immediate shutdown and maintenance suggestion are output. Finally, the failure mode determination results, risk level classification results, and handling suggestions are integrated to form a complete pipeline failure risk level and safety status assessment result.

[0059] This application's embodiments dynamically correct the fracture toughness of pipeline steel using real-time hydrogen partial pressure data and simultaneously adjust the safety boundary envelope curve of the failure assessment diagram. This solves the problems of traditional failure assessment methods, such as the failure to quantify hydrogen embrittlement leading to material toughness degradation and the large assessment deviation caused by the mismatch between the assessment benchmark and actual operating conditions, ensuring the relevance and accuracy of the assessment benchmark. Through precise mapping of coordinate points and comparison of relative positions, the two core failure modes of the pipeline—hydrogen-induced embrittlement fracture and plastic collapse—can be clearly distinguished. At the same time, accurate classification of safety risk levels is completed, and operation and maintenance suggestions adapted to on-site operating conditions are output. This realizes the full-process implementation from crack propagation rate prediction to risk quantification assessment and engineering treatment plan output, further improving the accuracy and engineering practicality of hydrogen pipeline failure risk assessment.

[0060] In some embodiments, the mechanical response data is transformed based on the target modified crack propagation model to obtain crack propagation rate data, and the crack propagation rate data is paired and integrated to obtain a model training dataset, including: transforming and calculating the mechanical response data based on the target modified crack propagation model to obtain crack propagation rate data.

[0061] Specifically, the target modified crack propagation model is a modified Paris fatigue crack propagation model adapted to the hydrogen transport environment and whose material parameters can be dynamically adjusted according to the hydrogen concentration at the crack tip; the mechanical response data are the effective stress intensity factor range data at the crack tip under different operating conditions and the average hydrogen concentration data in the corresponding crack tip plastic zone obtained from hydrogen-force coupled numerical simulation; the effective stress intensity factor range data under each set of operating conditions is substituted into the target modified crack propagation model corresponding to the crack tip hydrogen concentration under that operating condition, and the fatigue crack propagation rate data corresponding to the pipeline crack under that set of operating conditions is obtained through dynamic adaptation calculation of the model; after completing the batch calculation of the full set of operating conditions, the full set of crack propagation rate data covering the full hydrogen doping ratio, the full pressure fluctuation range, and the full crack depth gradient is obtained.

[0062] Furthermore, the operating parameters and crack propagation rate data are bound together to obtain multiple sets of structured data units.

[0063] Specifically, the operating parameters are the variable input parameters corresponding to each set of numerical simulations, including three core parameters: crack size, pipeline internal pressure amplitude, and hydrogen partial pressure. The operating parameters corresponding to each set of operating conditions are bound one-to-one with the crack propagation rate data calculated under that set of operating conditions, so that each set of operating parameters matches a unique crack propagation rate result, forming a set of independent and complete structured data units. By completing the pairing and binding of all operating conditions in the above manner, hundreds of sets of structured data units covering the entire range of operating conditions are finally obtained.

[0064] Furthermore, outlier removal and format standardization are performed on multiple sets of structured data units to obtain a regularized dataset.

[0065] Specifically, the validity of all structured data units is first verified, and invalid data units are removed due to non-convergence of numerical simulation calculations, abnormal mechanical response data caused by parameter mismatch, and crack propagation rate values ​​exceeding the reasonable range of engineering. At the same time, incomplete data units with missing parameters or disordered pairings are also removed to avoid invalid data interfering with the subsequent model training effect. After the outlier removal is completed, the format of all remaining valid structured data units is standardized to unify the parameter format, numerical precision and arrangement rules of all data units, eliminate the format differences between different data units, and finally obtain a well-organized dataset with uniform format, valid data and uniform distribution.

[0066] Furthermore, the regularized dataset is matched and bound to obtain the model training dataset.

[0067] Specifically, for each structured data unit in the regularized dataset, feature and label decomposition is performed. The crack size, pipeline pressure amplitude, and hydrogen partial pressure within the data unit are set as the three operating condition parameters as model input features, and the corresponding crack propagation rate data within the data unit is set as the model output label. The input features and output labels of each set of data are matched and bound one by one to clarify the corresponding mapping relationship and ensure that each set of features matches a unique label. After the matching and binding of the entire dataset is completed, the correspondence between features and labels in the dataset is fully verified to ensure that the mapping relationship is accurate. Finally, a model training dataset that meets the training requirements of machine learning regression models is obtained.

[0068] This application's embodiments utilize a target-corrected crack propagation model adapted to the hydrogen transportation environment to transform mechanical response data, ensuring the physical accuracy of crack propagation rate data. It fully integrates the hydrogen-force coupling mechanism, solving the problems of lack of physical consistency in traditional training data and poor model generalization ability. Through one-to-one binding of operating parameters and crack propagation rates, outlier removal, format standardization, and feature label matching, the integrity, validity, and standardization of the dataset are guaranteed, eliminating the interference of invalid data on model training. This provides a high-fidelity, high-quality data source for training intelligent prediction models, further improving the efficiency of model training and the final prediction accuracy.

[0069] In some embodiments, material property parameters and pipeline geometric parameters are acquired, and the material property parameters and pipeline geometric parameters are modeled and configured to obtain a hydrogen transport pipeline model library including pre-cracks. This includes: collecting and standardizing the material property parameters and pipeline geometric parameters to obtain basic parameters.

[0070] Specifically, material property parameters of the pipeline steel used in the target hydrogen transport pipeline are collected, including basic mechanical and material property data such as elastic modulus, Poisson's ratio, yield strength, fracture toughness, and hydrogen diffusion-related performance parameters; geometric parameters of the target pipeline are collected simultaneously, including structural dimensional data such as pipeline outer diameter, wall thickness, and pipe length; all collected parameters are standardized, unifying the units of measurement, numerical precision, and data format of all parameters, and eliminating outliers and invalid values ​​generated during the parameter collection process to ensure the consistency and validity of all parameters, ultimately obtaining basic parameters that can be directly used for 3D model construction.

[0071] Furthermore, a three-dimensional solid model is constructed based on the basic parameters, and the three-dimensional solid model is subjected to fine mesh division to obtain the initial pipeline simulation model.

[0072] Specifically, the aforementioned basic parameters are imported into a general-purpose finite element analysis software to construct a three-dimensional solid model of a hydrogen pipeline containing a semi-elliptical pre-fabricated crack in the inner wall, fully restoring the actual structural morphology of the pipeline and the geometric characteristics of the pre-fabricated crack. After completing the geometric construction of the model, the three-dimensional solid model is meshed. For the core region of stress concentration and hydrogen atom enrichment at the crack tip, a high-precision mesh is used for fine meshing to ensure the accuracy of subsequent numerical simulation calculations. For non-critical areas of the pipeline body, a conventional-scale mesh is used to balance calculation accuracy and simulation efficiency. At the same time, the boundary conditions, load application methods, and hydrogen diffusion boundary conditions of the model are configured and verified to ensure that the model can stably converge to complete the numerical calculations, and finally, the initial pipeline simulation model is obtained.

[0073] Furthermore, the range of values ​​for the operating parameters is configured to obtain parameterized control rules.

[0074] Specifically, the operating parameters include three core variable parameters: crack size, pipeline internal pressure amplitude, and hydrogen partial pressure. Considering the full range of actual engineering operations of hydrogen pipelines, the value ranges and gradient settings for these three parameters are configured. Crack size uses the crack depth ratio as the core indicator, with a value range covering the entire gradient from 0.1 to 0.7, corresponding to the entire lifecycle of a crack from initiation to deep propagation. The pipeline internal pressure amplitude range covers a fluctuation range of 4 MPa to 8 MPa, adapting to the operating pressure fluctuation range of conventional gas pipelines. The hydrogen partial pressure range covers the entire range from 0 MPa to 1.0 MPa, adapting to all scenarios from pure natural gas transportation to high-proportion hydrogen blending and pure hydrogen transportation. After configuring the parameter value ranges, parameterized control rules are established. When adjusting single or multiple parameters, the model can automatically update the corresponding geometric dimensions, load conditions, and boundary conditions without manual model reconstruction, providing an automated control foundation for subsequent batch simulation calculations.

[0075] Furthermore, the initial pipeline simulation model is adapted and updated based on parameterized control rules to obtain a hydrogen pipeline model library.

[0076] Specifically, based on the aforementioned parameterized control rules, the initial pipeline simulation model undergoes parameter adaptation and automatic update processing across the entire operating range. When adjusting the crack depth ratio parameter, the model automatically updates the geometric dimensions of the pre-fabricated crack; when adjusting the pipeline internal pressure amplitude parameter, the model automatically updates the load application conditions; when adjusting the hydrogen partial pressure parameter, the model automatically updates the boundary conditions for hydrogen diffusion. For each model with a set of parameters that has been adapted, convergence and effectiveness verification are performed to ensure that all models can stably complete subsequent hydrogen-force coupling numerical simulations. Finally, all verified and batch-callable parameterized pipeline models are integrated to form a hydrogen transport pipeline model library covering the full hydrogen doping ratio, the full pressure fluctuation range, and the full crack depth gradient.

[0077] This application's embodiments ensure the consistency and effectiveness of basic modeling parameters through standardized processing of material properties and pipeline geometric parameters. Combined with refined mesh generation of the crack tip region, it balances the computational accuracy and simulation efficiency of subsequent hydrogen-force coupled numerical simulations, solving the problem of insufficient accuracy in calculating crack tip stress and hydrogen concentration in traditional pipeline models. Through the configuration of parameterized control rules, automated adaptation and updating of the pipeline model are achieved, enabling batch simulation calculations across the entire operating range without manual model reconstruction, significantly improving the efficiency of modeling and simulation. The constructed full-condition model library replaces the specimen preparation of traditional physical experiments with digital models, further mitigating the safety risks of physical experiments, reducing experimental costs, and providing a stable and reliable model foundation for subsequent batch numerical simulations.

[0078] In some embodiments, hydrogen embrittlement influence correction processing is performed on the initial fatigue crack propagation model based on hydrogen concentration distribution data to obtain the target corrected crack propagation model, including: performing analytical processing on the initial fatigue crack propagation model to obtain the initial model baseline framework.

[0079] Specifically, the initial fatigue crack propagation model is the classic Paris fatigue crack propagation model commonly used in the field of oil and gas pipeline engineering, which is widely used in the calculation of crack propagation rate of conventional natural gas pipelines. The core calculation logic and structural composition of the initial model are comprehensively analyzed, and the fixed calculation structural terms, constant terms related to the inherent properties of pipeline steel materials, and correctable coefficient terms affected by service environment factors are separated. The coefficients that are directly affected by hydrogen embrittlement are identified as the parameters to be corrected. The basic calculation logic and fixed structure that are not affected by hydrogen environment are removed, and finally, an initial model benchmark framework that can be used for environmental correction is obtained.

[0080] Furthermore, the hydrogen concentration distribution data was fitted to obtain the hydrogen embrittlement influencing factor.

[0081] Specifically, the hydrogen concentration distribution data are the average hydrogen concentration data in the plastic zone at the crack tip under different working conditions obtained through previous hydrogen-force coupling numerical simulations. The fatigue performance degradation law of pipeline steel under the corresponding hydrogen concentration is matched synchronously. The quantitative correlation between the hydrogen concentration at the crack tip and the crack propagation acceleration effect caused by hydrogen embrittlement is fitted to clarify the acceleration of fatigue crack propagation by hydrogen embrittlement under different hydrogen concentrations. Finally, a hydrogen embrittlement influence factor that can dynamically and continuously change with the value of hydrogen concentration at the crack tip is obtained. This factor can fully cover the entire working condition range from zero hydrogen environment to high proportion hydrogen doping environment.

[0082] Furthermore, the initial model baseline framework was modified based on the hydrogen embrittlement influence factor to obtain the crack propagation model.

[0083] Specifically, the aforementioned hydrogen embrittlement influencing factor is introduced into the initial model baseline framework, and the analytically determined correctable coefficient terms affected by hydrogen embrittlement are dynamically adapted and corrected; the original fixed material constant terms in the model are replaced with dynamic parameter terms that can be adjusted synchronously with the hydrogen concentration at the crack tip. The specific expressions for both are as follows:

[0084]

[0085] in, The first material constant of the target pipeline steel according to the Paris formula under hydrogen-free environment (pure natural gas transportation conditions) is a fixed value obtained through conventional air environment fatigue tests. The first hydrogen embrittlement effect fitting coefficient is dimensionless and is obtained by fitting the crack tip hydrogen concentration with the material fatigue performance degradation law. The second material constant of the target pipeline steel, defined by the Paris formula, is a fixed value obtained through conventional air environment fatigue testing under hydrogen-free conditions (pure natural gas transportation). The second hydrogen embrittlement effect fitting coefficient is dimensionless and is obtained by fitting the hydrogen concentration at the crack tip with the material fatigue performance degradation law.

[0086] The model can automatically adjust the relevant parameters in the calculation logic according to the hydrogen concentration at the crack tip under different working conditions, accurately quantify the accelerating effect of hydrogen embrittlement on crack propagation, complete the initial environmental adaptation correction of the model, and obtain the initially corrected crack propagation model.

[0087] Furthermore, the crack propagation model is subjected to adaptability verification and parameter optimization to obtain the target modified crack propagation model.

[0088] Specifically, multiple sets of operating condition data covering the full hydrogen doping ratio, the full pressure fluctuation range, and the full crack depth gradient were selected. The crack tip hydrogen concentration and mechanical response data corresponding to each set of operating conditions were input into the preliminarily corrected crack propagation model to obtain the crack propagation rate results calculated by the model. The model calculation results were compared with the real crack propagation rate data obtained from the hydrogen-force coupling numerical simulation under the corresponding operating conditions to verify the model's adaptability to the hydrogen transportation environment and the accuracy of the calculation. For operating conditions where the calculation deviation exceeded the reasonable range of engineering, the dynamic parameter terms in the model were fine-tuned and optimized. Iterative verification and parameter optimization were carried out repeatedly until the calculation deviation of the model in the full operating condition range was within the allowable range of engineering, which could accurately reflect the fatigue crack propagation law of pipeline steel under the hydrogen transportation environment. Finally, a target-corrected crack propagation model that is fully adapted to the service environment of hydrogen transportation pipelines was obtained.

[0089] This application's embodiments, through analysis of the initial fatigue crack propagation model, accurately locate the correctable coefficient terms affected by hydrogen embrittlement, ensuring the targetedness and rationality of model correction. Simultaneously, it retains the basic computational logic and engineering applicability of the classic model, avoiding the problem of poor engineering adaptability caused by model structural modifications. By fitting crack tip hydrogen concentration data, dynamically changing hydrogen embrittlement influencing factors are obtained, accurately quantifying the accelerating effect of hydrogen embrittlement on crack propagation under different hydrogen environments. Combined with adaptability verification and parameter optimization, this ensures that the corrected model is fully adaptable to the full-condition service environment of hydrogen pipelines. It solves the core problems of traditional classic models not considering hydrogen embrittlement and having large calculation deviations under hydrogen transportation environments, significantly improving the accuracy of crack propagation rate calculation and providing a reliable core computational basis for subsequent dataset construction and accurate crack propagation prediction.

[0090] While this application provides method operation steps as shown in the embodiments or flowcharts, more or fewer operation steps may be included based on conventional or non-inventive labor. The order of steps listed in this embodiment is merely one possible execution order among many and does not represent the only execution order. In actual device or client product execution, the method can be executed sequentially according to this embodiment or the accompanying drawings, or in parallel (e.g., in a parallel processor or multi-threaded processing environment).

[0091] like Figure 3 As shown in the illustration, this application also provides a device 300 for predicting the crack propagation rate and assessing the failure of a hydrogen transportation pipeline. The device includes: The acquisition module 301 is used to acquire material property parameters and pipeline geometric parameters, and to model and configure the material property parameters and pipeline geometric parameters to obtain a hydrogen transport pipeline model library including pre-cracked ones.

[0092] The processing module 302 is used to perform hydrogen-force coupling numerical simulation processing on hydrogen pipelines based on the hydrogen pipeline model library to obtain mechanical response data and hydrogen concentration distribution data.

[0093] The processing module 302 is also used to perform hydrogen embrittlement correction processing on the initial fatigue crack propagation model based on hydrogen concentration distribution data to obtain the target corrected crack propagation model.

[0094] The processing module 302 is also used to transform the mechanical response data based on the target modified crack propagation model to obtain crack propagation rate data, and to perform pairing and integration processing on the crack propagation rate data to obtain the model training dataset.

[0095] The training module 303 is used to train and verify the model on the model training dataset by taking the operating condition parameters as input features and the crack propagation rate as output label, so as to obtain an intelligent prediction model. The operating condition parameters include at least one of crack size, pipeline internal pressure and hydrogen partial pressure.

[0096] The acquisition module 301 is also used to acquire real-time operating parameters of the hydrogen pipeline to be evaluated, input the real-time operating parameters into the intelligent prediction model for prediction processing, and obtain the crack propagation rate prediction result.

[0097] The processing module 302 is also used to perform failure assessment map mapping and safety boundary comparison processing on the crack propagation rate prediction results and real-time operating parameters to obtain the pipeline failure risk level and safety status assessment results.

[0098] Some modules in the apparatus described in this application can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, classes, etc., that perform a specific task or implement a specific abstract data type. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0099] The apparatus or module described in the above embodiments can be implemented by a computer chip or physical entity, or by a product with a certain function. For ease of description, the above apparatus is described by dividing it into various modules according to their functions. When implementing the embodiments of this application, the functions of each module can be implemented in one or more software and / or hardware. Of course, a module that implements a certain function can also be implemented by combining multiple sub-modules or sub-units.

[0100] The methods, apparatus, or modules described in this application can be implemented in a computer-readable program code manner. The controller can be implemented in any suitable manner, such as a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicon Labs C8051F320. A memory controller can also be implemented as part of the control logic of a memory. Those skilled in the art will also recognize that, in addition to implementing the controller in purely computer-readable program code manner, the same functionality can be achieved by logically programming the method steps to make the controller take the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the means included within it for implementing various functions can also be considered as structures within the hardware component. Alternatively, the device used to implement various functions can be viewed as either a software module implementing the method or a structure within a hardware component.

[0101] This application also provides an apparatus, the apparatus comprising: a processor; a memory for storing processor-executable instructions; wherein, when the processor executes the executable instructions, it implements the method described in this application.

[0102] This application also provides a non-volatile computer-readable storage medium storing a computer program or instructions thereon, which, when executed, enables the method described in this application embodiment to be implemented.

[0103] Furthermore, in the various embodiments of the present invention, each functional module can be integrated into a processing module, or each module can exist independently, or two or more modules can be integrated into a single module.

[0104] The aforementioned storage media include, but are not limited to, Random Access Memory (RAM), Read-Only Memory (ROM), Cache, Hard Disk Drive (HDD), or Memory Card. The memory can be used to store computer program instructions.

[0105] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary hardware. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product, or it can be embodied in the process of data migration. The computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, mobile terminal, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of this application.

[0106] The various embodiments described in this specification are presented in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on its differences from other embodiments. All or part of this application can be used in numerous general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, mobile communication terminals, multiprocessor systems, microprocessor-based systems, programmable electronic devices, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices, etc.

[0107] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of this application.

Claims

1. A method for predicting crack propagation rate and assessing failure in hydrogen transportation pipelines, characterized in that, include: Obtain material property parameters and pipeline geometric parameters, and perform modeling and configuration processing on the material property parameters and pipeline geometric parameters to obtain a hydrogen transportation pipeline model library including pre-cracked ones; Based on the hydrogen pipeline model library, hydrogen-force coupling numerical simulation was performed on the hydrogen pipeline to obtain mechanical response data and hydrogen concentration distribution data. Based on the hydrogen concentration distribution data, the initial fatigue crack propagation model is modified to address the hydrogen embrittlement effect, resulting in the target modified crack propagation model. Based on the target modified crack propagation model, the mechanical response data is transformed to obtain crack propagation rate data. The crack propagation rate data is then paired and integrated to obtain the model training dataset. Using operating condition parameters as input features and the crack propagation rate as output label, the model training dataset is subjected to model training and accuracy verification processing to obtain an intelligent prediction model. The operating condition parameters include at least one of crack size, pipeline internal pressure, and hydrogen partial pressure. The real-time operating parameters of the hydrogen pipeline to be evaluated are obtained, and the real-time operating parameters are input into the intelligent prediction model for prediction processing to obtain the crack propagation rate prediction result. The crack propagation rate prediction results and the real-time operating parameters are subjected to failure assessment map mapping and safety boundary comparison processing to obtain the pipeline failure risk level and safety status assessment results.

2. The method according to claim 1, characterized in that, The intelligent prediction model is obtained by using operating condition parameters as input features and crack propagation rate as output label to train and verify the model training dataset, resulting in a model training and accuracy verification process. This includes: The model training dataset is preprocessed to obtain a standardized training dataset; The standardized training dataset is shuffled and split to obtain training and test subsets. An initial training model is constructed, and the training subset is input into the initial training model for updating and optimization to obtain an initial prediction model. The initial prediction model is subjected to accuracy verification based on the test subset to obtain an intelligent prediction model.

3. The method according to claim 1, characterized in that, The failure assessment map mapping and safety boundary comparison processing of the crack propagation rate prediction results and the real-time operating parameters yields the pipeline failure risk level and safety status assessment results, including: A two-dimensional evaluation coordinate system composed of fracture ratio and load ratio is constructed to obtain the failure evaluation benchmark diagram; Based on the hydrogen partial pressure data in the real-time operating parameters of the hydrogen pipeline to be evaluated, the fracture toughness value is dynamically corrected to obtain the fracture toughness correction value. Based on the fracture toughness correction value, the failure assessment benchmark map is adjusted to obtain a dedicated failure assessment map. Based on the crack propagation rate prediction results and real-time operating parameters, the fracture ratio and load ratio corresponding to the current crack state are calculated and processed to obtain the coordinate point data to be evaluated. The coordinate point data to be evaluated is mapped onto the dedicated failure evaluation map, and the relative position of the coordinate point data to be evaluated and the safety boundary envelope curve is compared to obtain the position comparison result. Based on the location comparison results, failure mode determination and safety risk level classification are performed to obtain the pipeline failure risk level and safety status assessment results.

4. The method according to claim 1, characterized in that, The mechanical response data is transformed based on the target modified crack propagation model to obtain crack propagation rate data. The crack propagation rate data is then paired and integrated to obtain a model training dataset, including: Based on the target modified crack propagation model, the mechanical response data is transformed and calculated to obtain crack propagation rate data; The operating parameters and the crack propagation rate data are bound together to obtain multiple sets of structured data units; Outlier removal and format standardization are performed on the multiple sets of structured data units to obtain a regularized dataset; The regularized dataset is matched and bound to obtain the model training dataset.

5. The method according to claim 1, characterized in that, The process involves acquiring material property parameters and pipeline geometric parameters, modeling and configuring these parameters to obtain a hydrogen transport pipeline model library including pre-existing cracks. The material property parameters and the pipe geometric parameters are collected and standardized to obtain the basic parameters; A three-dimensional solid model is constructed based on the aforementioned basic parameters. The three-dimensional solid model is then subjected to fine mesh division to obtain an initial pipeline simulation model. The value range of the operating condition parameters is configured to obtain parameterized control rules; Based on the parametric control rules, the initial pipeline simulation model is adapted and updated to obtain a hydrogen transportation pipeline model library.

6. The method according to claim 1, characterized in that, The hydrogen embrittlement effect correction process applied to the initial fatigue crack propagation model based on the hydrogen concentration distribution data yields the target corrected crack propagation model, including: The initial fatigue crack propagation model is analyzed to obtain the initial model baseline framework; The hydrogen concentration distribution data is fitted to obtain the hydrogen embrittlement influencing factor; Based on the hydrogen embrittlement influencing factor, the initial model baseline framework is modified to obtain the crack propagation model; The crack propagation model is subjected to adaptability verification and parameter optimization to obtain the target modified crack propagation model.

7. The method according to claim 1, characterized in that, The crack depth ratio of the crack size ranges from 0.1 to 0.7, the internal pressure of the pipeline ranges from 4 MPa to 8 MPa, and the partial pressure of hydrogen ranges from 0 MPa to 1.0 MPa.

8. A device for predicting the crack propagation rate and assessing the failure of a hydrogen transportation pipeline, characterized in that, include: The acquisition module is used to acquire material property parameters and pipeline geometric parameters, and to perform modeling and configuration processing on the material property parameters and pipeline geometric parameters to obtain a hydrogen transport pipeline model library including pre-cracked structures. The processing module is used to perform hydrogen-force coupling numerical simulation processing on the hydrogen pipeline based on the hydrogen pipeline model library to obtain mechanical response data and hydrogen concentration distribution data. The processing module is also used to perform hydrogen embrittlement correction processing on the initial fatigue crack propagation model based on the hydrogen concentration distribution data to obtain the target corrected crack propagation model. The processing module is further configured to transform the mechanical response data based on the target modified crack propagation model to obtain crack propagation rate data, and to perform pairing and integration processing on the crack propagation rate data to obtain a model training dataset. The training module is used to train and verify the model on the model training dataset by taking the operating condition parameters as input features and the crack propagation rate as the output label, so as to obtain an intelligent prediction model. The operating condition parameters include at least one of crack size, pipeline internal pressure and hydrogen partial pressure. The acquisition module is also used to acquire real-time operating parameters of the hydrogen pipeline to be evaluated, input the real-time operating parameters into the intelligent prediction model for prediction processing, and obtain the crack propagation rate prediction result. The processing module is also used to perform failure assessment map mapping and safety boundary comparison processing on the crack propagation rate prediction results and the real-time operating parameters to obtain the pipeline failure risk level and safety status assessment results.

9. A computer device comprising a memory and a processor, the memory storing a computer program executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

Images