Evaluation method for predicting crack propagation and residual service life of hydrogen-doped natural gas pipeline

Abstract

The invention discloses an evaluation method for predicting crack propagation and residual service life of a natural gas hydrogen-doped pipeline, which comprises the following steps of: firstly, performing numerical simulation on crack depth detection data by using a numerical simulation method, and simulating a crack depth increasing process to obtain crack depth detection data; then, according to crack depth detection data information, crack propagation parameter prior distribution obtained through statistics of a historical database is combined, crack propagation parameters are calculated through a Bayesian updating method, a likelihood function of Bayesian updating is constructed through a Forman equation, and posterior distribution of the crack propagation parameters is calculated; substituting the posterior probability distribution obtained in the current loop as prior into the next loop for iterative updating; finally, the critical crack depth of the pipeline is estimated according to the failure criterion, the number of remaining use cycles of the pipeline and the remaining service life of the pipeline are obtained, the more accurate parameter posterior probability can be obtained, and the influence of uncertain information in the operation environment is relieved.

Description

Technical Field

[0001] The present invention belongs to the field of hydrogen pipeline transportation, and in particular relates to an evaluation method for predicting crack propagation and remaining service life of a natural gas hydrogen-blended pipeline. Background Art

[0002] Natural gas blended with hydrogen for transportation has become a key path to large-scale hydrogen energy application. However, hydrogen embrittlement, which causes accelerated crack growth in pipelines, poses a serious threat to the safety of hydrogen storage and transportation systems. Currently, the mechanism of hydrogen embrittlement is unclear, and pipeline failure models are often based on empirical formulas and experimental data. Crack growth in hydrogen environments is influenced by many factors, including pipeline material type, hydrogen concentration, environmental parameters, and pipeline operating conditions. This information often exhibits a degree of randomness and volatility, is difficult to accurately measure in actual operation, and contains a certain amount of noise, posing a challenge to crack growth prediction.

[0003] There is little research on crack propagation in natural gas hydrogen-blended pipelines in China, and most crack propagation models for natural gas hydrogen-blended pipelines are generally based on the assumption of constant operating conditions and ignore the impact of hydrogen on the fracture toughness of pipelines. They are unable to cope with the information randomness of the operating environment and conditions in the actual operation of natural gas hydrogen-blended pipelines. Summary of the Invention

[0004] To address the above technical issues, the present invention proposes an assessment method for predicting crack growth and remaining service life in natural gas hydrogen-blended pipelines. This method utilizes a hydrogen-induced crack evolution model and life prediction method based on Bayesian updating, considers the deteriorating effect of hydrogen on the fracture toughness of pipeline materials, and incorporates the uncertainty of multi-source information under complex operating conditions. This method dynamically updates the crack growth rate based on real-time detection data, ultimately predicting the remaining service life of the pipeline. In actual operating conditions, the Bayesian updating method allows for continuous and iterative updating of crack growth parameters based on detection data, resulting in more accurate parameter posterior probabilities and mitigating the impact of uncertain information in the operating environment.

[0005] In order to achieve the above technical objectives, the present invention adopts the following technical means:

[0006] A method for predicting crack growth and remaining service life of a natural gas hydrogen-blended pipeline comprises the following steps:

[0007] S1. Numerical simulation of detection: Based on the preset pipeline system parameters and considering the effect of hydrogen on the fracture toughness of the pipeline, a numerical simulation method is used to perform numerical simulation of crack depth detection data, simulate the crack depth growth process, and obtain crack depth detection data;

[0008] S2. Bayesian update process: specifically includes the following sub-steps:

[0009] S21, according to the crack depth detection data information obtained in step S1, combined with the historical database statistics to obtain the prior distribution of the crack propagation parameter m;

[0010] S22. Calculate the crack growth parameter using the Bayesian update method, construct the likelihood function of the Bayesian update using the Forman equation, and calculate the posterior distribution of the crack growth parameter m;

[0011] S23, substituting the posterior probability distribution obtained in this cycle as the prior into the next cycle for iterative update, step by step approaching the crack growth parameter m value in the actual working condition;

[0012] S3. Estimation of critical crack depth of pipeline: According to the failure criterion, the pipeline failure is defined when the crack depth reaches the critical value determined by the NG-18 standard. The critical crack depth of the pipeline is estimated using the ln-secant equation.

[0013] S4. Prediction of the remaining service life of the pipeline: Based on the critical crack depth of the pipeline estimated in step S3, the Forman equation is used to calculate the remaining service cycles of the pipeline, and the time dimension of the remaining service life of the pipeline is obtained by converting the remaining service cycles of the pipeline with the loading frequency.

[0014] Technical effect: This invention is based on the numerical simulation detection data of pipeline crack depth under multi-source information uncertainty conditions, takes into account the influence of hydrogen on the fracture toughness of pipeline materials, adopts the Bayesian update algorithm, and innovatively uses the Forman equation to calculate the likelihood function of Bayesian update, and converts the fracture toughness K C Taking this into account, the uncertainty parameter m is updated when hydrogen concentration changes dynamically. This quantitatively characterizes the accelerating effect of hydrogen concentration on crack growth rate and more accurately predicts the remaining service life of pipelines. This provides a basis for predicting the life of natural gas hydrogen-blended pipelines under multi-source information uncertainty and provides a reference for the accuracy standards of actual pipeline hydrogen blending ratios and operating condition data collection.

[0015] In an optional embodiment, in step S1, the preset pipeline system parameters include: pipeline material, hydrogen concentration, stress amplitude, initial crack depth, loading frequency, wall thickness, outer diameter and operating pressure.

[0016] Beneficial effects: Hydrogen concentration will greatly affect the fracture toughness of pipeline materials, stress amplitude will greatly affect the crack growth rate, initial crack depth will greatly affect the remaining service life of the pipeline, and other parameters will also have a certain impact on crack growth and the remaining service life of the pipeline.

[0017] In an optional embodiment, a numerical simulation of crack depth detection data is performed to simulate the crack depth growth process according to the Forman equation, which is in the form of the following equation:

[0018]

[0019] in,

[0020]

[0021]

[0022] Where a is the crack depth, N is the number of cycles, C and m are crack growth parameters, and K c is the fracture toughness fracture toughness, ΔK is the stress amplitude, R is the stress ratio, σ max is the maximum stress on the pipeline, σ min is the minimum stress on the pipeline, and Y is the shape factor.

[0023] Technical effect: The present invention combines the Forman equation with the Bayesian update algorithm and introduces the fracture toughness K C The item can quantitatively characterize the accelerating effect of hydrogen concentration on the crack growth rate and more accurately predict the remaining service life of the pipeline.

[0024] In an optional embodiment, in step S1, the number of cycles N is not less than 5×10 7 .

[0025] Beneficial effect: When the number of cycles is small, the crack growth rate is slow, the crack depth changes very slightly, and the crack depth detection value is greatly affected by the measurement error, resulting in an insignificant crack growth trend and a large error in the crack growth parameter m obtained in step S2.

[0026] In an optional embodiment, in step S1 , calculation is performed assuming that 1000 cycles are one crack depth growth period to reduce the amount of calculation.

[0027] In an optional embodiment, in step S22, the Bayesian updated posterior probability of the fusion Forman equation is calculated as follows:

[0028]

[0029] in

[0030]

[0031] Where m is the crack growth parameter, is the crack depth detection value of the i-th cycle, is the observed crack depth value of the i-1th cycle, K C (i) is the current cyclic fracture toughness, R iis the current cyclic stress ratio, C is a fixed value, Δσ i is the current cyclic stress amplitude.

[0032] In an optional embodiment, in step S4, the relationship between the critical crack depth of the pressurized pipeline and the pipeline system parameters is defined by the ln-secant equation in the NG-18 standard.

[0033] In an optional embodiment, in step S4, according to the critical crack depth value a of the pipeline c , the remaining service cycles of the pipeline are obtained by the Forman equation as follows:

[0034]

[0035] Where T left is the remaining service life of the pipeline; R is the stress ratio, K c is the fracture toughness fracture toughness, ΔK is the stress amplitude, C, m are the crack growth parameters, a is the crack depth, a c is the critical crack depth of the pipeline. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 Implement a flowchart for the case. DETAILED DESCRIPTION

[0037] The technical solution of the present invention is further discussed in detail below with reference to specific embodiments and the accompanying drawings.

[0038] like Figure 1 As shown, the present invention provides an evaluation method for predicting crack growth and remaining service life of a natural gas hydrogen-blended pipeline. First, a numerical simulation method is used to perform numerical simulation of crack depth detection data, simulate the crack depth growth process, and obtain crack depth detection data. Then, based on the crack depth detection data information and in combination with the prior distribution of crack growth parameters obtained from historical database statistics, the crack growth parameters are calculated using a Bayesian update method. The likelihood function of the Bayesian update is constructed using the Forman equation, and the posterior distribution of the crack growth parameters is calculated. The posterior probability distribution obtained in this cycle is substituted as the prior into the next cycle for iterative updating. Finally, the critical crack depth of the pipeline is estimated according to the failure criterion to obtain the remaining number of service cycles and the remaining service life of the pipeline.

[0039] Example 1: Under complex working conditions with rapid dynamic fluctuations in hydrogen concentration H0, traditional methods are unable to calculate the true value of the crack growth parameter m. The Bayesian update method can effectively solve this problem.

[0040] The present invention provides an evaluation method for predicting crack growth and remaining service life of a natural gas hydrogen-blended pipeline to achieve:

[0041] The pipeline material is X80 pipeline steel. The hydrogen concentration is randomly stepped between three thresholds of 5 vol%, 10 vol%, and 15 vol% during each cyclic loading. The measurement error is respectively ε1~N(0, 0.01 2 ),ε2~N(0,0.1 2 ) to conduct a control experiment; the maximum stress σ max ~N(300,0.1 2 )MPa, minimum stress σ min ~N(299,0.1 2 )MPa; pipeline crack depth a0 = 1.2mm; crack shape factor Y = 1.12; loading frequency f = 1Hz; wall thickness t = 20mm; pipeline outer diameter D = 1000mm; operating pressure P = 12MPa:

[0042] (1) Numerical simulation of detection: MATLAB is used to perform numerical simulation of crack depth detection data. According to the Forman equation, the crack depth growth process is simulated. A total of 10×10 8 cycles. The Forman equation is shown below:

[0043]

[0044] in,

[0045]

[0046] Where a is the crack depth, N is the number of cycles, C and m are crack growth parameters, ΔK is the stress amplitude, R is the stress ratio, and σ max is the maximum stress, σ min is the minimum stress and Y is the shape factor.

[0047] In the specific calculation, since the crack depth growth in each cycle is extremely small, it can be assumed that 1000 cycles is a crack depth growth period to reduce the amount of calculation;

[0048] (2) Bayesian update process: Based on the crack depth information obtained by detection and combined with the historical database statistics, the prior distribution of the crack extension parameter m is obtained. The posterior distribution of the crack extension parameter m is calculated using the Bayesian update method. The Bayesian updated posterior probability is calculated as shown in the following formula:

[0049]

[0050] Subsequently, the posterior probability distribution of m obtained in this cycle is substituted as the prior into the next cycle for the next update, gradually approaching the m value in the actual working condition. The Bayesian update result prediction is shown in Table 2:

[0051] Table 1 Bayesian update m value

[0052]

[0053] (3) Estimation of critical crack depth of pipeline: The failure criterion is defined as the crack depth reaching the critical value a determined by the NG-18 standard. c The relationship between the critical crack depth of a pressurized pipeline and these parameters of the pipeline wall thickness is expressed by the ln-secant equation as shown below:

[0054] in,

[0055]

[0056] Where a c is the critical crack depth of the pipeline, D is the pipeline diameter, t is the pipe wall thickness, E is the elastic modulus, L e is the effective defect length, flow stress σ f is the average of the yield strength and ultimate tensile strength, σ H is the hoop stress caused by the internal operating pressure of the pipeline, P is the pipeline operating pressure, K C is the fracture toughness fracture toughness, A c is the cross-sectional area of ​​the simply supported beam impact specimen; M S 、M t , z is an auxiliary variable;

[0057] (4) Prediction of remaining service life of pipeline: According to the critical crack depth a of pipeline c The value of the remaining service life of the pipeline can be obtained from the Forman equation as shown below:

[0058]

[0059] The time dimension of the remaining service life of the pipeline can be realized by converting the number of cycles and the loading frequency: Where T left The remaining service life of the pipeline is shown in Table 3.

[0060] Table 2 Prediction of remaining service life of pipelines

[0061]

[0062]

[0063] From the data in Table 3, we can see that the probability distribution of the remaining useful life calculated by the Bayesian update results gradually approaches the true value. The error of the remaining useful life calculated by the prior probability before the initial update is 128.1%, which is reduced to 1.859% after 4 updates (ε1~N(0,0.01 2 )) and 2.620%(ε2~N(0,0.1 2 )). From this, we can conclude that the Bayesian update is more effective, especially when the measurement error is small and the crack growth rate is large.

[0064] In the actual operation of natural gas hydrogen-blended pipelines, the hydrogen concentration inside the pipeline often does not change as drastically as assumed in this section, causing the crack growth parameter m and fracture toughness K to change dramatically. C The fluctuations in m are relatively small, so the Bayesian update method often provides a more accurate estimate of m. Ultimately, the standard deviation of the parameter m and the remaining useful life prediction is smaller, and the probability distribution range is narrower, resulting in superior results compared to Example 1. Compared to most existing methods, the present invention integrates the Forman equation with the Bayesian update algorithm, taking into account the effect of hydrogen on fracture toughness, resulting in a more accurate prediction of pipeline crack growth.

[0065] Example 2: A numerical simulation model of crack growth depth of X80 steel with a sudden change in hydrogen concentration at a certain moment was established. In the implementation, the measurement error ε~N(0,0.012)mm; the maximum stress σ max ~N(300,0.12)MPa, minimum stress σ min ~N(299,0.12)MPa; initial crack depth a0~N(2.0,0.22)mm. Crack shape factor Y = 1.12, loading frequency f = 1Hz, wall thickness t = 20mm, pipe outer diameter D = 1000mm, operating pressure P = 12MPa. The hydrogen concentration H0 and its corresponding m, KC changes with the number of cycles N are shown in Table 4:

[0066] Table 3 Data table of operating condition changes of natural gas hydrogen blending pipeline

[0067]

[0068] (1) Numerical simulation of detection: MATLAB was used to perform numerical simulation of crack depth detection data. Ten initial crack depths were randomly generated to obtain ten crack depth growth curves. According to the Forman equation, the crack depth growth process was simulated. A total of 10×10 8 cycles. The Forman equation is shown in the following form:

[0069] in,

[0070]

[0071] Where a is the crack depth, N is the number of cycles, C and m are crack growth parameters, ΔK is the stress amplitude, R is the stress ratio, and σ max is the maximum stress, σ min is the minimum stress and Y is the shape factor.

[0072] In the specific calculation, since the crack depth growth in each cycle is extremely small, it can be assumed that 1000 cycles is a crack depth growth period to reduce the amount of calculation;

[0073] (2) Bayesian update process: 7 of the 10 crack depth extension curves are randomly selected as the training set, and the remaining 3 are used as the test set. Based on the crack depth information obtained by detection and the historical database statistics, the prior distribution of the crack extension parameter m is obtained. The posterior distribution of the crack extension parameter m is calculated using the Bayesian update method. The Bayesian updated posterior probability is calculated as follows:

[0074]

[0075] Subsequently, the posterior probability distribution of m obtained in this cycle is substituted into the next cycle as a priori for the next update, gradually approaching the m value in the actual working condition. The Bayesian update results of the crack growth parameter m based on the training set data are shown in Table 4:

[0076] Table 4 Bayesian update m value

[0077]

[0078]

[0079] (3) Estimation of critical crack depth of pipeline: The failure criterion is defined as the crack depth reaching the critical crack depth value a of the pipeline determined by the NG-18 standard. c The relationship between the critical crack depth of a pressurized pipeline and these parameters of the pipeline wall thickness is expressed by the ln-secant equation as shown below:

[0080] in,

[0081]

[0082] Among them, a c is the critical crack depth of the pipeline, D is the pipeline diameter, t is the pipe wall thickness, E is the elastic modulus, L e is the effective defect length, flow stress σ f is the average of the yield strength and ultimate tensile strength, σ His the hoop stress caused by the internal operating pressure of the pipeline, P is the pipeline operating pressure, K C is the fracture toughness fracture toughness, A c is the cross-sectional area of ​​the simply supported beam impact specimen, M S 、M t , z is an auxiliary variable;

[0083] (4) Prediction of remaining service life of pipeline: According to the critical crack depth value a of pipeline c , the remaining number of pipeline service cycles can be obtained from the Forman equation as shown below:

[0084]

[0085] The time dimension of the remaining service life of the pipeline can be realized by converting the number of cycles and the loading frequency: Where T left Remaining service life of the pipeline. A curve data from the test set is randomly selected for verification, as shown in Table 5:

[0086] Table 5 Remaining useful life prediction of test set

[0087]

[0088]

[0089] As shown in Table 5, with the increase of the number of updates, the probability distribution of the crack growth parameter m gradually approaches the true value, and the error of the mean gradually decreases from 6.76% to 5.23%, 0.8%, 0.26%, and 0.25%. Therefore, the predicted remaining service life will also become closer to the true value, and the distribution range will become narrower, and the prediction will become more accurate. At the same time, it can be observed that due to the increase of hydrogen at 2.5×10 8 After the first cycle, it is added to the inside of the pipeline, making the fracture toughness K C The crack growth rate increases, and the remaining service life decreases sharply.

[0090] Compared with most existing methods, the present invention integrates the Forman equation and the Bayesian update algorithm, taking into account the effect of hydrogen on fracture toughness, making the prediction of pipeline crack propagation more accurate.

Claims

1. A method for predicting crack growth and remaining service life of a natural gas hydrogen-blended pipeline, characterized in that: The following steps are involved: S1. Numerical simulation of detection: Based on the preset pipeline system parameters and considering the effect of hydrogen on the fracture toughness of the pipeline, a numerical simulation method is used to perform numerical simulation of crack depth detection data, simulate the crack depth growth process, and obtain crack depth detection data; S2. Bayesian update process: specifically includes the following sub-steps: S21, according to the crack depth detection data information obtained in step S1, combined with the historical database statistics to obtain the prior distribution of the crack propagation parameter m; S22. Calculate the crack growth parameter using the Bayesian update method, construct the likelihood function of the Bayesian update using the Forman equation, and calculate the posterior distribution of the crack growth parameter m; S23, substituting the posterior probability distribution obtained in this cycle as the prior into the next cycle for iterative update, step by step approaching the crack growth parameter m value in the actual working condition; S3. Estimation of critical crack depth of pipeline: According to the failure criterion, the pipeline failure is defined when the crack depth reaches the critical value determined by the NG-18 standard. The critical crack depth of the pipeline is estimated using the ln-secant equation. S4. Prediction of the remaining service life of the pipeline: Based on the critical crack depth of the pipeline estimated in step S3, the Forman equation is used to calculate the remaining service cycles of the pipeline, and the time dimension of the remaining service life of the pipeline is obtained by converting the remaining service cycles of the pipeline with the loading frequency.

2. The method for predicting crack growth and remaining service life of a natural gas hydrogen-blended pipeline according to claim 1, characterized in that: In step S1, the preset pipeline system parameters include: pipeline material, hydrogen concentration, stress distribution, initial crack depth, loading frequency, wall thickness, outer diameter and operating pressure.

3. The method for predicting crack growth and remaining service life of a natural gas hydrogen-blended pipeline according to claim 2, characterized in that: In step S1, a numerical simulation of crack depth detection data is performed, and the crack depth growth process is simulated according to the Forman equation. The form of the Forman equation is shown as follows: in, Where a is the crack depth, N is the number of cycles, C and m are crack growth parameters, and K c is the fracture toughness fracture toughness, ΔK is the stress amplitude, R is the stress ratio, σ max is the maximum stress on the pipeline, σ min is the minimum stress on the pipeline, and Y is the shape factor.

4. The method for predicting crack growth and remaining service life of a natural gas hydrogen-blended pipeline according to claim 3, characterized in that: In step S1, the number of cycles N is not less than 5×10 7 .

5. The method for predicting crack growth and remaining service life of a natural gas hydrogen-blended pipeline according to claim 1, characterized in that: In step S1 , the calculation is performed assuming that 1000 cycles are one crack depth growth period to reduce the amount of calculation.

6. The method for predicting crack growth and remaining service life of a natural gas hydrogen-blended pipeline according to claim 1, characterized in that: In step S22, the Bayesian updated posterior probability of the integrated Forman equation is calculated as follows: in, Where m is the crack growth parameter, is the crack depth detection value of the i-th cycle, is the observed crack depth value of the i-1th cycle, K C (i) is the current cyclic fracture toughness, R i is the current cyclic stress ratio, C is a fixed value, Δσ i is the current cyclic stress amplitude.

7. The method for predicting crack growth and remaining service life of a natural gas hydrogen-blended pipeline according to claim 1, characterized in that: In step S4, according to the critical crack depth value a of the pipeline c , the remaining service cycles of the pipeline are obtained by the Forman equation as follows: Among them, T left is the remaining service life of the pipeline; R is the stress ratio, K c is the fracture toughness fracture toughness, ΔK is the stress amplitude, C, m are the crack growth parameters, a is the crack depth, a c is the critical crack depth of the pipeline.

Images