Method and system for predicting residual life of ultra-high sulfur-containing bimetallic gas transmission pipeline

Abstract

The invention relates to the technical field of data prediction processing, and discloses a residual life prediction method and system for an extra-high sulfur-containing bimetallic gas transmission pipeline, and the method comprises the steps: obtaining strain data, temperature data, outer layer corrosion data and sulfur deposition data; performing dynamic feature extraction on the data to obtain corrosion environment features and material state features; establishing a synergistic corrosion kinetic model based on the characteristic data, and establishing a corrosion rate equation by considering the synergistic effect of H2S and CO2; based on the multi-physics field coupling model, combining the comprehensive material data, the monitoring data and the gas data to establish a comprehensive life prediction model; predicting the residual life, and outputting a prediction result; and comparing and analyzing a prediction result and field detection data, and adjusting and optimizing model parameters according to an analysis result. By constructing the multi-physical field coupling model and the multi-modal intelligent sensing system, the precision of predicting the life of the bimetallic pipeline under the ultrahigh sulfur-containing working condition is improved, and the adaptability and reliability of the prediction method are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data prediction processing, in particular to a method and system for predicting the residual life of a double-metal gas pipeline with extremely high sulfur content. BACKGROUND

[0002] With the expansion of the development scale of global high-sulfur (H2S content ≥ 30 g / m 3 ) gas fields, the corrosion resistance of gas pipelines has become a critical requirement. In this context, double-metal composite pipes have gradually replaced traditional single-metal pipelines and become an important solution in high-sulfur environments due to their excellent corrosion resistance. However, engineering practice shows that such pipelines still face the risk of sudden rupture caused by interfacial failure and sulfur deposition corrosion during service, which seriously threatens production safety. To ensure pipeline operation, the industry usually relies on non-destructive testing and online monitoring technologies, combined with intelligent diagnostic systems for condition assessment and life prediction.

[0003] However, the existing technical system still has significant shortcomings and technical bottlenecks in dealing with extremely high-sulfur (H2S > 50 g / m 3 ) environments, mainly in the following three aspects: First, in terms of detection and perception, the adaptability to extreme environments is insufficient. Conventional piezoelectric ultrasonic sensors will be hydrogen sulfide embrittled in an ultra-high concentration of hydrogen sulfide environment, leading to a sharp decline in performance or even damage. The detection instrument also needs to meet the sealing requirements of pressure resistance, temperature resistance, and H2S permeation resistance. The existing materials are difficult to achieve effective compatibility between environmental tolerance and signal transmission reliability, causing essential obstacles in data acquisition, making it difficult to obtain high-quality and continuous monitoring data.

[0004] Secondly, in terms of diagnostic analysis, the data fusion and analysis capabilities of intelligent algorithms are bottlenecked. The existing intelligent diagnostic systems generally face difficulties in "multi-source data fusion" when dealing with monitoring data from different sources and types, and the data island phenomenon is serious. This makes it difficult for the diagnostic model to fully and accurately grasp the true health status of the pipeline, directly leading to low accuracy of residual life prediction, and the prediction results are difficult to serve as an effective basis for maintenance decisions.

[0005] Finally, in terms of prediction models, the consideration of key mechanisms is severely lacking. Current mainstream life prediction methods are based on macroscopic corrosion rates, ignoring two crucial microscopic and chemical mechanisms: first, the concentration gradient of key alloying elements such as Cr and Ni in the double-metal interfacial diffusion layer and their evolution rules under long-term service are not considered, which are the core factors affecting the interfacial stability and crack resistance; second, the complex synergistic corrosion effect of H2S and CO2 in actual working conditions is ignored, leading to a large deviation between the estimated corrosion rate and the actual situation.

[0006] Therefore, the prior art is difficult to accurately and reliably predict the remaining life of the extra-high sulfur-containing double-metal gas pipeline, thereby restricting the improvement of the safe and efficient operation and maintenance level. SUMMARY

[0007] The present application aims to provide a method and system for predicting the remaining life of an extra-high sulfur-containing double-metal gas pipeline, which combines multi-physical field modeling and intelligent prediction algorithm based on deep learning to improve the prediction accuracy.

[0008] To achieve the above-mentioned purpose, the present application adopts the following technical solution: a method for predicting the remaining life of an extra-high sulfur-containing double-metal gas pipeline, comprising the following steps, (1) Data acquisition: acquiring strain data, temperature data, outer corrosion data and sulfur deposition data respectively; (2) Feature extraction: performing dynamic feature extraction on the collected data to obtain corrosion environment features and material state features; (3) Multi-physical field coupling model construction: Based on the obtained feature data, a collaborative corrosion dynamics model is established, considering the synergistic effect of H2S and CO2, a corrosion rate equation is established, and a multi-physical field coupling model is obtained, expressed as ; In the formula, is the corrosion depth; is the reaction rate constant; is the H2S partial pressure; is the CO2 molar concentration; E a is the activation energy; is the absolute temperature; is the interfacial diffusion coefficient; is the diffusion layer thickness; is the empirical index; (4) Remaining life prediction: based on the multi-physical field coupling model, combined with comprehensive material data, monitoring data and gas data, a comprehensive life prediction model is established; the remaining life is predicted, and the prediction result is output; the comprehensive life prediction model is expressed as ; In the formula, is the critical corrosion depth; is the initial corrosion depth; is the yield strength; is the wall thickness loss; A is a material constant; B is a medium sensitivity coefficient; is the H2S molar concentration; (5) Model optimization: comparing and analyzing the prediction result with the field detection data, and adjusting the model parameters according to the analysis result.

[0009] Meanwhile, the scheme also provides a kind of extra-high sulfur content bimetallic gas pipeline remaining life prediction system, applied to the above-mentioned kind of extra-high sulfur content bimetallic gas pipeline remaining life prediction method, comprising, The data acquisition layer: a distributed optical fiber network is used to collect strain data and temperature data at a sampling frequency of 1KHz; a pulse eddy current array is used to collect outer layer corrosion data at a resolution of 0.5mm 2 ; a sulfur deposition laser radar is used to scan and obtain sulfur deposition data; The feature extraction layer: the collected data is fused, and corrosion environment features and material state features are dynamically extracted; The model construction layer: based on the feature data, an interface diffusion factor is introduced to establish a multi-physical field coupling model to quantize the acceleration effect of sulfur deposition on interface diffusion; The life prediction layer: based on the evolution process of the bimetallic interface diffusion layer, a phase field method is used to simulate the interface crack initiation process to obtain the diffusion layer thickness change rate and stress concentration coefficient, and the remaining life is predicted, and the prediction result is output; The intelligent correction layer: ResNet-50 is used to obtain the on-site pipeline ultrasonic detection image, and compared with the prediction result, the model parameters are corrected and optimized according to the comparison result.

[0010] The principle of the scheme is: The scheme is aimed at the non-stop production detection technology of high sulfur content bimetallic pipeline, and systematically analyzes from three dimensions of detection principle, environmental adaptability and data analysis, and fully considers the synergistic corrosion effect and interface diffusion layer problem. Through the deep integration of "mechanism driven" and "data intelligent", a full-chain, high-precision prediction system from micro mechanism to macro monitoring, from data perception to intelligent diagnosis is constructed.

[0011] The scheme establishes a H2S-CO2-H2O three-phase synergistic corrosion kinetics equation, introduces a nonlinear corrosion contribution coefficient model, scientifically quantizes the interaction and saturation effect of multiple corrosion media, and fundamentally corrects the deviation of the traditional model. Secondly, a bimetallic interface diffusion layer evolution algorithm based on the Fick's law is developed, and an interface diffusion factor is introduced to quantize the acceleration effect of sulfur deposition, so as to accurately predict the long-term performance degradation of the weak link of bimetallic bonding interface. At the same time, the phase field method is used to simulate the whole process from crack initiation to expansion, and the life prediction is improved from the macroscopic empirical formula to the micro damage mechanics level.

[0012] In addition, the scheme discards the traditional piezoelectric ultrasonic probe which is easy to fail in extreme environment, innovatively uses a distributed optical fiber network to perceive strain / temperature, a pulse eddy current array to perceive outer layer corrosion, and introduces a sulfur deposition laser radar to build a three-dimensional model, and constructs a corrosion-resistant, full-coverage, high-precision stereo monitoring network.

[0013] And for the unique problems of various monitoring signals, the scheme designs special intelligent compensation and analysis algorithm, and through the construction of a mixed prediction model, the intelligent identification of corrosion images and the trend prediction of sulfur deposition are respectively processed, so that the deep mining and fusion of multi-source information are realized.

[0014] The advantages of the scheme are: Precise breakthrough the prominent contradictions and conflict points in the prior art.

[0015] 1、In the prior art, the synergistic effect of H2S / C02 usually leads to signal attenuation, and the traditional eddy current / ultrasonic calibration curve is easy to fail. The scheme establishes an electrochemical-hydrodynamic coupling model and a nonlinear corrosion contribution coefficient model, which fundamentally reflects the chemical kinetics nature of the corrosion process, so that the model still maintains high precision under complex and variable actual working conditions, and completely gets rid of the dependence on the traditional calibration curve of single environmental factor and linear extrapolation.

[0016] 2、Due to the interference of the bimetallic interface, the electromagnetic signal jump of the metal interface is easy to lead to misjudgment of corrosion defects, and the signal-to-noise ratio is reduced by ≥40%. The scheme effectively filters out the inherent signal of the interface through the wavelet packet decomposition-reconstruction algorithm, and significantly improves the signal-to-noise ratio. Combined with the high-resolution detection capability of 0.5mm 2 of the pulse eddy current array, the real corrosion defects and interface structure interference can be clearly distinguished, and the false positive rate is greatly reduced.

[0017] 3、Sulfur deposition shielding effect, sulfur deposition layer on the pipe wall is easy to cause ultrasonic echo signal distortion, so that the wall thickness measurement error is ±1.5mm. The scheme adopts sulfur deposition laser radar to carry out three-dimensional modeling (accuracy ±0.1mm), and accurately obtains the thickness and morphology of the deposition layer. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 It is a flow chart of the method for predicting the residual life of the ultra-high sulfur-containing bimetallic gas pipeline; Figure 2 It is a multi-field coupling relationship diagram in the method for predicting the residual life of the ultra-high sulfur-containing bimetallic gas pipeline; Figure 3 It is a dynamic prediction and optimization flow chart in the method for predicting the residual life of the ultra-high sulfur-containing bimetallic gas pipeline; Figure 4 It is a structural schematic diagram of the residual life prediction system for the ultra-high sulfur-containing bimetallic gas pipeline; Figure 5 It is a multi-source data acquisition flow chart in the residual life prediction system for the ultra-high sulfur-containing bimetallic gas pipeline. DETAILED DESCRIPTION

[0019] The specific embodiments are further described in detail below: The method and system for predicting the residual life of a high-sulfur bimetallic gas pipeline in the embodiment focus on the residual life prediction technology of a high-sulfur bimetallic gas pipeline, and mainly solve the problems of H2S / C02 synergistic corrosion and interface diffusion layer evolution, while strengthening the applicability in production sites. Through the triple innovation of "multi-physical field modeling + intelligent algorithm + extreme environment adaptation", the problem of bimetallic pipeline interface failure prediction under high-sulfur working conditions is systematically solved for the first time.

[0020] Scheme one A method for predicting the residual life of a high-sulfur bimetallic gas pipeline is provided, as shown in the accompanying Figure 1 The method comprises the following steps, (1) Data acquisition: respectively acquiring strain data, temperature data, outer corrosion data and sulfur deposition data.

[0021] In the embodiment, strain data and temperature data are acquired through a distributed optical fiber network, outer corrosion data are detected by using a pulse eddy current array, and sulfur deposition data are acquired by using a sulfur deposition laser radar scan to construct a three-dimensional model.

[0022] (2) Feature extraction: dynamically extracting features from the collected data to obtain corrosion environment features and material state features.

[0023] In the embodiment, the real-time strain data, sulfur deposition scanning data and medium composition obtained are comprehensively analyzed to perform dynamic feature extraction to obtain corrosion environment features and material state features.

[0024] The corrosion environment features are used to establish an H2S-C02- synergistic effect quantification model, which is represented as (1); In the formula, is the H2S corrosion contribution coefficient (g / m 3 ) -0.8 , which is determined by fitting an electrochemical polarization experiment; is the C02 corrosion contribution coefficient (MPa -0.6 ), which is calibrated by a high-pressure axe weight loss experiment; is a temperature coupling coefficient (ppm -1 ·K -1 , which is measured by cyclic voltammetry. When the model is established, the variables are first normalized by experimental data (to eliminate the dimension effect), and then the coefficient is fitted to obtain the coefficient, which has included the dimension compensation of temperature coupling.

[0025] is the H2S volume concentration (g / m 3), measured by an online gas chromatograph; The partial pressure of CO2 (MPa) is monitored in real time by an infrared sensor; for Concentration (ppm) was detected using an ion-selective electrode. The absolute temperature (K) is acquired using a PT100 temperature sensor. The temperature acceleration factor is dimensionless and is based on a correction of the Arrhenius equation. [1] In this embodiment, the principle of dimensional consistency is followed to deduce the result. Dimensions = Corrosion Index Dimensions / dimension( It is the volume concentration, dimensionless, therefore (The dimensions are consistent with the corrosion index). Dimensions = Corrosion index dimensions / CO2 partial pressure dimensions (partial pressure is in MPa, therefore...) The dimension is the corrosion index dimension. MPa -1 ); Dimensions = Corrosion Index Dimensions / ( dimension· dimension)( The concentration is expressed in mol / L. Since it is dimensionless, The dimension is the corrosion index dimension. (mol / L) -1 Based on the above principles, the corrosion coefficients can be obtained respectively. , , The physical dimensions are respectively (g / m 3 ) -0.8 ;(MPa -0.6 (ppm) -1 ·K -1 . The dimensions have been integrated The coupling effect of concentration and temperature ultimately ensures · , , All three have the same dimensions and are consistent with the corrosion index.

[0026] The H2S concentration is expressed as a power of 0.8 to reflect the saturation effect of corrosion rate under high H2S concentrations; while the CO2 partial pressure is expressed as a power of 0.6 to characterize the equilibrium limitation of carbonic acid ionization. Temperature coupling term ( This reflects the effect of temperature on The corrosion rate is exponentially enhanced; for every 10°C increase, the corrosion rate increases by approximately 1.5 times.

[0027] Meanwhile, in practical applications, cross term correction needs to be added, expressed as (2); in which, is the diffusion layer thickness.

[0028] The material state characteristics include interface diffusion layer characteristic extraction.

[0029] The interface diffusion layer characteristics include electrochemical-stress coupling field, element diffusion-corrosion coupling field, phase field-corrosion coupling field, and mechanical property degradation.

[0030] In this embodiment, the electrochemical-stress coupling field modified Scully equation is expressed as (3); in which, is the exchange current density; is the stress generated by external load; is the yield strength; is the limiting diffusion current density; is the overpotential change under the action of stress; is the Tafel slope.

[0031] The element diffusion-corrosion coupling field adopts phase field-finite element joint simulation.

[0032] The phase field-corrosion coupling field is expressed as[2] (4); in which, is the Cr-based diffusion coefficient (m 2 / s), which is calibrated on site through high-temperature diffusion couple experiments; is the local corrosion rate (m / s), which is calibrated on site through ultrasonic thickness inversion, 0.2 is the dimension correction coefficient, is the dimensionless normalization coefficient, is the "dimension conversion coefficient" calibrated through experimental data, and its value has included dimension conversion, ensuring that 1 and can be added; is the Cr element concentration gradient (mol / m 4 ), so that the dimension is (m² / s)·(mol / m 4 )=mol / (m²·s); is the Cr concentration, with the dimension of mol / m 3 , which is calibrated on site through LIBS laser composition analysis; is the dimensionless correction term, which does not change the overall dimension; is the interface chemical reaction source term mol / (m 3• s), in-situ calibrated by electrochemical impedance spectroscopy fitting and dimensionally unified with the diffusion term (D) with dimension of mol / (m³ s).

[0033] The degradation of mechanical properties, i.e. the elastic modulus decay, is expressed as (5). wherein, G0is the initial elastic modulus of the material (GPa), measured by tensile testing machine; C0is the interface Fe element concentration, dimensionless, obtained by EDS energy spectrum area scanning.

[0034] (3) Multi-physical field coupling model construction: In this embodiment, combined with the multi-field coupling relationship diagram shown in FIG. 1, the mathematical relationship corresponding to the flow chart is: Figure 2 Corrosion to diffusion: (6). wherein, is the time rate of change of concentration ; D0is the reference diffusion coefficient, representing the material diffusion ability without corrosion; is the local corrosion rate (m / s); is the Laplace operator, describing the spatial distribution gradient of concentration , which is the spatial driving term of diffusion; in this embodiment, 1 and constitute a “corrosion influence correction factor”, and 0.2 is a dimension normalization coefficient, which ensures that the two terms are dimensionless and can be added, is the dimensionless influence degree of corrosion rate on diffusion, 1 is the reference diffusion weight without corrosion, and the two together describe the correction effect of corrosion on the diffusion coefficient, which is expressed by equation (6) as “the diffusion coefficient changes linearly with the corrosion rate”, and 1 ensures that when there is no corrosion (i.e. =0), the diffusion coefficient is , which meets the reference definition.

[0035] Diffusion to stiffness: (7). Stress to corrosion: (8). wherein, G is the elastic modulus of the material; G0is the initial elastic modulus of the material; is the CO2 corrosion contribution coefficient; is the reference corrosion rate under no stress (or initial stress state); ​R is the volumetric stress; R is the gas constant (value is 8.314 J / (mol)). K); T is the thermodynamic temperature.

[0036] Based on the obtained feature data, a synergistic corrosion kinetic model is established, considering the synergistic effect of H2S and CO2, and a corrosion rate equation is established, resulting in a multiphysics coupling model, expressed as follows: (9); In the formula, Corrosion depth (mm); The reaction rate constant ( ); H2S partial pressure (MPa); CO2 molar concentration (mol / m 3 ); E a Activation energy (J / mol); Absolute temperature; The interfacial diffusion coefficient (m) 2 / s); For the thickness of the diffusion layer ( ); It is an empirical index, calibrated through material testing.

[0037] In this embodiment, corrosion experiments were conducted on the material under a standard corrosion environment (with known H2S partial pressure, CO2 concentration, temperature, etc.) to obtain data on the relationship between corrosion depth and time. .

[0038] This represents the change in corrosion depth over time, with dimensions in mm / s, assuming an empirical exponent that ( The dimension of ) is MPa mol / m 3 ,but The dimensions are (mm / s) / [(MPa) mol / m 3 ) [(m / s)]=mm m -1 MPa -1 m 3 mol -1 =mm m 2 MPa -1 mol -1 (The final adjustment needs to be made based on the specific value of the experience index).

[0039] The above formula describes the change in corrosion depth over time under the synergistic effect of H2S and CO2, integrating the chemical reaction rate ( ) and interface diffusion rate ( The coupling effect of ).

[0040] The evolution equation of the interface diffusion layer is obtained based on the multiphysics coupling model. In this embodiment, taking the 316L / X70 bimetallic interface as an example, it is expressed as follows: (10); In the formula, The diffusion kinetic coefficient (m) 2 / s); Residual stress (MPa); The viscosity coefficient of the material (Pa·s); This is plastic strain.

[0041] In this embodiment, the construction of the multiphysics coupling model also includes dynamically calibrating the diffusion layer thickness using in-situ TEM observation data, expressed as follows: (11); In the formula, Let t be the thickness of the diffusion layer (m) at time t, representing the thickness of the layered region formed by the diffusion of chromium atoms at time t. These are dimensionless calibration coefficients; The basic diffusion coefficient of Cr; This refers to the duration of pipeline use.

[0042] In this embodiment, The basic diffusion coefficient of Cr (m) 2 / s), where t is a time unit of seconds, then The dimension of m 2 Then the whole The dimensionless quantity is m, and Consistent. Among them, This is a dimensionless coefficient calibrated using in-situ TEM observation data, used to correct the deviation between the theoretical diffusion thickness and the actual observed value.

[0043] (4) Remaining Life Prediction: In this embodiment, a comprehensive life prediction model is established based on a multiphysics coupling model, combined with comprehensive material data (yield strength), monitoring data (wall thickness loss), and gas data (H2S concentration); the remaining life is predicted, and the prediction results are output. The comprehensive life prediction model is expressed as follows: (12); In the formula, The critical corrosion depth is taken as 20% of the wall thickness. This represents the initial corrosion depth. Yield strength; For wall thickness loss; A is a material constant, which in this embodiment is 2.1 x 10⁻⁶ for X70 steel. -4 3616L is taken as 1.7 x 10 -5 B is the medium sensitivity coefficient, obtained through regression analysis of gas field test data. The molar concentration of H2S.

[0044] In this embodiment, the remaining lifetime prediction also includes incorporating production data, such as flow rate and temperature, to form a sulfur deposition corrosion correction term, which corrects for sulfur deposition corrosion. The correction term is then expressed as follows: (13); In the formula, The deposition rate coefficient was determined through downhole corrosion coupon experiments. For flow rate; For temperature; This refers to the duration of pipeline use.

[0045] In this embodiment, to meet the parameter and visualization requirements of the remaining life prediction algorithm for ultra-high sulfur bimetallic gas pipelines, the remaining life prediction results are output, mainly including interactive dashboard display and failure probability radar chart display.

[0046] (5) Model optimization: Compare and analyze the prediction results with the field detection data, and adjust the model parameters according to the analysis results.

[0047] In this embodiment, the optimization strategy for key parameters is shown in Table 1 below. Table 1

[0048] Specifically, in conjunction with the appendix Figure 3 As shown, lifespan prediction is performed in real time based on on-site monitoring data. The prediction results are compared with the on-site monitoring data to determine if a threshold has been triggered. If not, monitoring continues; if the threshold is triggered, an alarm is triggered and model retraining is initiated, i.e., parameter optimization training. In this embodiment, Kalman filtering is used to smooth the monitoring data, and Weibull distribution is used to update the failure probability, which participates in the model parameter optimization process, forming a monitoring-prediction closed loop.

[0049] Option 2 A system for predicting the remaining life of ultra-high sulfur bimetallic gas pipelines is provided, which is applied to the aforementioned method for predicting the remaining life of ultra-high sulfur bimetallic gas pipelines, as shown in the attached figure. Figure 4 As shown, including, Data acquisition layer: A distributed fiber optic network is used to acquire strain and temperature data at a sampling frequency of 1 kHz; a pulsed eddy current array is used at a sampling frequency of 0.5 mm. 2 The outer corrosion data was acquired at a high resolution; sulfur deposition data was obtained by sulfur deposition lidar scanning.

[0050] In this embodiment, as shown in the appendix Figure 5 As shown, strain / temperature data are acquired via a distributed fiber optic network, outer layer corrosion data is acquired via a pulsed eddy current array, 3D modeling data is obtained via a sulfur deposition lidar, H2S / CO2 concentrations are obtained via an online gas chromatograph, and localized corrosion signals are obtained via an electrochemical noise probe. The acquired data are transmitted to a data preprocessing center for multi-source data fusion processing, and the processed data is then standardized and stored.

[0051] Specifically, in this embodiment, the parameter data and their sources are mainly shown in Table 2 below.

[0052] Table 2

[0053] Feature extraction layer: The collected data is fused and the corrosion environment features and material state features are dynamically extracted.

[0054] Model building layer: Based on feature data, an interface diffusion factor is introduced to establish a multiphysics coupling model to quantify the accelerating effect of sulfur deposition on interface diffusion.

[0055] Lifetime prediction layer: Based on the evolution process of the diffusion layer at the bimetallic interface, the phase field method is used to simulate the interface crack initiation process, obtain the diffusion layer thickness change rate and stress concentration factor, predict the remaining lifetime, and output the prediction results.

[0056] Intelligent correction layer: ResNet-50 is used to acquire ultrasonic inspection images of pipelines on site, and the images are compared with the prediction results. The model parameters are corrected and optimized based on the comparison results.

[0057] It also includes a visualization output layer: In this embodiment, the remaining lifetime curve and confidence interval are displayed in real time through an interactive dashboard, and the sensitivity analysis of parameters such as the impact of H2S concentration on lifetime is slidably adjusted; and a failure probability radar chart is output to compare the risk weights of different failure modes (sulfur corrosion / interface peeling / stress cracking).

[0058] This solution employs a distributed fiber optic network to synchronously acquire strain / temperature monitoring data, and utilizes a pulsed eddy current array to achieve an outer layer corrosion detection resolution of 0.5 mm. 2 Using a sulfur deposition lidar, a 3D modeling accuracy of ±0.1mm was achieved.

[0059] In terms of algorithm architecture, a dynamic compensation mechanism is adopted, LSTM neural network compensates for ultrasonic guided wave dispersion effect, wavelet packet decomposition-reconstruction algorithm solves the signal ambiguity problem of pulse eddy current interface, combined with a hybrid prediction model, ultrasonic image is processed by ResNet-50, and LSTM predicts sulfur deposition growth trend, thereby overcoming the conflicts and difficulties in traditional prediction methods, improving the accuracy of life prediction, facilitating timely maintenance or replacement of pipelines, and improving safety.

[0060] The following is a detailed description through specific implementation examples.

[0061] Calculations were performed based on production data and gas quality data of a certain gas pipeline operating in Dongguan 100.

[0062] Among them, the H2S content in the gas chromatography data reached 43.315~44.590 g / m³. 3 (>30g / m 3 Threshold); significant synergistic corrosion effect of H2S / CO2 (CO2 content 76.035~87.431 g / m³). 3 ).

[0063] The corrosion factor is then automatically calculated as follows: In the formula, Corrosion factor (dimensionless); Hydrogen sulfide mass concentration (g / m³) 3 ); Carbon dioxide mass concentration (g / m³) 3 ).

[0064] The thickness of the dynamic diffusion layer is calculated as follows: In the formula, The thickness of the diffusion layer is (mm). The current year (e.g., 2025); The base year (e.g., 2020).

[0065] The remaining useful life prediction calculation process is as follows: ; In the formula, Remaining lifespan (years); The base life (in years) needs to be determined based on the initial parameters of the pipe.

[0066] If we substitute the data from 2023, we know that: ; ; ; Assumption Year.

[0067] Corrosion factor: ; Diffusion layer thickness: ; Remaining lifespan: .

[0068] Meanwhile, actual monitoring revealed a serious safety hazard in the pipeline, necessitating immediate cessation of its use and replacement or emergency repair. The calculation results are consistent with the actual monitoring findings.

[0069] In this embodiment, to address the problem of insufficient accuracy in predicting the remaining life of ultra-high sulfur bimetallic pipelines, a three-phase synergistic corrosion kinetic equation of H2S-CO2-H2O was established for the first time. An electrochemical-hydrodynamic coupling model was introduced to quantify the accelerating effect of sulfur deposition on interfacial diffusion. A bimetallic interface diffusion layer evolution algorithm based on Fick's law was developed, and the crack initiation process was simulated by the phase field method, effectively establishing a multi-physics coupling model.

[0070] Meanwhile, abandoning conventional sensors susceptible to H2S embrittlement, this innovative approach employs distributed optical fibers for intrinsically safe strain / temperature monitoring, utilizes pulsed eddy current arrays for high-resolution outer layer corrosion imaging, and introduces sulfur deposition lidar for 3D modeling. This fundamentally solves the two major challenges of sensor failure in extreme environments and the shielding effect of sulfur deposition. The outer layer corrosion detection resolution reaches 0.5 mm. 2 The accuracy of the constructed 3D model was controlled within ±0.1 mm. This was achieved even under extreme conditions (H2S > 50 g / m³). 3 This solution increases the sensor's hydrogen sulfide embrittlement resistance lifespan from 200 hours to 2000 hours. The H2S penetration corrosion rate of the metal seals is reduced from >0.5 mm / year to <0.1 mm / year.

[0071] A diagnostic engine driven by physical mechanisms and artificial intelligence was constructed. The physical model is used as the prior knowledge and constraint of the intelligent algorithm. LSTM is used to dynamically compensate for ultrasonic dispersion, wavelet packet decomposition is used to specifically analyze interface interference signals, and a hybrid model of ResNet-50 and LSTM is constructed to process images and trend prediction respectively. This achieves deep fusion and intelligent analysis of multi-source heterogeneous data, which can "know not only what" but also "know why". The crack identification accuracy reaches 98.7%, and the error of sulfur deposition growth trend prediction is controlled within ±3%.

[0072] The above descriptions are merely embodiments of the present invention, and common knowledge such as specific technical solutions and / or characteristics are not described in detail here. It should be noted that those skilled in the art can make various modifications and improvements without departing from the technical solutions of the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A method for predicting the remaining life of a bimetallic gas pipeline with ultra-high sulfur content, characterized in that, Includes the following steps, (1) Data acquisition: acquire strain data, temperature data, outer corrosion data and sulfur deposition data respectively; (2) Feature extraction: Dynamic feature extraction is performed on the collected data to obtain the characteristics of the corrosion environment and the material state; (3) Construction of multiphysics coupling model: Based on the obtained feature data, a synergistic corrosion kinetic model is established, considering the synergistic effect of H2S and CO2. A corrosion rate equation is then established, resulting in a multiphysics coupling model, expressed as follows: ； In the formula, 'Depth of corrosion; It is the reaction rate constant; For H2S voltage division; CO2 molar concentration; E a It is the activation energy; Absolute temperature; The interfacial diffusion coefficient; The thickness of the diffusion layer; It is an experience index; (4) Remaining lifetime prediction: Based on a multiphysics coupling model, combined with comprehensive material data, monitoring data, and gas data, a comprehensive lifetime prediction model is established; the remaining lifetime is predicted, and the prediction results are output; the comprehensive lifetime prediction model is expressed as follows: ； In the formula, This represents the critical corrosion depth. This represents the initial corrosion depth. Yield strength; A represents wall thickness loss; B represents material constant; and C represents medium sensitivity coefficient. The molar concentration of H2S; (5) Model optimization: Compare and analyze the prediction results with the field detection data, and adjust the model parameters according to the analysis results.

2. The method for predicting the remaining life of an ultra-high sulfur bimetallic gas pipeline according to claim 1, characterized in that: The corrosive environmental characteristics are used to establish H2S-CO2- Synergistic effect quantification model, expressed as ； In the formula, Contribution coefficient to H2S corrosion; Contribution coefficient to CO2 corrosion; for Coupling coefficient with temperature; H2S volume concentration; For CO2 voltage divider; for concentration; Absolute temperature; This is the temperature acceleration factor.

3. The method for predicting the remaining life of an ultra-high sulfur bimetallic gas pipeline according to claim 1, characterized in that: The material state characteristics include the extraction of interface diffusion layer features; these features include electrochemical-stress coupling field, element diffusion-corrosion coupling field, phase field-corrosion coupling field, and mechanical property degradation; the evolution equation of the interface diffusion layer is obtained based on the multiphysics coupling model, expressed as follows: ； In the formula, The diffusion kinetic coefficient; This is residual stress; The viscosity coefficient of the material; This is plastic strain.

4. The method for predicting the remaining life of an ultra-high sulfur bimetallic gas pipeline according to claim 3, characterized in that: The electrochemical-stress coupling field-corrected scully equation is expressed as follows: ； In the formula, For exchange current density; Stress generated by external loads; Yield strength; This represents the limiting diffusion current density; This represents the change in overpotential under stress. This is the Tafel slope.

5. The method for predicting the remaining life of an ultra-high sulfur bimetallic gas pipeline according to claim 3, characterized in that: The phase field-corrosion coupling field is expressed as: ； In the formula, The basic diffusion coefficient of Cr; This represents the localized corrosion rate. This refers to the concentration of Cr element. The Cr element concentration gradient; This is the source term for interfacial chemical reactions.

6. The method for predicting the remaining life of an ultra-high sulfur bimetallic gas pipeline according to claim 3, characterized in that: The mechanical property degradation is expressed as ； In the formula, This is the initial elastic modulus of the material; The concentration of Fe element at the interface.

7. The method for predicting the remaining life of an ultra-high sulfur bimetallic gas pipeline according to claim 1, characterized in that: The remaining service life prediction also includes incorporating production data to correct for sulfur deposition corrosion; the correction term is expressed as follows: ； In the formula, This is the deposition rate coefficient; For flow rate; For temperature; This refers to the duration of pipeline use.

8. The method for predicting the remaining life of an ultra-high sulfur bimetallic gas pipeline according to claim 5, characterized in that: The construction of the multiphysics coupling model also includes dynamically calibrating the diffusion layer thickness using in-situ TEM observation data, expressed as... In the formula, Let be the thickness of the diffusion layer at time t, representing the thickness of the layered region formed by the diffusion of chromium atoms at time t. These are dimensionless calibration coefficients; The basic diffusion coefficient of Cr; This refers to the duration of pipeline use.

9. A system for predicting the remaining life of a bimetallic gas pipeline with ultra-high sulfur content, characterized in that, The method for predicting the remaining life of a bimetallic gas pipeline with ultra-high sulfur content, as described in any one of claims 1-8, is applicable to this method. include, Data acquisition layer: A distributed fiber optic network is used to acquire strain and temperature data at a sampling frequency of 1 kHz; a pulsed eddy current array is used at a sampling frequency of 0.5 mm. 2 The outer corrosion data was acquired at a high resolution; sulfur deposition data was obtained using sulfur deposition lidar scanning. Feature extraction layer: The collected data is fused and the corrosion environment features and material state features are dynamically extracted; Model building layer: Based on feature data, an interface diffusion factor is introduced to establish a multiphysics coupling model to quantify the accelerating effect of sulfur deposition on interface diffusion; Lifetime prediction layer: Based on the evolution process of the diffusion layer at the bimetallic interface, the phase field method is used to simulate the interface crack initiation process, obtain the diffusion layer thickness change rate and stress concentration factor, predict the remaining lifetime, and output the prediction results; Intelligent correction layer: ResNet-50 is used to acquire ultrasonic inspection images of pipelines on site, and the images are compared with the prediction results. The model parameters are corrected and optimized based on the comparison results.

10. A system for predicting the remaining life of an ultra-high sulfur bimetallic gas pipeline according to claim 9, characterized in that, It also includes a visualization output layer: real-time display of remaining lifetime curves and confidence intervals via an interactive dashboard, and sensitivity analysis of sliding adjustment parameters; It also outputs a failure probability radar chart and compares the risk weights of different failure modes.

Images