Pub Date : 2025-10-28DOI: 10.1016/j.ijpvp.2025.105689
Kai Yu , Xinhong Li , Ziyue Han , Xiuquan Liu , Yuanjiang Chang , Guoming Chen
Pressure shells are widely employed in subsea energy development. The manufacturing process and harsh marine environments may lead to the damage of shells, which may reduce the strength, and even lead to the instablility of subsea shell. This study aims to investigate the buckling behavior of capsule-shaped pressure shells with defects under hydrostatic pressure corresponding to a 2000 m water depth, with a focus on understanding the effects of single and multiple defects on the critical buckling load. A FE model of a capsule-shaped pressure shell is developed, and nonlinear buckling analyses are performed using the Riks method. Two types of defects are considered, i.e., initial geometric defects, e.g., out-of-roundness, and damage defects, e.g., corrosion or cracks. It is observed that the effect of initial geometric defects on critical buckling load is negligible. For single defects, corrosion area, corrosion depth, and crack length are dominant factors affecting buckling resistance. In cases of the double corrosion defects, the critical buckling load gradually recovers with increasing corrosion distance. For coupled crack-corrosion defects, most significant reduction in critical buckling load occurs when crack boundary just comes into contact with the corrosion pit. This study quantitatively investigates the coupled effects of defects on structural stability, and the outcomes can be applied for integrity management of capsule-shaped subsea pressure shells.
{"title":"Buckling failure assessment of capsule-shaped subsea pressure shell containing defects","authors":"Kai Yu , Xinhong Li , Ziyue Han , Xiuquan Liu , Yuanjiang Chang , Guoming Chen","doi":"10.1016/j.ijpvp.2025.105689","DOIUrl":"10.1016/j.ijpvp.2025.105689","url":null,"abstract":"<div><div>Pressure shells are widely employed in subsea energy development. The manufacturing process and harsh marine environments may lead to the damage of shells, which may reduce the strength, and even lead to the instablility of subsea shell. This study aims to investigate the buckling behavior of capsule-shaped pressure shells with defects under hydrostatic pressure corresponding to a 2000 m water depth, with a focus on understanding the effects of single and multiple defects on the critical buckling load. A FE model of a capsule-shaped pressure shell is developed, and nonlinear buckling analyses are performed using the Riks method. Two types of defects are considered, i.e., initial geometric defects, e.g., out-of-roundness, and damage defects, e.g., corrosion or cracks. It is observed that the effect of initial geometric defects on critical buckling load is negligible. For single defects, corrosion area, corrosion depth, and crack length are dominant factors affecting buckling resistance. In cases of the double corrosion defects, the critical buckling load gradually recovers with increasing corrosion distance. For coupled crack-corrosion defects, most significant reduction in critical buckling load occurs when crack boundary just comes into contact with the corrosion pit. This study quantitatively investigates the coupled effects of defects on structural stability, and the outcomes can be applied for integrity management of capsule-shaped subsea pressure shells.</div></div>","PeriodicalId":54946,"journal":{"name":"International Journal of Pressure Vessels and Piping","volume":"219 ","pages":"Article 105689"},"PeriodicalIF":3.5,"publicationDate":"2025-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145416693","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-27DOI: 10.1016/j.ijpvp.2025.105693
Chih-Hsuan Lee
The nozzles in RPV systems are critical components due to their high-stress concentrations, which can significantly affect the system's structural integrity. Despite the relatively minor impact of material radiation embrittlement, the stress concentration at nozzle corners should be closely monitored and precisely evaluated to ensure long-term operational safety. As different RPV nozzle geometries result in varying stress distributions and stress intensity factors (SIFs) for specific crack depths along a 45° path from the high-stress concentration point, identifying the geometry with the lowest SIFs represents the optimal design. Recently, artificial intelligence (AI) algorithms have been used to assist in calculating the stress distribution in finite element analysis (FEA), which can rapidly acquire a solution without any convergence issues during FEA. In this work, the verified finite element models (FEMs) of the RPV nozzle are established to generate extensive datasets, which correspond to various geometry sizes with the SIFs. These results are then applied to a machine learning model, support vector regression (SVR), which includes a kernel function that is suitable for high-dimensional cases. After training the SVR model, it was applied to particle swarm optimization (PSO) to identify the optimal design for the RPV nozzle geometry. The results demonstrate that the PSO with the trained SVR model can find an optimal design of nozzle geometry, which is better than the FEA results.
{"title":"Prediction and optimization of stress intensity factors for reactor pressure vessel nozzles using support vector regression and particle swarm optimization","authors":"Chih-Hsuan Lee","doi":"10.1016/j.ijpvp.2025.105693","DOIUrl":"10.1016/j.ijpvp.2025.105693","url":null,"abstract":"<div><div>The nozzles in RPV systems are critical components due to their high-stress concentrations, which can significantly affect the system's structural integrity. Despite the relatively minor impact of material radiation embrittlement, the stress concentration at nozzle corners should be closely monitored and precisely evaluated to ensure long-term operational safety. As different RPV nozzle geometries result in varying stress distributions and stress intensity factors (SIFs) for specific crack depths along a 45° path from the high-stress concentration point, identifying the geometry with the lowest SIFs represents the optimal design. Recently, artificial intelligence (AI) algorithms have been used to assist in calculating the stress distribution in finite element analysis (FEA), which can rapidly acquire a solution without any convergence issues during FEA. In this work, the verified finite element models (FEMs) of the RPV nozzle are established to generate extensive datasets, which correspond to various geometry sizes with the SIFs. These results are then applied to a machine learning model, support vector regression (SVR), which includes a kernel function that is suitable for high-dimensional cases. After training the SVR model, it was applied to particle swarm optimization (PSO) to identify the optimal design for the RPV nozzle geometry. The results demonstrate that the PSO with the trained SVR model can find an optimal design of nozzle geometry, which is better than the FEA results.</div></div>","PeriodicalId":54946,"journal":{"name":"International Journal of Pressure Vessels and Piping","volume":"219 ","pages":"Article 105693"},"PeriodicalIF":3.5,"publicationDate":"2025-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145465517","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-27DOI: 10.1016/j.ijpvp.2025.105695
Hyun-Jae Lee , Hune-Tae Kim , Seok-Pyo Hong
To investigate the guidance for existing methods to estimate stress intensity factors (KI) and the J-integrals for circumferentially cracked pipes in the presence of weld residual stress (WRS), finite element (FE) analysis is conducted. The axial component of the Level 3 WRS profiles provided in R6 is considered. It is found that the weight function method is applicable for KI estimation, and Vo can be taken as unity for J estimation as advised in R6. Furthermore, the interaction, within elastic-plastic regime, between the Level 3 profiles and axial tension as mechanical loading can be addressed using the no elastic follow-up V-factor, V(2). Noting that reconstruction of WRS for fracture mechanics FE analysis is demonstrated, and an extension to the Level 2 profiles, upper-bound profiles, is discussed as a means to reduce conservatism.
{"title":"FE validation of R6 J estimation for circumferentially cracked pipes under combined residual stress and mechanical loading: Reconstruction of R6 Level 3 axial residual stress for pipe butt weld","authors":"Hyun-Jae Lee , Hune-Tae Kim , Seok-Pyo Hong","doi":"10.1016/j.ijpvp.2025.105695","DOIUrl":"10.1016/j.ijpvp.2025.105695","url":null,"abstract":"<div><div>To investigate the guidance for existing methods to estimate stress intensity factors (<em>K</em><sub><em>I</em></sub>) and the <em>J</em>-integrals for circumferentially cracked pipes in the presence of weld residual stress (WRS), finite element (FE) analysis is conducted. The axial component of the Level 3 WRS profiles provided in R6 is considered. It is found that the weight function method is applicable for <em>K</em><sub><em>I</em></sub> estimation, and <em>V</em><sub><em>o</em></sub> can be taken as unity for <em>J</em> estimation as advised in R6. Furthermore, the interaction, within elastic-plastic regime, between the Level 3 profiles and axial tension as mechanical loading can be addressed using the no elastic follow-up <em>V</em>-factor, <em>V</em><sup><em>(2)</em></sup>. Noting that reconstruction of WRS for fracture mechanics FE analysis is demonstrated, and an extension to the Level 2 profiles, upper-bound profiles, is discussed as a means to reduce conservatism.</div></div>","PeriodicalId":54946,"journal":{"name":"International Journal of Pressure Vessels and Piping","volume":"219 ","pages":"Article 105695"},"PeriodicalIF":3.5,"publicationDate":"2025-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145528019","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-25DOI: 10.1016/j.ijpvp.2025.105675
Yuman Sun , Wenhong Ding , Zhonghai Zang , Hongyuan Ding , Zhuang Chen , Chenxu Wang
This study investigates the degradation of fracture toughness in AISI 4130 steel exposed to high-pressure hydrogen through integrated experimental testing and finite element modeling. Slow strain rate tensile (SSRT) and elastic-plastic fracture toughness tests were performed at 5 MPa, 10 MPa, and 30 MPa hydrogen pressures, with ambient air serving as a reference. Experimental results revealed a pronounced deterioration of mechanical properties, evidenced by a significant reduction in elastic-plastic fracture toughness (JIC) from 60.34 kJ/m2 in ambient air to 9.99 kJ/m2 under 30 MPa hydrogen pressure. Concurrently, the hydrogen embrittlement index (IHE) increased from 70.68 % at 5 MPa to 86.26 % at 30 MPa. Fractographic analysis further demonstrated a progressive transition from ductile microvoid coalescence (MVC) in ambient air to a mixed mode of quasi-cleavage (QC) and martensitic lath decohesion (MLD) at intermediate pressures, and ultimately to intergranular (IG) fracture at 30 MPa. Coupled finite element simulations elucidated hydrogen diffusion and accumulation at the crack tip under stress gradients. The numerical analysis confirmed that elevated hydrogen pressure enhanced both lattice and trapped hydrogen enrichment, leading to intensified strain localization and a reduction in the critical stress required for crack propagation. These numerical results corroborated the experimental observations, confirming that the synergistic effects of hydrogen-enhanced localized plasticity (HELP) and hydrogen-enhanced decohesion (HEDE) mechanisms inhibit crack-tip blunting, reduce energy dissipation, and accelerate the transition from ductile to brittle fracture. These findings provide a mechanistic foundation for predicting fracture behavior in high-pressure hydrogen environments and underscore the necessity of incorporating fracture toughness degradation into structural integrity assessments of high-pressure hydrogen storage systems.
{"title":"Influence of hydrogen pressure on fracture toughness degradation of AISI 4130 Steel: Experimental and finite element study","authors":"Yuman Sun , Wenhong Ding , Zhonghai Zang , Hongyuan Ding , Zhuang Chen , Chenxu Wang","doi":"10.1016/j.ijpvp.2025.105675","DOIUrl":"10.1016/j.ijpvp.2025.105675","url":null,"abstract":"<div><div>This study investigates the degradation of fracture toughness in AISI 4130 steel exposed to high-pressure hydrogen through integrated experimental testing and finite element modeling. Slow strain rate tensile (SSRT) and elastic-plastic fracture toughness tests were performed at 5 MPa, 10 MPa, and 30 MPa hydrogen pressures, with ambient air serving as a reference. Experimental results revealed a pronounced deterioration of mechanical properties, evidenced by a significant reduction in elastic-plastic fracture toughness (<em>J</em><sub>IC</sub>) from 60.34 kJ/m<sup>2</sup> in ambient air to 9.99 kJ/m<sup>2</sup> under 30 MPa hydrogen pressure. Concurrently, the hydrogen embrittlement index (<em>I</em><sub>HE</sub>) increased from 70.68 % at 5 MPa to 86.26 % at 30 MPa. Fractographic analysis further demonstrated a progressive transition from ductile microvoid coalescence (MVC) in ambient air to a mixed mode of quasi-cleavage (QC) and martensitic lath decohesion (MLD) at intermediate pressures, and ultimately to intergranular (IG) fracture at 30 MPa. Coupled finite element simulations elucidated hydrogen diffusion and accumulation at the crack tip under stress gradients. The numerical analysis confirmed that elevated hydrogen pressure enhanced both lattice and trapped hydrogen enrichment, leading to intensified strain localization and a reduction in the critical stress required for crack propagation. These numerical results corroborated the experimental observations, confirming that the synergistic effects of hydrogen-enhanced localized plasticity (HELP) and hydrogen-enhanced decohesion (HEDE) mechanisms inhibit crack-tip blunting, reduce energy dissipation, and accelerate the transition from ductile to brittle fracture. These findings provide a mechanistic foundation for predicting fracture behavior in high-pressure hydrogen environments and underscore the necessity of incorporating fracture toughness degradation into structural integrity assessments of high-pressure hydrogen storage systems.</div></div>","PeriodicalId":54946,"journal":{"name":"International Journal of Pressure Vessels and Piping","volume":"219 ","pages":"Article 105675"},"PeriodicalIF":3.5,"publicationDate":"2025-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145416653","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-25DOI: 10.1016/j.ijpvp.2025.105694
Ping Wang , Yankai Wang , Zhihao Chen , Banglong Yu , Yong Liu , Hongliang Qian
Multiple repair welds significantly affect the microstructure and mechanical properties of aluminum alloy joints. In the study, 0 to 3 repair welds were conducted on 5083-O aluminum alloy using melt inert-gas (MIG) welding. Through metallographic observation, hardness testing, and fatigue analysis, a systematic comparison and analysis of the changes in the microstructure and mechanical properties of the repaired welded joints were carried out. Finite element analysis (FEA), validated by the blind-hole method, was used to analyze the residual stress redistribution. The results show that repair welding aggravates grain boundary liquefaction in the partial melting zone (PMZ) and grain coarsening in the heat-affected zone (HAZ), causing a slight decrease in hardness from 87 MPa in the as-welded state to 81 MPa after three repairs. Additionally, the peak residual stress increases notably with the number of repair welds, reducing the fatigue life. Compared with the as-welded joints, the fatigue life of joints after one, two, and three repairs decreased by 11.5 %, 41.2 %, and 35.4 %, respectively. Finally, the study developed the master S-N curve, considering the influence of residual stress and angular misalignment, and provided a theoretical framework for the fatigue life assessment for repaired welded aluminum alloy structures.
{"title":"Fatigue property analysis of 5083-O aluminum alloy joints considering multiple repaired welded residual stress","authors":"Ping Wang , Yankai Wang , Zhihao Chen , Banglong Yu , Yong Liu , Hongliang Qian","doi":"10.1016/j.ijpvp.2025.105694","DOIUrl":"10.1016/j.ijpvp.2025.105694","url":null,"abstract":"<div><div>Multiple repair welds significantly affect the microstructure and mechanical properties of aluminum alloy joints. In the study, 0 to 3 repair welds were conducted on 5083-O aluminum alloy using melt inert-gas (MIG) welding. Through metallographic observation, hardness testing, and fatigue analysis, a systematic comparison and analysis of the changes in the microstructure and mechanical properties of the repaired welded joints were carried out. Finite element analysis (FEA), validated by the blind-hole method, was used to analyze the residual stress redistribution. The results show that repair welding aggravates grain boundary liquefaction in the partial melting zone (PMZ) and grain coarsening in the heat-affected zone (HAZ), causing a slight decrease in hardness from 87 MPa in the as-welded state to 81 MPa after three repairs. Additionally, the peak residual stress increases notably with the number of repair welds, reducing the fatigue life. Compared with the as-welded joints, the fatigue life of joints after one, two, and three repairs decreased by 11.5 %, 41.2 %, and 35.4 %, respectively. Finally, the study developed the master S-N curve, considering the influence of residual stress and angular misalignment, and provided a theoretical framework for the fatigue life assessment for repaired welded aluminum alloy structures.</div></div>","PeriodicalId":54946,"journal":{"name":"International Journal of Pressure Vessels and Piping","volume":"219 ","pages":"Article 105694"},"PeriodicalIF":3.5,"publicationDate":"2025-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145416694","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-24DOI: 10.1016/j.ijpvp.2025.105688
Chen Liu , Yuanze Ma , Shuo Zhang , Changyun Li , Liqun Li , Jing Wang , Haoyue Li , Caiwang Tan , Hongbo Xia , Peng He
This study investigates the influence of laser power on weld formation, microstructure, and mechanical performance of 6-mm-thick Q235B steel joints fabricated by laser–CMT hybrid welding, combined with finite element thermal simulations. Results showed that increasing laser power decreased the cooling rate and promoted grain coarsening in all weld regions. The WZ consisted mainly of FA, FSP, FP, and bainite; the CGHAZ was dominated by Widmanstätten; while the FGHAZ contained refined ferrite–pearlite compared with the BM. EBSD analysis indicated that higher laser power enhanced texture intensity, while the fraction of HAGBs decreased from 69.8 % to 54.2 % and the KAM value dropped from 0.68° to 0.53°, reflecting a reduction in geometrically necessary dislocation density. At 5100 W, the joint exhibited optimal properties, with tensile strength of 730 MPa, elongation of 13.12 %, strength-ductility balance of 9.6 × 103 MPa% and peak microhardness of 204.6 HV. These superior properties were attributed to defect-free weld morphology, moderate grain size, and a favorable balance of texture and grain boundary characteristics. Numerical simulations accurately reproduced thermal cycles and weld profiles, confirming that higher power extends cooling time and reduces cooling rate, thereby accelerating grain growth. Overall, appropriate heat input is essential for controlling microstructural evolution and achieving a superior strength–ductility synergy in laser–CMT hybrid welded joints.
{"title":"Influence of laser power on weld formation, microstructure, and mechanical properties of Q235B steel joined by Laser-CMT hybrid welding process","authors":"Chen Liu , Yuanze Ma , Shuo Zhang , Changyun Li , Liqun Li , Jing Wang , Haoyue Li , Caiwang Tan , Hongbo Xia , Peng He","doi":"10.1016/j.ijpvp.2025.105688","DOIUrl":"10.1016/j.ijpvp.2025.105688","url":null,"abstract":"<div><div>This study investigates the influence of laser power on weld formation, microstructure, and mechanical performance of 6-mm-thick Q235B steel joints fabricated by laser–CMT hybrid welding, combined with finite element thermal simulations. Results showed that increasing laser power decreased the cooling rate and promoted grain coarsening in all weld regions. The WZ consisted mainly of FA, FSP, FP, and bainite; the CGHAZ was dominated by Widmanstätten; while the FGHAZ contained refined ferrite–pearlite compared with the BM. EBSD analysis indicated that higher laser power enhanced texture intensity, while the fraction of HAGBs decreased from 69.8 % to 54.2 % and the KAM value dropped from 0.68° to 0.53°, reflecting a reduction in geometrically necessary dislocation density. At 5100 W, the joint exhibited optimal properties, with tensile strength of 730 MPa, elongation of 13.12 %, strength-ductility balance of 9.6 × 10<sup>3</sup> MPa% and peak microhardness of 204.6 HV. These superior properties were attributed to defect-free weld morphology, moderate grain size, and a favorable balance of texture and grain boundary characteristics. Numerical simulations accurately reproduced thermal cycles and weld profiles, confirming that higher power extends cooling time and reduces cooling rate, thereby accelerating grain growth. Overall, appropriate heat input is essential for controlling microstructural evolution and achieving a superior strength–ductility synergy in laser–CMT hybrid welded joints.</div></div>","PeriodicalId":54946,"journal":{"name":"International Journal of Pressure Vessels and Piping","volume":"219 ","pages":"Article 105688"},"PeriodicalIF":3.5,"publicationDate":"2025-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145416697","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-24DOI: 10.1016/j.ijpvp.2025.105691
O. Muránsky , M.N. Tran , W. Payten
This study explores the potential of PINN-style physics-regularised neural networks to improve long-term creep life predictions for 2.25Cr-1Mo (Grade 22) steel using only short-term experimental data. Three modelling approaches were evaluated: a purely physics-based semi-empirical model based on the Stress-Modified Ductility Exhaustion (SMDE) formulation, a purely data-driven neural network (NNN), and PINN-style models combining empirical learning with physics-based regularisation via a dual-loss approach. Model performance was assessed in both interpolation (within the short-term training domain) and extrapolation (on unseen long-term data). The SMDE model served as a physics-based baseline, exhibiting stable interpolation and extrapolation behaviour. In contrast, the NNN model overfitted the short-term data and failed to generalise to long-term conditions. Through systematic exploration of physics–data weighting, two PINN-style configurations with physics weighting of 0.70 and 0.75 were identified as optimal, based solely on interpolation performance. These models subsequently outperformed both the SMDE and NNN baselines in extrapolation, demonstrating stable, conservative predictions beyond the training range. It should be noted, however, that unlike classical PINNs that embed partial differential equation (PDE) or ordinary differential equation (ODE) residuals via automatic differentiation, the present framework employs the semi-empirical SMDE creep damage formulation as a constitutive physics-based model. We therefore describe it as a PINN-style, physics-regularised neural network, which balances empirical fidelity with mechanistic regularisation. The novelty of this work lies in applying such a PINN-style dual-loss framework to creep rupture life prediction for the first time, integrating a mechanistic creep damage model directly into neural network training and demonstrating improved extrapolation capability when only short-term data are available. These findings highlight the value of the PINN-style framework in enhancing model generalisation when only limited experimental data are available, particularly in contexts where physical mechanisms are well understood.
{"title":"On the application of PINN-style physics-regularised neural networks to high-temperature creep rupture life prediction","authors":"O. Muránsky , M.N. Tran , W. Payten","doi":"10.1016/j.ijpvp.2025.105691","DOIUrl":"10.1016/j.ijpvp.2025.105691","url":null,"abstract":"<div><div>This study explores the potential of PINN-style physics-regularised neural networks to improve long-term creep life predictions for 2.25Cr-1Mo (Grade 22) steel using only short-term experimental data. Three modelling approaches were evaluated: a purely physics-based semi-empirical model based on the Stress-Modified Ductility Exhaustion (SMDE) formulation, a purely data-driven neural network (NNN), and PINN-style models combining empirical learning with physics-based regularisation via a dual-loss approach. Model performance was assessed in both interpolation (within the short-term training domain) and extrapolation (on unseen long-term data). The SMDE model served as a physics-based baseline, exhibiting stable interpolation and extrapolation behaviour. In contrast, the NNN model overfitted the short-term data and failed to generalise to long-term conditions. Through systematic exploration of physics–data weighting, two PINN-style configurations with physics weighting of 0.70 and 0.75 were identified as optimal, based solely on interpolation performance. These models subsequently outperformed both the SMDE and NNN baselines in extrapolation, demonstrating stable, conservative predictions beyond the training range. It should be noted, however, that unlike classical PINNs that embed partial differential equation (PDE) or ordinary differential equation (ODE) residuals via automatic differentiation, the present framework employs the semi-empirical SMDE creep damage formulation as a constitutive physics-based model. We therefore describe it as a PINN-style, physics-regularised neural network, which balances empirical fidelity with mechanistic regularisation. The novelty of this work lies in applying such a PINN-style dual-loss framework to creep rupture life prediction for the first time, integrating a mechanistic creep damage model directly into neural network training and demonstrating improved extrapolation capability when only short-term data are available. These findings highlight the value of the PINN-style framework in enhancing model generalisation when only limited experimental data are available, particularly in contexts where physical mechanisms are well understood.</div></div>","PeriodicalId":54946,"journal":{"name":"International Journal of Pressure Vessels and Piping","volume":"219 ","pages":"Article 105691"},"PeriodicalIF":3.5,"publicationDate":"2025-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145416696","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This article mainly focuses on the hydrogen induced fracture behavior of austenitic stainless steel AISI 304. Experiments have shown that plastic deformation induces α′ martensite after entering the elastoplastic stage, which significantly affects hydrogen embrittlement sensitivity to varying degrees through the stress states caused by the specimen's geometries. The results indicate that stress concentration accelerates hydrogen diffusion, and the overall failure mode is more prone to hydrogen induced fracture. The lower the notch constraint, the greater the hydrostatic stress gradient, and the shear component increases the local hydrogen concentration by increasing the plastic strain. The hydrogen embrittlement sensitivity (IHE) increases firstly and then decreases with notch radius by considering both stress concentration and failure mode for normal notched specimen. The IHE of V-notched specimen is almost independent of the inclination angle, while the IHE of U-notched specimen is mainly affected by the failure mode by considering all the factors including stress concentration, failure mode, constraint and shear stress.
{"title":"Research on the factors of hydrogen induced fracture behavior of AISI 304: loading level, stress concentration and stress state","authors":"Xinting Miao , Hao Xin , Jinbo Zhang , Ping Tao , Jian Peng","doi":"10.1016/j.ijpvp.2025.105687","DOIUrl":"10.1016/j.ijpvp.2025.105687","url":null,"abstract":"<div><div>This article mainly focuses on the hydrogen induced fracture behavior of austenitic stainless steel AISI 304. Experiments have shown that plastic deformation induces α′ martensite after entering the elastoplastic stage, which significantly affects hydrogen embrittlement sensitivity to varying degrees through the stress states caused by the specimen's geometries. The results indicate that stress concentration accelerates hydrogen diffusion, and the overall failure mode is more prone to hydrogen induced fracture. The lower the notch constraint, the greater the hydrostatic stress gradient, and the shear component increases the local hydrogen concentration by increasing the plastic strain. The hydrogen embrittlement sensitivity (<em>I</em><sub>HE</sub>) increases firstly and then decreases with notch radius by considering both stress concentration and failure mode for normal notched specimen. The <em>I</em><sub>HE</sub> of V-notched specimen is almost independent of the inclination angle, while the <em>I</em><sub>HE</sub> of U-notched specimen is mainly affected by the failure mode by considering all the factors including stress concentration, failure mode, constraint and shear stress.</div></div>","PeriodicalId":54946,"journal":{"name":"International Journal of Pressure Vessels and Piping","volume":"219 ","pages":"Article 105687"},"PeriodicalIF":3.5,"publicationDate":"2025-10-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145362497","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-22DOI: 10.1016/j.ijpvp.2025.105690
Muhammad Bilal Jan, Mengyu Chai
Creep deformation is a critical concern in high-temperature materials and structural components operating under prolonged stress, such as those used in pressure vessels. Accurate prediction of creep rupture life is essential for ensuring the safety and longevity of components in aerospace, power generation, and other high-temperature industries. Traditionally, creep life has been predicted by widely-used empirical models such as the Larson–Miller parameter, the Monkman–Grant relationship, and the θ projection method. These conventional empirical methods suffer from inherent limitations, including over-simplified assumptions, poor generalizability, an inability to capture complex, non-linear relationships, and restricted use of input features, which limits their applicability to modern material systems. To address these limitations, Machine learning (ML), being data-driven, can simultaneously handle many input features, learn nonlinear and complex interactions, capture dependencies among multiple features, and exhibit robust generalizability. To support this capability, this review outlines a comprehensive ML workflow—from data acquisition to deployment—by summarizing existing studies in the literature, tailored for creep life prediction to guide future efforts in data-driven creep modeling. To provide clarity and depth, the literature is systematically reviewed and grouped into four key themes: hybrid modeling approaches, creep mechanism-aware models, feature selection techniques for enhanced creep life, and alloy design strategies for improved creep resistance. Finally, this review identifies key challenges such as data scarcity, a lack of physics integration and constraints in ML models, unavailability of a real-time in-situ creep life prediction framework, and difficulties in interpretability and explainability, and proposes potential future directions for these challenges, informed by both existing literature and original analytical insights. The field ML-based creep rupture life prediction lacks a review that synthesizes the available literature, outlines the limitations of existing approaches, and identifies future research directions. This review paper addresses this need by providing a clear overview to guide researchers and support further progress in the field.
{"title":"Machine learning approaches for creep rupture life prediction of metallic materials: A comprehensive review","authors":"Muhammad Bilal Jan, Mengyu Chai","doi":"10.1016/j.ijpvp.2025.105690","DOIUrl":"10.1016/j.ijpvp.2025.105690","url":null,"abstract":"<div><div>Creep deformation is a critical concern in high-temperature materials and structural components operating under prolonged stress, such as those used in pressure vessels. Accurate prediction of creep rupture life is essential for ensuring the safety and longevity of components in aerospace, power generation, and other high-temperature industries. Traditionally, creep life has been predicted by widely-used empirical models such as the Larson–Miller parameter, the Monkman–Grant relationship, and the θ projection method. These conventional empirical methods suffer from inherent limitations, including over-simplified assumptions, poor generalizability, an inability to capture complex, non-linear relationships, and restricted use of input features, which limits their applicability to modern material systems. To address these limitations, Machine learning (ML), being data-driven, can simultaneously handle many input features, learn nonlinear and complex interactions, capture dependencies among multiple features, and exhibit robust generalizability. To support this capability, this review outlines a comprehensive ML workflow—from data acquisition to deployment—by summarizing existing studies in the literature, tailored for creep life prediction to guide future efforts in data-driven creep modeling. To provide clarity and depth, the literature is systematically reviewed and grouped into four key themes: hybrid modeling approaches, creep mechanism-aware models, feature selection techniques for enhanced creep life, and alloy design strategies for improved creep resistance. Finally, this review identifies key challenges such as data scarcity, a lack of physics integration and constraints in ML models, unavailability of a real-time in-situ creep life prediction framework, and difficulties in interpretability and explainability, and proposes potential future directions for these challenges, informed by both existing literature and original analytical insights. The field ML-based creep rupture life prediction lacks a review that synthesizes the available literature, outlines the limitations of existing approaches, and identifies future research directions. This review paper addresses this need by providing a clear overview to guide researchers and support further progress in the field.</div></div>","PeriodicalId":54946,"journal":{"name":"International Journal of Pressure Vessels and Piping","volume":"219 ","pages":"Article 105690"},"PeriodicalIF":3.5,"publicationDate":"2025-10-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145362500","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-16DOI: 10.1016/j.ijpvp.2025.105682
Gang Li , Yichao Zhu
The study of the stress corrosion cracking (SCC) behaviour of pipeline steel is of great significance for the safe operation in the oil and gas industry. However, current experimental studies, being costly in both economic and temporal terms, can only deliver data suggesting the instantaneous SCC behaviour of pipeline steel, while quantities of actual interest, such as the lifespan against SCC, cannot be measured directly. To address this issue, a semi-analytical model based on partial differential equations is developed to model the SCC kinetics for steels making oil and gas pipeline. With the effect of stress gradient on ion transportation near crack tips taken into account, the mechanism of repeated rupture of the oxide film can be mimicked. With only one parameter needing calibration, the model proposed in this study is shown to make predictions, within a few seconds on a laptop computer, over SCC indices that are difficult to experimentally measure, such as the crack incubation period under various mechanical and chemical environments. It is predicted by the model that for a 56 mm-thick API 5L X70 steel segment with a 2 mm surface scratch, it takes roughly 90 years for the scratch to become an active crack under a tensile load of 120 MPa and with an environmental pH value of 6.8 and a chloride ion concentration of 0.004 mol/L, and it takes another 30 years for SCC evolution before the final material failure.
{"title":"Full-life simulation of the stress corrosion cracking behaviour of the pipeline steel for oil and gas","authors":"Gang Li , Yichao Zhu","doi":"10.1016/j.ijpvp.2025.105682","DOIUrl":"10.1016/j.ijpvp.2025.105682","url":null,"abstract":"<div><div>The study of the stress corrosion cracking (SCC) behaviour of pipeline steel is of great significance for the safe operation in the oil and gas industry. However, current experimental studies, being costly in both economic and temporal terms, can only deliver data suggesting the instantaneous SCC behaviour of pipeline steel, while quantities of actual interest, such as the lifespan against SCC, cannot be measured directly. To address this issue, a semi-analytical model based on partial differential equations is developed to model the SCC kinetics for steels making oil and gas pipeline. With the effect of stress gradient on ion transportation near crack tips taken into account, the mechanism of repeated rupture of the oxide film can be mimicked. With only one parameter needing calibration, the model proposed in this study is shown to make predictions, within a few seconds on a laptop computer, over SCC indices that are difficult to experimentally measure, such as the crack incubation period under various mechanical and chemical environments. It is predicted by the model that for a 56 mm-thick API 5L X70 steel segment with a 2 mm surface scratch, it takes roughly 90 years for the scratch to become an active crack under a tensile load of 120 MPa and with an environmental pH value of 6.8 and a chloride ion concentration of 0.004 mol/L, and it takes another 30 years for SCC evolution before the final material failure.</div></div>","PeriodicalId":54946,"journal":{"name":"International Journal of Pressure Vessels and Piping","volume":"219 ","pages":"Article 105682"},"PeriodicalIF":3.5,"publicationDate":"2025-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145362499","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号