基于物理信息神经网络的油气管道内腐蚀预测方法

doi:10.11902/1005.4537.2024.363

中国腐蚀与防护学报

2025, Vol. 45

Issue (5): 1320-1330 CSTR: 32134.14.1005.4537.2024.363 DOI: 10.11902/1005.4537.2024.363

研究报告

本期目录 | 过刊浏览

基于物理信息神经网络的油气管道内腐蚀预测方法

周涛涛¹^,²^,³, 刘迎正¹^,²^,³(

), 郑文培¹^,²^,³, 姜恒良⁴, 刘海鹏⁴, 夏刚⁴

1 中国石油大学(北京)安全与海洋工程学院北京 102249
2 油气生产安全与应急技术应急管理部重点实验室北京 102249
3 国家市场监督管理总局重点实验室(油气生产装备质量检测与健康诊断) 北京 102249
4 中国石油国际勘探开发有限公司北京 102100

Pipeline Corrosion Prediction Method Based on Physics-informed Neural Networks

ZHOU Taotao¹^,²^,³, LIU Yingzheng¹^,²^,³(

), ZHENG Wenpei¹^,²^,³, JIANG Hengliang⁴, LIU Haipeng⁴, XIA Gang⁴

1 College of Safety and Ocean Engineering, China University of Petroleum (Beijing), Beijing 102249, China
2 Key Laboratory of Oil and Gas Safety and Emergency Technology, Ministry of Emergency Management, Beijing 102249, China
3 Key Laboratory of Oil and Gas Production Equipment Quality Inspection and Health Diagnosis, State Administration for Market Regulation, Beijing 102249, China
4 China National Oil and Gas Exploration and Development Corporation Limited, Beijing 102100, China

引用本文:

周涛涛, 刘迎正, 郑文培, 姜恒良, 刘海鹏, 夏刚. 基于物理信息神经网络的油气管道内腐蚀预测方法[J]. 中国腐蚀与防护学报, 2025, 45(5): 1320-1330.
Taotao ZHOU, Yingzheng LIU, Wenpei ZHENG, Hengliang JIANG, Haipeng LIU, Gang XIA. Pipeline Corrosion Prediction Method Based on Physics-informed Neural Networks[J]. Journal of Chinese Society for Corrosion and protection, 2025, 45(5): 1320-1330.

全文: PDF(3465 KB) HTML

摘要:

为了对管道内腐蚀预测进行准确分析，解决传统机器学习在预测管道腐蚀速率时的解释能力不足和泛化能力不足的问题。通过将温度、CO₂分压与腐蚀速率之间的物理特性嵌入到神经网络中，使得模型服从给定的机理约束，同时考虑结构损失，缓解了模型过拟合和欠拟合，建立了基于物理信息神经网络(PINN)管道内腐蚀预测模型。结果表明：PINN模型效果优于支持向量机(SVM)、极端梯度提升(XGBoost)、人工神经网络(ANN)等模型，严格遵循着相关变量的单调性关系，确保了预测结果的物理一致性，同时该模型也表现出卓越的泛化能力。

关键词 ：内腐蚀, 速率预测, 物理信息神经网络, 物理单调性

Abstract：

With the increasing oil and gas exploration, the number of service pipelines is growing, making the accurate prediction of internal corrosion crucial for pipeline integrity management. To address the limitations of traditional machine learning methods in interpreting and generalizing corrosion rate predictions, the monotonic relationships between temperature, CO₂ partial pressure, and corrosion rate are incorporated into a Physics-Informed Neural Network (PINN). This approach is designed to adhere to mechanistic constraints, avoid overfitting and underfitting, and ensure physical consistency. The PINN model is shown to outperform Support Vector Machines (SVM), Extreme Gradient Boosting (XGBoost), and Artificial Neural Networks (ANN), demonstrating superior accuracy and generalization.

Key words： internal corrosion rate prediction physics-informed neural networks monotonicity in physics

收稿日期: 2024-11-06 32134.14.1005.4537.2024.363

ZTFLH:

TE985

基金资助:国家自然科学基金(72301294);中国石油天然气集团有限公司-中国石油大学(北京)战略合作科技专项(ZLZX2020-05);中国石油大学(北京)科研启动基金(2462023BJRC016)

通讯作者: 刘迎正，E-mail：liuyingzheng2024@126.com，研究方向为油气管道腐蚀预测

Corresponding author: LIU Yingzheng, E-mail: liuyingzheng2024@126.com

作者简介: 周涛涛，男，1990年生，博士，副教授

	服务

	把本文推荐给朋友
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	周涛涛
	刘迎正
	郑文培
	姜恒良
	刘海鹏
	夏刚
44.8K

链接本文:

https://www.jcscp.org/CN/10.11902/1005.4537.2024.363 或 https://www.jcscp.org/CN/Y2025/V45/I5/1320

Labels	Variable	Variable name	Unit	Range	Average
X₁	T	Temperature	℃	41.7-69.4	58.5
X₂	P	System pressure	Pa	1.52 × 10⁶-3.23 × 10⁶	2.28 × 10⁶
X₃	$p C O 2$	CO₂ partial pressure	Pa	2.9 × 10⁴-3.8 × 10⁴	3.3 × 10⁴
X₄	pH	pH	-	4.2-6.2	5.8
X₅	Vf	Fluid flow rate	m·s^-1	0.37-0.69	0.53
X₆	Cl^-	Cl^- concentration	mg·L^-1	3420-8280	6227
X₇	CO₂ eq	CO₂ concentration	mg·L^-1	15.8-27.4	22.7
X₈	HCO $3 -$	HCO $3 -$ concentration	mg·L^-1	107-208	150
X₉	WC	Water content	%	52.5-63.9	58.6
Y	V_corr	Internal corrosion rate	mm·a^-1	2.289-3.105	2.681

表1 管道内腐蚀预测模型的相关变量

图1 基于PINN的内腐蚀预测模型框架

图2 不同模型预测结果比较

表2 各模型误差

图3 不同模型预测的管道在不同CO2分压下的腐蚀速率随温度的变化

图4 各个模型中预测点对温度和pCO2偏导数的分布图

图5 不同模型预测的管道在不同温度下的腐蚀速率随pCO2的变化

[1]	Wasim M, Djukic M B. External corrosion of oil and gas pipelines: A review of failure mechanisms and predictive preventions [J]. J. Nat. Gas Sci. Eng., 2022, 100: 104467
[2]	Shuai Y, Wang X H, Cheng Y F. Buckling resistance of an X80 steel pipeline at corrosion defect under bending moment [J]. J. Nat. Gas Sci. Eng., 2021, 93: 104016
[3]	El Amine Ben Seghier M, Keshtegar B, Tee K F, et al. Prediction of maximum pitting corrosion depth in oil and gas pipelines [J]. Eng. Fail. Anal., 2020, 112: 104505
[4]	Ding J L, Xi X, Zhang D H. Development and outlook of China's pipeline transportation technologies under energy security strategy [J]. Oil Gas Storage Transp., 2022, 41: 632
[4]	丁建林, 西昕, 张对红. 能源安全战略下中国管道输送技术发展与展望 [J]. 油气储运, 2022, 41: 632
[5]	Xiang Z H, Gao S H. Efficient management and safe operation of gas pipelines under the background of "Dual Carbon" [J]. Chem. Eng. Manag., 2024, (3): 123
[5]	相泽厚, 高胜寒. “双碳”背景下输气管道的高效管理与安全运行 [J]. 化工管理, 2024, (3): 123
[6]	Di C Y, Cui G S. Analysis of oil and gas pipeline corrosion causes and influencing factors [J]. Petrochem. Ind. Technol., 2024, 31(7): 69
[6]	邸春雨, 崔国森. 浅析油气管道腐蚀原因及影响因素 [J]. 石化技术, 2024, 31(7): 69
[7]	Wu F Y, Dai L B, Zhang X Z, et al. Analysis of chemical corrosion mechanism of oil and gas pipelines and its protective measures [J]. Contemp. Chem. Ind., 2024, 53: 1394
[7]	吴付洋, 戴立斌, 张学忠等. 油气管道化学腐蚀机理分析及其防护措施 [J]. 当代化工, 2024, 53: 1394
[8]	Zhu H H, Guo H L, Yan S D, et al. Study on bearing capacity of buried pipeline containing corrosion defects under external force [J]. J. Saf. Sci. Technol., 2018, 14(11): 70
[8]	朱豪豪, 郭海林, 闫生栋等. 含腐蚀缺陷埋地管道在外力作用下的承载力研究 [J]. 中国安全生产科学技术, 2018, 14(11): 70
[9]	Xie M J, Tian Z G. A review on pipeline integrity management utilizing in-line inspection data [J]. Eng. Fail. Anal., 2018, 92: 222
[10]	De Waard C, Milliams D E. Carbonic acid corrosion of steel [J]. Corrosion, 1975, 31(5): 177
[11]	De Waard C, Lotz U, Milliams D E. Predictive model for CO₂ corrosion engineering in wet natural gas pipelines [J]. Corrosion, 1991, 47: 976
[12]	De Waard C, Lotz U. Prediction of CO₂ corrosion of carbon steel [A]. Corrosion National Association of Corrosion [C]. Houston, Taxes, 1993
[13]	De Waard C, Lotz U, Dugstad A. Influence of liquid flow velocity on CO₂ corrosion: a semi-empirical model [R]. Houston: NACE International, 1995
[14]	De Waard C, Smith L, Bartlett P, et al. Modelling Corrosion Rates in Oil Production Tubing [M]. Eurocorr: Riva del Garda, 2001, 254
[15]	Nesic S, Xiao Y, Pots B F M. A quasi 2-D localized corrosion model [A]. NACE Corrosion [C]. New Orleans, Louisiana, 2004
[16]	Nesic S, Wang S H, Cai J Y, et al. Integrated CO₂ corrosion-multiphase flow model [A]. SPE International Oilfield Corrosion Conference and Exhibition [C]. Aberdeen, United Kingdom, 2004
[17]	Abdar P S, Hariri M B, Kahyarian A, et al. A revision of mechanistic modeling of mild steel corrosion in H₂S environments [J]. Electrochim. Acta, 2021, 382: 138231
[18]	Ma H N, Zhang W D, Wang Y, et al. Advances in corrosion growth modeling for oil and gas pipelines: a review [J]. Process Saf. Environ. Prot., 2023, 171: 71
[19]	Al-Fakih A M, Algamal Z Y, Lee M H, et al. Quantitative structure-activity relationship model for prediction study of corrosion inhibition efficiency using two-stage sparse multiple linear regression [J]. J. Chemom., 2016, 30: 361
[20]	Velázquez J C, Caleyo F, Valor A, et al. Predictive model for pitting corrosion in buried oil and gas pipelines [J]. Corrosion, 2009, 65: 332
[21]	Ossai C I, Boswell B, Davies I. Markov chain modelling for time evolution of internal pitting corrosion distribution of oil and gas pipelines [J]. Eng. Fail. Anal., 2016, 60: 209
[22]	Qin M, Liao K X, He G X, et al. Main control factors and prediction model of flow-accelerated CO₂/H₂S synergistic corrosion for X65 steel [J]. Process Saf. Environ. Prot., 2022, 160: 749
[23]	El-Abbasy M S, Senouci A, Zayed T, et al. Artificial neural network models for predicting condition of offshore oil and gas pipelines [J]. Automat. Construct., 2014, 45: 50
[24]	Zhou Y X, Peng X Y, Geng Y H, et al. Internal corrosion rate prediction of shale gas gathering pipeline based on KPCA-GA-BP model [J]. Corros. Prot., 2024, 45(4): 63
[24]	周逸轩, 彭星煜, 耿月华等. 基于KPCA-GA-BP模型的页岩气集输管道的内腐蚀速率预测 [J]. 腐蚀与防护, 2024, 45(4): 63
[25]	Liu J H, Tang H G, Fu J, et al. Pipeline corrosion rate prediction based on SSA-SVR model [J]. Hot Work. Technol., 2025, 54(4): 142
[25]	刘军衡, 唐海光, 付军等. 基于ISSA-SVR模型的管道腐蚀速率预测 [J]. 热加工工艺, 2025, 54(4): 142
[26]	Zhou C Y, Zhang Z, Wang Y, et al. Prediction on CO₂ corrosion rate of tubing based on neural oblivious decision ensemble [J]. J. Saf. Sci. Technol., 2023, 19(6): 127
[26]	周驰宇, 张智, 王杨等. 基于神经遗忘决策集成的油管CO₂腐蚀速率预测 [J]. 中国安全生产科学技术, 2023, 19(6): 127
[27]	Karpatne A, Atluri G, Faghmous J H, et al. Theory-guided data science: a new paradigm for scientific discovery from data [J]. IEEE Trans. Knowl. Data Eng., 2017, 29: 2318
[28]	Azodi C B, Tang J L, Shiu S H. Opening the black box: interpretable machine learning for geneticists [J]. Trends Genet., 2020, 36: 442 doi: S0168-9525(20)30069-X pmid: 32396837
[29]	Raissi M, Yazdani A, Karniadakis G E. Hidden fluid mechanics: learning velocity and pressure fields from flow visualizations [J]. Science, 2020, 367: 1026 doi: 10.1126/science.aaw4741 pmid: 32001523
[30]	Chen L E, Zeng X Y, Zeng Z H, et al. Temperature prediction of lithium-ion batteries based on physical information and deep neural network [J]. Electr. Power, 2024, 57(11): 18
[30]	陈来恩, 曾小勇, 曾子豪等. 基于物理信息与深度神经网络的锂离子电池温度预测 [J]. 中国电力, 2024, 57(11): 18
[31]	Karniadakis G E, Kevrekidis I G, Lu L, et al. Physics-informed machine learning [J]. Nat. Rev. Phys., 2021, 3: 422 doi: 10.1038/s42254-021-00314-5
[32]	Zhou T T, Jiang S, Han T, et al. A physically consistent framework for fatigue life prediction using probabilistic physics-informed neural network [J]. Int. J. Fatigue, 2023, 166: 107234
[33]	Ren S J, Zhu B Y, Weng Q H, et al. Forecasting method for NO _x emission in coal fired boiler based on physics-informed neural network [J]. Proceedings of the CSEE, 2024, 44: 8157
[33]	任少君, 朱保宇, 翁琪航等. 基于物理信息神经网络的燃煤锅炉NO _x 排放浓度预测方法 [J]. 中国电机工程学报, 2024, 44: 8157
[34]	Deng Z P, Ren S J, Weng Q H, et al. Prediction method for biomass gasification product distribution based on physics-informed neural network [J]. J. Chin. Soc. Power Eng., 2024, 44: 719
[34]	邓志平, 任少君, 翁琪航等. 基于物理信息神经网络的生物质气化产物分布预测方法 [J]. 动力工程学报, 2024, 44: 719 doi: 10.19805/j.cnki.jcspe.2024.240004
[35]	Chen Q, Huang W, Zhang C H, et al. Physics-guided prediction of corrosion rate inside gathering and transportation [J]. J. Chin. Soc. Corros. Prot., 2025, 45: 720
[35]	陈潜, 黄伟, 张昌会等. 基于物理引导的集输管道内腐蚀速率预测及可解释性分析 [J]. 中国腐蚀与防护学报, 2025, 45: 720
[36]	Liu G Y, Zhi S J, Liu X, et al. Research progress on internal corrosion of supercritical/dense-phase CO₂ pipelines for CCUS [J]. Oil Gas Storage Transp., 2024, 43: 510
[36]	刘广瑜, 支树洁, 柳歆等. CCUS超临界/密相CO₂管道内腐蚀研究进展 [J]. 油气储运, 2024, 43: 510
[37]	Hu L H, Yi H L, Yang W J, et al. Effect of water content on corrosion behavior of X65 pipeline steel in supercritical CO₂ fluids [J]. J. Chin. Soc. Corros. Prot., 2024, 44: 576
[37]	胡丽华, 衣华磊, 杨维健等. 水含量对超临界CO₂输送管道腐蚀的影响 [J]. 中国腐蚀与防护学报, 2024, 44: 576 doi: 10.11902/1005.4537.2023.227
[38]	Song Y Y. Research for internal corrosion rate prediction of submarine oil and gas pipelines in service[D]. Xi'an: Xi'an University of Architecture and Technology, 2020
[38]	宋莹莹. 在役海底油气管道内腐蚀速率预测研究 [D]. 西安: 西安建筑科技大学, 2020

[1]	陈潜, 黄伟, 张昌会, 管奥成, 张宸, 叶晓芃. 基于物理引导的集输管道内腐蚀速率预测及可解释性分析[J]. 中国腐蚀与防护学报, 2025, 45(3): 720-730.
[2]	郑文培, 刘迎正, 张娅茹, 周涛涛, 李兴涛, 禹胜阳, 蒋璐朦. 基于全局与局部误差融合的集成腐蚀速率预测模型：一种稳健优化策略[J]. 中国腐蚀与防护学报, 2025, 45(2): 523-532.
[3]	王文泉, 崔宇, 薛蕴鹏, 刘莉, 王福会. 潮湿空气中应力耦合固态NaCl作用下GH4169合金的中温腐蚀行为研究[J]. 中国腐蚀与防护学报, 2024, 44(6): 1399-1411.
[4]	张斐, 王海涛, 何勇君, 张天遂, 刘宏芳. 成品油输送管道微生物腐蚀案例分析[J]. 中国腐蚀与防护学报, 2021, 41(6): 795-803.

Viewed

Full text

Abstract

Cited

Shared

Discussed