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

doi:10.11902/1005.4537.2024.242

中国腐蚀与防护学报

2025, Vol. 45

Issue (2): 523-532 CSTR: 32134.14.1005.4537.2024.242 DOI: 10.11902/1005.4537.2024.242

研究报告

本期目录 | 过刊浏览

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

郑文培¹^,²^,³, 刘迎正¹^,²^,³(

), 张娅茹¹^,²^,³, 周涛涛¹^,²^,³, 李兴涛⁴, 禹胜阳⁴, 蒋璐朦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

An Integrated Corrosion Rate Prediction Model Based on Global and Local Error Fusion: A Robust Optimization Strategy

ZHENG Wenpei¹^,²^,³, LIU Yingzheng¹^,²^,³(

), ZHANG Yaru¹^,²^,³, ZHOU Taotao¹^,²^,³, LI Xingtao⁴, YU Shengyang⁴, JIANG Lumeng⁴

^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, Beijing 102100, China

引用本文:

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.
Wenpei ZHENG, Yingzheng LIU, Yaru ZHANG, Taotao ZHOU, Xingtao LI, Shengyang YU, Lumeng JIANG. An Integrated Corrosion Rate Prediction Model Based on Global and Local Error Fusion: A Robust Optimization Strategy[J]. Journal of Chinese Society for Corrosion and protection, 2025, 45(2): 523-532.

全文: PDF(1917 KB) HTML

摘要:

油气管道腐蚀速率的准确预测对维护管道安全至关重要，腐蚀速率预测模型的多样性和选择难度给实际应用带来了挑战。本文旨在改进传统的线性集成策略，首先综合全局与局部误差，提出了综合误差评价指标，并据此发展了一种新的线性集成策略，经5个标准测试函数的严格测试，证实了其广泛的适用性和优越性；将此策略应用于管道腐蚀速率预测，融合Fick定律(Fick's law algorithm，FLA)优化的BP神经网络(Backpropagation neural network)、克里金(Kriging)代理模型和极限学习机(Extreme learning machine，ELM)模型，成功构建了一个高效的集成预测模型。结果表明，与单一模型和传统集成模型相比，新策略在预测精度和稳定性上表现更优，当局部误差调节因子设定为0.5时，集成模型达到最佳性能；3个模型集成之后对比单个模型预测效果良好，预测效果有显著提升。该研究对于管道系统的可靠性评估和维护决策具有重要的工程应用价值。

关键词 ：综合误差, 线性集成, 管道腐蚀, 速率预测

Abstract：

Accurate prediction of the corrosion rate of oil and gas pipelines is crucial for maintaining pipeline safety, and the diversity and difficulty of choosing corrosion rate prediction models have brought challenges to practical applications. Herein, to improve the traditional linear integration strategy, firstly the global error and local error are synthesized and a comprehensive error evaluation index is postulated, then a novel linear integration strategy is established accordingly, its wide applicability and superiority are confirmed by rigorous testing with five standard test functions. By applying this strategy to the prediction of pipeline corrosion rate, an efficient integrated prediction model was successfully constructed via combining the FLA-optimized BP neural network, Kriging and ELM model. The results show that the new strategy performs better in terms of prediction accuracy and stability compared with the single model and the traditional integrated model, while the integrated model reaches the best performance when the local error adjustment factor is set to 0.5; In case the combination of the three models is adopted for prediction, it can produce better results rather than any single model, therefore the prediction effect is significantly improved. This study has important engineering application value for the reliability assessment and maintenance decision of pipeline systems.

Key words： aggregate error linear integrated corrosion of pipeline rate prediction

收稿日期: 2024-08-02 32134.14.1005.4537.2024.242

ZTFLH:

TE985

基金资助: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：“一带一路”海外长输管道完整性关键技术研究与应用(ZLZX2020-05);国家自然科学基金青年科学基金项目(72301294);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: 郑文培，男，1981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
	郑文培
	刘迎正
	张娅茹
	周涛涛
	李兴涛
	禹胜阳
	蒋璐朦
44.8K

链接本文:

https://www.jcscp.org/CN/10.11902/1005.4537.2024.242 或 https://www.jcscp.org/CN/Y2025/V45/I2/523

Sample	Displayed formula
1	y = x₁exp (-(x $12$ + x $22$ )), x_i ∈[-2, 2]
2	y = 2x $12$ - 1.05x $14$ + $x 16 6$ + x₁x₂ + x $22$ , x_i ∈[-5, 5]
3	y = $x 2 - 5.1 x 12 4 π 2 + 5 x 1 π - 6 2$ + 10 $1 - 1 8 π$ cosx₁ + 10, x₁∈[5, 10], x₂∈[0, 15]
4	y = 10sin[2(x₁ - 0.6 $π$ )] + x₂ + x₃ + x₄ + x₁x₂ + x₃x₄ + x $13$ + x $43$ , x_i ∈[0, 1]
5	y = $∑ i = 1 7 310 + s i n 1615 x i - 1 + s i n 2 1615 x i - 1$ , x_i ∈[-1, 1]

表1 基准功能

表2 集成模型预测的比较

图1 集成模型方差对比图

图2 集成模型平均相对误差对比图

Sample	X1	X2	X3	X4	X5	X6	X7	X8	V_corr
1	109.03	6.5100	0.0717	1.1824	35.887	0.7636	0.9033	0.1050	0.1898
2	115.12	6.3578	0.0370	1.6696	37.932	0.7561	0.8283	0.1240	0.1497
3	115.05	6.2169	0.0838	3.2597	37.306	0.7701	0.8673	0.1102	0.1525
4	110.56	6.3860	0.1065	1.4988	36.520	0.7601	0.8581	0.1127	0.1746
5	110.66	6.5301	0.0535	2.5883	37.896	0.7622	0.8601	0.1232	0.1641
6	114.09	6.5905	0.0565	4.8084	37.097	0.7712	0.8809	0.1149	0.1528
$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$
150	113.18	6.3849	0.0608	4.3186	37.839	0.7645	0.8497	0.1150	0.1485

表3 管道腐蚀监测部分数据

Sample	X1	X2	X3	X4	X5	X6	X7	X8	V_corr
1	0.9621	0.8157	-0.5651	-0.2026	0.1234	0.6291	-0.1607	0.1975	1.0000
2	0.6001	0.1622	0.3896	-0.4472	0.1574	0.0654	0.9870	0.0096	0.5044
3	0.8462	0.3619	-0.0432	0.0965	-0.1003	0.5975	0.2789	-0.0343	0.5412
4	-0.5925	0.4254	-0.2301	-0.4480	0.2983	-0.8951	-0.3238	0.1194	0.8190
5	0.3094	-0.1121	-0.3212	-0.0539	0.3443	-0.6101	-0.0699	-0.6429	0.6901
6	-0.1407	-0.2053	-0.3387	-0.1544	0.0101	-0.9614	0.2104	-0.4095	0.6533
$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$
150	0.2237	0.4388	-0.5234	0.0295	0.0112	0.2865	-0.2689	0.6539	0.5467

表4 标准化后部分数据

图3 模型流程图

图4 3种模型及其集合误差对比

表5 3种模型以及它们的集合预测效果对比

表6 不同集成模型预测效果对比

1	Wang Z, Lu J W, Zhou L L, et al. Corrosion analysis and protective measures of pressure piping of atmospheric and vacuum unit [J]. Sichuan Chem. Ind., 2022, 25(1): 26
1	王峥, 卢俊文, 周璐璐等. 常减压装置压力管道腐蚀分析及防护措施研究 [J]. 四川化工, 2022, 25(1): 26
2	Wang F, Li J, Li J L, et al. Corrosion behavior of metallic pipes in CO₂/H₂S environment [J]. Hot Work. Technol., 2021, 50(4): 1
2	王帆, 李娟, 李金灵等. 金属管道在CO₂/H₂S环境中的腐蚀行为 [J]. 热加工工艺, 2021, 50(4): 1
3	Yang T, Xu L, Wang J C, et al. Research progress on CO₂ corrosion and protection of tubing and casing [J]. J. Chin. Soc. Corros. Prot., 2024, 44(5): 1134
3	杨涛, 许磊, 王建春等. 油套管CO₂腐蚀和防护研究进展 [J]. 中国腐蚀与防护学报, 2024, 44(5): 1134
4	Vanaei H R, Eslami A, Egbewande A. A review on pipeline corrosion, in-line inspection (ILI), and corrosion growth rate models [J]. Int. J. Pressure Vessels Piping, 2017, 149: 43
5	Norsok M. CO₂corrosion rate calculation model. Majorstural [S]. Norway: Norwegian Technological Standards Institute Oscarsgt, 2005: 20
6	De Waard C, Milliams D E. Carbonic acid corrosion of steel [J]. Corrosion, 1975, 31: 177
7	Nesic S, Postlethwaite J, Olsen S. An electrochemical model for prediction of corrosion of mild Steel in aqueous carbon dioxide solutions [J]. Corrosion, 1996, 52: 280
8	Zhao G X, Liu R R, Ding L Y, et al. Effect of temperature on CO₂ corrosion behavior of 5Cr steel in simulated high temperature and high pressure environment of oilfield [J]. J. Chin. Soc. Corros. Prot., 2024, 44(1): 175
8	赵国仙, 刘冉冉, 丁浪勇等. 温度对5Cr钢在模拟油田高温高压环境中CO₂腐蚀行为的影响 [J]. 中国腐蚀与防护学报, 2024, 44(1): 175
9	Li X R. Research on residual strength of corroded pipeline based on machine learning [D]. Hangzhou: Hangzhou Dianzi University, 2024
9	李绪尧. 基于机器学习的腐蚀管道剩余强度预测 [D]. 杭州: 杭州电子科技大学, 2024
10	Xiao R G, Wang D, Wang Q X. Prediction of corrosion rate of submarine oil and gas pipelines based on ASO-BP neural network [J]. Chem. Ind. Eng., 2022, 39(6): 109
10	肖荣鸽, 王栋, 王勤学. 基于ASO-BP神经网络的海底油气管道腐蚀速率预测 [J]. 化学工业与工程, 2022, 39(6): 109
11	Xiao R G, Jin S S, Zhuang Q, et al. Prediction of corrosion rate of oil and gas pipeline based on grey theory [J]. Hot Work. Technol., 2022, 51(18): 53
11	肖荣鸽, 靳帅帅, 庄琦等. 基于灰色理论的油气管道腐蚀速率预测 [J]. 热加工工艺, 2022, 51(18): 53
12	Chen Q, Huang W, Zhang C H, et al. Physics-guided prediction of corrosion rate inside gathering and transportation pipelines and explanatory analysis [J/OL]. J. Chin. Soc. Corros. Prot., 1-15[2024-08-02].
12	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/OL]. 中国腐蚀与防护学报, 1-15[2024-08-02].
13	Guang Y. Research on remaining life prediction method of external corrosion of buried pipeline based on deep learning [D]. Chongqing: Chongqing University of Science and Technology, 2023
13	光宇. 基于深度学习的埋地管道外腐蚀剩余寿命预测方法研究 [D]. 重庆: 重庆科技学院, 2023
14	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
14	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
15	Zhang X S, Chang Y G. Prediction of external corrosion rate of offshore oil and gas pipelines based on FA-BAS-ELM [J]. China Saf. Sci. J., 2022, 32(2): 99
15	张新生, 常潆戈. 基于FA-BAS-ELM的海洋油气管道外腐蚀速率预测 [J]. 中国安全科学学报, 2022, 32(2): 99 doi: 10.16265/j.cnki.issn1003-3033.2022.02.014
16	Zheng D K, Cheng Y P, Li H R, et al. Application of IAFSA-GRNN in CO₂ corrosion rate prediction of oil gathering and transportation pipelines [J]. China Saf. Sci. J., 2022, 32(1): 110
16	郑度奎, 程远鹏, 李昊燃等. IAFSA-GRNN在油田集输管道CO₂腐蚀速率预测中的应用 [J]. 中国安全科学学报, 2022, 32(1): 110 doi: 10.16265/j.cnki.issn1003-3033.2022.01.015
17	Lee L H, Rajkumar R, Lo L H, et al. Oil and gas pipeline failure prediction system using long range ultrasonic transducers and Euclidean-Support Vector Machines classification approach [J]. Expert Syst. Appl., 2013, 40: 1925
18	Bastian B T, Jaspreeth N, Ranjith S, et al. Visual inspection and characterization of external corrosion in pipelines using deep neural network [J]. NDT E Int., 2019, 107: 102134
19	Zhang J B. Research on prediction method of casing damage based on ensembling learning [D]. Daqing: Northeast Petroleum University, 2022
19	张珺博. 基于集成学习的套损井预测方法研究 [D]. 大庆: 东北石油大学, 2022
20	Zhang X, Men J J, Rong Q, et al. Research on prediction model of flexural bearing capacity of corroded RC beams based on ensemble learning [J/OL]. Ind. Constr., 1-13[2024-08-02].
20	张雪, 门进杰, 荣强等. 基于集成学习的锈蚀RC梁抗弯承载力预测模型研究 [J/OL]. 工业建筑, 1-13[2024-08-02].
21	Cai B Q. Research on corrosion rate prediction of marine pipeline based on ensemble learning [D]. Xi'an: Xi'an University of Architecture and Technology, 2021
21	蔡宝泉. 基于集成学习的海洋管道腐蚀速率预测研究 [D]. 西安: 西安建筑科技大学, 2021
22	Hu Y, Qu B Y, Liang J, et al. A survey on evolutionary ensemble learning algorithm [J]. Chin. J. Intell. Sci. Technol., 2021, 3: 18
22	胡毅, 瞿博阳, 梁静等. 进化集成学习算法综述 [J]. 智能科学与技术学报, 2021, 3: 18
23	Chen J, Zhao X Y, Cai L L, et al. Application of BP neural network in internal corrosion direct assessment methodology for pipelines [J]. Total Corros. Control, 2023, 37(6): 97
23	陈君, 赵晓云, 蔡乐乐等. BP神经网络在管道内腐蚀直接评价中的应用 [J]. 全面腐蚀控制, 2023, 37(6): 97
24	Zhang Y R. Research on corrosion rate prediction model of overseas oil and gas pipelines [D]. Beijing: China University of Petroleum (Beijing), 2023
24	张娅茹. 海外油气管道腐蚀速率预测模型研究 [D]. 北京: 中国石油大学(北京), 2023
25	Lv L L, Wang J, Qi Q F, et al. Corrosion rate prediction model of oil-gas mixed transportation pipelines based on KPCA-IGOA-ELM [J]. Oil Gas Storage Transp., 2023, 42: 785
25	吕林林, 王杰, 祁庆芳等. 基于KPCA-IGOA-ELM的油气混输管道腐蚀速率预测模型 [J]. 油气储运, 2023, 42: 785
26	Zhou Y X, Peng X Y, Geng Y H. Research on internal corrosion prediction model of shale gas gathering pipeline based on LHS-Kriging-DW [J]. Hot Work. Technol., 2024, 53(16): 113
26	周逸轩, 彭星煜, 耿月华. 基于LHS-Kriging-DW的页岩气集输管道内腐蚀预测模型研究 [J]. 热加工工艺, 2024, 53(16): 113
27	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
27	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
28	Gao J, Cui H B, Fan T, et al. A structural reliability calculation method based on adaptive Kriging ensemble model [J]. China Mech. Eng., 2024, 35: 83
28	高进, 崔海冰, 樊涛等. 一种基于自适应Kriging集成模型的结构可靠性分析方法 [J]. 中国机械工程, 2024, 35: 83
29	Li L. Study on prediction of CO₂ internal corrosion rate of submarine multiphase flow pipeline [D]. Xi'an: Xi'an University of Architecture and Technology, 2023
29	李蕾. 海底多相流管道CO₂内腐蚀速率预测研究 [D]. 西安: 西安建筑科技大学, 2023
30	Ma C Q, Wang Q X. BP neural network research based on local and global error [J]. J. Wuhan Univ. Technol., 2009, 31(20): 99
30	马成前, 王庆喜. 基于局部及全局误差的BP神经网络研究 [J]. 武汉理工大学学报, 2009, 31(20): 99
31	Guan E D. Prediction model for internal corrosion rate of multiphase flow gathering pipeline based on IGSA-RFR [J]. Oil Gas Storage Transp., 2022, 41: 1448
31	管恩东. 基于IGSA-RFR的多相流集输管道内腐蚀速率预测模型 [J]. 油气储运, 2022, 41: 1448
32	Zhou Y, Wang S X. Prediction of internal corrosion rate of gas field gathering pipelines based on GRA-IFA-LSSVM model [J]. Corros. Prot., 2022, 43(8): 86
32	周阳, 王寿喜. 基于GRA-IFA-LSSVM模型的气田集输管道内腐蚀速率预测 [J]. 腐蚀与防护, 2022, 43(8): 86
33	He T L, Li H R, Cheng Y P, et al. Prediction of pipeline corrosion rate based on new GM(1,N) model [J]. Corros. Prot., 2021, 42(10): 79
33	何天隆, 李昊燃, 程远鹏等. 基于新型GM(1,N)模型的油气管道腐蚀速率预测 [J]. 腐蚀与防护, 2021, 42(10): 79
34	Teng Q. Research on human motion recognition based on local error model and convolutional neural network [D]. Nanjing: Nanjing Normal University, 2021
34	滕起. 基于局部误差模型和卷积神经网络的人体运动识别研究 [D]. 南京: 南京师范大学, 2021
35	Krige D G. A statistical approach to some basic mine valuation problems on the Witwatersrand [J]. J. South. Afr. Inst. Min. Metall., 1951, 52: 119
36	Zhang L G. Research on local and moment-independent sensitivity analysis for structures with uncertainty [D]. Xi'an: Northwestern Polytechnical University, 2015: 98
36	张磊刚. 不确定性结构的局部和矩独立灵敏度方法研究 [D]. 西安: 西北工业大学, 2015: 98
37	Yu M G, Pan Z K, Jiang R C, et al. Multi-objective optimization design of the high-speed train head based on the approximate model [J]. J. Mech. Eng., 2019, 55(24): 178 doi: 10.3901/JME.2019.24.178
37	于梦阁, 潘振宽, 蒋荣超等. 基于近似模型的高速列车头型多目标优化设计 [J]. 机械工程学报, 2019, 55(24): 178 doi: 10.3901/JME.2019.24.178
38	Safaee A M. Performance of the multi-tiered GMSE walls under seismic conditions: comparison of physical and numerical simulations [J]. Soil Dyn. Earthquake Eng., 2022, 159: 107316
39	Ma M T, Zhao Z. Prediction of corrosion rate of process pipelines based on KPCA-CSO-RVM model [J]. Saf. Environ. Eng., 2021, 28(4): 1
39	马梦桐, 赵琢. 基于KPCA-CSO-RVM模型的工艺管道腐蚀速率预测 [J]. 安全与环境工程, 2021, 28(4): 1
40	Zhe N, Yang J F, Liu W B, et al. Prediction of corrosion rate of process pipeline based on KPCA and SVM [J]. Corros. Prot., 2019, 40(1): 56
40	者娜, 杨剑锋, 刘文彬等. 基于KPCA和SVM的工艺管道腐蚀速率预测 [J]. 腐蚀与防护, 2019, 40(1): 56

[1]	贾世超, 高佳祺, 郭浩, 王超, 陈杨杨, 李旗, 田一梅. 再生水水质因素对铸铁管道的腐蚀研究[J]. 中国腐蚀与防护学报, 2020, 40(6): 569-576.
[2]	张新生,曹乃宁,李亚云. 基于Gumbel极值I型分布埋地油气管道的剩余寿命预测[J]. 中国腐蚀与防护学报, 2016, 36(4): 370-374.

Viewed

Full text

Abstract

Cited

Shared

Discussed