基于PCA-BPNN模型的埋地管道腐蚀速率预测研究

兰州理工大学学报 ›› 2024, Vol. 50 ›› Issue (4): 60-68.

基于PCA-BPNN模型的埋地管道腐蚀速率预测研究

于扬¹, 孙东亮^*2

1.华东理工大学机械与动力工程学院, 上海 200237;
2.华东理工大学资源与环境工程学院, 上海 200237

收稿日期:2023-02-15 出版日期:2024-08-28 发布日期:2024-08-30
通讯作者: 孙东亮(1984-),男,黑龙江齐齐哈尔人.博士,副教授.Email:dongliangsun@126.com
基金资助:
城市安全风险评估系统产品开发项目(B100-41901)

Research on corrosion rate prediction of buried pipeline based on PCA-BPNN model

YU Yang¹, SUN Dong-liang²

1. School of Mechanical and Power Engineering, East China University of Science and Technology, Shanghai 200237, China;
2. School of Resources and Environmental Engineering, East China University of Science and Technology, Shanghai 200237, China

Received:2023-02-15 Online:2024-08-28 Published:2024-08-30

摘要/Abstract

摘要： 为了更加准确可靠地预测埋地管道的腐蚀速率,融合PCA分析法和多隐层BP人工神经网络模拟方法进行研究.选取陕西省某油气公司的埋地输油管道,构建8维度外腐蚀指标体系,在PCA-多隐层BPNN模型中模拟训练得到结果.通过PCA预处理将外腐蚀指标体系降为3维,以便减少多元素信息带来的耦合影响,模拟得到隐藏层参数最优的BPNN模型,预测腐蚀速率,求出预测值精确度,统计得到改进后方法精确度大于95%的个数是单一BP方法的2.5倍.为了检验PCA-多隐层BPNN方法的鲁棒性,另取20组数据代入验证,再次证实了PCA-多隐层BPNN模型所得的误差更小,更能满足实际工程需要.

关键词: 埋地管道, 腐蚀速率, PCA-多隐层BPNN模型

Abstract: In order to predict the corrosion rate of buried pipelines more accurately and reliably, the method integrating PCA analysis and BP artificial neural network simulation is studied. A buried oil pipeline of an oil and gas company in Shaanxi Province was selected to construct an 8-dimensional external corrosion index system, and the simulation training results were obtained in the PCA-multi-hidden layer BPNN model. The external corrosion index system was reduced to three dimensions by PCA pretreatment, so as to reduce the coupling influence brought by multi-element information. The BPNN model with the optimal hidden layer parameters was trained to predict the corrosion rate. The accuracy of the predicted value was obtained, and it was found that the number of accuracy greater than 95% of the improved method was 2.5 times that of the single BP method. In order to test the robustness of the PCA-multi-hidden layer BPNN method, an additional set of 20 groups of data are substituted for verification, reaffirming that the error of PCA-multi-hidden layer BPNN model is smaller and can better meet the needs of practical engineering.

Key words: buried pipes, corrosion rate, PCA-multiple hidden layer BPNN model

中图分类号:

TQ050.9

于扬, 孙东亮. 基于PCA-BPNN模型的埋地管道腐蚀速率预测研究[J]. 兰州理工大学学报, 2024, 50(4): 60-68.

YU Yang, SUN Dong-liang. Research on corrosion rate prediction of buried pipeline based on PCA-BPNN model[J]. Journal of Lanzhou University of Technology, 2024, 50(4): 60-68.

参考文献

[1] 杨立伟,张晓瑞.长输管道工艺站场埋地管线腐蚀原因及防护途径探讨 [J].中国石油和化工标准与质量,2022,42:38-39.
[2] 于力晗.第十届全国腐蚀大会摘要集 [C].南昌:中国腐蚀与防护学会,2019:39-41.
[3] WANG X M,TAN Y F,ZHANG T T.et al.Numerical study on the diffusion process of pinhole leakage of natural gas from underground pipelines to the soil [J].Journal of Natural Gas Science and Engineering,2021,87:103792.
[4] KIM J G,KIM Y W,KANG M C.Corrosion characteristics of rigid polyurethane thermally insulated pipeline with insulation defects [J].The Journal of Science and Engineering,2002,58(2):28-32.
[5] 陈健飞,王孟,江文军,等.基于灰色关联度的埋地管道检测数据分析 [J].表面技术,2017,46(4):264-269.
[6] 张杰.基于主成分-聚类分析法的管道风险评价方法 [J].油气储运,2014,33(2):139-143.
[7] SWATI J.Corrosion conference and expo [C].Dallas:NACE International,2015:4398-4412.
[8] 孟倩,马小平,周延.改进的粒子群支持向量机预测瓦斯涌出量 [J].矿业安全与环保,2015,42(2):1-5.
[9] 骆正山,姚梦月,骆济豪,等.基于KPCA-BAS-GRNN的埋地管道外腐蚀速率预测 [J].表面技术,2018,47(11):173-180.
[10] 邓志安,李姝仪,李晓坤,等.基于模糊神经网络的海洋管线腐蚀速率预测新方法 [J].中国腐蚀与防护学报,2015,35(6):571-576.
[11] 崔梦天,梁杰,龙松林,等.基于PFA-BP神经网络的软件缺陷预测模型的研究 [J].西南民族大学学报,2021,47(5):510-514.
[12] 王琬,张俊文,林铭,等.基于BP人工神经网络理论对埋地燃气管线腐蚀剩余寿命研究 [J].现代机械,2010,25(3):44-46.
[13] ROMANOFF M.Underground corrosion [M].Houston:National Association of Corrosion Engineers,1957.
[14] CALEYO F,VELAZQUEZ J C,VALOR A,et al.Probability distribution of pitting corrosion depth and rate in underground pipelines:a Monte Carlo study [J].Corrosion Science,2009,51(9):1925-1934.
[15] 林海明.对主成分分析法运用中十个问题的解析 [J].统计与决策,2007,28(16):16-18.
[16] 潘林生.长输管道工艺站场埋地管线腐蚀原因及防护对策探讨 [J].化工管理,2020(15):187-188.
[17] 张文跃.主成分分析法在地下水分析中的应用探究 [J].技术与市场,2016,23(5):234.
[18] 陈宁.基于主成分分析和BP神经网络的软件工作量估算方法 [J].信息与电脑,2022,34(3):80-83.
[19] FAN C H,HE Y T,ZHANG H X,et al.Predictive model based on genetic algorithm-neural network for fatigue performances of pre-corroded aluminum alloys [J].Key Engineering Materials,2007(353/354/355/356/357/358):1029-1032.
[20] WANG H,YAJIMA A,LIANG R Y,et al.Bayesian modeling of external corrosion in underground pipelines based on the integration of Markov Chain Monte Carlo techniques and clustered inspection data [J].Computer-aided civil and infrastructure engineering,2015,30(4):300-316.