PDF下载

[1]赵赏鑫,田野.基于PP-YOLOE与MTL的管道腐蚀缺陷识别量化方法[J].油气储运,2024,43(12):1355-1364.[doi:10.6047/j.issn.1000-8241.2024.12.004]
　ZHAO Shangxin,TIAN Ye.Identification and size quantification of pipeline corrosion defects based on PP-YOLOE and MTL[J].Oil & Gas Storage and Transportation,2024,43(12):1355-1364.[doi:10.6047/j.issn.1000-8241.2024.12.004]

点击复制

基于PP-YOLOE与MTL的管道腐蚀缺陷识别量化方法

《油气储运》[ISSN:1000-8241/CN:13-1093/TE] 卷: 43 期数: 2024年12期页码: 1355-1364 栏目: 检测与完整性出版日期: 2024-12-25

Title:: Identification and size quantification of pipeline corrosion defects based on PP-YOLOE and MTL

文章编号:: 1000-8241（2024）12-1355-10

作者:: 赵赏鑫; 田野; 国家管网集团西部管道有限责任公司 ? 新疆多介质管道安全输送重点实验室

Author(s):: ZHAO Shangxin; TIAN Ye; PipeChina West Pipeline Co. Ltd.//Xinjiang Key Laboratory for Transportation Safety of Multi-Medium Pipelines

关键词:: 管道; 目标检测; 深度学习; 漏磁内检测; 缺陷量化; 多任务

Keywords:: pipeline; target detection; deep learning; magnetic flux leakage technique for in-line inspection; size quantification of defects; multi-task

分类号:: TE832

DOI:: 10.6047/j.issn.1000-8241.2024.12.004

文献标志码:: A

摘要:: 【目的】管道漏磁内检测技术因无需耦合剂、易于实现自动化等优点已广泛应用于管道缺陷在线检测中。利用检测数据准确预测缺陷尺寸并开展适用性评价，对于后续的修复决策至关重要。【方法】以采集的漏磁数据为基础，根据缺陷数据特点，提出一种基于PP-YOLOE（Paddle Paddle-You Only Look Once Evolved）的深度学习目标检测模型，通过将三轴漏磁数据转换为彩色图像，输入模型中进行目标检测训练，实现缺陷的快速定位与数据提取，为缺陷量化提供准确可靠的数据集。结合多任务学习（Multi-Task Learning,MTL）模型，将管道缺陷处的周向、轴向、径向漏磁数据作为模型的输入，根据漏磁数据特征对缺陷的长度、宽度、深度进行并行输出，实现对缺陷的尺寸评估与预测。【结果】与实际牵拉实验数据对比后发现，通过PP-YOLOE与MTL相结合的识别量化模型提高了管道缺陷识别效率，改善了缺陷尺寸预测结果，缺陷目标检测的召回率达到0.87、准确率达到0.94；NADAM（Nesterov-accelerated Adaptive Moment）算法与SGD（Stochastic Gradient Descent）算法相比，预测腐蚀缺陷长度、深度的精度分别提升了9％、4％。【结论】该方法适用于海量管道漏磁检测数据的分析，提高了判读效率，可为管道缺陷的剩余强度评估和剩余寿命预测提供可靠的基础数据支持。（图10，表2，参30）

Abstract:: [Objective] The magnetic flux leakage technique for in-line inspection of pipelines is widely used for the online detection of pipeline defects, due to its advantages including being couplant-free and easy to automate. Subsequent accurate prediction of defect sizes and evaluation of applicability using inspection data are essential for making informed repair decisions. [Methods] A deep learning target detection model based on Paddle Paddle-You Only Look Once Evolved (PP-YOLOE) is proposed, developed from collected magnetic flux leakage data and the characteristics of defect data. Color images converted from triaxial magnetic flux leakage data are input into the model for training in target detection, facilitating rapid defect positioning and data extraction, and creating accurate and reliable datasets for quantifying defect sizes. Additionally, a Multi-Task Learning (MTL) model is introduced, taking circumferential, axial, and radial magnetic flux leakage data of pipeline defects as input. This model outputs data regarding the length, width, and depth of defects in parallel, leveraging the characteristics of magnetic flux leakage data to achieve size evaluation and prediction of defects. [Results] Comparing the data from pulling experiments revealed that the combination of PP-YOLOE and MTL improved defect identification efficiency and the prediction accuracy of defect sizes. The recall rate and accuracy rate for defect target detection reached 0.87 and 0.94, respectively. Compared with the application of Stochastic Gradient Descent (SGD), the Nesterov-accelerated Adaptive Moment Estimation (NADAM) algorithm achieved an increase in quantification accuracy for corrosion defects by 9% in length and 4% in depth. [Conclusion] The proposed method is applicable for analyzing massive pipeline magnetic flux leakage testing data and improving the identification efficiency. It provides reliable foundational data support for assessing residual strength and predicting the remaining life of pipelines with defects. (10 Figures, 2 Tables, 30 References)

参考文献/References:

[1] 张丽珍，王文奥. 天然气管道阀门微小泄漏声发射检测及缺陷识别方法[J]. 油气储运，2024，43（6）：656?664. 10.6047/j.issn.1000-8241.2024.06.007. ZHANG L Z，WANG W A. Acoustic emission detection and defect identification method for micro-leakage in natural gas pipeline valves[J]. Oil & Gas Storage and Transportation，2024，43(6):656?664.
[2] 张艳博，任瑞峰，梁鹏，姚旭龙，王帅. 基于热成像的埋地热力管道缺陷检测试验研究[J]. 仪器仪表学报，2020，41（6）：161?170. 10.19650/j.cnki.cjsi.J2006239. ZHANG Y B，REN R F，LIANG P，YAO X L，WANG S. Experimental study on flaw detection of buried heat pipeline based by infrared thermal[J]. Chinese Journal of Scientific Instrument，2020，41(6): 161?170.
[3] 王竹筠. 基于深度学习的管道漏磁信号智能识别方法研究[D].沈阳：沈阳工业大学，2019. WANG Z J. Study on intelligent recognition of pipeline magnetic flux leakage signal based on deep learning[D]. Shenyang:Shenyang University of Technology，2019.
[4] 郑远攀，李广阳，李晔. 深度学习在图像识别中的应用研究综述[J]. 计算机工程与应用，2019，55（12）：20?36. 10.3778/j.issn.1002-8331.1903-0031. ZHENG Y P，LI G Y，LI Y. Survey of application of deep learning in image recognition[J]. Computer Engineering and Applications，2019，55(12): 20?36.
[5] 万勇，王宇，杨勇，戴永寿. 管道缺陷类型多特征量阈值识别方法[J]. 油气储运，2020，39（3）：268?276. 10.6047/j.issn.1000-8241.2020.03.003. WAN Y，WANG Y，YANG Y，DAI Y S. Multi-characteristic threshold identification method for pipeline defect types[J]. Oil &Gas Storage and Transportation，2020，39(3): 268?276.
[6] 陈一诺，董绍华. 基于随机森林算法的管道缺陷预测方法[J].油气储运，2018，37（9）：975?979. 10.6047/j.issn.1000-8241. 2018.09.003. CHEN Y N，DONG S H. Pipeline defect prediction method based on Random Forests algorithm[J]. Oil & Gas Storage and Transportation，2018，37(9): 975?979.
[7] 赵东升，杨理践，耿浩，郑福印，唐权宇. 一种面向管道缺陷检测的漏磁数据增强方法[J]. 仪表技术与传感器，2023（12）：111?116. 10.3969/j.issn.1002-1841.2023.12.020. ZHAO D S，YANG L J，GENG H，ZHENG F Y，TANG Q Y. MFL data augmentation method for pipeline defect detection[J]. Instrument Technique and Sensor，2023(12): 111?116.
[8] 王国庆，李璇，杨理践，高松巍，耿浩. 基于改进YOLOv5算法的管道漏磁信号识别方法[J]. 计算机测量与控制，2022，30（8）：147?154. 10.16526/j.cnki.11-4762/tp.2022.08.024. WANG G Q，LI X，YANG L J，GAO S W，GENG H. Pipeline magnetic flux leakage signal recognition method based on improved YOLOv5 algorithm[J]. Computer Measurement & Control，2022，30(8): 147?154.
[9] 时亚南，陈志远，刘兆英，陈迎春，张婷，范效礼，等. 改进YOLOv5n的管道DR缺陷图像检测方法[J]. 计算机工程与应用，2024，60（12）：366?372. 10.3778/j.issn.1002-8331.2304-0397. SHI Y N，CHEN Z Y，LIU Z Y，CHEN Y C，ZHANG T，FAN X L，et al. Improved YOLOv5n pipeline DR defect image detection method[J]. Computer Engineering and Applications，2024，60(12):366?372.
[10] LYU C G，ZHANG M Q，LI B H，LIU Y G，LIN X J. High reliability pipeline leakage detection based on machine vision in complex industrial environment[J]. IEEE Sensors Journal，2022，22(21): 20748?20760. DOI: 10.1109/JSEN.2022.3206456.
[11] YANG L J，WANG Z J，GAO S W. Pipeline magnetic flux leakage image detection algorithm based on multiscale SSD network[J]. IEEE Transactions on Industrial Informatics，2020，16(1): 501?509. DOI: 10.1109/TII.2019.2926283.
[12] NGUYEN H T，NGUYEN M N，PHUNG L D，PHAM L T T. Anomalies detection in chest X-rays images using faster R-CNN and YOLO[J]. Vietnam Journal of Computer Science，2023，10(4):499?515. DOI: 10.1142/S2196888823500094.
[13] ZHANG W K，HONG Z Y，XIONG L P，ZENG Z Q，CAI Z S，TAN K Y. Sinextnet: a new small object detection model for aerial images based on PP-Yoloe[J]. Journal of Artificial Intelligence and Soft Computing Research，2024，14(3): 251?265. DOI: 10.2478/jaiscr-2024-0014.
[14] SHETTY A K，SAHA I，SANGHVI R M，SAVE S A，PATEL Y J. A review: object detection models[C]. Maharashtra: 2021 6th International Conference for Convergence in Technology (I2CT)，2021: 1-8.
[15] 崔国宁，杨理践，耿浩，刘屹然. 基于卷积神经网络的管道缺陷量化识别方法[J]. 仪表技术与传感器，2022（10）：99?103. 10.3969/j.issn.1002-1841.2022.10.018. CUI G N，YANG L J，GENG H，LIU Y R. Quantitative identification method of pipeline defects based on convolutional neural network[J]. Instrument Technique and Sensor，2022(10):99?103.
[16] 邢海燕，王松弘泽，弋鸣，杨健平，朱孔阳，刘超. 基于IPSO-GRU深度学习算法的海底管道缺陷尺寸磁记忆定量反演模型[J].工程科学学报，2022，44（5）：911?919. 10.3321/j.issn.1001-053X.2022.5.bjkjdxxb202205010.XING H Y，WANG S H Z，YI M，YANG J P，ZHU K Y，LIU C. Metal magnetic memory quantitative inversion model based on IPSO-GRU algorithm for detecting submarine pipeline defect[J]. Chinese Journal of Engineering，2022，44(5): 911?919.
[17] ZHANG Y，YANG Q. A survey on Multi-Task learning[J]. IEEE Transactions on Knowledge and Data Engineering，2022，34(12):5586?5609. DOI: 10.1109/TKDE.2021.3070203.
[18] CRAWSHAW M. Multi-task learning with deep neural networks:a survey[DB/OL]. (2020-09-10)[2024-07-05]. https://arxiv.org/abs/2009.09796.
[19] 欧正宇，韩赞东，岑佶松，李夏喜. 埋地管道检测技术现状与展望[J]. 测控技术，2023，42（6）：1?9. 10.19708/j.ckjs.2022. 11.323. OU Z Y，HAN Z D，CEN J S，LI X X. Present status and perspectives of inspection technology for buried pipeline[J]. Measurement & Control Technology，2023，42(6): 1?9.
[20] 陈兆凡，赵春阳，李博. 一种改进IoU损失的边框回归损失函数[J]. 计算机应用研究，2020，37（增刊2）：293?296. CHEN Z F，ZHAO C Y，LI B. Bounding box regression loss function based on improved IoU loss[J]. Application Research of Computers，2020，37(S2): 293?296.
[21] 任伟建，陈明文，康朝海，霍凤财，任璐，张永丰. 基于YOLOv3的金属表面缺陷检测研究[J]. 控制工程，2024，31（7）：1219?1228. 10.14107/j.cnki.kzgc.20220773. REN W J，CHEN M W，KANG C H，HUO F C，REN L，ZHANG Y F. Research on metal surface defect detection based on YOLOv3[J]. Control Engineering of China，2024，31(7): 1219?1228.
[22] WANG Y S，HE M Y，SUN L，WU D，WANG Y，QING X L. Weighted adaptive Kalman filtering-based diverse information fusion for hole edge crack monitoring[J]. Mechanical Systems and Signal Processing，2022，167(Part A): 108534.
[23] FANG H，GONG C L，LI C N，ZHANG Y W，DA RONCH A. A sequential optimization framework for simultaneous design variables optimization and probability uncertainty allocation[J]. Structural and Multidisciplinary Optimization，2021，63(3):1307?1325. DOI: 10.1007/s00158-020-02759-1.
[24] SU S S W，KEK S L. An improvement of stochastic gradient descent approach for mean-variance portfolio optimization pro-blem[J]. Journal of Mathematics，2021，2021: 8892636. DOI: 10.1155/2021/8892636.
[25] YANG J X，LONG Q. A modification of adaptive moment estimation (Adam) for machine learning[J]. Journal of Industrial and Management Optimization，2024，20(7): 2516?2540. DOI: 10.3934/jimo.2024014.
[26] PARHI R，NOWAK R D. The role of neural network activation functions[J]. IEEE Signal Processing Letters，2020，27: 1779?1783. DOI: 10.1109/LSP.2020.3027517.
[27] 张乃雪，钟羽中，赵涛，佃松宜. 基于Smooth-DETR的产品表面小尺寸缺陷检测算法[J]. 计算机应用研究，2022，39（8）：2520?2525. 10.19734/j.issn.1001-3695.2022.01.0011. ZHANG N X，ZHONG Y Z，ZHAO T，DIAN S Y. Detection method for small-size surface defects based on Smooth-DETR[J]. Application Research of Computers，2022，39(8): 2520?2525.
[28] 田野，高涛，许光达，丁融. 长输管道漏磁内检测缺陷识别量化技术研究[J]. 油气田地面工程，2018，37（10）：51?54，59. 10.3969/j.issn.1006-6896.2018.10.012. TIAN Y，GAO T，XU G D，DING R. Quantitative technology research on the defect identification of magnetic leakage internal detection in long distance pipelines[J]. Oil-Gas Field Surface Engineering，2018，37(10): 51?54,59.
[29] 赵帅，秦林，林冬，高健，刘畅，李潮浪，等. 油气管道漏磁数据处理和缺陷识别量化方法的研究进展[J]. 腐蚀与防护，2024，45（2）：27?35. 10.11973/fsyfh-202402004. ZHAO S，QIN L，LIN D，GAO J，LIU C，LI C L，et al. Progress in research on magnetic flux leakage data processing and defect identification quantification methods[J]. Corrosion & Protection，2024，45(2): 27?35.
[30] 李睿. 油气管道内检测技术与数据分析方法发展现状及展望[J].油气储运，2024，43（3）：241?256. 10.6047/j.issn.1000-8241. 2024.03.001. LI R. Current progress and prospects of in-line inspection techniques and data analysis methods for oil and gas pipelines[J]. Oil & Gas Storage and Transportation，2024，43(3): 241?256.

相似文献/References:

[1]曾忠刚　温秀荷　索杏兰　孔军　姜敏　陈凯　赵庆兵　曾朗峻.塔河凝析油管道结蜡计算[J].油气储运,2012,31(8):622.[doi:10.6047/j.issn.1000-8241.2012.08.018]
　Zeng Zhongang,Wen Xiuhe,Suo Xinglan,et al.Wax deposition calculation for Tahe Condensate Pipeline[J].Oil & Gas Storage and Transportation,2012,31(12):622.[doi:10.6047/j.issn.1000-8241.2012.08.018]
[2]夏文鹤茹黎南李明曹谢东.基于卫星物联网技术的油气管道远程监控[J].油气储运,2012,31(12):898.[doi:10.6047/j.issn.1000-8241.2012.12.006]
　Xia Wenhe,Ru Linan,Li Ming,et al. Remote monitoring on oil and gas pipelines with satellite Internet of Things network[J].Oil & Gas Storage and Transportation,2012,31(12):898.[doi:10.6047/j.issn.1000-8241.2012.12.006]
[3]董绍华韩忠晨杨毅曹兴.物联网技术在管道完整性管理中的应用[J].油气储运,2012,31(12):906.[doi:10.6047/j.issn.1000-8241.2012.12.008]
　Dong Shaohua,Han Zhongchen,Yang Yi,et al.The application of Internet of Things technology in the integrity management of pipeline[J].Oil & Gas Storage and Transportation,2012,31(12):906.[doi:10.6047/j.issn.1000-8241.2012.12.008]
[4]李睿,侯宇,刘淑聪,等.锚纹特征对管道外防腐环氧涂层附着力的影响[J].油气储运,2011,30(05):355.[doi:10.6047/j.issn.1000-8241.2011.05.010]
　Li Rui,Hou Yu,Li Jingmiao,et al.Impact of pattern on the adhesion of pipe external epoxy coating[J].Oil & Gas Storage and Transportation,2011,30(12):355.[doi:10.6047/j.issn.1000-8241.2011.05.010]
[5]段纪淼,宫敬,张宇,等.多相混输管道蜡沉积研究进展[J].油气储运,2011,30(04):241.[doi:10.6047/j.issn.1000-8241.2011.04.001]
　Duan Jimiao,Gong Jing,Zhang Yu,et al.Research progress of wax deposition in multiphase mixed transmission pipeline[J].Oil & Gas Storage and Transportation,2011,30(12):241.[doi:10.6047/j.issn.1000-8241.2011.04.001]
[6]宋晓琴,刘广文,王雯娟.管道油品泄漏原因及其对环境的影响[J].油气储运,2011,30(04):297.[doi:10.6047/j.issn.1000-8241.2011.04.016]
　Song Xiaoqin,Liu Guangwen,Wang Wenjuan. The influence of oil pipeline leakage on environmental pollution [J].Oil & Gas Storage and Transportation,2011,30(12):297.[doi:10.6047/j.issn.1000-8241.2011.04.016]
[7]蔡亮学,何利民,吕宇玲,等.水平定向钻管道穿越孔底泥浆的力学特性[J].油气储运,2011,30(01):25.[doi:10.6047/j.issn.1000-8241.2011.01.007]
　Cai Liangxue,He Limin,Lv Yuling,et al.Hole-bottom slurry mechanical properties of horizontal directional drilling in pipeline crossing project[J].Oil & Gas Storage and Transportation,2011,30(12):25.[doi:10.6047/j.issn.1000-8241.2011.01.007]
[8]樊三新,刘玉玲,楚威威.高寒地区管道低温水试压方法[J].油气储运,2011,30(01):73.[doi:10.6047/j.issn.1000-8241.2011.01.020]
　Fan Sanxin,Liu Yuling,Chu Weiwei.Pipeline hydrotest method of cold water in frigid zone[J].Oil & Gas Storage and Transportation,2011,30(12):73.[doi:10.6047/j.issn.1000-8241.2011.01.020]
[9]杜明俊,张振庭,张朝阳,等.多相混输管道90°弯管冲蚀破坏应力分析[J].油气储运,2011,30(06):427.[doi:10.6047/j.issn.1000-8241.2011.06.006]
　Du Mingjun,Chao Ling,Fu Xiaodong,et al.Analysis of erosion fracture stress of 90° elbow in multi-phase mixed transmission pipeline[J].Oil & Gas Storage and Transportation,2011,30(12):427.[doi:10.6047/j.issn.1000-8241.2011.06.006]
[10]高河东,祁志江,任金岭,等.川气东送管道野三河跨越方案比选与实施[J].油气储运,2011,30(06):445.[doi:10.6047/j.issn.1000-8241.2011.06.011]
　Gao Hedong,Qi Zhijiang,Ren Jinling,et al.Programs comparison and construction of the Yesanhe River Overhead Crossing NG Project of Sichuan-to-East Gas Pipeline[J].Oil & Gas Storage and Transportation,2011,30(12):445.[doi:10.6047/j.issn.1000-8241.2011.06.011]

备注/Memo

赵赏鑫，男，1979年生，教授级高工，2006年博士毕业于中国石油大学（北京）机械设计与理论专业，现主要从事油气管道建设管理工作。地址：新疆乌鲁木齐市高新区（新市区）喀什东路街道四平路2288号，830010。电话：0991-7683051。Email：tianye04@pipechina.com.cn
● Received: 2024-07-29● Revised: 2024-08-23● Online: 2024-10-09

更新日期/Last Update: 2024-12-25