管道缺陷类型判别和参数分析的研究
|
摘要
管道漏磁检测技术是当今无损检测技术领域的研究热点之一。漏磁检测技术具有较高的检测可靠性和检测速度,已经被广泛地运用于管道的检测中。
本文首先研究了管道漏磁检测原理和当前管道漏磁检测系统中的缺陷判别技术,并分析了模板匹配法缺陷识别的特点和不足。
接着从裂纹和孔洞这两种重要缺陷的漏磁信号分析比较入手,探讨并研究了能够区分这两种缺陷的漏磁信号特征量。
本文重点对BP和RBF两种神经网络的结构和算法做了详细的研究,并分别利用这两种网络对裂纹和孔洞进行类型判别尝试,对两种网络的识别结果进行比较。结果显示,通过神经网络可以对裂纹和孔洞这两类重要缺陷进行初步的类型判别。
其后,本文用实验探讨了人工刻槽深度和长度与漏磁信号的关系,重点研究了与人工刻槽参数有关的漏磁信号特征量。
然后,研究采用BP神经网络对人工刻槽参数分析的方法。在MATLAB中建立BP神经网络模型,网络以漏磁信号特征量为输入,刻槽的深度和长度为输出,并利用人工刻槽缺陷来对分析结果进行验证。结果显示,BP神经网络对人工刻槽参数分析具有一定的效果。
最后对全文做出了总结,并对管道缺陷类型判别和参数分析的研究进行了展望。
Pipeline magnetic leak detection technology is one of the hottest spots in the nondestructive testing field. MFL detection technology has a high reliability and speed, so it has been widely used in the detection of pipeline.
At first, this paper researches the pipeline leak detection principle and defects type determination technology in current pipeline leak detection system. Then analysis the deficiencies and shortcomings of template matching for the characteristics identified.
Then start with the two major defects cracks and holes' leakage magnetic signal analysis and comparison, to explore and study the signal characteristics of magnetic flux leakage which can distinguish between the two of defects.
This article focus on BP and RBF neural network' structure and algorithms, and use these two networks to try to identify cracks and holes and compare the results. The results showed that neural network can preliminary determine the type between cracks and holes which are two types of important defects.
Later, this paper experiment on the artificial groove with the depth and length. In order to study the relationship between the MFL signal and artificial groove parameters. Then this paper focuses on the groove the signal characteristics of magnetic flux leakage which are related with parameters.
Then, this paper explores the artificial groove parameters analysis through BP neural network. We establish BP neural network model in MATLAB. Characteristics of leak magnetic signal as the network's input, the groove's depth and length as the network's output. We use artificial groove defects to analysis the results' validation. The results showed that the analysis to parameters of artificial groove though BP neural network gets a certain effect. Finally, we make a summary of the transcript and outlook the research of pipeline defects type determination and parameters analysis.
本文首先研究了管道漏磁检测原理和当前管道漏磁检测系统中的缺陷判别技术,并分析了模板匹配法缺陷识别的特点和不足。
接着从裂纹和孔洞这两种重要缺陷的漏磁信号分析比较入手,探讨并研究了能够区分这两种缺陷的漏磁信号特征量。
本文重点对BP和RBF两种神经网络的结构和算法做了详细的研究,并分别利用这两种网络对裂纹和孔洞进行类型判别尝试,对两种网络的识别结果进行比较。结果显示,通过神经网络可以对裂纹和孔洞这两类重要缺陷进行初步的类型判别。
其后,本文用实验探讨了人工刻槽深度和长度与漏磁信号的关系,重点研究了与人工刻槽参数有关的漏磁信号特征量。
然后,研究采用BP神经网络对人工刻槽参数分析的方法。在MATLAB中建立BP神经网络模型,网络以漏磁信号特征量为输入,刻槽的深度和长度为输出,并利用人工刻槽缺陷来对分析结果进行验证。结果显示,BP神经网络对人工刻槽参数分析具有一定的效果。
最后对全文做出了总结,并对管道缺陷类型判别和参数分析的研究进行了展望。
Pipeline magnetic leak detection technology is one of the hottest spots in the nondestructive testing field. MFL detection technology has a high reliability and speed, so it has been widely used in the detection of pipeline.
At first, this paper researches the pipeline leak detection principle and defects type determination technology in current pipeline leak detection system. Then analysis the deficiencies and shortcomings of template matching for the characteristics identified.
Then start with the two major defects cracks and holes' leakage magnetic signal analysis and comparison, to explore and study the signal characteristics of magnetic flux leakage which can distinguish between the two of defects.
This article focus on BP and RBF neural network' structure and algorithms, and use these two networks to try to identify cracks and holes and compare the results. The results showed that neural network can preliminary determine the type between cracks and holes which are two types of important defects.
Later, this paper experiment on the artificial groove with the depth and length. In order to study the relationship between the MFL signal and artificial groove parameters. Then this paper focuses on the groove the signal characteristics of magnetic flux leakage which are related with parameters.
Then, this paper explores the artificial groove parameters analysis through BP neural network. We establish BP neural network model in MATLAB. Characteristics of leak magnetic signal as the network's input, the groove's depth and length as the network's output. We use artificial groove defects to analysis the results' validation. The results showed that the analysis to parameters of artificial groove though BP neural network gets a certain effect. Finally, we make a summary of the transcript and outlook the research of pipeline defects type determination and parameters analysis.
引文
[1]Mandache C,Clapham L.A model for magnetic flux leakage signal predictions[J].Jounral of Physics D-Applied Physics.2003,36(20):2427-2431
[2]Hwang J H,Lord W.Finite element analysis of the magnetic field distribution inside a rotating ferromagneitc bar[J].IEEE Transactions on Magnetics,1974,10(4):1113-1118
[3]Jones D,Peet S,Espinosa L,et al.High resolution pipeline inspection-cost-effective pipe rehabilitation[J].International Gas Engineering and Management.2000,40(5):8-10
[4]A.L.Crouch,Richard Anglisano,Mohammed Naarah.Quantitative field evaluation of magnetic flux-leakage and ultrasonic in-line inspection[J].Pipe&Pipelines International,1996,41(3):23-31.
[5]W.Lord,J.H.Hwang.Defect Characterization firm Magnetic Leakage Fields[J].British Tournalof NDT,1997:14-18
[6]Liu Z,Kang Y,Wu X,et al.Study on local magnetization of magnetic flux leakage testing for storage tank floors[J].Insight,2003,45(5):328-331
[7]Yarotskii V A.Estimation of magnetic dipole detection parameters[J].Measurement Techniques,1997,40(7):684-688
[8]Jaein Lim.Data Fusion for NDE Signal Characterization[D].Iowa:Ph.D Dissertation of Iowa Station University,2001:423-425
[9]Forster F.New Findings in the Field of Nondestructive Magnetic Leakage[J].Field InspectionNDT,1988,19(1):3-14
[10]Antonio Petraglia,UFRJ,Rio de Janeiro,RJ,Brasil,Detection and Identification of Defects in Oil Pipelines by Magnetic Flux Leakage[J],NDTISS 99-International symposium on NDT contribution to the infrastructure Safety Systems,Nov 1999 in Tores Brazil.
[11]何辅云.无缝钢管折叠缺陷检测方法及装置[J].合肥工业大学学报(自然科学版),1999,22(3):77-80
[12]D.L.A therton,Finite element calculations and computer measurements of magnetic flux leakage patterns for pits.British Journal of NDT[J],1988,30(2),159-160.
[13]Eduardo Altschuler.Monte carlo simulation of false alarms and detection reliability in magnetic flux leakage inspection of steel tubes[J].Materials Evaluation,1996,55(9):1032
[14]何辅云.钢管二维漏磁检测系统的再研究[J].无损检测,2000,22(2):68-70
[15]杨理践,陈晓春,魏兢.油气管道漏磁检测的信号处理技术[J].沈阳工业大学学报,1999,21(6):516-518
[16]王亚东.钢管漏磁检测技术研究[D].硕士学位论文,大庆石油学院,2004.
[17]何辅云,张艳等.大口径钢管检测系统[J].合肥工业大学学报,2005,28(9): 1093-1098
[18]何辅云.石油管道的高速检测与缺陷识别[J],无损检测,2000,22(5):206-208
[19]解源,康宜华.钢管管缝缺陷漏磁场模型及其应用[J].武汉理工大学学报,2002,24(1):42-45
[20]黄作英,胭沛文.计算机仿真在管道漏磁检测中的应用[J].计算机仿真,2004,21(10):174-177.
[21]汪友生,潘孟贤,何辅云.缺陷参数与漏磁信号相互关系的实验研究[J].合肥工业大学学报(自然科学版)1998,21(5):28-31
[22]胡险峰.法向漏磁通随外磁化场和矩形槽尺寸的变化[J].无损检测,200123(7):227-232
[23]高炬,汪来富,杨海青等.缺陷长度对漏磁场的影响[J].科学技术与工程.2002,2(4):52-53.
[24]李路明,郑鹏,黄松岭,张家骏.表面裂纹宽度对漏磁场Y分量的影响[J].清华大学学报(自然科学版),1999,39(2):43-45
[25]李明忠,丁克勤,张有忱.表面裂纹宽度对漏磁场的影响分析[J].无损探伤,2003,27(1):13-14,30
[26]刘美全,徐章遂.表面裂纹间漏磁场的相互作用及其识别[J].自动化与仪表,2003,18(4):1-4
[27]王贤琴,阮江军,杜志叶.钢管漏磁探伤中孔洞缺陷的漏磁场分析[J].无损检测,2005,27(12):649-651
[28]金涛.海底管道漏磁检测信号处理及缺陷识别研究[D].博士学位论文,上海交通大学,2005
[29]王长龙,陈鹏,刘美全等漏磁信号特征提取及检测研究[J].军械工程学院学报.2004,16(4):1-4
[30]郭罗军,刘丽川,江志,廖异勇,沈光新.漏磁检测缺陷信号特征分析[J1.无损检测,2005,27(12):646-647
[31]袁曾任.人工神经元网络及其应用[M].北京:清华大学出版社,1999
[32]Maritn Hagan,HowardB.Demuth,Mark H.Beale等.神经网络设计[M].北京:机械工业出版社,2002
[33]魏海神.神经网络结构设计的理论与方法[M].北京:国防工业出版社,2005
[34]李家伟,陈积愚主编.无损检测手册[M].北京:机械工业出版社,2002
[35]殷勤业,杨宗凯,谈正等.模式识别与神经网络[M].北京:机械工业出版社,1992
[36]黄松龄.基于神经网络的水冷壁管缺陷识别和分类方法[J].电测与仪表,2006,28(6):281-284
[37]飞思科技产品研发中心.神经网络理论与MATLAB7实现[M].北京:电子工业出版社,2005
[38]周开利,康耀红.神经网络模型及其MATLAB仿真程序设计[M].北京:清华大学出版社,2005.7
[39]BP改进算法在MATLAB7.0中的实现及比较研究[J].甘肃科技,2007,23(11):43-44
[40]国家质量监督检验检疫总局.无损检测—钢管自动漏磁检测系统综合性能测试[S].2008
[41]麦小琴.管道缺陷智能识别的研究[D].硕士学位论文,沈阳工业大学,2005
[42]刘美全,徐章遂.一种基于漏磁峰值相互作用的裂纹定量计算方法[J].测控技术,2003,22(11):28-31
[43]王长龙,刘兵,纪风珠,陈正阁.基于BP神经网络的漏磁定量检测[J].兵器材料科学与工程,2007,30(1):8-10
[44]宋小春,黄松领,康宜华,赵伟.漏磁无损检测中的缺陷信号定量解释方法[J].无损检测,2007,29(7):407-411
[45]飞思科技产品研发中心.MATLAB7基础与提高[M].北京:电子工业出版社,2005
[2]Hwang J H,Lord W.Finite element analysis of the magnetic field distribution inside a rotating ferromagneitc bar[J].IEEE Transactions on Magnetics,1974,10(4):1113-1118
[3]Jones D,Peet S,Espinosa L,et al.High resolution pipeline inspection-cost-effective pipe rehabilitation[J].International Gas Engineering and Management.2000,40(5):8-10
[4]A.L.Crouch,Richard Anglisano,Mohammed Naarah.Quantitative field evaluation of magnetic flux-leakage and ultrasonic in-line inspection[J].Pipe&Pipelines International,1996,41(3):23-31.
[5]W.Lord,J.H.Hwang.Defect Characterization firm Magnetic Leakage Fields[J].British Tournalof NDT,1997:14-18
[6]Liu Z,Kang Y,Wu X,et al.Study on local magnetization of magnetic flux leakage testing for storage tank floors[J].Insight,2003,45(5):328-331
[7]Yarotskii V A.Estimation of magnetic dipole detection parameters[J].Measurement Techniques,1997,40(7):684-688
[8]Jaein Lim.Data Fusion for NDE Signal Characterization[D].Iowa:Ph.D Dissertation of Iowa Station University,2001:423-425
[9]Forster F.New Findings in the Field of Nondestructive Magnetic Leakage[J].Field InspectionNDT,1988,19(1):3-14
[10]Antonio Petraglia,UFRJ,Rio de Janeiro,RJ,Brasil,Detection and Identification of Defects in Oil Pipelines by Magnetic Flux Leakage[J],NDTISS 99-International symposium on NDT contribution to the infrastructure Safety Systems,Nov 1999 in Tores Brazil.
[11]何辅云.无缝钢管折叠缺陷检测方法及装置[J].合肥工业大学学报(自然科学版),1999,22(3):77-80
[12]D.L.A therton,Finite element calculations and computer measurements of magnetic flux leakage patterns for pits.British Journal of NDT[J],1988,30(2),159-160.
[13]Eduardo Altschuler.Monte carlo simulation of false alarms and detection reliability in magnetic flux leakage inspection of steel tubes[J].Materials Evaluation,1996,55(9):1032
[14]何辅云.钢管二维漏磁检测系统的再研究[J].无损检测,2000,22(2):68-70
[15]杨理践,陈晓春,魏兢.油气管道漏磁检测的信号处理技术[J].沈阳工业大学学报,1999,21(6):516-518
[16]王亚东.钢管漏磁检测技术研究[D].硕士学位论文,大庆石油学院,2004.
[17]何辅云,张艳等.大口径钢管检测系统[J].合肥工业大学学报,2005,28(9): 1093-1098
[18]何辅云.石油管道的高速检测与缺陷识别[J],无损检测,2000,22(5):206-208
[19]解源,康宜华.钢管管缝缺陷漏磁场模型及其应用[J].武汉理工大学学报,2002,24(1):42-45
[20]黄作英,胭沛文.计算机仿真在管道漏磁检测中的应用[J].计算机仿真,2004,21(10):174-177.
[21]汪友生,潘孟贤,何辅云.缺陷参数与漏磁信号相互关系的实验研究[J].合肥工业大学学报(自然科学版)1998,21(5):28-31
[22]胡险峰.法向漏磁通随外磁化场和矩形槽尺寸的变化[J].无损检测,200123(7):227-232
[23]高炬,汪来富,杨海青等.缺陷长度对漏磁场的影响[J].科学技术与工程.2002,2(4):52-53.
[24]李路明,郑鹏,黄松岭,张家骏.表面裂纹宽度对漏磁场Y分量的影响[J].清华大学学报(自然科学版),1999,39(2):43-45
[25]李明忠,丁克勤,张有忱.表面裂纹宽度对漏磁场的影响分析[J].无损探伤,2003,27(1):13-14,30
[26]刘美全,徐章遂.表面裂纹间漏磁场的相互作用及其识别[J].自动化与仪表,2003,18(4):1-4
[27]王贤琴,阮江军,杜志叶.钢管漏磁探伤中孔洞缺陷的漏磁场分析[J].无损检测,2005,27(12):649-651
[28]金涛.海底管道漏磁检测信号处理及缺陷识别研究[D].博士学位论文,上海交通大学,2005
[29]王长龙,陈鹏,刘美全等漏磁信号特征提取及检测研究[J].军械工程学院学报.2004,16(4):1-4
[30]郭罗军,刘丽川,江志,廖异勇,沈光新.漏磁检测缺陷信号特征分析[J1.无损检测,2005,27(12):646-647
[31]袁曾任.人工神经元网络及其应用[M].北京:清华大学出版社,1999
[32]Maritn Hagan,HowardB.Demuth,Mark H.Beale等.神经网络设计[M].北京:机械工业出版社,2002
[33]魏海神.神经网络结构设计的理论与方法[M].北京:国防工业出版社,2005
[34]李家伟,陈积愚主编.无损检测手册[M].北京:机械工业出版社,2002
[35]殷勤业,杨宗凯,谈正等.模式识别与神经网络[M].北京:机械工业出版社,1992
[36]黄松龄.基于神经网络的水冷壁管缺陷识别和分类方法[J].电测与仪表,2006,28(6):281-284
[37]飞思科技产品研发中心.神经网络理论与MATLAB7实现[M].北京:电子工业出版社,2005
[38]周开利,康耀红.神经网络模型及其MATLAB仿真程序设计[M].北京:清华大学出版社,2005.7
[39]BP改进算法在MATLAB7.0中的实现及比较研究[J].甘肃科技,2007,23(11):43-44
[40]国家质量监督检验检疫总局.无损检测—钢管自动漏磁检测系统综合性能测试[S].2008
[41]麦小琴.管道缺陷智能识别的研究[D].硕士学位论文,沈阳工业大学,2005
[42]刘美全,徐章遂.一种基于漏磁峰值相互作用的裂纹定量计算方法[J].测控技术,2003,22(11):28-31
[43]王长龙,刘兵,纪风珠,陈正阁.基于BP神经网络的漏磁定量检测[J].兵器材料科学与工程,2007,30(1):8-10
[44]宋小春,黄松领,康宜华,赵伟.漏磁无损检测中的缺陷信号定量解释方法[J].无损检测,2007,29(7):407-411
[45]飞思科技产品研发中心.MATLAB7基础与提高[M].北京:电子工业出版社,2005