船舶动力设备振动评估及故障特性提取研究
|
摘要
船舶机舱动力设备是船舶的心脏,其运行状况的好坏直接影响到船舶的正常航行。因此确保船舶动力设备良好的运行状态对保障海上人命安全及避免财产损失具有十分重要的意义。
振动是设备运行所伴随的必然现象,同时也是反映其运行状态的重要指标。国际标准乃至船级社评估指南又采用振动响应的有效值来表征船舶结构或设备的振动状态。因此通过振动信号的获取及处理分析对船舶动力设备进行状态评估和故障诊断是简单可行的办法,这对老龄船舶显得尤为重要。对设备状态进行正确评估后,如果判断为有故障,需要对故障进行精密诊断。特征信号的获取是故障诊断的主要问题,独立分量分析为故障特征的提取提供了一种十分有效的方法,并在工程应用中取得了成功。
本文首先阐述了船舶动力设备状态评估及故障诊断的国内外研究进展和背景。并介绍了目前常用的评估和诊断方法。讨论了用于处理图像和语音信号的独立分量分析方法在振动信号分离中应用的可行性。
其次,学习了振动信号获取的详细过程和测试系统的选用原则。对信号的调理、采集、量化、截断和加窗通过实例进行说明。通过编写的程序验证理论的正确性和完整性。重点对测试获取的振动时域信号的检验和预处理进行了学习和程序的编写。同时为了便于应用ISO-10816对设备状态进行评估,本文推导证明了时域统计量有效值的频域计算方法,大大提高了计算标准中所规定频率范围内有效值的准确性。
船舶机舱设备由于特殊的布置特点,彼此相互振动影响较大,所测得的振动信号会引入其他设备的干扰信号以及外部环境的噪声信号。因此去掉这些干扰信号与背景噪声对于提高机械设备状态检测与故障诊断的准确性具有重要意义。
最后,本文系统地介绍了独立分量分析方法。应用这种信号处理方法,可以在测试条件满足要求的情况下对原始振源信号进行分离和提取。本文选取了适合分离振动信号的概率密度模型进行程序的编写,同时通过数值模拟和实测语音信号进行验证。最后将算法应用于实测船用动力设备振动信号的分离上取得了良好的效果。仿真和实验结果表明,将该方法用于故障特征提取是有效的。
The state of power equipment, which is the core part of ships, affects safety at sea. To guarantee the good condition of power equipment is highly significant To be a companion to operation of machine, vibration is the important index which reports the situation accurately. Both of ISO and classification societies use root mean square of vibration to describe the vibration level of hull structure and equipment. Consequently, it is a simple and useable method to assess and diagnosis power equipment by vibration monitoring and analysis. Especially for old age ships. If it is detected that the machine is faulty after assessment, precise diagnosis should be taken. The characteristic signal abstraction is the main problem of fault diagnosis. Independent component analysis provides a very effective method for abstraction and engineering application is successful.
In this thesis, the development of machine state assessment and fault diagnosis are being talked about first. And some typical methods are introduced. Then discuss if ICA, which is used in image processing and speech separation, can be applied in vibration signal analysis.
Then the detail process of vibration signal test and the principle for testing system selection are being studied. For the platform, acquisition, quantization, truncation and window, some example is enumerated. Computer program verifies the correctness and integrity of theory. The focal point of this part is inspection and preprocessing of test signal. Assessing the state of equipment by ISO-10816.Author deduced the frequency method to calculate root mean square and increases the calculation efficiency.Power equipment exist inter-acting vibration due to layout features. Because of this, the result signal contains noise signal. It is necessary to get rid of disturbance to increase the accuracy.
Finally, independent component analysis is introduced systematically.Source signal can be separated and extracted by this signal process method. Probability density modal which is suit to separate vibration signal has been established. The result is satisfied when the method is used in vibration signal of power equipment which is tested during trial voyage. This method is effective to abstract fault characteristic by simulation and experiment.
振动是设备运行所伴随的必然现象,同时也是反映其运行状态的重要指标。国际标准乃至船级社评估指南又采用振动响应的有效值来表征船舶结构或设备的振动状态。因此通过振动信号的获取及处理分析对船舶动力设备进行状态评估和故障诊断是简单可行的办法,这对老龄船舶显得尤为重要。对设备状态进行正确评估后,如果判断为有故障,需要对故障进行精密诊断。特征信号的获取是故障诊断的主要问题,独立分量分析为故障特征的提取提供了一种十分有效的方法,并在工程应用中取得了成功。
本文首先阐述了船舶动力设备状态评估及故障诊断的国内外研究进展和背景。并介绍了目前常用的评估和诊断方法。讨论了用于处理图像和语音信号的独立分量分析方法在振动信号分离中应用的可行性。
其次,学习了振动信号获取的详细过程和测试系统的选用原则。对信号的调理、采集、量化、截断和加窗通过实例进行说明。通过编写的程序验证理论的正确性和完整性。重点对测试获取的振动时域信号的检验和预处理进行了学习和程序的编写。同时为了便于应用ISO-10816对设备状态进行评估,本文推导证明了时域统计量有效值的频域计算方法,大大提高了计算标准中所规定频率范围内有效值的准确性。
船舶机舱设备由于特殊的布置特点,彼此相互振动影响较大,所测得的振动信号会引入其他设备的干扰信号以及外部环境的噪声信号。因此去掉这些干扰信号与背景噪声对于提高机械设备状态检测与故障诊断的准确性具有重要意义。
最后,本文系统地介绍了独立分量分析方法。应用这种信号处理方法,可以在测试条件满足要求的情况下对原始振源信号进行分离和提取。本文选取了适合分离振动信号的概率密度模型进行程序的编写,同时通过数值模拟和实测语音信号进行验证。最后将算法应用于实测船用动力设备振动信号的分离上取得了良好的效果。仿真和实验结果表明,将该方法用于故障特征提取是有效的。
The state of power equipment, which is the core part of ships, affects safety at sea. To guarantee the good condition of power equipment is highly significant To be a companion to operation of machine, vibration is the important index which reports the situation accurately. Both of ISO and classification societies use root mean square of vibration to describe the vibration level of hull structure and equipment. Consequently, it is a simple and useable method to assess and diagnosis power equipment by vibration monitoring and analysis. Especially for old age ships. If it is detected that the machine is faulty after assessment, precise diagnosis should be taken. The characteristic signal abstraction is the main problem of fault diagnosis. Independent component analysis provides a very effective method for abstraction and engineering application is successful.
In this thesis, the development of machine state assessment and fault diagnosis are being talked about first. And some typical methods are introduced. Then discuss if ICA, which is used in image processing and speech separation, can be applied in vibration signal analysis.
Then the detail process of vibration signal test and the principle for testing system selection are being studied. For the platform, acquisition, quantization, truncation and window, some example is enumerated. Computer program verifies the correctness and integrity of theory. The focal point of this part is inspection and preprocessing of test signal. Assessing the state of equipment by ISO-10816.Author deduced the frequency method to calculate root mean square and increases the calculation efficiency.Power equipment exist inter-acting vibration due to layout features. Because of this, the result signal contains noise signal. It is necessary to get rid of disturbance to increase the accuracy.
Finally, independent component analysis is introduced systematically.Source signal can be separated and extracted by this signal process method. Probability density modal which is suit to separate vibration signal has been established. The result is satisfied when the method is used in vibration signal of power equipment which is tested during trial voyage. This method is effective to abstract fault characteristic by simulation and experiment.
引文
[1]袁宏义.设备振动诊断技术基础[M].北京：国防工业出版社,1991.
[2]盛兆顺.设备状态监测与故障诊断技术及应用[M].北京：化学工业出版社,2003.
[3]高金吉.机泵群实时监测网络和故障诊断专家系统[J].中国工程科学,2001,3(9)：41-47.
[4]王若冲.大型散货船动力设备振动评价与诊断[C].第四届中国船舶与海洋工程学术会议,江苏镇江,2008：84-89.
[5]汪红艳.盲源分离及其在2D12型往复式压缩机故障诊断中的应用[D].哈尔滨：哈尔滨工业大学航天学院,2007.
[6]王宇,迟毅林,伍星,等.基于盲信号处理的机械噪声监测与故障诊断[J].振动与冲击,2009,28(6)：32-41.
[7]纪康宝.船舶故障诊断与检测检验检修标准规范[M].安徽：安徽文化音像出版社,2004.
[8]陈长征,胡立新,周勃,等.设备振动分析与故障诊断技术[M].北京：科学出版社,2007.
[9]张梅军.机械状态检测与故障诊断[M].北京：国防工业出版社,2008.
[10]石来德,丁玉兰.机械设备状态监测和故障诊断技术[J].建筑机械化,1987(7)：40-44.
[11]Bell A J, Sejnowski T J. An Information-Maximization Approach to Blind Separation and Blind Deconvolution[J]. Neural Computation,1995,7:1129-1159.
[12]Cichocki A, Unbehaucn R, Moszczynski L. A new on line adaptive learning algorithm for blind separation of source signals[C]. ISANN94, Taiwan,1994,406-411.
[13]Yang H H, Amari S. Adaptive on-line learning algorithms for blind separation Maximum entropy and minimum mutual information[J]. Neural Computation,1997,9:1457-1482.
[14]Cardoso J F, Laheld B. Equivariant adaptive source separation[J]. Signal Processing,1996,44:3017-3029.
[15]Amari S. Natural gradient works efficiently in learning[J]. Neural Computation,1998,10:251-276.
[16]Hyvarinen A. Fast and Robust Fixed-Point Algorithms for Independent Component Analysis[J]. Neural Networks,1999,10(3):626-634.
[17]Gelle G, Colas M, Delaunay G. Blind sources separation applied to rotating machines monitoring by acoustic and vibrations analysis[J] Mechanical Systems and Signal Processing,2000,14 (3),427-442.
[18]Antoni J. Blind separation of vibration components:Principles and demonstrations[J]. Mechanical Systems and Signal Processing,2005,19:1166-1180.
[19]Cho K S,Lee S Y. Implementation of INFOMAX ICA algorithm with analog CMOS circuits[C]. ICA2001, San Diego,2001:70-73.
[20]Lyon R H. Machinery noise and diagnostics[M]. Boston:Butterworths,1957.
[21]何清波,孔凡让,朱忠奎等.盲卷积分离及其在机械振动信号消噪中的应用研究[J].振动与冲击,2006,25(2)：31-34.
[22]Ypma A, Leshem A, Duin P W. Blind separation of rotating machine sources:Bilinear forms and convolutive mixtures[J]. Neurocomputing,2002,49:349-368.
[23]Gelle G, Colas M, Delaunay G. Blind sources separation applied to rotating machines monitoring by acoustical and vibrations analysis[J]. Mechanical Systems and Signal Processing,2000,14 (3),427-442.
[24]焦卫东.基于独立分量分析的旋转机械故障诊断方法研究[D].浙江：浙江大学机械与能源工程学院,2003.
[25]操炼.振动信号在船舶设备检验中的应用探讨[D].上海：上海海事大学,2005
[26]樊新海.工程测试技术基础[M].北京：国防工业出版社,2007.
[27]孔德仁.工程测试技术[M].北京：科学出版社,2004.
[28]尤德斐.数字化测量技术及仪器[M].北京：机械工业出版社,1988
[29]J.T.Broch.机械振动与冲击测量[M].北京：国防工业出版社,1972.
[30]姜大正.环境激励下船舶结构模态分析实验与理论研究[D].大连：大连理工大学船舶工程学院,2009.
[31]刘利生.外测数据事后处理[M].北京：国防工业出版社,2000.
[32]J.S.贝达特,A.G.皮尔索.随机数据分析方法[M].北京：国防工业出版社,1976.
[33]李彤.基于LabVIEW的船舶舱室甲板虚拟振动测量系统设计[D].大连：大连理工大学船舶工程学院,2008.
[34]张善文,雷英杰,冯有前MATLAB在时间序列分析中的应用[M].西安：西安电子科技大学出版社,2007.
[35]樊新海,赵智勇,安钢,等.机械振动烈度的频域算法研究[J].装甲兵工程学院学报,2008,22(1)：42-45.
[36]杨叔子,吴雅,轩建平等.时间序列分析的工程应用[M].武汉：华中理工大学出版社,2007.
[37]管致中,夏恭格.信号与线性系统[M].北京：高等教育出版社,1992.
[38]同济大学应用数学系.高等数学[M].北京：高等教育出版社,2004.
[39]王齐MATLAB在振动信号处理中的应用[M].北京：高等教育出版社,2004.
[40]薛年喜MATLAB在数字信号处理中的应用[M].北京：清华大学出版社,2003.
[41]张贤达,保铮.盲信号分离[J].电子学报,2001,29(12)：1766-1771.
[42]李志农,范涛,刘立州,等.基于变分贝叶斯理论的机械故障源盲分离方法研究[J].振动与冲击,2009,28(6)：12-16.
[43]周宗潭,董华国等.独立成分分析[M].北京：电子工业出版社,2007.
[44]Yang H H, Amari S. Adaptive on-line learning algorithms for blind separation Maximum entropy and minimum mutual information[J]. Neural Computation, 1997,9:1457-1482.
[45]张永祥,李军,孙云岭,等.基于遗传算法和峰度最佳的滚动轴承故障诊断[J].振动与冲击,2007,26(8)：122-124.
[46]A. Hyvarinen, E. Oja. Independent Component Analysis:Algorithms and Applications [J]. Neural Network,2000,13(4-5):411-430.
[47]马建仓,牛亦龙,等.盲信号处理[M].北京：国防工业出版社,2006.
[48]苏永生,王永生,段向阳.一种基于改进的傅里叶变换的弱故障信号特征提取新方法研究[J].振动与冲击,2009,28(8)：172-175.
[49]Lee T W, Girolami M. Independent Component Analysis Using an Extended Information Algorithm for Mixed Sub Gaussian and Super-Gaussian Sources[J]. Neural Computation,1999,11 (2):469-433.
[50]Girolami M. An alternative perspective on adaptive independent component analysis algorithms[J]. Neural Computation,1998,10:2103-2114.
[51]胡广书.数字信号处理：理论、算法与实现[M].北京：清华大学出版社,2003
[52]Lee T W, Girolami M. Independent Component Analysis Using an Extended Information Algorithm for Mixed Sub Gaussian and Super-Gaussian Sources[J]. Neural Computation,1999,11 (2):409-433.
[53]杨福生,洪波.独立分量分析的原理与应用[M].北京：清华大学出版社,2006.
[54]Cheung Y M, Liu H L. A new approach to blind source separation with global optimal property[C]. Processing of the IASTED internaltional Conference of Neural Networks and Computational Intelligence.2004:137-141.
[55]张发启.盲信号处理及应用[M].西安：西安电子科技大学出版社,2006.
[56]Makino S, Lee T W, Sawada H. Bland Speech Separation [M]. Dordrecht, the Netherlands Springer,2007.
[2]盛兆顺.设备状态监测与故障诊断技术及应用[M].北京：化学工业出版社,2003.
[3]高金吉.机泵群实时监测网络和故障诊断专家系统[J].中国工程科学,2001,3(9)：41-47.
[4]王若冲.大型散货船动力设备振动评价与诊断[C].第四届中国船舶与海洋工程学术会议,江苏镇江,2008：84-89.
[5]汪红艳.盲源分离及其在2D12型往复式压缩机故障诊断中的应用[D].哈尔滨：哈尔滨工业大学航天学院,2007.
[6]王宇,迟毅林,伍星,等.基于盲信号处理的机械噪声监测与故障诊断[J].振动与冲击,2009,28(6)：32-41.
[7]纪康宝.船舶故障诊断与检测检验检修标准规范[M].安徽：安徽文化音像出版社,2004.
[8]陈长征,胡立新,周勃,等.设备振动分析与故障诊断技术[M].北京：科学出版社,2007.
[9]张梅军.机械状态检测与故障诊断[M].北京：国防工业出版社,2008.
[10]石来德,丁玉兰.机械设备状态监测和故障诊断技术[J].建筑机械化,1987(7)：40-44.
[11]Bell A J, Sejnowski T J. An Information-Maximization Approach to Blind Separation and Blind Deconvolution[J]. Neural Computation,1995,7:1129-1159.
[12]Cichocki A, Unbehaucn R, Moszczynski L. A new on line adaptive learning algorithm for blind separation of source signals[C]. ISANN94, Taiwan,1994,406-411.
[13]Yang H H, Amari S. Adaptive on-line learning algorithms for blind separation Maximum entropy and minimum mutual information[J]. Neural Computation,1997,9:1457-1482.
[14]Cardoso J F, Laheld B. Equivariant adaptive source separation[J]. Signal Processing,1996,44:3017-3029.
[15]Amari S. Natural gradient works efficiently in learning[J]. Neural Computation,1998,10:251-276.
[16]Hyvarinen A. Fast and Robust Fixed-Point Algorithms for Independent Component Analysis[J]. Neural Networks,1999,10(3):626-634.
[17]Gelle G, Colas M, Delaunay G. Blind sources separation applied to rotating machines monitoring by acoustic and vibrations analysis[J] Mechanical Systems and Signal Processing,2000,14 (3),427-442.
[18]Antoni J. Blind separation of vibration components:Principles and demonstrations[J]. Mechanical Systems and Signal Processing,2005,19:1166-1180.
[19]Cho K S,Lee S Y. Implementation of INFOMAX ICA algorithm with analog CMOS circuits[C]. ICA2001, San Diego,2001:70-73.
[20]Lyon R H. Machinery noise and diagnostics[M]. Boston:Butterworths,1957.
[21]何清波,孔凡让,朱忠奎等.盲卷积分离及其在机械振动信号消噪中的应用研究[J].振动与冲击,2006,25(2)：31-34.
[22]Ypma A, Leshem A, Duin P W. Blind separation of rotating machine sources:Bilinear forms and convolutive mixtures[J]. Neurocomputing,2002,49:349-368.
[23]Gelle G, Colas M, Delaunay G. Blind sources separation applied to rotating machines monitoring by acoustical and vibrations analysis[J]. Mechanical Systems and Signal Processing,2000,14 (3),427-442.
[24]焦卫东.基于独立分量分析的旋转机械故障诊断方法研究[D].浙江：浙江大学机械与能源工程学院,2003.
[25]操炼.振动信号在船舶设备检验中的应用探讨[D].上海：上海海事大学,2005
[26]樊新海.工程测试技术基础[M].北京：国防工业出版社,2007.
[27]孔德仁.工程测试技术[M].北京：科学出版社,2004.
[28]尤德斐.数字化测量技术及仪器[M].北京：机械工业出版社,1988
[29]J.T.Broch.机械振动与冲击测量[M].北京：国防工业出版社,1972.
[30]姜大正.环境激励下船舶结构模态分析实验与理论研究[D].大连：大连理工大学船舶工程学院,2009.
[31]刘利生.外测数据事后处理[M].北京：国防工业出版社,2000.
[32]J.S.贝达特,A.G.皮尔索.随机数据分析方法[M].北京：国防工业出版社,1976.
[33]李彤.基于LabVIEW的船舶舱室甲板虚拟振动测量系统设计[D].大连：大连理工大学船舶工程学院,2008.
[34]张善文,雷英杰,冯有前MATLAB在时间序列分析中的应用[M].西安：西安电子科技大学出版社,2007.
[35]樊新海,赵智勇,安钢,等.机械振动烈度的频域算法研究[J].装甲兵工程学院学报,2008,22(1)：42-45.
[36]杨叔子,吴雅,轩建平等.时间序列分析的工程应用[M].武汉：华中理工大学出版社,2007.
[37]管致中,夏恭格.信号与线性系统[M].北京：高等教育出版社,1992.
[38]同济大学应用数学系.高等数学[M].北京：高等教育出版社,2004.
[39]王齐MATLAB在振动信号处理中的应用[M].北京：高等教育出版社,2004.
[40]薛年喜MATLAB在数字信号处理中的应用[M].北京：清华大学出版社,2003.
[41]张贤达,保铮.盲信号分离[J].电子学报,2001,29(12)：1766-1771.
[42]李志农,范涛,刘立州,等.基于变分贝叶斯理论的机械故障源盲分离方法研究[J].振动与冲击,2009,28(6)：12-16.
[43]周宗潭,董华国等.独立成分分析[M].北京：电子工业出版社,2007.
[44]Yang H H, Amari S. Adaptive on-line learning algorithms for blind separation Maximum entropy and minimum mutual information[J]. Neural Computation, 1997,9:1457-1482.
[45]张永祥,李军,孙云岭,等.基于遗传算法和峰度最佳的滚动轴承故障诊断[J].振动与冲击,2007,26(8)：122-124.
[46]A. Hyvarinen, E. Oja. Independent Component Analysis:Algorithms and Applications [J]. Neural Network,2000,13(4-5):411-430.
[47]马建仓,牛亦龙,等.盲信号处理[M].北京：国防工业出版社,2006.
[48]苏永生,王永生,段向阳.一种基于改进的傅里叶变换的弱故障信号特征提取新方法研究[J].振动与冲击,2009,28(8)：172-175.
[49]Lee T W, Girolami M. Independent Component Analysis Using an Extended Information Algorithm for Mixed Sub Gaussian and Super-Gaussian Sources[J]. Neural Computation,1999,11 (2):469-433.
[50]Girolami M. An alternative perspective on adaptive independent component analysis algorithms[J]. Neural Computation,1998,10:2103-2114.
[51]胡广书.数字信号处理：理论、算法与实现[M].北京：清华大学出版社,2003
[52]Lee T W, Girolami M. Independent Component Analysis Using an Extended Information Algorithm for Mixed Sub Gaussian and Super-Gaussian Sources[J]. Neural Computation,1999,11 (2):409-433.
[53]杨福生,洪波.独立分量分析的原理与应用[M].北京：清华大学出版社,2006.
[54]Cheung Y M, Liu H L. A new approach to blind source separation with global optimal property[C]. Processing of the IASTED internaltional Conference of Neural Networks and Computational Intelligence.2004:137-141.
[55]张发启.盲信号处理及应用[M].西安：西安电子科技大学出版社,2006.
[56]Makino S, Lee T W, Sawada H. Bland Speech Separation [M]. Dordrecht, the Netherlands Springer,2007.