基于能量耗损的机械设备故障诊断理论与方法研究
|
摘要
目前，机械设备故障诊断方法主要有振动分析和油液分析等方法。无论是基于振动分析的设备故障诊断，还是基于油液分析（磨损信息）的设备故障诊断，它们有一个共同点，就是设备故障时都会伴随有系统能量耗损变化。针对机械设备在发生故障时都伴随能量耗损变化这一特征，开创性地提出了一种基于能量耗损的机械设备故障诊断理论与方法。这种方法通过获取摩擦学系统的能量耗损信息，建立能量耗损信息的相关性，提取能量耗损信息特征并进行故障模式识别，建立基于能量耗损的故障规则。
首先，论文提出基于能量耗损的机械设备故障诊断新方法。研究摩擦学系统的能量耗损理论与摩擦过程的能量耗损信息流，研究输入能量耗损信息特征、磨损信息特征、振动信号特征。提出了能量耗损信息的相对标度和能量耗损信息的累计相对标度，输入能量耗损采用功率或者油耗等特征量，磨损能量耗损采用光谱元素指标；振动耗能采用振动速度信号时域均方值与振动加速度的峭度等指标，建立能量耗损信息的特征集。研究能量耗损信息的相关性，建立基于能量耗损的机械设备相关性模型。
其次，论证了基于能量耗损的机械设备故障诊断方法是可行的。齿轮模拟故障实验研究表明，齿轮发生点蚀、剥落、断齿等不同故障时，输入的功率耗损波动特性不同；磨损能耗信息磨损量和严重磨损指数表明故障的剧烈程度，振动时域信号通过小波包分析提取了各频带的能量分布。齿轮疲劳故障诊断相关性研究表明，输入功率耗损与磨损特征信息与振动特征信息变化规律具有一致性，具有较强的相关性。柴油机活塞缸套疲劳性实验研究表明，瞬时油耗随着活塞磨损故障程度的增加而增加，磨损能耗信息磨损量和严重磨损量指数一直递增。振动能量的变化具有随故障程度增加而增大的趋势，三者能量耗损信息表现规律具有一致性。从而验证本文提出的基于能量耗损的机械设备故障诊断理论和方法是可行的。
再次，提出一种基于流形学习算法与支持向量机结合的故障模式识别方法。研究局部线性嵌入LLE、局部切空间排列算法LTSA流形学习算法，并对算法进行了改进。采用流形学习算法对齿轮和柴油机能量耗损数据降维，然后采用多类分类器进行分类，通过分类识别率来判断模式识别的效果。仿真和实验表明流形学习是一种有效的非线性特征提取方法，改进的算法使邻域较好保持了曲面数据的原有对应关系，使得投影后的特征保持了样本间的差异信息和同类样本之间的相似信息。改进的流形学习算法的识别率得到了提高。流形学习与支持向量机结合的模式识别方法是一种有效的特征提取和模式识别方法。
然后，建立了能量耗损信息的故障诊断规则。研究了粗糙集与模糊理论的故障规则提取方法，利用粗糙集理论中的不可分辨关系把齿轮能量耗损信息的故障论域划分等价类，生成粗糙集的上近似关系和下近似关系，通过属性重要性分析和属性约简导出故障决策知识和故障分类规则，建立了齿轮能量耗损信息的故障规则；采用模糊理论与神经网络结合的方法，应用自适应模糊控制规则提取方法，输入柴油机能量耗损信息的模糊量，能自动对模糊控制规则进行修改，建立了能量耗损信息的柴油机活塞磨损模糊的故障规则。
最后，研究了能量耗损信息监测与诊断系统的基本结构。设计了基于虚拟仪器技术的能量耗损信息监测与诊断系统结构，使用LabVIEW虚拟化图形化用图标代替文本创建应用程序的计算机编程语言，开发了能量耗损信息在线故障诊断监测系统，包括数据采集系统，信号分析系统，实现了能量耗损信息的采集与分析，初步建立了能量耗损信息的监测与诊断系统。
At present, there are mainly vibration analyses and oil analysis for mechanicalequipment fault diagnosis. They have one thing in common that system energy loss change isoften associated with equipment fault both in equipment fault diagnosis based on vibrationanalysis and based on oil analysis (wear) of equipment fault diagnosis.In view of energy lossalways changes with mechanical equipment fault, mechanical equipment fault diagnosistheory and method based on the energy loss was put forward. This method is that the relativityof energy loss information was studied by getting tribology system energy loss information,the characteristics of energy loss information was extracted for fault pattern recognition, toestablish fault rules based on energy loss.
First of all, a new fault diagnosis method based on energy loss was presented in thispaper. Tribology system energy loss theory and energy loss information flow from the processof friction and wear were studied. The characteristics of input energy loss, friction andvibration energy loss information were analyzed. Relative scale and cumulative relative scaleof energy loss information were presented. Input energy loss was used by the input power orfuel loss characteristics, vibration energy loss was used by vibration velocity signal timedomain and vibration acceleration mean square value of kurtosis index, wear energy loss wasused by spectral elements and iron spectrum to establish energy loss information feature set toestablish the correlation of energy loss information of fault diagnosis model.
Second, the feasibility of mechanical equipment fault diagnosis method based on theenergy loss was proved. Gear fault imitate fault experiments show that input powerattenuation volatility characteristics different kind of faults, such as gear pitting and spalling,broken teeth. Abrasion wear quantity and severe wear index of energy loss information meanfailure intensity. Energy distribution of vibration signal frequency band was extracted bywavelet packet analysis. Gear fatigue fault diagnosis correlation studies show that the inputpower loss has strong correlation with wear characteristics information and vibrationcharacteristics. Diesel engine piston liner fatigue experimental research show that theinstantaneous fuel consumption as the piston fault increase with the increase of the degree ofwear and tear, abrasion and serious energy loss wear abrasion index has been increasing.Thereis an increasing of vibration energy along with the degree of fault change. The results showthat input energy loss information, wear information and vibration information have highcorrelation of the tribological system fault diagnosis, which indicates that the proposed theory and methods based on energy loss in this thesis are feasible.
Then, feature extraction and pattern recognition of energy loss information based onmanifold learning combining support vector machine method was proposed. Locally linearembedding and local tangent space alignment were studied and were modified. With theimproved manifold learning algorithm for energy loss data dimension reduction, the effect ofpattern recognition to determined through the recognition rate from classifier.Simulation andexperiment studies show that manifold learning is a kind of effective nonlinear featureextraction method, improvement of algorithm keeps good neighborhood the surface data ofthe original corresponding relation, keeps well the difference between the sample information,and maintain the similarity information between the similar samples, extractes the categoryinformation for samples. Manifold learning recognition rate is higher than the recognition rateof principal component analysis and kernel methods. So, the fault pattern recognitionmanifold learning combined support vector machine method was effective method for faultfeature extraction and pattern recognition.
Next, fault rules based on energy loss was set. Rough set and fuzzy theory were studiedfor fault rules which using the indiscernibility relation divide discourse equivalence class,generate the relations between approximation and lower approximation of gear energy lossinformation rough set relationship, and attribute reduction are derived through the analysis ofthe attribute importance decision making fault knowledge and the classification fault rules,solve the fault diagnosis system in the discretization of continuous attribute value. Gear faultrules based on energy loss was established by analyzing attribute importance and attributereduction. The method of fuzzy theory and neural networks was study for diesel engine faultrules which was established by adaptive newwork-based fuzzy inference system.
Finally, structure of the energy loss information monitoring and diagnosis system wasstudied using virtual instrument technology. Energy loss information monitoring anddiagnosis system was designed based on virtual instrument. Energy loss information onlinemonitoring fault diagnosis system is developed by using LabVIEW which is by virtualgraphics paraphrase icon to create the application of computer programming language,including the data acquisition system and signal processing technology, so energy lossmonitoring and diagnosis system was established primarily.
首先，论文提出基于能量耗损的机械设备故障诊断新方法。研究摩擦学系统的能量耗损理论与摩擦过程的能量耗损信息流，研究输入能量耗损信息特征、磨损信息特征、振动信号特征。提出了能量耗损信息的相对标度和能量耗损信息的累计相对标度，输入能量耗损采用功率或者油耗等特征量，磨损能量耗损采用光谱元素指标；振动耗能采用振动速度信号时域均方值与振动加速度的峭度等指标，建立能量耗损信息的特征集。研究能量耗损信息的相关性，建立基于能量耗损的机械设备相关性模型。
其次，论证了基于能量耗损的机械设备故障诊断方法是可行的。齿轮模拟故障实验研究表明，齿轮发生点蚀、剥落、断齿等不同故障时，输入的功率耗损波动特性不同；磨损能耗信息磨损量和严重磨损指数表明故障的剧烈程度，振动时域信号通过小波包分析提取了各频带的能量分布。齿轮疲劳故障诊断相关性研究表明，输入功率耗损与磨损特征信息与振动特征信息变化规律具有一致性，具有较强的相关性。柴油机活塞缸套疲劳性实验研究表明，瞬时油耗随着活塞磨损故障程度的增加而增加，磨损能耗信息磨损量和严重磨损量指数一直递增。振动能量的变化具有随故障程度增加而增大的趋势，三者能量耗损信息表现规律具有一致性。从而验证本文提出的基于能量耗损的机械设备故障诊断理论和方法是可行的。
再次，提出一种基于流形学习算法与支持向量机结合的故障模式识别方法。研究局部线性嵌入LLE、局部切空间排列算法LTSA流形学习算法，并对算法进行了改进。采用流形学习算法对齿轮和柴油机能量耗损数据降维，然后采用多类分类器进行分类，通过分类识别率来判断模式识别的效果。仿真和实验表明流形学习是一种有效的非线性特征提取方法，改进的算法使邻域较好保持了曲面数据的原有对应关系，使得投影后的特征保持了样本间的差异信息和同类样本之间的相似信息。改进的流形学习算法的识别率得到了提高。流形学习与支持向量机结合的模式识别方法是一种有效的特征提取和模式识别方法。
然后，建立了能量耗损信息的故障诊断规则。研究了粗糙集与模糊理论的故障规则提取方法，利用粗糙集理论中的不可分辨关系把齿轮能量耗损信息的故障论域划分等价类，生成粗糙集的上近似关系和下近似关系，通过属性重要性分析和属性约简导出故障决策知识和故障分类规则，建立了齿轮能量耗损信息的故障规则；采用模糊理论与神经网络结合的方法，应用自适应模糊控制规则提取方法，输入柴油机能量耗损信息的模糊量，能自动对模糊控制规则进行修改，建立了能量耗损信息的柴油机活塞磨损模糊的故障规则。
最后，研究了能量耗损信息监测与诊断系统的基本结构。设计了基于虚拟仪器技术的能量耗损信息监测与诊断系统结构，使用LabVIEW虚拟化图形化用图标代替文本创建应用程序的计算机编程语言，开发了能量耗损信息在线故障诊断监测系统，包括数据采集系统，信号分析系统，实现了能量耗损信息的采集与分析，初步建立了能量耗损信息的监测与诊断系统。
At present, there are mainly vibration analyses and oil analysis for mechanicalequipment fault diagnosis. They have one thing in common that system energy loss change isoften associated with equipment fault both in equipment fault diagnosis based on vibrationanalysis and based on oil analysis (wear) of equipment fault diagnosis.In view of energy lossalways changes with mechanical equipment fault, mechanical equipment fault diagnosistheory and method based on the energy loss was put forward. This method is that the relativityof energy loss information was studied by getting tribology system energy loss information,the characteristics of energy loss information was extracted for fault pattern recognition, toestablish fault rules based on energy loss.
First of all, a new fault diagnosis method based on energy loss was presented in thispaper. Tribology system energy loss theory and energy loss information flow from the processof friction and wear were studied. The characteristics of input energy loss, friction andvibration energy loss information were analyzed. Relative scale and cumulative relative scaleof energy loss information were presented. Input energy loss was used by the input power orfuel loss characteristics, vibration energy loss was used by vibration velocity signal timedomain and vibration acceleration mean square value of kurtosis index, wear energy loss wasused by spectral elements and iron spectrum to establish energy loss information feature set toestablish the correlation of energy loss information of fault diagnosis model.
Second, the feasibility of mechanical equipment fault diagnosis method based on theenergy loss was proved. Gear fault imitate fault experiments show that input powerattenuation volatility characteristics different kind of faults, such as gear pitting and spalling,broken teeth. Abrasion wear quantity and severe wear index of energy loss information meanfailure intensity. Energy distribution of vibration signal frequency band was extracted bywavelet packet analysis. Gear fatigue fault diagnosis correlation studies show that the inputpower loss has strong correlation with wear characteristics information and vibrationcharacteristics. Diesel engine piston liner fatigue experimental research show that theinstantaneous fuel consumption as the piston fault increase with the increase of the degree ofwear and tear, abrasion and serious energy loss wear abrasion index has been increasing.Thereis an increasing of vibration energy along with the degree of fault change. The results showthat input energy loss information, wear information and vibration information have highcorrelation of the tribological system fault diagnosis, which indicates that the proposed theory and methods based on energy loss in this thesis are feasible.
Then, feature extraction and pattern recognition of energy loss information based onmanifold learning combining support vector machine method was proposed. Locally linearembedding and local tangent space alignment were studied and were modified. With theimproved manifold learning algorithm for energy loss data dimension reduction, the effect ofpattern recognition to determined through the recognition rate from classifier.Simulation andexperiment studies show that manifold learning is a kind of effective nonlinear featureextraction method, improvement of algorithm keeps good neighborhood the surface data ofthe original corresponding relation, keeps well the difference between the sample information,and maintain the similarity information between the similar samples, extractes the categoryinformation for samples. Manifold learning recognition rate is higher than the recognition rateof principal component analysis and kernel methods. So, the fault pattern recognitionmanifold learning combined support vector machine method was effective method for faultfeature extraction and pattern recognition.
Next, fault rules based on energy loss was set. Rough set and fuzzy theory were studiedfor fault rules which using the indiscernibility relation divide discourse equivalence class,generate the relations between approximation and lower approximation of gear energy lossinformation rough set relationship, and attribute reduction are derived through the analysis ofthe attribute importance decision making fault knowledge and the classification fault rules,solve the fault diagnosis system in the discretization of continuous attribute value. Gear faultrules based on energy loss was established by analyzing attribute importance and attributereduction. The method of fuzzy theory and neural networks was study for diesel engine faultrules which was established by adaptive newwork-based fuzzy inference system.
Finally, structure of the energy loss information monitoring and diagnosis system wasstudied using virtual instrument technology. Energy loss information monitoring anddiagnosis system was designed based on virtual instrument. Energy loss information onlinemonitoring fault diagnosis system is developed by using LabVIEW which is by virtualgraphics paraphrase icon to create the application of computer programming language,including the data acquisition system and signal processing technology, so energy lossmonitoring and diagnosis system was established primarily.
引文
[1]温诗铸，黄平.摩擦学原理（第3版）[M].北京:清华大学出版社,2008:1-302.
[2]何正嘉，陈进，王太勇，等.机械设备故障诊断理论及应用[M].北京:高等教育出版社,2008:84-315.
[3]谢友柏.工程前沿（第2卷）—摩擦学科学与工程前沿[M].高等教育出版社,2005:141-155.
[4] Xie You Bo. Three Axioms in Tribology[C]. Beijing,China: Proceedings of IIIInternational Symposium on Tribo-Fatigue,2000.
[5]谢友柏.摩擦学系统理论研究和建模[J].摩擦学学报.2010,30(1):1-8.
[6]黄文虎，夏松波，刘瑞岩.机械故障诊断原理、技术及应用[M].上海:科学技术出版社,1996:47-48.
[7] Stachowiak W. G., Podsiadlog P. C. Characterization and classification of wear particlesand surfaces[J]. Wear.2001(241):149-200.
[8] Roylance B. J. Ferrography—then and now[J]. Tribology International.2005(38):857-862.
[9] Myshkin N. K., Kong H., Grigoriev A. The use of color in wear debris analysis[J]. Wear.2001(251):1218-1226.
[10] Roylance B. J., Williams J. A., Dwyer-Joyce R. Wear debris and associated wearphenomena-funamena research and practice[J]. Proceedings of the Institution ofMechanical Engineer,Joural of Engineering Tribology.2000(214):79-105.
[11] H Kuang J., D Lin A. The effect of tooth wear on the vibration spectrum of a spur gearpair[J]. ASME Journal of Vibration and Acoustics.2001,123(2):311-317.
[12] Majumdar A., Tien C. L. Fractal characterization and simulation of rough surface[J].Wear.1990,136(2):313-327.
[13]严新平，谢友柏，萧汉梁.摩擦学故障种类诊断的D-S信息融合研究[J].摩擦学学报.1999,19(02):145-150.
[14]吕植勇，严新平，彭雅芳，等.磨损磨粒的主成分聚类方法分析[J].摩擦学学报.2008,28(5):453-455.
[15]袁成清，严新平，彭中笑.磨粒的三维表面特征描述[J].摩擦学学报.2007,27(3):294-296.
[16]陈国安，葛世荣.基于分形理论的磨合磨损预测模型[J].机械工程学报.2000,36(2):29-33.
[17]曹一波，谢小鹏.基于D-S证据理论和集成神经网络的磨粒识别[J].润滑与密封.2006(5):64-67.
[18] S A Neild, P D Mcfadden, M S Williams. A review of time-frequency methods forstructural vibration analysis[J].2003(25):713-728.
[19] Wang Y. X., He Z., Zi Y. Y. Enhancement of signal denoising and multiple faultsignatures detecting in rotating machinery using duai-tree complex wavelet transform[J].2010,24(1):119-137.
[20]李志农，郝伟，韩捷.基于非线性时序模型盲辨识的因子隐Markov模型识别方法[J].机械工程学报.2007,43(1):191-195.
[21] Logan B., Moreno P. Factorial HMMs for acoustic modeling[J]. Proceedings of the IEEEInternational Conference on Acoustics, Speech and Signal Processing.1998(2):813-816.
[22]李晓虎，贾民平，许云飞.频谱分析法在齿轮箱故障诊断中的应用[J].振动、测试与诊断.2003,23(9):165-170.
[23]来五星，轩建平，史铁林，等. Wigner-Ville时频分布研究及其在齿轮故障诊断中的应用[J].振动工程学报.2003,16(2):247-250.
[24]丁康，李巍华，朱小勇.齿轮及齿轮箱故障诊断实用技术[M].北京:机械工业出版社,2005:28-33.
[25]高立新，吴丽娟，张建宇.基于EMD解调方法的齿轮早期故障诊断[J].北京工业大学学报.2009,35(7):876-881.
[26]丁康，朱小勇，陈亚华.齿轮箱典型故障振动特征与诊断策略[J].振动与冲击.2001,30(3):7-12.
[27]张雨，李岳，温熙森.基于二进小波变换的高速柴油机故障特征辨识[J].内燃机工程.1999(4):55-59.
[28]林京，屈梁生.基于连续小波变换的奇异性检测与故障诊断[J].振动工程学报.2000,36(12):95-100.
[29]李加庆，进陈，史重九.基于声全息的故障诊断方法[J].机械工程学报.2009,45(5):34-38.
[30] Deuszkiewicz P., Radkowski S. On-line condition monitoring of a power transmissionunit of a rail vehicle [J]. Mechanical Systems and Signal Processing.2003,17(6):1321-1334.
[31] Kuang J. H., D Lin A. The effect of tooth wear on the vibration spectrum of a spur gearpair[J]. ASME Journal of Vibration and Acoustics.2001,123(2):311-317.
[32] Sung C. K., Tai H M Chen. Locating defects of a gear system by the technique ofwavelet transform[J]. Mechanism and Machine Theory.2000(35):1169-1182.
[33] Baydar N., Ball A. Detection of gear deterioration under varying load conditions byusing the instantaneous power spectrum[J]. Mechanical System and Signal Processing.2000,14(6):907-921.
[34]程军圣，于德杰，杨宇.基于EMD的能量算子解调方法及其在机械设备故障诊断中的应用[J].机械工程学报.2004,40(8):115-118.
[35]杨宇，于德介，程军圣.基于EMD与神经网络的滚动轴承故障诊断方法[J].振动与冲击.2005,24(1):85-88.
[36]张雨，徐小林，张建华.设备状态监测与故障诊断的理论和实践[M].国防科技大学出版社,2000.
[37]袁东来.柴油机缸套—活塞组磨损状态识别研究[D].长沙:中南大学,2007.
[38] Jing Lin. Feature Extraction Based on Morlet Wavelet and it s Application forMechanical Fault Diagnosis[J]. Journal of Sound and Vibration.2000,234(1):135-148.
[39] Yun Jie Xu, Shu Dong Xiu. A New and Effective Method of Bearing Fault DiagnosisUsing Wavelet Packet Transform Combined with Support Vector Machine[J].2011(11):2502-2509.
[40]厉成恩.高速柴油机缸套-活塞环监测诊断技术试验研究[D].武汉:武汉理工大学,2011.
[41]叶枫桦.缸套—活塞润滑对内燃机振动影响研究[D].武汉:武汉理工大学,2010.
[42]张斌.考虑间隙的机械系统的动力学研究[D].青岛:中国海洋大学,2007.
[43] Iang Quansheng, Jia Minping, Hu Jianzhong. Manifold Lapacian Eigenmap Method forFault Diagnosis [J]. Chinese Journal of Mechanical Engineering.2008,21(3):90-93.
[44]蒋全胜.基于流形学习的机械设备故障诊断理论与方法研究[D].南京:东南大学,2009.
[45]张伟，周维佳，李斌.基于分维LLE和Fisher判别的故障诊断方法[J].仪器仪表学报.2010(2):325-333.
[46]黎敏，阳建宏，徐金梧，等.基于高维空间流形变化的设备状态趋势分析方法[J].机械工程学报.2009(2):213-218.
[47]栗茂林，王孙安，梁霖.利用非线性流形学习的轴承早期故障特征提取方法[J].西安交通大学学报.2010(5):45-49.
[48]张妮，田学民.基于等距离映射的非线性动态故障检测方法[J].上海交通大学学报.2011(8):1202-1206.
[49]邓蕾，李锋，姚金宝.基于流形学习和隐Markov模型的故障诊断[J].计算机集成制造系统.2010(10):2153-2159.
[50]李锋，汤宝平，董绍江.基于正交邻域保持嵌入特征约简的故障诊断模型[J].仪器仪表学报.2011(3):621-627.
[51] Jiang Quansheng, Jia Mi Ping, Hua Jianzhong, et al. Machinery fault diagnosis usingsupervied manifold learning[J]. Mechanical Systems and Signal Processing.2009(23):2301-2311.
[52] Li Benwei, Zhang Yun. Supervised locally liner embedding projection for machineryfault diagnosis[J]. Mechanical System and Signal Processing.2011,25(8):3125-3134.
[53] Zhang Shanwen, Chau Kwok-Wing. Dimension Reduction Using Semi-SupervisedLocally Linear Embedding for Plant Classification[J]. Emerging Intelligent ComputingTechnology and Application.2009:948-955.
[54] V. K Jain, Shashikala Tapaswi, Anupam Shukla. Location Estimation Based onSemi-Supervised Locally Liner Embedding Approach for Indoor Wireless Networks[J].Wireless Personal Communication.2012,67(4):879-893.
[55] Peng Z., Kessissoglou N. J. An integrated approach to fault diagnosis of machinery usingwear debris and vibration analysis[J]. Wear.2003(255):1221-1231.
[56] Ebersbach S., Z Peng, Kessissoglou N. J. The investigation of the condition and faults ofa spur gearbox using vibration and wear debris analysis techniques[J]. Wear.2006(240):16-24.
[57]曹一波.融合油液分析和振动分析的齿轮磨损故障诊断研究[D].广州:华南理工大学,2007.
[58]冯伟.基于基于摩擦学与动力学的齿轮系统故障诊断相关性研究[D].华南理工大学,2010.
[59] Luo R. C., Yih C. C., Su K. L. Multisensor Fusion and Integration: Approaches,Application and Future Research Directions[J]. IEEE Sensors Journal.2002,2(2):107-119.
[60] Rangwala S., Domfeld D. A. Sensor intergration using neural networks for intelligenttool condition monitoring[J]. Engineering for Industry.1990,112(8):219-228.
[61] Basir O., Yuan X. Engine fault diagnosis based on multi-sensor information fusion usingDempster-Shafer evidence theory[J]. Information Fusion.2007,8(4):379-386.
[62]郭文勇，朴甲哲，张永祥.基于多信息的柴油机缸套磨损诊断研究[J].2005,17(1):68-70.
[63]于德介，杨宇，程军圣.一种基于SVM和EMD的齿轮故障诊断方法[J].机械工程学报.2005,41(1):140-144.
[64]蒋玲莉，刘义伦，李学军，等.基于SVM与多振动信息融合的齿轮故障诊断[J].中南大学学报（自然科学版）.2010,41(6):2184-2188.
[65] Urbakh M., Klafter J., Gourdon D. The nonlinear nature of friction[J]. Nature.2004(430):520-528.
[66]许中明，黄平.摩擦微观能量耗散机理的复合振子模型研究[J].物理学报.2006,55(5):2427-2432.
[67]成濑长太郎.论齿轮的摩擦损失[J].朱永明译.机械设计与研究.1984(2):118-128.
[68] Diab Y., Ville F., Velex P. Investigations on power losses in highspeed gears[Z].2006:220,191-198.
[69]孙进才.统计能量分析（SAE）研究进展[J].自然科学进展.1998,8(2):129-136.
[70] Langley R. S. Analysis of energy flow in beams and frame and frameworks using thedirect-dynamic stiffness method [J]. Sound and Vibration.1990(136):439-452.
[71] Beshara M., Keane A. J. Vibrational energy flows between plates with compliant anddissipative coupling[J]. Sound and Vibration.1998,213(3):511-535.
[72] Park D. H., Y Hong S. Power flow models and analysis of in-plant waves in finitecoupled thin plates [J]. Sound and Vibration.2001,244(4):651-668.
[73] Nefske D. J., Sung S. H. Power flow finite element analysis of dynamic system:Basiedtheory and application to beam[J]. Transaction of thr ASME.1989,111(1):94-100.
[74] Wang A., Vlahopoulos N. Development of an energy boundary element formulation forcomputin high-frquency sound radiation from incoherent intensity boundary condition[J].Sound And Vibration.2004(278):413-436.
[75]谢小鹏，冯伟，黄墩烈.基于能量耗损的摩擦学系统状态识别方法研究[J].润滑与密封.2010,35(2):27-31.
[76]冯伟，谢小鹏，刘粲.基于能量耗损的齿轮磨损与振动相关性建模研究[J].振动、测试与诊断.2010,30(4):458-461.
[77]谢小鹏，肖海兵，冯伟.基于能量耗损的发动机故障诊断方法研究[J].润滑与密封.2011,36(5):1-3.
[78]严新平，周强.机械系统工况监测与故障诊断[M].武汉:武汉理工大学出版社,2009.
[79]郝腾飞，陈果.基于贝叶斯最优核判别分析的机械故障诊断[J].振动与冲击.2012,31(13):26-30.
[80] Bo L., Wang L., Jiao L. Feature scaling for kernel fisher discrimnant analysis usingleave-one-out cross validation[J]. Neural Compution.2006,18(4):961-978.
[81] V Hodge, J Austin. A Survey of Outlier Detection Methodologies[J]. ArtificalIntelligence Review.2004(22):85-126.
[82] Gao J., Cheng H. B., Tan P. N. Semi-supervised Outlier Detection[J]. Proceedings of theACM Symposium on Applied Computing.2006(1):635-636.
[83] Widodo A., Yang B. S. Support vector machine in machine condition monitoring andfault diagnosis[J]. Mechanical System Signal Processing.2007,21(6):2560-2574.
[84] Jie Zhang, Morris Julian. Process modelling and diagnosis using fuzzy neuralnetworks[J]. Fuzzy Sets And Systems.1996,79:127-140.
[85] Chengcai Ma, Xiaodong Gu, Yuanyuan Wang. Fault diagnosis of power electronicsystem based on fault gration and network group[J]. Neurocomputing.2009,72:2909-2914.
[86]胡寿松，周川，王源.基于小波神经网络的组合故障模式识别[J].自动化学报.2002,28(4):540-543.
[87] Dempster A. P. Upper and lower probabilities induced by a multivalued mapping[J]. TheAnnals of Mathematical Statistics.1967,38(4):325-339.
[88] Shafer G. A. Mathematical theory of evidence [M]. Princeton NJ: Princeton UniversityPress:1976:19-63.
[89] Vapnik V. An overview of statistical learning theory[J]. IEEE Transactions On NeuralNetworks.1999,10(5):988-999.
[90] Suykens J. A. K., Brabanter J. D., Lukas L., et al. Weighted least squares support vectormachine: robustness and sparse approximation[J]. Neurocomputing.2002(48):85-105.
[91] Castero D., Ranson A., Matheus J., et al. Fault dection and identification in chemicalprocesses using multivariable statistcal techniques and SVM for classification[Z]. NorthCarolina:2002165-175.
[92] Hsu C., Lin C. J. A comparison of methods for mult-class support vector machine[J].IEEE Transactions on Neural Networks.2002,13(2):415-425.
[93] Achmad Widodo, Bo-Suk Yang. Wavelect support vector machine for induction machinefault diagnosis based on transient current signal[J]. Expert Systems with Applications.2008,35:307-316.
[94]赵文浩.基于半监督学习的故障诊断研究[D].上海交通大学,2010.
[95]肖健华.智能模式与识别方法[M].广州:华南理工大学出版社,2005.
[96]肖嵘.基于支持向量机的模式识别技术中若干问题的研究[D].南京:南京大学,2001.
[97]蒋玲莉.基于核方法的旋转机械故障诊断技术与模式分析方法研究[D].长沙:中南大学,2010.
[98] C T Lin, C S G Lee. Neural-network-based fuzzy logic control and decision system[J].IEEE Transation Computer.1991,40(12):1320-1336.
[99] C J Lin, C T Lin. An ART-based fuzzy adaptive learing control network[J]. IEEETrans.Fuzzy Syst.1997,5(4):477-496.
[100] Wang Jeen-Shing, Lee C. S. George. Self-adaptive neuro-fuzzy inference systems forclassification application[J]. IEEE Trans.Fuzzy Syst.2002,10(6):790-802.
[101] H. Han C. Y. Su, U. Stepanenko. Adaptive control of a class of nonlinear systems withnonlinearly parameterized fuzzy approximators[J]. IEEE,Trans.Fuzzy Syst.2001,9(2):315-322.
[102] Z Pawlak. Rough sets and intelligent data analysis[J]. Information Sciences.2002,47(1/4):1-12.
[103] Z Pawlak. Rough Sets[J]. Internation Journal of Information and Computer Science.1982(11):341-356.
[104] Z Pawlak. Rough Sets, Rough relation and rough fundament information [J].1996,27(2,3):103-108.
[105] Mi J. S., Wu W., Z Zhang. Approaches to knowledge reduction based on variableprecision rough set model[J]. Information Sciences.2004(159):255-272.
[106] Grabmeier J., Rudolph A. Techniques of cluster algorithms in data mining[J]. DataMining and Knowledge Discovery.2002,6(4):303-360.
[107]冯志鹏，杜金莲，宋希梗，等.粗糙集与神经网络集成在故障诊断中应用研究[J].大连理工大学学报.,43(1):70-76.
[108]陈果.神经网络提取及其在转子故障中的应用研究[J].振动与冲击.2009,28(3):59-62.
[109]张绍兵.基于神经网络的规则提取与分类算法的研究[D].哈尔滨:哈尔滨工程大学,2006.
[110]戴振东，薛群基.摩擦体系结构分析和定量建模的熵探索[J].南京航空航天大学学报.2003,25(6):585-589.
[111]王瑞洲，谢小鹏，彭韦盛，等.汽车电动门锁闭锁器耐久性实验及优化研究[J].润滑与密封.2011,36(9):34-37.
[112] Chen S., Zhu Y., Zhang D., et al. Feature extraction approaches based on matrixpettern:MatPCA and MatFLDA[J].2005,26(8):1157-1167.
[113] Yang J., F Frangi A., Yang J. Y., et al. KPCA plus LDA:a complete kernel Fisherdiscriminant framework for feature extraction and recognition[J]. IEEE Transaction onPattern Analysis and Machine Intelligence.2005,27(2):230-244.
[114] Li Y. Alignment of overlapping locally scaled patches for multidimensional scaling anddimensionality reduction[J]. IEEE Transaction on Pattern Analysis and MachineIntelligence.2008,30(3):438-450.
[115] G Mark. Mercer kernel-based clustering in feature space[J]. IEEE Transactions onNeural Networks.2002,13(3):780-784.
[116]詹宇斌，殷建平，刘新旺.局部切空间与对齐的核主成分分析解释[J].计算机工程与科学.2010,32(6):158-161.
[117] Williams I. C. K. On a Connection Between Kernel PCA and Metric MultidimensionalScaling[J]. Machine Learning.2004,46(1-3):11-19.
[118] Scholkopf B., Smola A., Mullt J. Nonlinear Component Analysis as a KernelEigen-value Problem[J]. Neural Compution.1998,10(5):1299-1319.
[119] Billings S. A., Lee K. L. Nonlinear Fisher Discriminant Analysis Using a MinimumSquared Error Cost Function and the Orthogonal Least Squares Algorithm[J]. NeuralNetworks.2002,15(2):263-270.
[120] Xu Yong, Jingyu Yang, Liu Jianfeng. An Efficient Renovation on Kernel FisherDiscriminant Analysis and Face Recognition Experiment[J]. Pattern Recognition.2004,37(10):2091-2094.
[121]徐勇，陆建锋，金忠，等.改进的KMSE方法及其实现[J].模式识别与人工智能.2007,20(3):394-397.
[122] Fritzke B. Growing Cell Structures-A Self-organizing Network for Unsupervised andSupervised Learning[J]. Neural Network.1994,5(4):561-575.
[123] Hussin M. F., Kamel M. Document Clustering Using Hierarchical SOMART NeuralNetwork[J].2003,3(6):2238-2242.
[124] Tenenbaum J. B., De Silva V., Langford J. C. A global geometric frame-work fornonlinear dimensionality reduction[J]. Science.2000,290(5500):2319-2323.
[125] Roweis S., Saul L. K. Nonlinear dimensionality reduction by locally linearembedding[J]. Science.2000,290(5500):2323-2326.
[126] H S Seung, D D Lee. The manifold ways of perception[J]. Science.2000,290(5500):2268-2269.
[127] Huang Hong, Feng Hailiang, He Tongdi. Face Image Retrieval Integrating ManifoldLearning with Relevance Feedback [J]. Journal of South China University ofTechnology (Natural Science Edition).2011,39(5):91-96.
[128]赵文浩.基于半监督学习的故障诊断研究[D].上海:上海交通大学,2010.
[129]杨剑，李伏欣，王珏.一种改进的局部空切排列算法[J].软件学报.2005,16(9):1584-1590.
[130] G X Chen, Z R Zhou. Experiment observation of initiation process of friction-inducedvibration under reciprocating sliding condition[J]. Wear.2005(259):277-281.
[131] N. Hinrichs M. Oestreich K. Popp. Dynamics of oscillators with impact andfriction.Chaos [J]. Solitons and Fractals.1997,8(4):535-558.
[132] P L Ko, M C Taponat, R Pfaifer. Friction-induced vibration-with and without externaldisturbance [J]. Tribology Internation.2001(34):7-24.
[133]陈立平，张云清，任卫群，等.机械系统动力学分析及ADAMS应用教程[M].北京:清华大学出版社,2005:3-75.
[134] M Desbazeille, Randall R. B., F Guillet, et al. Model-based diagnosis of large dieselengines based on angular speed variations of the crankshaft[J]. Mechanical Systems andSignal Processing.2010(24):1529-1541.
[135] Liu Zhenyong, Cheng Xin, Ren Shiwei. Angular and temporal determinism of rotatingmachine signals: The diesel engine case[J]. Mechanical Systems and Signal Processing.2010(24):2012-2020.
[136]李葆文，张翠风，王胜.机电设备诊断原理与技术[M].广州:华南理工大学出版社,1996:190-227.
[137]肖云魁.汽车故障诊断学[M].北京:北京理工大学出版社,2001:1-235.
[138]徐勇，杨健，张大鹏.模式识别中的核方法及其应用[M].北京:国防工业出版社,2010.
[139]成署，张振任.发动机现代诊断技术[M].西安:西安交通大学,2006.
[140]李亮.基于虚拟仪器技术的柴油机智能故障诊断系统研究[D].南京:南京农业大学,2004.
[141] Thomas P. Fault detection and diagnosis in engineering systems.[J].2002,10(9):1037-1038.
[142] Zhao F., Koutsoukos X., Haussecker H., et al. Monitoring and Fault Diagnosis ofHybrid Systems[J]. Systems, Man, and Cybernetics, Part B: Cybernetics, IEEETransactions on.2005,35(6):1225-1240.
[2]何正嘉，陈进，王太勇，等.机械设备故障诊断理论及应用[M].北京:高等教育出版社,2008:84-315.
[3]谢友柏.工程前沿（第2卷）—摩擦学科学与工程前沿[M].高等教育出版社,2005:141-155.
[4] Xie You Bo. Three Axioms in Tribology[C]. Beijing,China: Proceedings of IIIInternational Symposium on Tribo-Fatigue,2000.
[5]谢友柏.摩擦学系统理论研究和建模[J].摩擦学学报.2010,30(1):1-8.
[6]黄文虎，夏松波，刘瑞岩.机械故障诊断原理、技术及应用[M].上海:科学技术出版社,1996:47-48.
[7] Stachowiak W. G., Podsiadlog P. C. Characterization and classification of wear particlesand surfaces[J]. Wear.2001(241):149-200.
[8] Roylance B. J. Ferrography—then and now[J]. Tribology International.2005(38):857-862.
[9] Myshkin N. K., Kong H., Grigoriev A. The use of color in wear debris analysis[J]. Wear.2001(251):1218-1226.
[10] Roylance B. J., Williams J. A., Dwyer-Joyce R. Wear debris and associated wearphenomena-funamena research and practice[J]. Proceedings of the Institution ofMechanical Engineer,Joural of Engineering Tribology.2000(214):79-105.
[11] H Kuang J., D Lin A. The effect of tooth wear on the vibration spectrum of a spur gearpair[J]. ASME Journal of Vibration and Acoustics.2001,123(2):311-317.
[12] Majumdar A., Tien C. L. Fractal characterization and simulation of rough surface[J].Wear.1990,136(2):313-327.
[13]严新平，谢友柏，萧汉梁.摩擦学故障种类诊断的D-S信息融合研究[J].摩擦学学报.1999,19(02):145-150.
[14]吕植勇，严新平，彭雅芳，等.磨损磨粒的主成分聚类方法分析[J].摩擦学学报.2008,28(5):453-455.
[15]袁成清，严新平，彭中笑.磨粒的三维表面特征描述[J].摩擦学学报.2007,27(3):294-296.
[16]陈国安，葛世荣.基于分形理论的磨合磨损预测模型[J].机械工程学报.2000,36(2):29-33.
[17]曹一波，谢小鹏.基于D-S证据理论和集成神经网络的磨粒识别[J].润滑与密封.2006(5):64-67.
[18] S A Neild, P D Mcfadden, M S Williams. A review of time-frequency methods forstructural vibration analysis[J].2003(25):713-728.
[19] Wang Y. X., He Z., Zi Y. Y. Enhancement of signal denoising and multiple faultsignatures detecting in rotating machinery using duai-tree complex wavelet transform[J].2010,24(1):119-137.
[20]李志农，郝伟，韩捷.基于非线性时序模型盲辨识的因子隐Markov模型识别方法[J].机械工程学报.2007,43(1):191-195.
[21] Logan B., Moreno P. Factorial HMMs for acoustic modeling[J]. Proceedings of the IEEEInternational Conference on Acoustics, Speech and Signal Processing.1998(2):813-816.
[22]李晓虎，贾民平，许云飞.频谱分析法在齿轮箱故障诊断中的应用[J].振动、测试与诊断.2003,23(9):165-170.
[23]来五星，轩建平，史铁林，等. Wigner-Ville时频分布研究及其在齿轮故障诊断中的应用[J].振动工程学报.2003,16(2):247-250.
[24]丁康，李巍华，朱小勇.齿轮及齿轮箱故障诊断实用技术[M].北京:机械工业出版社,2005:28-33.
[25]高立新，吴丽娟，张建宇.基于EMD解调方法的齿轮早期故障诊断[J].北京工业大学学报.2009,35(7):876-881.
[26]丁康，朱小勇，陈亚华.齿轮箱典型故障振动特征与诊断策略[J].振动与冲击.2001,30(3):7-12.
[27]张雨，李岳，温熙森.基于二进小波变换的高速柴油机故障特征辨识[J].内燃机工程.1999(4):55-59.
[28]林京，屈梁生.基于连续小波变换的奇异性检测与故障诊断[J].振动工程学报.2000,36(12):95-100.
[29]李加庆，进陈，史重九.基于声全息的故障诊断方法[J].机械工程学报.2009,45(5):34-38.
[30] Deuszkiewicz P., Radkowski S. On-line condition monitoring of a power transmissionunit of a rail vehicle [J]. Mechanical Systems and Signal Processing.2003,17(6):1321-1334.
[31] Kuang J. H., D Lin A. The effect of tooth wear on the vibration spectrum of a spur gearpair[J]. ASME Journal of Vibration and Acoustics.2001,123(2):311-317.
[32] Sung C. K., Tai H M Chen. Locating defects of a gear system by the technique ofwavelet transform[J]. Mechanism and Machine Theory.2000(35):1169-1182.
[33] Baydar N., Ball A. Detection of gear deterioration under varying load conditions byusing the instantaneous power spectrum[J]. Mechanical System and Signal Processing.2000,14(6):907-921.
[34]程军圣，于德杰，杨宇.基于EMD的能量算子解调方法及其在机械设备故障诊断中的应用[J].机械工程学报.2004,40(8):115-118.
[35]杨宇，于德介，程军圣.基于EMD与神经网络的滚动轴承故障诊断方法[J].振动与冲击.2005,24(1):85-88.
[36]张雨，徐小林，张建华.设备状态监测与故障诊断的理论和实践[M].国防科技大学出版社,2000.
[37]袁东来.柴油机缸套—活塞组磨损状态识别研究[D].长沙:中南大学,2007.
[38] Jing Lin. Feature Extraction Based on Morlet Wavelet and it s Application forMechanical Fault Diagnosis[J]. Journal of Sound and Vibration.2000,234(1):135-148.
[39] Yun Jie Xu, Shu Dong Xiu. A New and Effective Method of Bearing Fault DiagnosisUsing Wavelet Packet Transform Combined with Support Vector Machine[J].2011(11):2502-2509.
[40]厉成恩.高速柴油机缸套-活塞环监测诊断技术试验研究[D].武汉:武汉理工大学,2011.
[41]叶枫桦.缸套—活塞润滑对内燃机振动影响研究[D].武汉:武汉理工大学,2010.
[42]张斌.考虑间隙的机械系统的动力学研究[D].青岛:中国海洋大学,2007.
[43] Iang Quansheng, Jia Minping, Hu Jianzhong. Manifold Lapacian Eigenmap Method forFault Diagnosis [J]. Chinese Journal of Mechanical Engineering.2008,21(3):90-93.
[44]蒋全胜.基于流形学习的机械设备故障诊断理论与方法研究[D].南京:东南大学,2009.
[45]张伟，周维佳，李斌.基于分维LLE和Fisher判别的故障诊断方法[J].仪器仪表学报.2010(2):325-333.
[46]黎敏，阳建宏，徐金梧，等.基于高维空间流形变化的设备状态趋势分析方法[J].机械工程学报.2009(2):213-218.
[47]栗茂林，王孙安，梁霖.利用非线性流形学习的轴承早期故障特征提取方法[J].西安交通大学学报.2010(5):45-49.
[48]张妮，田学民.基于等距离映射的非线性动态故障检测方法[J].上海交通大学学报.2011(8):1202-1206.
[49]邓蕾，李锋，姚金宝.基于流形学习和隐Markov模型的故障诊断[J].计算机集成制造系统.2010(10):2153-2159.
[50]李锋，汤宝平，董绍江.基于正交邻域保持嵌入特征约简的故障诊断模型[J].仪器仪表学报.2011(3):621-627.
[51] Jiang Quansheng, Jia Mi Ping, Hua Jianzhong, et al. Machinery fault diagnosis usingsupervied manifold learning[J]. Mechanical Systems and Signal Processing.2009(23):2301-2311.
[52] Li Benwei, Zhang Yun. Supervised locally liner embedding projection for machineryfault diagnosis[J]. Mechanical System and Signal Processing.2011,25(8):3125-3134.
[53] Zhang Shanwen, Chau Kwok-Wing. Dimension Reduction Using Semi-SupervisedLocally Linear Embedding for Plant Classification[J]. Emerging Intelligent ComputingTechnology and Application.2009:948-955.
[54] V. K Jain, Shashikala Tapaswi, Anupam Shukla. Location Estimation Based onSemi-Supervised Locally Liner Embedding Approach for Indoor Wireless Networks[J].Wireless Personal Communication.2012,67(4):879-893.
[55] Peng Z., Kessissoglou N. J. An integrated approach to fault diagnosis of machinery usingwear debris and vibration analysis[J]. Wear.2003(255):1221-1231.
[56] Ebersbach S., Z Peng, Kessissoglou N. J. The investigation of the condition and faults ofa spur gearbox using vibration and wear debris analysis techniques[J]. Wear.2006(240):16-24.
[57]曹一波.融合油液分析和振动分析的齿轮磨损故障诊断研究[D].广州:华南理工大学,2007.
[58]冯伟.基于基于摩擦学与动力学的齿轮系统故障诊断相关性研究[D].华南理工大学,2010.
[59] Luo R. C., Yih C. C., Su K. L. Multisensor Fusion and Integration: Approaches,Application and Future Research Directions[J]. IEEE Sensors Journal.2002,2(2):107-119.
[60] Rangwala S., Domfeld D. A. Sensor intergration using neural networks for intelligenttool condition monitoring[J]. Engineering for Industry.1990,112(8):219-228.
[61] Basir O., Yuan X. Engine fault diagnosis based on multi-sensor information fusion usingDempster-Shafer evidence theory[J]. Information Fusion.2007,8(4):379-386.
[62]郭文勇，朴甲哲，张永祥.基于多信息的柴油机缸套磨损诊断研究[J].2005,17(1):68-70.
[63]于德介，杨宇，程军圣.一种基于SVM和EMD的齿轮故障诊断方法[J].机械工程学报.2005,41(1):140-144.
[64]蒋玲莉，刘义伦，李学军，等.基于SVM与多振动信息融合的齿轮故障诊断[J].中南大学学报（自然科学版）.2010,41(6):2184-2188.
[65] Urbakh M., Klafter J., Gourdon D. The nonlinear nature of friction[J]. Nature.2004(430):520-528.
[66]许中明，黄平.摩擦微观能量耗散机理的复合振子模型研究[J].物理学报.2006,55(5):2427-2432.
[67]成濑长太郎.论齿轮的摩擦损失[J].朱永明译.机械设计与研究.1984(2):118-128.
[68] Diab Y., Ville F., Velex P. Investigations on power losses in highspeed gears[Z].2006:220,191-198.
[69]孙进才.统计能量分析（SAE）研究进展[J].自然科学进展.1998,8(2):129-136.
[70] Langley R. S. Analysis of energy flow in beams and frame and frameworks using thedirect-dynamic stiffness method [J]. Sound and Vibration.1990(136):439-452.
[71] Beshara M., Keane A. J. Vibrational energy flows between plates with compliant anddissipative coupling[J]. Sound and Vibration.1998,213(3):511-535.
[72] Park D. H., Y Hong S. Power flow models and analysis of in-plant waves in finitecoupled thin plates [J]. Sound and Vibration.2001,244(4):651-668.
[73] Nefske D. J., Sung S. H. Power flow finite element analysis of dynamic system:Basiedtheory and application to beam[J]. Transaction of thr ASME.1989,111(1):94-100.
[74] Wang A., Vlahopoulos N. Development of an energy boundary element formulation forcomputin high-frquency sound radiation from incoherent intensity boundary condition[J].Sound And Vibration.2004(278):413-436.
[75]谢小鹏，冯伟，黄墩烈.基于能量耗损的摩擦学系统状态识别方法研究[J].润滑与密封.2010,35(2):27-31.
[76]冯伟，谢小鹏，刘粲.基于能量耗损的齿轮磨损与振动相关性建模研究[J].振动、测试与诊断.2010,30(4):458-461.
[77]谢小鹏，肖海兵，冯伟.基于能量耗损的发动机故障诊断方法研究[J].润滑与密封.2011,36(5):1-3.
[78]严新平，周强.机械系统工况监测与故障诊断[M].武汉:武汉理工大学出版社,2009.
[79]郝腾飞，陈果.基于贝叶斯最优核判别分析的机械故障诊断[J].振动与冲击.2012,31(13):26-30.
[80] Bo L., Wang L., Jiao L. Feature scaling for kernel fisher discrimnant analysis usingleave-one-out cross validation[J]. Neural Compution.2006,18(4):961-978.
[81] V Hodge, J Austin. A Survey of Outlier Detection Methodologies[J]. ArtificalIntelligence Review.2004(22):85-126.
[82] Gao J., Cheng H. B., Tan P. N. Semi-supervised Outlier Detection[J]. Proceedings of theACM Symposium on Applied Computing.2006(1):635-636.
[83] Widodo A., Yang B. S. Support vector machine in machine condition monitoring andfault diagnosis[J]. Mechanical System Signal Processing.2007,21(6):2560-2574.
[84] Jie Zhang, Morris Julian. Process modelling and diagnosis using fuzzy neuralnetworks[J]. Fuzzy Sets And Systems.1996,79:127-140.
[85] Chengcai Ma, Xiaodong Gu, Yuanyuan Wang. Fault diagnosis of power electronicsystem based on fault gration and network group[J]. Neurocomputing.2009,72:2909-2914.
[86]胡寿松，周川，王源.基于小波神经网络的组合故障模式识别[J].自动化学报.2002,28(4):540-543.
[87] Dempster A. P. Upper and lower probabilities induced by a multivalued mapping[J]. TheAnnals of Mathematical Statistics.1967,38(4):325-339.
[88] Shafer G. A. Mathematical theory of evidence [M]. Princeton NJ: Princeton UniversityPress:1976:19-63.
[89] Vapnik V. An overview of statistical learning theory[J]. IEEE Transactions On NeuralNetworks.1999,10(5):988-999.
[90] Suykens J. A. K., Brabanter J. D., Lukas L., et al. Weighted least squares support vectormachine: robustness and sparse approximation[J]. Neurocomputing.2002(48):85-105.
[91] Castero D., Ranson A., Matheus J., et al. Fault dection and identification in chemicalprocesses using multivariable statistcal techniques and SVM for classification[Z]. NorthCarolina:2002165-175.
[92] Hsu C., Lin C. J. A comparison of methods for mult-class support vector machine[J].IEEE Transactions on Neural Networks.2002,13(2):415-425.
[93] Achmad Widodo, Bo-Suk Yang. Wavelect support vector machine for induction machinefault diagnosis based on transient current signal[J]. Expert Systems with Applications.2008,35:307-316.
[94]赵文浩.基于半监督学习的故障诊断研究[D].上海交通大学,2010.
[95]肖健华.智能模式与识别方法[M].广州:华南理工大学出版社,2005.
[96]肖嵘.基于支持向量机的模式识别技术中若干问题的研究[D].南京:南京大学,2001.
[97]蒋玲莉.基于核方法的旋转机械故障诊断技术与模式分析方法研究[D].长沙:中南大学,2010.
[98] C T Lin, C S G Lee. Neural-network-based fuzzy logic control and decision system[J].IEEE Transation Computer.1991,40(12):1320-1336.
[99] C J Lin, C T Lin. An ART-based fuzzy adaptive learing control network[J]. IEEETrans.Fuzzy Syst.1997,5(4):477-496.
[100] Wang Jeen-Shing, Lee C. S. George. Self-adaptive neuro-fuzzy inference systems forclassification application[J]. IEEE Trans.Fuzzy Syst.2002,10(6):790-802.
[101] H. Han C. Y. Su, U. Stepanenko. Adaptive control of a class of nonlinear systems withnonlinearly parameterized fuzzy approximators[J]. IEEE,Trans.Fuzzy Syst.2001,9(2):315-322.
[102] Z Pawlak. Rough sets and intelligent data analysis[J]. Information Sciences.2002,47(1/4):1-12.
[103] Z Pawlak. Rough Sets[J]. Internation Journal of Information and Computer Science.1982(11):341-356.
[104] Z Pawlak. Rough Sets, Rough relation and rough fundament information [J].1996,27(2,3):103-108.
[105] Mi J. S., Wu W., Z Zhang. Approaches to knowledge reduction based on variableprecision rough set model[J]. Information Sciences.2004(159):255-272.
[106] Grabmeier J., Rudolph A. Techniques of cluster algorithms in data mining[J]. DataMining and Knowledge Discovery.2002,6(4):303-360.
[107]冯志鹏，杜金莲，宋希梗，等.粗糙集与神经网络集成在故障诊断中应用研究[J].大连理工大学学报.,43(1):70-76.
[108]陈果.神经网络提取及其在转子故障中的应用研究[J].振动与冲击.2009,28(3):59-62.
[109]张绍兵.基于神经网络的规则提取与分类算法的研究[D].哈尔滨:哈尔滨工程大学,2006.
[110]戴振东，薛群基.摩擦体系结构分析和定量建模的熵探索[J].南京航空航天大学学报.2003,25(6):585-589.
[111]王瑞洲，谢小鹏，彭韦盛，等.汽车电动门锁闭锁器耐久性实验及优化研究[J].润滑与密封.2011,36(9):34-37.
[112] Chen S., Zhu Y., Zhang D., et al. Feature extraction approaches based on matrixpettern:MatPCA and MatFLDA[J].2005,26(8):1157-1167.
[113] Yang J., F Frangi A., Yang J. Y., et al. KPCA plus LDA:a complete kernel Fisherdiscriminant framework for feature extraction and recognition[J]. IEEE Transaction onPattern Analysis and Machine Intelligence.2005,27(2):230-244.
[114] Li Y. Alignment of overlapping locally scaled patches for multidimensional scaling anddimensionality reduction[J]. IEEE Transaction on Pattern Analysis and MachineIntelligence.2008,30(3):438-450.
[115] G Mark. Mercer kernel-based clustering in feature space[J]. IEEE Transactions onNeural Networks.2002,13(3):780-784.
[116]詹宇斌，殷建平，刘新旺.局部切空间与对齐的核主成分分析解释[J].计算机工程与科学.2010,32(6):158-161.
[117] Williams I. C. K. On a Connection Between Kernel PCA and Metric MultidimensionalScaling[J]. Machine Learning.2004,46(1-3):11-19.
[118] Scholkopf B., Smola A., Mullt J. Nonlinear Component Analysis as a KernelEigen-value Problem[J]. Neural Compution.1998,10(5):1299-1319.
[119] Billings S. A., Lee K. L. Nonlinear Fisher Discriminant Analysis Using a MinimumSquared Error Cost Function and the Orthogonal Least Squares Algorithm[J]. NeuralNetworks.2002,15(2):263-270.
[120] Xu Yong, Jingyu Yang, Liu Jianfeng. An Efficient Renovation on Kernel FisherDiscriminant Analysis and Face Recognition Experiment[J]. Pattern Recognition.2004,37(10):2091-2094.
[121]徐勇，陆建锋，金忠，等.改进的KMSE方法及其实现[J].模式识别与人工智能.2007,20(3):394-397.
[122] Fritzke B. Growing Cell Structures-A Self-organizing Network for Unsupervised andSupervised Learning[J]. Neural Network.1994,5(4):561-575.
[123] Hussin M. F., Kamel M. Document Clustering Using Hierarchical SOMART NeuralNetwork[J].2003,3(6):2238-2242.
[124] Tenenbaum J. B., De Silva V., Langford J. C. A global geometric frame-work fornonlinear dimensionality reduction[J]. Science.2000,290(5500):2319-2323.
[125] Roweis S., Saul L. K. Nonlinear dimensionality reduction by locally linearembedding[J]. Science.2000,290(5500):2323-2326.
[126] H S Seung, D D Lee. The manifold ways of perception[J]. Science.2000,290(5500):2268-2269.
[127] Huang Hong, Feng Hailiang, He Tongdi. Face Image Retrieval Integrating ManifoldLearning with Relevance Feedback [J]. Journal of South China University ofTechnology (Natural Science Edition).2011,39(5):91-96.
[128]赵文浩.基于半监督学习的故障诊断研究[D].上海:上海交通大学,2010.
[129]杨剑，李伏欣，王珏.一种改进的局部空切排列算法[J].软件学报.2005,16(9):1584-1590.
[130] G X Chen, Z R Zhou. Experiment observation of initiation process of friction-inducedvibration under reciprocating sliding condition[J]. Wear.2005(259):277-281.
[131] N. Hinrichs M. Oestreich K. Popp. Dynamics of oscillators with impact andfriction.Chaos [J]. Solitons and Fractals.1997,8(4):535-558.
[132] P L Ko, M C Taponat, R Pfaifer. Friction-induced vibration-with and without externaldisturbance [J]. Tribology Internation.2001(34):7-24.
[133]陈立平，张云清，任卫群，等.机械系统动力学分析及ADAMS应用教程[M].北京:清华大学出版社,2005:3-75.
[134] M Desbazeille, Randall R. B., F Guillet, et al. Model-based diagnosis of large dieselengines based on angular speed variations of the crankshaft[J]. Mechanical Systems andSignal Processing.2010(24):1529-1541.
[135] Liu Zhenyong, Cheng Xin, Ren Shiwei. Angular and temporal determinism of rotatingmachine signals: The diesel engine case[J]. Mechanical Systems and Signal Processing.2010(24):2012-2020.
[136]李葆文，张翠风，王胜.机电设备诊断原理与技术[M].广州:华南理工大学出版社,1996:190-227.
[137]肖云魁.汽车故障诊断学[M].北京:北京理工大学出版社,2001:1-235.
[138]徐勇，杨健，张大鹏.模式识别中的核方法及其应用[M].北京:国防工业出版社,2010.
[139]成署，张振任.发动机现代诊断技术[M].西安:西安交通大学,2006.
[140]李亮.基于虚拟仪器技术的柴油机智能故障诊断系统研究[D].南京:南京农业大学,2004.
[141] Thomas P. Fault detection and diagnosis in engineering systems.[J].2002,10(9):1037-1038.
[142] Zhao F., Koutsoukos X., Haussecker H., et al. Monitoring and Fault Diagnosis ofHybrid Systems[J]. Systems, Man, and Cybernetics, Part B: Cybernetics, IEEETransactions on.2005,35(6):1225-1240.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |