基于小波包分解与机器学习的汽车调光电机异音识别
|
摘要
为实现汽车调光电机装置异音检测的自动化,本文采用机器学习的方法开展产品异音识别研究。在分析确定产品异音来源的基础上,采集正常件和异音件的振动信号,利用小波包分解,结合时频域分析,在能量谱和时域特征中提取10个特征向量,基于BP神经网络对200个信号样本进行机器学习分类。并对20个样件进行试验,识别汽车调光电机异音的正确率达到96.7%。研究表明,采用机器学习的方法能够有效地识别电机异音,此研究具有工程应用价值。
In order to realize the automation of abnormal sound detection in automotive dimming motor device, the method of machine learning is used to carry out the research of abnormal sound recognition. On the basis of analyzing and determining the source of distorted sound, vibration signals of normal and distorted parts are collected, and 10 feature vectors are extracted from energy spectrum and time domain features by wavelet packet decomposition combined with time-frequency domain analysis. Finally, BP neural network are used to classify 200 signal samples for machine learning. 20 samples are tested, and the correct rate of identifying the abnormal sound of the dimming motor is 96.7%. The research shows that the machine learning method can effectively recognize the motor abnormal sound, and the research has the value in engineering application.
In order to realize the automation of abnormal sound detection in automotive dimming motor device, the method of machine learning is used to carry out the research of abnormal sound recognition. On the basis of analyzing and determining the source of distorted sound, vibration signals of normal and distorted parts are collected, and 10 feature vectors are extracted from energy spectrum and time domain features by wavelet packet decomposition combined with time-frequency domain analysis. Finally, BP neural network are used to classify 200 signal samples for machine learning. 20 samples are tested, and the correct rate of identifying the abnormal sound of the dimming motor is 96.7%. The research shows that the machine learning method can effectively recognize the motor abnormal sound, and the research has the value in engineering application.
引文
[1]易子馗,谭建平,闫涛.基于BP神经网络的电机噪声在线评价系统[J].噪声与振动控制,2017,37(1):142-148.YI Z K, TAN J P, YAN T. Analysis of an online evaluation system of motor noises based on BP neural network[J]. Noise and Vibration Control, 2017, 37(1):142-148.
[2]张智胜,张云鹏,刘青.支持向量机和小波包分析下的轴承故障诊断[J].机械设计与制造,2017(3):204-207.ZHANG Z S, ZHANG Y P, LIU Q. Fault diagnosis on bearing by support vector machine and wavelet analysis[J]. Machinery Design and Manufacture, 2017, 1(3):142-148.
[3]刘东霞,史天录.微型电机噪声研究[J].噪声与振动控制,2008(1):145-147.LIU D X, SHI T L. Investigation on noise of micro-motor[J].Noise and Vibration Control, 2008(1):145-147.
[4]张宏,范琳琳.声发射信号处理及去噪方法的算法实现[J].林业机械与木工设备,2009,37(11):26-28.ZHANG H, FAN L L. Acoustic emission signal processing and realization of de-noising algorithm[J]. Forestry Machinery&Woodworking Equipment, 2009, 37(11):26-28.
[5]陈国强,彭文静.微电机噪声信号的小波包分析与故障检测[J].湖南工程学院学报,2010,4(20):24-27.CHEN G Q, PENG W J. Wavelet packet analysis and fault detection of micro-motor noise signal[J]. Journal of Hunan Institute of Engineering, 2010,4(20):24-27.
[6]孔玲军.MATLAB小波分析超级学习手册[M].北京:人民邮电出版社,2014.KONG L J. MATLAB wavelet analysis super learning manual[M].Beijing:Posts and Telecom Press, 2014.
[7]BROWN C L, REED S E, DIETZ M S, et al. Detection and classification of motor vehicle noise in a forested landscape[J].Environmental Management, 2013, 52(5):1262-1270.
[8]吴德会.一种基于支持向量机的齿轮箱故障诊断方法[J].振动、测试与诊断,2008,28(4):338-342.WU D H. A gearbox fault diagnosis method based on support vector machine[J]. Journal of Vibration, Measurement and Diagnosis,2008, 28(4):338-342.
[9]刘俊焱,尹婷婷,陈云凤,等.基于改进的NCSPSO优化SVM的木材缺陷图像分割[J].林业机械与木工设备,2014,42(10):26-29.LIU J Y, YIN T T, CHEN Y F, et al. Wood defect image segmentation based on improved NCSPSO's optimization of SVM[J]. Forestry Machinery&Woodworking Equipment, 2014, 42(10):26-29.
[10]王文剑,门昌骞.支持向量机建模与应用[M].北京:科学出版社,2014.WANG W J, MEN C Q. Support vector machine modeling and application[M]. Beijing:Science Press, 2014.
[11]曹婧华,赵飞,李聪,等.状态参数识别在齿轮箱故障诊断中的应用[J].哈尔滨理工大学报,2013,18(3):95-101.CAO J H, ZHAO F, LI C,et al. The identification of status parameters in gearbox fault diagnosis[J]. Journal of Harbin University of Science and Technology, 2013, 18(3):95-101.
[2]张智胜,张云鹏,刘青.支持向量机和小波包分析下的轴承故障诊断[J].机械设计与制造,2017(3):204-207.ZHANG Z S, ZHANG Y P, LIU Q. Fault diagnosis on bearing by support vector machine and wavelet analysis[J]. Machinery Design and Manufacture, 2017, 1(3):142-148.
[3]刘东霞,史天录.微型电机噪声研究[J].噪声与振动控制,2008(1):145-147.LIU D X, SHI T L. Investigation on noise of micro-motor[J].Noise and Vibration Control, 2008(1):145-147.
[4]张宏,范琳琳.声发射信号处理及去噪方法的算法实现[J].林业机械与木工设备,2009,37(11):26-28.ZHANG H, FAN L L. Acoustic emission signal processing and realization of de-noising algorithm[J]. Forestry Machinery&Woodworking Equipment, 2009, 37(11):26-28.
[5]陈国强,彭文静.微电机噪声信号的小波包分析与故障检测[J].湖南工程学院学报,2010,4(20):24-27.CHEN G Q, PENG W J. Wavelet packet analysis and fault detection of micro-motor noise signal[J]. Journal of Hunan Institute of Engineering, 2010,4(20):24-27.
[6]孔玲军.MATLAB小波分析超级学习手册[M].北京:人民邮电出版社,2014.KONG L J. MATLAB wavelet analysis super learning manual[M].Beijing:Posts and Telecom Press, 2014.
[7]BROWN C L, REED S E, DIETZ M S, et al. Detection and classification of motor vehicle noise in a forested landscape[J].Environmental Management, 2013, 52(5):1262-1270.
[8]吴德会.一种基于支持向量机的齿轮箱故障诊断方法[J].振动、测试与诊断,2008,28(4):338-342.WU D H. A gearbox fault diagnosis method based on support vector machine[J]. Journal of Vibration, Measurement and Diagnosis,2008, 28(4):338-342.
[9]刘俊焱,尹婷婷,陈云凤,等.基于改进的NCSPSO优化SVM的木材缺陷图像分割[J].林业机械与木工设备,2014,42(10):26-29.LIU J Y, YIN T T, CHEN Y F, et al. Wood defect image segmentation based on improved NCSPSO's optimization of SVM[J]. Forestry Machinery&Woodworking Equipment, 2014, 42(10):26-29.
[10]王文剑,门昌骞.支持向量机建模与应用[M].北京:科学出版社,2014.WANG W J, MEN C Q. Support vector machine modeling and application[M]. Beijing:Science Press, 2014.
[11]曹婧华,赵飞,李聪,等.状态参数识别在齿轮箱故障诊断中的应用[J].哈尔滨理工大学报,2013,18(3):95-101.CAO J H, ZHAO F, LI C,et al. The identification of status parameters in gearbox fault diagnosis[J]. Journal of Harbin University of Science and Technology, 2013, 18(3):95-101.