基于MRSVD与Elman神经网络的供输弹系统早期故障识别
|
摘要
对于供输弹系统早期故障信息易湮没在强噪声中,潜在故障特征难以提取的问题,提出1种基于MRSVD与Elman神经网络的早期故障识别方法。供输弹系统振动信号经过预处理后采用双树复小波进行降噪,之后利用多分辨奇异值分解对降噪信号有效分解为若干分量,提取各分量的能量,归一化后将其作为特征值,运用Elman神经网络对供输弹系统早期故障有效识别。结果表明,该方法能有效识别供输弹系统早期故障,识别率为94.44%,证明该方法对自动供输弹系统早期故障识别的有效性。
Aiming at the problem that the early failure information of the ammunition supply system is easily annihilated in strong noise,and the potential fault features are difficult to extract,an early fault identification method based on MRSVD and Elman neural network is proposed. After preprocessing,the vibration signal of the ammunition supply system is de-noised by the dual-tree complex wavelet,and then the multi-resolution singular value decomposition is used to effectively decompose the noise-reduced signal into several components. The energy of each component is extracted and normalized and used as a feature. Value,using Elman neural network to effectively identify the early failure of the ammunition supply system. The results show that this method can effectively identify the early failure of the ammunition supply system,and the recognition rate is 94.44%. It is proved that the method is effective for the early fault identification of the ammunition supply system.
Aiming at the problem that the early failure information of the ammunition supply system is easily annihilated in strong noise,and the potential fault features are difficult to extract,an early fault identification method based on MRSVD and Elman neural network is proposed. After preprocessing,the vibration signal of the ammunition supply system is de-noised by the dual-tree complex wavelet,and then the multi-resolution singular value decomposition is used to effectively decompose the noise-reduced signal into several components. The energy of each component is extracted and normalized and used as a feature. Value,using Elman neural network to effectively identify the early failure of the ammunition supply system. The results show that this method can effectively identify the early failure of the ammunition supply system,and the recognition rate is 94.44%. It is proved that the method is effective for the early fault identification of the ammunition supply system.
引文
[1]付志敏,潘宏侠,许昕,等.基于PCA-KLD的供输弹系统早期故障识别[J].机械设计与研究,2018,34(2):192-195.
[2]康郦,冯德朝.某中口径舰炮自动机设计思想与特点分析[J].火炮发射与控制学报,2010(2):61-65.
[3]李葵,范玉刚,吴建德.基于MRSVD和VPMCD的轴承故障智能诊断方法研究[J].计算机工程与应用,2016,52(8):153-157.
[4]XIE H B,ZHENG Y P,GUO J Y.Classification of the mechanomyogram signal using a wavelet packet transform and singular value decomposition for multifunction prosthesis control[J].Physiological Measurement,2009,30(5):441-457.
[5]AUSSEM A.Dynamical recurrent neural networks towards prediction and modeling of dynamical systems[J].Neurocomputing,1999,28(1/3):207-232.
[6]赵洁.基于第二代小波分析的汽轮机组故障诊断研究[D].天津:天津理工大学,2017.
[7]罗忠辉,吴百海.大型复杂机械早期故障特征提取技术研究[J].机电工程技术,2004(1):31-32.
[8]GOLAFSHAN R,SANLITURK K Y.SVD and Hankel matrix based de-noising approach for ball bearing fault detection and its assessment using artificial faults[J].Mechanical Systems&Signal Processing,2016,(S70-71):36-50.
[9]赵学智,叶邦彦,林颖.奇异值分解对轴承振动信号中调幅特征信息的提取[J].北京理工大学学报,2011,31(5):572-577.
[10]胥永刚,谢志聪,孟志鹏,等.基于奇异值分解的磁记忆信号特征提取方法[J].振动、测试与诊断,2014,34(6):1105-11.
[11]汤宝平,习建民,李锋.基于Elman神经网络的旋转机械故障诊断[J].计算机集成制造系统,2010,16(10):2148-2152.
[12]刘敏娜.改进的Elman神经网络在齿轮箱故障诊断中的应用[D].太原:中北大学,2012.
[13]田园.基于PCA-KLD与深度学习的供输弹系统故障预示研究[D].太原:中北大学,2017.
[14]康晓敏,王伟超,陈应显.基于小波能量谱的刨煤机刨链动态张力信号分析[J].机械设计,2018,35(1):46-50.
[15]张前图,房立清.基于奇异值差分谱和对称极坐标法的轴承故障特征提取[J].机械设计与研究,2015,31(5):58-61.
[2]康郦,冯德朝.某中口径舰炮自动机设计思想与特点分析[J].火炮发射与控制学报,2010(2):61-65.
[3]李葵,范玉刚,吴建德.基于MRSVD和VPMCD的轴承故障智能诊断方法研究[J].计算机工程与应用,2016,52(8):153-157.
[4]XIE H B,ZHENG Y P,GUO J Y.Classification of the mechanomyogram signal using a wavelet packet transform and singular value decomposition for multifunction prosthesis control[J].Physiological Measurement,2009,30(5):441-457.
[5]AUSSEM A.Dynamical recurrent neural networks towards prediction and modeling of dynamical systems[J].Neurocomputing,1999,28(1/3):207-232.
[6]赵洁.基于第二代小波分析的汽轮机组故障诊断研究[D].天津:天津理工大学,2017.
[7]罗忠辉,吴百海.大型复杂机械早期故障特征提取技术研究[J].机电工程技术,2004(1):31-32.
[8]GOLAFSHAN R,SANLITURK K Y.SVD and Hankel matrix based de-noising approach for ball bearing fault detection and its assessment using artificial faults[J].Mechanical Systems&Signal Processing,2016,(S70-71):36-50.
[9]赵学智,叶邦彦,林颖.奇异值分解对轴承振动信号中调幅特征信息的提取[J].北京理工大学学报,2011,31(5):572-577.
[10]胥永刚,谢志聪,孟志鹏,等.基于奇异值分解的磁记忆信号特征提取方法[J].振动、测试与诊断,2014,34(6):1105-11.
[11]汤宝平,习建民,李锋.基于Elman神经网络的旋转机械故障诊断[J].计算机集成制造系统,2010,16(10):2148-2152.
[12]刘敏娜.改进的Elman神经网络在齿轮箱故障诊断中的应用[D].太原:中北大学,2012.
[13]田园.基于PCA-KLD与深度学习的供输弹系统故障预示研究[D].太原:中北大学,2017.
[14]康晓敏,王伟超,陈应显.基于小波能量谱的刨煤机刨链动态张力信号分析[J].机械设计,2018,35(1):46-50.
[15]张前图,房立清.基于奇异值差分谱和对称极坐标法的轴承故障特征提取[J].机械设计与研究,2015,31(5):58-61.