童鞋论文网专业提供代写本科论文、代写硕士论文、代写职称论文、代写毕业论文、论文发表等服务,上万成功案例,欢迎咨询。 网站地图  |  加入收藏  |  设为首页
童鞋论文网
热门搜索
行政 教育 营销 法律 经济 护理 心理 会计 文学 工商 计算机 旅游
站内搜索:
童鞋论文网
 当前位置:童鞋论文网 > 论文发表 > 文章正文
 
基于振动信号分析的齿轮状态监测方法研究
论文作者:童鞋论文网  论文来源:www.txlunwenw.com  发布时间:2019/9/26 10:28:31  

摘要:随着科学技术的不断进步,智能制造日益成为未来制造业发展引领趋势,因而在机械生产过程中,国家对机械设备的安全性、可靠性等各方面综合性能提出了新要求。齿轮作为机械传动系统不可或缺的重要部件,因其发生故障而导致的机械设备损坏及人员伤亡事件时有发生。因此,若能够及时有效地监测齿轮的运行状态,对保障机械系统的安全运行以及避免重大事故的发生具有非常重要的现实意义。本文以齿轮故障为研究对象,开展了深入的分析研究,主要内容包括以下三个方面:

1.基于小波包和独立成分分析相结合的特征提取以及基于变分模态分解模糊熵和样本熵相结合的齿轮振动信号特征提取方法研究。本文从信号特征提取优化的角度提出了一种基于小波包(Wavelet packet,简称WP)和独立成分分析(Independent component analysis,简称ICA)相结合的特征提取方法,该方法充分融合WP和ICA特征提取方法的优势特点;同时提出了一种基于变分模态分解(VMD)模糊熵和样本熵相结合的特征提取融合方法,通过将两种熵值进行融合,从而提取更能够全面反映齿轮运行状态的特征信息。

2.多方向-多方法的特征融合方法研究。针对单一方向特征难以全面表征齿轮运行状态的问题,提出了一种基于多方向-多方法的特征融合方法,该方法充分利用不同方向信号以及不同特征提取方法的优势特点。首先,利用WP、VMD、EEMD(Ensemble empirical mode decomposition)三种特征提取方法分别提取X、Y、Z三个振动方向上齿轮振动信号的特征,从而可获得9组基特征向量;其次,将任意两个故障诊断辨识率以及不同振动方向之间和不同特征提取方法之间的互补性作为模糊逻辑的输入参数,根据模糊逻辑值的大小确定两方向上最佳的融合模式;最后,根据振动方向上信号的振动强弱确定主轴和辅助轴,根据平均故障诊断辨识率的值确定主轴的特征提取方法,考虑不同振动方向之间以及不同特征提取方法之间的互补性和提取的特征维度大小,从而确定辅助轴的特征提取方法,最终确定三方上最佳的融合模式。

3.基于核函数融合的SVM分类方法以及基于参数优化的BP神经网络的故障诊断分类方法研究。从分类网络训练模型融合角度提出了一种基于核函数融合的SVM分类方法,用训练样本训练SVM核函数网络分类模型;将具有最佳自分类精度的两种核函数网络分类模型挑选出来,并以具有最高自分类精度的核函数网络模型对待测试样本进行分类;然后利用SVM工具箱的输出参数所设定的取值范围,挑选出满足设定条件的测试样本;最后将挑选出的测试样本用另一种核函数模型重新进行分类。同时提出了一种基于参数优化的BP神经网络分类方法,对于四分类的故障分类问题,通过构造六个二分类网络模型,从六个二分类模型中挑选更佳的初始化权值和阈值参数组合,进而进一步优化四分类BP神经网络分类模型。

为充分验证上述所提方法的有效性,进行了相关实验分析研究,实验研究表明,所提方法能够实现更高精度的齿轮故障诊断。

With the continuous innovation of sciencetechnology, intelligent manufacturing has become the leading trend ofmanufacturing fields in the future. Therefore, in the process of mechanicalproduction, our government has put forward new requirements for thecomprehensive performance of mechanical equipment, such as safety andreliability, and so on. Gear is an indispensable part of the mechanicaltransmission system, and mechanical equipment damage and casualties usuallyoccur due to malfunctions. Therefore, if the running state of the gear can bemonitored timely, it is of great practical significance to ensure the safeoperation of the mechanical system and to avoid the occurrence of majoraccidents. In this paper, the gear fault is taken as the research object anddeep researches have been carried out. The main research contents include thethree aspects.

1.Feature extraction combination based onwavelet packet(WP)andindependent component analysis(ICA)and gear vibration signal feature extraction based on combination offuzzy entropy and sample entropy of variation mode decomposition(VMD). From the perspective of signal featureextraction optimization, a feature extraction method that is based on thecombination of WP and ICA is proposed. The method fully combines the advantagesof WP and ICA feature extraction method. In addition, a feature extractionfusion method of fuzzy entropy and sample entropy is proposed on the basis ofVMD method. By the way of entropy values fusion, the acquired fusion featurecan fully reflect the running state of the gear.

Multi-directional-multi-method featurefusion method research. It is difficult to fully present gear running state byusing the single vibration direction feature. Therefor, amulti-directional-multi-method feature fusion method is proposed, which makesfull use of the advantages of different direction signals and different featureextraction methods. Firstly, WP, VMD and EEMD feature extraction methods areadopted to extract the feature of the gear vibration signals of the X, Y and Zaxes respectively, so 9 groups of base feature vectors can be obtained.Secondly, any two of the fault diagnosis identification rates and thecomplementary performance between different vibration directions and differentfeature extraction methods are used for input parameters of fuzzy logic unit,and the best fusion mode is determined by output value of fuzzy logic in twodirections. Finally, according to the vibration strength to determine the mainaxis and the auxiliary axis, the feature extraction method of the main axis isdetermined according to the value of the average fault diagnosis identificationrate, and the complementary performance between the different vibrationdirections and different feature extraction methods , also in the view of thefeature dimension, the feature extraction methods of auxiliary axes aredetermined , by which the best fusion mode can be obtained in the threedirections.

3.The method research of SVM classificationmethod based on kernel function fusion and BP neural network classificationmethod based on parameter optimization. From the perspective of classificationnetwork training model fusion, the kernel function fusion of SVM is proposed.The SVM kernel function network classification models are trained with trainingsamples, then the two kernel function network classification models with thebest self-classification accuracy are selected and the test samples areclassified by the kernel function network model with the highestself-classification accuracy. According to the range values of the outputparameters of SVM toolbox, the test samples that satisfy the set conditions areselected. Finally, the selected test samples are reclassified by the otherkernel function network model. At the same time, from the perspective ofparameter optimization, the BP neural network classification method based onparameter optimization is proposed.

关键词:齿轮;故障诊断;状态监测;特征提取;模式分类

gear;faultdiagnosis;condition monitoring;featureextraction;mode classification

上一篇:阀体气密性智能检测与监控系统的研发     下一篇:超声波气体流量计整流器技术研究
 
论文客服 95702844
论文热线18796337551
本站网址 www.txlunwenw.com
郑重承诺 原创,包修改,包通过!
童鞋论文网真诚欢迎新老客户的光临与惠顾!
 
 热门文章
· 住宅建设用地使用权续期研究
· 用户锁定效应视角下的网易云音
· 众筹产品创新性对融资绩效的影
· 南京市城镇居民冷鲜肉认知度问
· 摩拜共享单车可持续盈利研究
· 催生地方政府隐性债务风险的原
· 社交媒体中不实信息转发主体心
· 基于网络平台的SPA模式研究——
· 基于游客感知的云台山景区微信
· 新媒体时代背景下乡村旅游企业
 
 推荐文章
· 部编版小学语文文言文类文章编
· F模型下“嘉欣丝绸”财务风险研
· 基于4V理论的云南嘉华饼屋营销
· 论商业性小屏直播的主持互动策
· 新兴偶像经济发展现状及发展建
· 辅助线在小学生空间观念培育中
· 罗浮山国家级风景名胜区探险旅
· 基于《预算法》的地方政府债务
· 商品房买卖合同“霸王条款”存
· 中泰水果贸易竞争力比较分析
 
 相关文章
· 环保验收的噪声监测方法研究
· 大气环境无人机监测方法探讨
· 社会历史之谜的解答——《<政治
· 近代西方哲学中的自我观念及其
· 文化产权交易市场研究综述
· 汉代“街弹”问题研究
· 海峡两岸妈祖文化互动交流意义
· 群众戏剧需要走出自娱自乐的怪
· 戏剧表演中形体语言的重要地位
· 《地心引力》——一部有故事的

论文原创,准时交稿,包修改,包通过,八年信誉,品质保证
童鞋论文网提供代写本科论文,代写专科论文,代写毕业论文,代写职称论文,代写硕士论文,论文发表及查重服务
备案号:苏ICP备11011187号-1 苏公网安备32030502000117号
版权归童鞋论文网所有 禁止一切商用盗用