改进的Elman神经网络在齿轮箱故障诊断中的应用

发布时间：2018-03-13 04:23

本文选题：齿轮箱　切入点：故障诊断　出处：《中北大学》2012年硕士论文　论文类型：学位论文

【摘要】：作为机械设备中关键部件的齿轮箱，在连接及传递动力方面起着至关重要的作用，齿轮箱一旦发生故障将会直接关系到整个设备的运转，甚至危及人身安全，因此齿轮箱故障诊断技术的研究具有重要的安全和实际工程意义。目前，随着人们对故障诊断准确度等要求的提高，使得智能领域与故障诊断的结合越来越紧密，从而为故障诊断技术的研究提供了一个新的重要手段。而以Elman人工神经网络为代表的回归神经网络( recurrent neural network, RNN)因其具有较好的动态特性而被广泛地应用于故障诊断。本文以齿轮箱为研究对象，在深入研究其内部结构和工作过程的基础上，建立了齿轮箱故障诊断系统，，利用小波包理论实现了对测量到的齿轮箱振动信号进行降噪处理，借鉴信息融合的思想成功地提取了能够反映齿轮箱运转信息的融合特征量作为神经网络的输入参量。在认真研究了Elman网络的理论框架的基础上，总结分析了Elman神经网络用于故障诊断的不足，提出了改进Elman神经网络，并将其应用于齿轮箱故障的故障诊断。本文选取了齿轮箱经常容易发生的五种故障模式，作为研究对象。建立了改进Elman神经网络、OHF Elman和Elman神经网络三种模型并进行仿真。经过一系列训练与测试，结果表明：改进后的Elman网络在稳定性及收敛速度等方面有明显优势，可以在实际工程故障诊断中应用，为齿轮箱的故障诊断技术提供了一种更有效的方法。
[Abstract]:As a key component of mechanical equipment, gearbox plays an important role in connection and transmission of power. Once the gearbox breaks down, it will directly affect the operation of the whole equipment and even endanger the personal safety. Therefore, the research of gearbox fault diagnosis technology has important safety and practical engineering significance. At present, with the improvement of the accuracy of fault diagnosis, the combination of intelligent field and fault diagnosis is becoming more and more close. Thus, it provides a new and important means for the research of fault diagnosis technology, and recurrent neural Network (RNN), represented by Elman artificial neural network, is widely used in fault diagnosis because of its good dynamic characteristics. In this paper, a gearbox fault diagnosis system is established on the basis of deeply studying the internal structure and working process of the gearbox, and the noise reduction of the measured vibration signal of the gearbox is realized by using the wavelet packet theory. Based on the idea of information fusion, the fusion characteristic quantity, which can reflect the gearbox operation information, is successfully extracted as the input parameter of the neural network. The theoretical framework of the Elman network is carefully studied. In this paper, the shortcomings of Elman neural network in fault diagnosis are summarized and analyzed, and an improved Elman neural network is put forward, which is applied to the fault diagnosis of gearbox. In this paper, five fault modes of gearbox are selected. As the research object, three models of improved Elman neural network, Elman Elman and Elman neural network, are established and simulated. After a series of training and testing, the results show that the improved Elman neural network has obvious advantages in terms of stability and convergence speed. It can be used in practical engineering fault diagnosis and provides a more effective method for gearbox fault diagnosis.
【学位授予单位】：中北大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH165.3

【相似文献】