论文总字数：30647字

摘 要

随着能源危机的加剧以及社会对能源需求的扩大，风能因为它的资源丰富、可再生性和成本低的优点而受到人们的重视。风力发电就是利用风力设备将风能转变成机械能再转变成电能的一种发电方式。在整个发电过程中，风机传动系统起着重要的作用。随着风电系统的装机容量和发电量的不断增大，对风电传动系统的要求也愈来愈高。传动系统主要包括齿轮箱和滚动轴承。传统的计划维修和事后维修方式需要机组停机检修，这样造成的后续影响大，而且检修的成本高，安全性低。因为风电齿轮箱是风电机组容易发生故障的部件之一，所以进行风电齿轮箱故障特征信号提取研究很有必要。

本毕业设计课题在了解风电机组的结构和常见故障形式的前提下，研究了故障的特点，进而展开对风电齿轮箱故障特征信号提取方法的研究。本课题采用的是基于改进的EEMD技术和声发射技术的风电齿轮箱故障特征信号提取研究的方式，为风电机组齿轮箱故障特征信号的诊断识别提供了技术支持。在进行实验时，首先确保以齿轮箱故障特征信号的非平稳特性为起点，使用时域分析方法对振动信号进行分析处理；接着通过对实验数据的处理；在理解EMD分解技术原理的基础上，进一步提出用改进后的EEMD技术提取故障特征信号；最后，在研究了实验数据所得边际谱图的基础上，分析EEMD分解技术在提取特征信号方面的作用，得出该方法具有实用便捷，数据准确的优点，证明其确实是一种非常有效的故障特征信号提取方式。

关键词：故障诊断；齿轮箱；EEMD技术；声发射技术

ABSTRACT

With the intensification of the energy crisis and the expansion of the social demand for energy, wind energy has attracted people's attention because of its rich resources and the nature of sustainable development. Wind power generation is a power generation process that uses wind power equipment to convert wind energy into mechanical energy and then into electrical energy. Wind power system plays an important role in this process. With the increasing installed capacity and power generation of wind turbines, the frequency of failures of wind turbines is also increasing. The traditional method of regular inspection and post-accident maintenance requires the unit to be shut down for maintenance, which has a great impact on the follow - up. moreover, the maintenance cost is high and the safety is low. Moreover, because wind turbine gearbox is one of the most prone components of wind turbine, it is necessary to extract the fault characteristic signals of wind turbine gearbox.
On the premise of deeply understanding the structure and common fault States of wind turbine generators, this graduation project studied the characteristics of fault States, and then carried out research on the extraction method of fault characteristic signals of wind turbine gearbox. This topic adopts the method of wind turbine gearbox fault characteristic signal extraction research based on improved eemd technology and acoustic emission technology, which provides technical support for the diagnosis and identification of wind turbine gearbox fault characteristic signal. During the experiment, firstly, it is ensured that the vibration signal is analyzed and processed by using the time domain analysis method starting from the non-stationary characteristic of the gearbox fault characteristic signal. Secondly, through the processing of experimental data, the differences between vibration technology and acoustic emission technology in the fault signal judgment of gearbox are compared. Furthermore, EMD decomposition technology is studied and an improved eemd technology is proposed to extract fault feature signals. Finally, based on the research of experimental data, the function of eemd decomposition technology in reducing fault error and extracting characteristic signals is analyzed. the results show that the method has the advantages of practicality, convenience and accurate data, proving that it is indeed a very effective fault characteristic signal extraction scheme.

Key words:Fault diagnosis; gearbox EEMD technology; acoustic emission technology

目 录

摘要 I

ABSTRACT II

第一章 绪论 1

1.1 课题研究的背景及意义 1

1.1.1 风力发电系统的基本组成和结构 1

1.1.2 风力发电面临的问题 1

1.1.3 课题研究的意义 2

1.2 国内外课题研究现状 3

1.2.1 国外研究现状 3

1.2.2 国内研究现状 4

1.3 论文主要内容及安排 5

第二章 风电齿轮箱的基本结构以及工作原理 6

2.1 基本结构及工作原理 6

2.1.1 基本机构 6

2.1.2 工作原理 7

2.2 齿轮箱故障类型和特点 8

2.2.1 齿轮故障类型及特点 8

2.2.2 滚动轴承故障类型及特点 9

2.3 本章小结 11

第三章 风电传动模拟实验台以及行星齿轮箱故障制备 12

3.1 风电传动实验台简介 12

3.1.1 实验台的主要组成 12

3.1.2 风电传动系统模拟实验台 12

3.1.3 电机转速以及变频器输出功率关系 14

3.2 行星齿轮箱故障制备 15

3.3 本章小结 15

第四章 信号采集系统和振动信号时域处理 16

4.1 振动和声发射信号采集系统的参数设置 16

4.2 齿轮箱振动信号时域分析方法 18

4.3 齿轮箱振动信号时域方法研究分析 19

4.4 齿轮箱振动信号与声发射信号对比分析 21

4.5 本章小结 21

第五章 基于改进的EEMD技术和声发射技术的风机齿轮箱故障特征提取 22

5.1 EMD的简介 22

5.2 EMD技术的原理 22

5.2.1 本征模态函数 22

5.2.2 EMD分解过程 23

5.2.3 分解举例 23

5.3 改进的EEMD方法 26

5.4 实验数据分析 27

5.4.1 正常及故障齿轮箱AE信号对比 27

5.4.2 正常齿轮箱信号 27

5.4.3 断齿齿轮箱信号 28

5.4.4 结论 28

5.5 本章小结 29

第六章 结论与展望 30

参考文献 31

致 谢 33

附 录 34

剩余内容已隐藏，请支付后下载全文，论文总字数：30647字