行星齿轮振动信号分离及故障诊断技术研究

发布时间：2018-04-21 10:39

本文选题：行星齿轮 + 故障诊断　；参考：《南昌航空大学》2013年硕士论文

【摘要】：行星齿轮传动在机械设备中应用广泛，行星齿轮的运行状态对整个设备起着重要的作用，因此对行星齿轮运行状态的在线监测和故障诊断尤为重要。基于振动信号的故障诊断是齿轮故障诊断的常用方法。为此，本文提出了一种时域同步平均和窗函数相结合与改进独立分量分析方法对行星齿轮振动信号进行提取和分离。主要内容如下： 1.对行星齿轮箱相关参数的计算，找出其传动规律，根据某个特定的行星齿轮转过最小圈数后回到起始点这一特性，利用光电编码器与行星齿轮箱主轴同轴转动以获取整周期采集信号，采用时域同步平均方法进行去噪，研究窗函数泄露机理，提出时域同步平均与加窗相结合的方法对振动信号进行相加平均，分离出完整的行星轮信号； 2.为分离出有用的故障信号，研究快速独立分量分析分离原理，对该算法中的目标函数下的非线性函数进行深入分析，发现原本算法中的三个非线性函数成分复杂，软件计算缓慢，影响分离时间和效果，由于Huber-M函数的乘加特性，简单易行，因此将其替换FastICA算法中的非线性函数，得到新的HM-FastICA算法，通过实验分析，验证其分离有效性； 3.优化行星齿轮箱振动信号采集系统，，分析行星齿轮整生命周期振动信号，为较快得到破损状态下的信号，本文通过人为损坏的方法破坏一个行星轮齿，为确保采样信号的准确性，采用三个传感器对其进行采集，采用时域同步平均和窗函数对采集的振动信号进行去噪和截取，实验结果表明，该方法不仅能有效获得36个行星轮信号，还能清晰反映当前行星齿的振动情况，说明该方法的有效性； 4.采用HM-FastICA分别对时域同步平均后的行星轮振动信号的时域和频域信号进行故障分离。分离结果可知，HM-FastICA算法比FastICA算法分离出的信号效果精度更高，算法更为简便且有用信号得到一定加强，从分离后的信号图中可以看出啮合频率及其倍频附近出现的调制信号可作为故障监测的指标，分离结果能判别行星轮的工作情况，最后结合无量纲参数进行分析，从结果可知，峭度能反映当前轮齿的振动能量分布。
[Abstract]:Planetary gear transmission is widely used in mechanical equipment. The running state of planetary gear plays an important role in the whole equipment, so on-line monitoring and fault diagnosis of planetary gear is particularly important. Fault diagnosis based on vibration signal is a common method for gear fault diagnosis. In this paper, an improved independent component analysis (ICA) method is proposed to extract and separate the vibration signals of planetary gears by combining the time-domain synchronous averaging and window functions together with an improved independent component analysis (ICA) method. The main contents are as follows: 1. By calculating the relative parameters of the planetary gearbox, the transmission law is found out, according to the characteristic that a certain planetary gear turns the minimum number of circles and returns to the starting point. Using the coaxial rotation of the photoelectric encoder and the planetary gearbox spindle to obtain the whole cycle acquisition signal, the time-domain synchronous averaging method is used to de-noise, and the window function leakage mechanism is studied. The method of time-domain synchronous averaging and windowing is proposed to add average vibration signal to separate the whole planetary gear signal. 2. In order to isolate useful fault signals, the separation principle of fast independent component analysis (FICA) is studied, and the nonlinear function under the objective function of the algorithm is deeply analyzed. It is found that the three nonlinear function components in the original algorithm are complex. The calculation of the software is slow, which affects the separation time and effect. Because of the multiplicative and additive characteristics of the Huber-M function, it is easy to replace the nonlinear function in the FastICA algorithm, and a new HM-FastICA algorithm is obtained. The experimental results show that the separation is effective. 3. The vibration signal acquisition system of planetary gear box is optimized, and the whole life cycle vibration signal of planetary gear is analyzed. In order to get the signal under damaged condition quickly, a planetary gear tooth is destroyed by artificial damage method in this paper. In order to ensure the accuracy of the sampling signal, three sensors are used to collect the signal, and the time domain synchronous averaging and window function are used to de-noise and intercept the vibration signal. The experimental results show that, This method can not only effectively obtain 36 planetary gear signals, but also clearly reflect the vibration of the current planetary teeth, which shows the effectiveness of the method. 4. The time-domain and frequency-domain signals of planetary gear vibration are separated by HM-FastICA. The separation results show that the HM-FastICA algorithm is more accurate than the FastICA algorithm, and the algorithm is simpler and more useful than the FastICA algorithm. From the separated signal diagram, it can be seen that the meshing frequency and modulation signals near the octave frequency can be used as indicators for fault monitoring, and the separation results can be used to judge the working condition of the planetary gear. Finally, the dimensionless parameters are used to analyze the meshing frequency and the modulation signal near the octave frequency. Kurtosis can reflect the vibration energy distribution of the current gear teeth.
【学位授予单位】：南昌航空大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TH132.425;TH165.3

【参考文献】