基于啮合振动信号的齿轮寿命预测

发布时间：2018-04-28 22:20

本文选题：啮合残余信号 + 寿命预测　；参考：《南昌航空大学》2012年硕士论文

【摘要】：齿轮传动作为机械传动的主要方式，在其稳定运行中发挥着重要作用。运行过程中齿轮的寿命预测，对于确定设备大修周期、预防事故发生具有重要意义。传统的寿命预测方法多是基于疲劳损伤累积假说，需要已知外载荷及其材料的疲劳寿命曲线，主要为齿轮设计服务，不能估计由于齿轮断齿所引起的齿轮寿命减少。齿轮寿命终结是由于齿轮发生了裂纹、断齿等故障，这些故障常是从齿轮微细裂纹或材料组织形态的细微改变开始的，并不是没有任何征兆，往往在啮合振动信号中有所反映。本文在故障诊断的基础上对齿轮的剩余寿命进行预测，以齿轮振动啮合残余信号为分析对象，它是将时域同步平均信号剔除啮合频率及其谐波余下的部分做傅里叶逆变换得到的，对齿轮微细裂纹及组织形态的变化更加敏感。具体做以下研究： 1．提出基于AR模型及两样本K-S检验的齿轮故障进展指标，设计基于AR模的线性滤波器处理齿轮振动啮合残余信号得到预测误差信号，将预测误差信号进行两样本K-S检验，得到K-S检验的统计量D，以其作为齿轮故障进展指标。 2．以得到的故障进展指标为基础进行寿命预测，借助于灰色理论及马尔科夫链相结合的方法建立寿命预测模型，灰色GM（1,1）模型对故障进展总体趋势进行预测，马尔科夫链模型修正其残差。 3．搭建电能回送式齿轮故障诊断试验台及基于LABVIEW的数据采集系统，进行齿轮全生命周期试验，即齿轮从正常状态运行至完全失效，间隔采集振动数据进行剩余寿命预测试验。结果表明，基于AR模型及两样本K-S检验的齿轮故障进展指标能够及时发现齿轮早期裂纹并能较好地反映其故障进展趋势；基于灰色马尔科夫的寿命预测模型能很好的预测齿轮故障进展趋势，便于剩余寿命的预测，，该预测方法对于在线齿轮的剩余寿命预测具有一定的工程意义。
[Abstract]:Gear transmission, as the main mode of mechanical transmission, plays an important role in its stable operation. The prediction of gear life during operation is of great significance for determining the overhaul period of equipment and preventing accidents. The traditional life prediction methods are mostly based on the fatigue damage accumulation hypothesis, which requires the fatigue life curve of known external loads and materials, which is mainly used for gear design, and can not estimate the reduction of gear life caused by gear tooth breaking. The end of gear life is due to the fault of the gear, such as crack and broken tooth, which usually start from the micro crack of gear or the slight change of material structure, and it is not without any sign, and it is often reflected in the meshing vibration signal. In this paper, the residual life of gear is predicted on the basis of fault diagnosis, and the residual signal of gear vibration meshing is taken as the analysis object. It is obtained by using Fourier inverse transform to remove the meshing frequency and the remaining part of harmonic from the time-domain synchronous average signal, and is more sensitive to the change of fine crack and microstructure of gear. Do the following research: 1. A gear fault progress index based on AR model and two samples K-S test is proposed. A linear filter based on AR mode is designed to process the residual signal of gear vibration meshing to get the prediction error signal. The prediction error signal is tested by two samples K-S test. The statistical quantity D of K-S test is obtained, which is regarded as the index of gear fault progress. 2. Based on the obtained index of fault progression, the life prediction model is established by means of grey theory and Markov chain method, and the general trend of fault progress is predicted by grey GM1 / 1) model. Markov chain model modifies its residuals. 3. A power return gear fault diagnosis test-bed and a data acquisition system based on LABVIEW were built. The gear life cycle test was carried out, that is, the gear running from normal state to complete failure, and the residual life prediction test was carried out by collecting vibration data at intervals. The results show that the gear fault progress index based on AR model and two-sample K-S test can detect the early crack in time and reflect the trend of fault development. The life prediction model based on grey Markov can predict the fault trend of gear, and it is convenient to predict the residual life. This prediction method has certain engineering significance for the prediction of the remaining life of on-line gear.
【学位授予单位】：南昌航空大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH132.41;TH165.3

【相似文献】