基于MEEMD与相关分析的行星齿轮箱测点优化

Optimization of measuring points of planetary gear box based on meemd and correlation analysis

机械传动设备行星齿轮箱极易发生故障,为了减少行星齿轮箱故障诊断中传感器的布置数量,进而降低成本。提出了一种基于改进的集成经验模态分解(MEEMD)信息熵与相关分析相结合的行星齿轮箱测点优化方法。首先,使用MEEMD算法分解5种工况的振动测试信号。其次,用分解出的各分量与原始数据之间的相关系数筛选出包含主要故障信息的IMF分量,并提取它们的信息熵特征,构造成样本特征向量。最后,控制同一工况不同测点以及同一测点不同工况的信息熵特征向量进行相关性分析。分析出相对冗杂多余的测点进行筛选剔除,达到测点优化的目标。

Planetary gearboxes are highly susceptible to failure as mechanical transmission equipment.In order to reduce the number of sensor arrangements in planetary gearbox fault diagnosis such that reduce the cost,planetary gearbox measurement point optimization method based on the multi-dimensional ensemble empirical mode decomposition(MEEMD)information entropy combined with correlation analysis is proposed.Firstly,the vibration test signals for the five operating conditions are decomposed using MEEMD.Secondly,the correlation coefficients between the decomposed components and the raw data are used to filter out the IMF components containing the main fault information,and their information entropy features are extracted to construct the sample feature vector.Finally,the information entropy eigenvectors of different measurement points of the same operating condition and different operating conditions of the same measurement point are controlled for correlation analysis.The relatively redundant measurement points are analyzed and eliminated to achieve the goal of measurement point optimization.