多域特征在行星齿轮箱局部故障识别中的应用

发布时间：2018-12-11 13:18

【摘要】：行星齿轮箱运动复杂,其振动过程中的调幅调频特性使得振动信号成分复杂且呈非线性,增加了行星齿轮箱故障识别的难度。采用相空间重构法分别构建信号时域及频域空间;通过局部均值分解及瞬时频率计算构建信号时频联合域空间,解决了经验模式分解方法对噪音敏感、存在负频率及虚假模态问题;利用奇异值分解法提取信号多域空间特征;在局部切空间排列降维基础上,改进虚假近邻点法来估计本征维数及最优k邻域,并通过局部切空间排列法对多域奇异值特征空间进行降维,获取低维流形特征,从而识别太阳轮、行星轮和齿圈的局部故障。上述理论方法经行星齿轮箱试验台及工程数据验证,有效识别出齿轮的局部故障。
[Abstract]:Because of the complex motion of planetary gearbox and the amplitude modulation and frequency modulation in the vibration process, the vibration signal is complex and nonlinear, which makes it more difficult to identify the fault of planetary gearbox. The phase space reconstruction method is used to construct the signal time domain and frequency domain respectively. The local mean decomposition and instantaneous frequency calculation are used to construct the signal time-frequency joint domain space, which solves the problem that the empirical mode decomposition method is sensitive to noise and has negative frequency and false modal problems, and the singular value decomposition method is used to extract the multi-domain spatial features of the signal. Based on the local tangent space arrangement, the pseudo-nearest neighbor method is improved to estimate the intrinsic dimension and the optimal k-neighborhood. The local tangent space permutation method is used to reduce the dimension of the multi-domain singular value feature space to obtain the low-dimensional manifold feature. To identify the sun wheel, planetary gear and gear ring local fault. The above theory and method are verified by the planetary gearbox test bench and engineering data, and the local faults of gears are effectively identified.
【作者单位】：北京科技大学机械工程学院;
【基金】：国家自然科学基金(51475038)
【分类号】：TH132.425

【相似文献】