VMD改进多尺度排列熵和LLTSA的列车车轮损伤诊断

Train Wheel Damage Diagnosis based on VMD Improved Multiscale Permutation Entropy and LLTSA

针对列车车轮损伤振动信号特征难以提取的问题,本文提出基于变分模态分解(VMD)改进多尺度排列熵和线性局部切空间排列算法(LLTSA)的车轮损伤诊断方法。首先利用VMD方法分解原始振动信号得到若干个模态分量,计算各模态分量的改进多尺度排列熵,然后采用LLTSA方法进行特征维数约简,并与等距映射流形算法(ISOMAP)降维结果对比,获得最优的低维特征向量,最后将低维特征向量作为核极限学习机(KELM)的输入进行分类辨识。实验分析结果表明,该方法能够成功识别出车轮损伤状态。

Aiming at the problem that it is difficult to extract vibration signal features of train wheel damage,a new diagnosis method of the wheel damage based on variational mode decomposition(VMD)improved multiscale permutation entropy and linear local tangent space alignment(LLTSA)is proposed in this paper.Firstly,the original vibration signal of train wheel with damage is decomposed into some modal components by VMD.The improved multiscale permutation entropy of each modal component is calculated.Then,the feature dimension is reduced using LLTSA and compared with the dimensionality reduction results of isometric mapping(ISOMAP),and the optimal low-dimensional feature vector is obtained.Finally,the feature vector is input into the kernel extreme learning machine(KELM)for classification and recognition.The experimental analysis results show that this method can identify wheel damage states successfully.