一种变分模态分解和频域积分相结合的故障特征提取方法

A Fault Feature Extraction Method Combining Variational Mode Decomposition and Frequency Domain Integration

在对振动加速度信号进行积分时,信号中存在的不可避免的直流分量以及积分过程中产生的趋势项和误差行为,使得积分结果的精度大大降低,尤其是二次积分后,信号的频域特性丢失严重。针对此问题,提出了一种基于变分模态分解(Variational mode decomposition, VMD)和频域积分相结合的故障特征提取方法。首先基于最大峭度准则对加速度信号进行变分模态分解,根据皮尔逊相关系数法从若干分量中选取相关系数最大的分量作为最优分量,最后对最优分量进行频域二次积分,得到振动位移信号。仿真及实验结果表明该方法有效降低了趋势项及噪声等带来的干扰,提高了故障信息的辨识精度,具有较好的优越性。

Due to the inevitable DC component in the signal and the trend term and error behavior generated during the integration process, the accuracy of the integration result is greatly reduced when integrating the signal of vibration acceleration. Especially after the secondary integration, the frequency domain characteristics of the displacement signal are lost seriously. In order to solve this problem, a fault feature extraction method combining variational mode decomposition(VMD) and frequency domain integration is proposed in this paper. Firstly, the acceleration signal is decomposed by VMD, and then the component with the largest correlation coefficient is selected as the optimal component from several components according to the pearson correlation coefficient method. Finally, the optimal component is subjected to frequency domain quadratic integration to obtain the vibration displacement signal. Simulation and experimental results show that the method can effectively reduce the interference caused by trend items and noise, and improve the accuracy of fault information, which has better superiority.