论文部分内容阅读
针对实测转子振动信号的非平稳特性和在现实条件中难以获得大量典型故障样本的问题,提出一种基于谐波窗分解(harmonic window decomposition,HWD)、样本熵与灰色关联度相结合的故障识别方法。首先,为了降低噪声的影响,引入循环统计学的思想对传统形态滤波方法进行改进,定义了顺序形态滤波器,并结合实际选用最简单的直线结构元素,对实测转子振动信号进行顺序形态滤波降噪预处理。然后,采用不分层分析的HWD来提取包含转子典型故障信息的6个特征频带,运用非线性动力学参数样本熵作为特征,计算转子正常、不平衡、不对中、油膜涡动、油膜振荡等5种工况的样本熵。最后,由于灰色关联度分析对小样本模式识别具有良好的分类效果,以特征频带的样本熵为元素构造特征向量,通过计算不同振动信号的灰色关联度来判断转子的工作状态和故障类型。试验分析结果表明,所提的方法能够有效地应用于转子系统的故障诊断。
In view of the non-stationary characteristics of measured rotor vibration signals and the difficulty of obtaining a large number of typical fault samples in real conditions, a fault identification based on harmonic window decomposition (HWD), sample entropy and gray relational degree is proposed. method. First of all, in order to reduce the influence of noise, the idea of cyclic statistics is introduced to improve the traditional morphological filtering method. The sequential morphological filter is defined. The most straightforward linear structure element is selected to perform the sequential morphological filtering on the measured rotor vibration signal Noise pretreatment. Then, using the HWD without stratification analysis, six characteristic bands including the typical fault information of the rotor are extracted, and the normal, unbalanced, misaligned, eddy current, oil film oscillation and so on are calculated by using the sample entropy of the nonlinear dynamic parameters as the characteristic Sample Entropy of Five Operating Conditions. Finally, the gray correlation analysis has a good classification effect for small sample pattern recognition. The sample entropy of the feature band is used as an element to construct the eigenvector. The gray relational degree of different vibration signals is calculated to judge the working status and fault type of the rotor. The experimental results show that the proposed method can be effectively applied to the fault diagnosis of the rotor system.