基于第二代小波的在线动平衡中振动信号处理方法的研究

发布时间：2018-05-07 23:17

本文选题：振动信号 + 特征提取　；参考：《合肥工业大学》2014年硕士论文

【摘要】：在数控机床电主轴测试系统中,需要从采集的振动信号中提取动不平衡量对不平衡的电主轴进行质量补偿。通过测振传感器采集得到的动不平衡振动信号往往受到噪声严重的污染,特别是当不平衡量很小时,振动信号的信噪比差,会导致动平衡的测试精度及稳定性降低。因此针对微弱周期振动信号进行消噪处理是从强噪声干扰的振动信号中提取特征信号的有效手段。本文分析了由转子不平衡引起的振动信号的特征,构造了振动信号的数学模型。文中研究了第二代小波变换和第二代小波包的频率混叠问题,详细阐述了无频率混叠现象的自适应冗余第二代小波的构造方法。针对该方法中存在的传统阈值函数造成小波系数不连续或恒定偏差的问题,引入双变量改进阈值函数,提出了一种改进的基于自适应冗余第二代小波的振动信号消噪方法,并结合互相关分析对仿真信号进行消噪,提取动不平衡信号特征值,验证了双变量改进阈值函数给消噪效果带来的改进。针对第二代小波包存在的频率混叠问题,提出了改进的第二代小波包算法：置零小波包分解中近似系数不满足采样定理的频率成分和细节系数中混入的相邻子带的频率成分同时移频细节系数中剩下的频率成分,并恢复移频小波包重构过程中的细节系数。通过仿真实验对比分析,验证了改进的第二代小波包算法能很好的保留原信号的特征并且有效的去除噪声,有助于普罗尼算法更加准确的提取动不平衡信号的频率和相位。
[Abstract]:In the electrical spindle test system of numerical control machine tools, the quality compensation of the unbalanced electric spindle should be extracted from the vibration signal from the vibration signal. The dynamic unbalance vibration signal obtained by the vibration sensor is often polluted seriously by the noise, especially when the unequal measurement is very small, the signal to noise ratio of the vibration signal is poor and will guide the vibration signal. The precision and stability of the dynamic balance are reduced. Therefore, the noise elimination for the weak periodic vibration signal is an effective means to extract the characteristic signals from the vibration signals with strong noise interference. This paper analyzes the characteristics of the vibration signals caused by the rotor unbalance, and constructs a mathematical model of the vibration signal. In this paper, the second generation of the vibration signals is studied. The frequency aliasing problem of the wavelet transform and the second Dai Xiaobo packet is discussed in detail. The construction method of adaptive redundancy second Dai Xiaobo is described in detail. In view of the problem that the traditional threshold function in this method causes the discontinuous or constant deviation of the wavelet coefficients, the two variable quantity is introduced to improve the threshold function, and an improved basis is proposed. In the adaptive redundancy second Dai Xiaobo vibration signal denoising method, combining the cross correlation analysis to denoise the simulation signal and extract the characteristic value of the dynamic unbalanced signal, the improvement of the bivariate improved threshold function is verified. The improved second Dai Xiaobo is proposed for the frequency aliasing problem in the second Dai Xiaobo packet. Packet algorithm: the approximate coefficients in the zero wavelet packet decomposition are not satisfied with the frequency components of the frequency components of the adjacent subbands mixed with the sampling theorem and the frequency components of the frequency shift detail coefficient at the same time, and the detail coefficients of the frequency shift wavelet packet reconstruction are restored. The improved second is verified by the comparison and analysis of the imitation true experiment. The Dai Xiaobo packet algorithm can preserve the characteristics of the original signal well and effectively remove the noise. It can help the prop algorithm to extract the frequency and phase of the dynamic unbalanced signal more accurately.

【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN911.7;TG659

【相似文献】