基于切片双谱分析的滚动轴承故障识别研究

Roller bearing fault identification based on slice bi-spectrum analysis

针对滚动轴承故障识别问题,提出基于共振解调理论和高阶谱分析相结合的新方法.讨论了双谱理论和双谱对角切片的特性,分析了故障滚动轴承振动信号的特点.利用希尔伯特(Hilbert)变换构造原始振动信号的解析信号,求取信号的包络,计算包络信号的对角切片双谱,提取由于滚动轴承二次相位耦合产生的非线性特征,得到滚动轴承的故障特征频率,识别滚动轴承的故障模式.试验和理论分析表明,利用该方法可以有效地将正常轴承和内滚道伤轴承、外滚道伤轴承的振动信号进行分离.该方法保留了信号的相位信息,并且能够有效地抑制噪声,在降维的同时简化了算法.

A new method based on resonance demodulation theory and high-order spectrum analysis was presented for identifying roller bearing fault. By discussing the bi-spectrum theory and diagonal slice attributes of bi-spectrum, the characteristics of defective roller bearing's vibration signal were analyzed. The original vibration signal's analytical signal and the original vibration signal's envelope signal were acquired by Hilbert transform. A diagonal slice of the envelope signal bi-spectrum was calculated and the non-linear feature deriving from the quadratic phase coupling was extracted. The fault characteristic frequency was obtained and the roller bearing's fault pattern was identified. Experiments analysis show that the roller bearing's vibration signal with normal, inner rings defect and outer rings defect are identified effectively. This method preserves the signal's phase information and inhibits the noise effectively. The calculation dimension is reduced and the algorithm is simplified by using this method.