基于FRFT的单分量阶比双谱提取微弱故障特征

提出了一种基于分数阶傅里叶变换(fractional Fourier transform,简称FRFT)的单分量阶比双谱分析方法,消除阶比双谱分析多分量信号时产生的交叉项,提取变速器齿轮微弱故障特征。根据变速器输入轴转速信号及传动比确定FRFT最佳阶次,对变速器升速过程振动信号进行最佳阶次FRFT,在该分数阶域分离目标阶比分量,对分离出的单分量信号分别进行阶比双谱分析,并累加各分量阶比双谱结果得到基于FRFT的单分量阶比双谱。试验结果表明,变速器变速过程振动信号为多阶比分量信号,直接对其进行阶比双谱分析会产生明显的交叉项,使阶比双谱阶次和幅值失真。基于FRFT的单分量阶比双谱方法能有效屏蔽其他分量和噪声干扰、消除交叉项,真实、准确反映被分析信号的阶比双谱,有效提取变速器齿轮微弱故障特征。

Weak Fault Characteristic Extraction by Single Component Order Bispectrum Based on Fractional Fourier Transform

A single component order bispectrum method based on Fractional Fourier Transform(FRFT) is presented and applied to eliminate the intercrossed items created by analysis of multiple components signal by order bispectrum, and the weak fault characteristics of gear box are extracted. The best FRFT transform order is calculated according to the rotate speed signal of the gearbox’s input axis and the gear ratio, and the vibration signal of gearbox’s accelerated running process is analyzed by FRFT with the best FRFT transform order, the order components are separated into single component signals in that FRFT field and analyzed by order bispectrum respectively, then those order bispectrum results are added to make the single component order bispectrum based on FRFT. The experimental results show that the vibration signal of gearbox contains multiple components, if that signal is analyzed by order bispectrum directly, intercrossed items will be obvious and the order and order bispectrum are inaccurate; if that vibration signal is analyzed by the single component order bispectrum based on FRFT, the other components and noise are shielded, the intercrossed items are eliminated, the order bispectrum of the vibration signal is true and accurate, the weak fault characteristics of gearbox are extracted effectively.