基于小波包分解与随机森林的离心泵故障诊断

基于小波包分解与随机森林的离心泵故障诊断

马飞,邵礼光,徐君,陶梦秋,袁沛,胡炳涛

Centrifugal pump fault diagnosis based on wavelet pack decomposition and random forest

Fei MA,Liguang SHAO,Jun XU,Mengqiu TAO,Pei YUAN,Bingtao HU

表1 振动信号时域统计特征

Table 1 Time-domain statistical features of vibration signal

特征量	计算式
有效值	$x_{r m s} = \sqrt[]{\frac{1}{M} \sum_{m = 1}^{M} x_{m}^{2}}$
均值	$\bar{x} = \frac{1}{M} {\sum_{m = 1}^{M} x}_{m}$
标准差	$σ_{x} = \sqrt[]{\frac{1}{M - 1} \sum_{m = 1}^{M} {(x_{m} - \bar{x})}^{2}}$
峰值因子	$C_{f} = \frac{m a x (\|x_{m}\|)}{x_{r m s}}$
波形因子	$S_{f} = \frac{x_{r m s}}{\frac{1}{M} \sum_{m = 1}^{M} \|x_{m}\|}$
峭度	$K_{r} = \frac{\sum_{m = 1}^{M} {(x_{m} - \bar{x})}^{4}}{M σ_{x}^{4}}$
偏度	$S_{k} = \frac{\sum_{m = 1}^{M} {(x_{m} - \bar{x})}^{3}}{M σ_{x}^{3}}$
扰动因子	$P_{f} = \frac{σ_{x}}{\bar{x}}$