A fault diagnosis method for rolling bearing was proposed based on MLMW denoised CWT gray moment vector-LSSVM.Sampled signals were denoised by MLMW, and the CWT scalogram of denoised signals was partitioned into several areas, whose gray moment were calculated to form a gray moment vector that was used as the input of LSSVM for fault classification. The experimental results show fault features of MLMW denoised signal’s scalogram is clearer and more distinctive than original,and its gray moment vector can describe bearing condition effectively.With the number of scalogram partitions increasing,the diagnosis accurate rate increases.Compared with diagnosis methods of original gray moment vector-LSSVM and wavelet denoised gray moment vector-LSSVM,the proposed method has higher accuracy,better generalization and requires fewer training samples,and can identify bearing fault patterns accurately.

gray moment vector, least square support vector machine(LSSVM), max-lifting morphological wavelet(MLMW), rolling bearing

提出了基于极大提升形态小波(MLMW)降噪的CWT灰度矩向量-LSSVM的轴承故障诊断方法。先利用MLMW对信号进行降噪处理,再将降噪信号的CWT灰度图划分为若干区域,计算各分区的灰度矩组成灰度矩向量,将其作为LSSVM的输入进行故障分类。试验结果表明:相对于原始信号的灰度图,MLMW降噪后的灰度图特征突出、区分显著,相应的灰度矩向量可有效刻画轴承状态;随着分区数增加,诊断准确率升高;相对于原始灰度矩向量-LSSVM方法和小波降噪的灰度矩向量-LSSVM方法,所提出方法准确率高、泛化性好、所需训练样本少,可准确识别轴承故障类型。

灰度矩向量, 最小二乘支持向量机(LSSVM), 极大提升形态小波(MLMW), 滚动轴承