机械故障特征信息提取的ICA信息融合方法

ICA Fusion for Feature Extraction of Mechanical Fault

峭度和负熵是盲信号独立性的两个自然测度,可以被用来捕捉机械振动信号信息的动态变化特征,并提取机械设备的故障特征信息。峭度和负熵是从两个不同的角度和层面阐释机械设备的故障特征信息,信息量是互补的。若将峭度信息和负熵信息融合,则必然能够更全面、更深刻地来表征机械设备的状态。因此引入信息融合的思想,提出基于ICA信息融合的机械故障特征信息提取方法,综合峭度和负熵信息来提取机械设备的故障特征信息。液压齿轮泵模式识别试验表明,该方法可以应用于机械设备的故障特征信息提取。

The kurtosis and negentropy,as two natural measures of independence for blind signals,can be utilized to capture the dynamic information characteristics of mechanical vibration signals. The dynamic information can be extracted as their fault features. The kurtosis and negentropy can explain the feature information of mechanical fault from two different viewpoints that the and their information contents are complementary mutually. If the kurtosis information and negentropy information were fused into one feature vector,it can definitely express the running states of machines more comprehensively and more profoundly. A feature extraction method based on ICA fusion for mechanical fault was proposed,which can fuse the kurtosis information and negentropy information as the final optimum feature information. The pattern recognition experiments of hydraulic gear pump indicate that this method can be applied to feature extraction of mechanical equipment.