基于深度置信网络和信息融合技术的轴承故障诊断

提出一种基于深度置信网络(DBN)和信息融合技术的轴承故障诊断新方法。首先采用集合经验模式分解将轴承振动时域信号分解为若干个固有模态函数,并分别输入至若干个DBN中进行故障状态识别,然后通过简单投票法将每个DBN识别的结果进行决策层信息融合,从而得到轴承故障的最终诊断结果。通过对单负载和多负载下不同类型和不同损伤程度的滚动轴承故障诊断进行实例分析,验证了本文方法的有效性和精确性。

Bearing fault diagnosis based on deep belief network and information fusion

A novel method for bearing fault diagnosis was presented on the basis of deep belief network (DBN) and information fusion technology. Firstly, ensemble empirical mode decomposition was applied to decompose the time domain signal of bearing vibration into several intrinsic mode functions which were separately input to DBNs for fault state identification.Then for the purpose of information fusion at decision level, the simple voting method was used to combine the diagnostic results by each DBN and obtain the final classification of bearing faults. Vibration signal datasets of rolling bearings with different fault types and damage degrees under single load and multi-load conditions were collected for algorithm validation. The fault recognition results verified the effectiveness and accuracy of the proposed method.