基于自注意机制胶囊网络的行星齿轮箱故障诊断

Fault Diagnosis of Planetary Gearboxes Based on Self-attentive Mechanism Capsule Network

针对实际工程中行星齿轮箱故障数据有限、诊断准确率不高的问题,提出一种基于自注意机制胶囊网络的故障诊断方法。直接将采集到的行星齿轮箱振动信号作为输入,用首层宽卷积层提取浅层特征,过滤输入中的高频噪声;引入自注意机制关注信号关键特征;再次将所提特征输入胶囊层,进一步提取特征并实现故障分类;采用行星齿轮箱实验平台数据对所提方法进行实验验证。实验结果表明:在样本数量有限的情况下,所提方法仍能取得不错的诊断准确率。

A fault diagnosis method based on self-attentive mechanism capsule network is proposed to solve the problems of limited fault data and low diagnosis accuracy for planetary gearboxes in practical engineering.The acquired planetary gearbox vibration signal is directly used as the input to extract primary features through the first wide convolutional layer and filter the high-frequency noise in the input.The self-attentive mechanism is introduced to focus on the key features of the signal.The proposed features are input into the capsule layer to further extract features and achieve fault classification.The proposed method is verified by the data of planetary gearbox experimental platform.The results show that the proposed method can still achieve good diagnostic accuracy with limited samples.