基于小样本音频信号的柱塞泵故障诊断  

Fault Diagnosis of Plunger Pumps Based on Audio Signal ofSmall Samples

作　　者：刘禹 戴永寿[1] 李立刚[1] LIU Yu;DAI Yongshou;LI Ligang(College of Oceanography and Space Informatics,China University of Petroleum(East China),Qingdao 266580,Shandong,China)

机构地区：[1]中国石油大学(华东)海洋与空间信息学院,山东青岛266580

出　　处：《噪声与振动控制》2023年第5期142-147,273,共7页Noise and Vibration Control

基　　金：国家自然科学基金资助项目(41974144)。

摘　　要：针对音频信号在柱塞泵故障诊断中存在的样本数量不足、故障特征微弱等问题,提出一种基于音频信号结合元迁移学习(Meta-transfer Learning,MTL)的柱塞泵故障诊断方法(Fault Diagnosis of Plunger Pump Based on MTL,MTL-PAFD)。该方法以柱塞泵的音频信号为样本,在单一传感器条件下,通过Gammatone滤波器组对信号进行处理,可有效提高强噪声干扰下音频信号的表征能力,然后结合元迁移学习,能实现小样本条件下的柱塞泵故障诊断。同时,根据柱塞泵故障诊断的实际需求,改进元迁移学习在故障诊断应用中的测试方法,能够自适应处理未知故障类。实验结果表明,MTL-PAFD仅对已知类别的故障诊断准确率可达到91.41%,而经过快速自适应学习后,其在识别未知故障类时准确率能达到89.64%。The problems of insufficient samples and weak fault features of audio signals in the fault diagnosis of plunger pumps are studied.A fault diagnosis method of plunger pumps based on audio signal and meta-transfer learning(MTL-PAFD)is proposed.The method takes the audio signals of the plunger pump as samples,which are acquired by a single sensor.Through the Gammatone filter bank processing,the representation ability of the audio signal under strong noise interference is effectively improved.Then,combined with meta-transfer learning,the few-shot fault diagnosis of the plunger pump is realized.In addition,according to the actual needs of fault diagnosis of plunger pump,the test method of meta-transfer learning applied in the fault diagnosis is improved,which can process unknown fault classes adaptively.Experimental results show that the accuracy of the MTL-PAFD method can reach 91.41%for diagnosis of known fault classes.After fast adaptive learning,this method can achieve an accuracy of 89.64%when identifying unknown fault classes.

关 键 词：故障诊断 柱塞泵 音频信号 小样本学习 元学习 

分 类 号：TH3[机械工程—机械制造及自动化] TH165.3