摘要
为有效降低生产实践中旋挖钻机钻杆发生掉钻事故的概率,基于室内模型对不同钻杆状态和转速下的旋挖钻机钻杆振动情况进行试验,利用经验模态分解(EMD)原理完成了信号特征分解,并基于反向传播(BP)神经网络对故障进行诊断识别。结果表明:钻杆状态和转速对振动信号的EMD能量和EMD奇异值有较大影响,其中EMD能量主要集中在低阶次,EMD奇异值随阶次增加而逐渐减小。EMD奇异值大小排序为:正常状态>松动状态>破坏状态。基于EMD能量分布进行故障识别时,正常状态、松动状态和破坏状态的平均诊断识别正确率分别为97.2%、83.3%和83.3%。基于EMD奇异值分布进行故障识别时,正常状态、松动状态和破坏状态的平均诊断识别正确率分别为88.9%、86.1%和77.7%。基于EMD能量分布特征的故障诊断效果优于基于EMD奇异值分布特征的故障诊断效果,且对于钻杆松动状态的诊断识别效果优于破坏状态的诊断识别效果。
In order to effectively reduce the probability of drill pipe dropping accidents in rotary drilling rigs in production practice,the vibration of rotary drilling rig drill pipe under different drill pipe states and rotating speeds is tested based on the indoor model.The signal feature decomposition is completed by using the principle of empirical mode decomposition(EMD),and the fault is diagnosed and identified based on back propagation(BP)neural network.The results show that the state and speed of the drill pipe have a great influence on the EMD energy and EMD singular value of the vibration signal.The EMD energy is mainly concentrated in the low order,and the EMD singular value gradually decreases with the increase of order.The order of EMD singular value is normal state>loose state>damaged state.When fault identification is carried out based on EMD energy distribution,the average diagnostic accuracy of the normal state,loose state and damaged state is 97.2%,83.3%and 83.3%respectively.When fault identification is carried out based on EMD singular value distribution,the average diagnostic accuracy of the normal state,loose state and damaged state is 88.9%,86.1%and 77.7%respectively.The fault diagnosis effect based on EMD energy distribution is better than that based on EMD singular value distribution,and the diagnosis and identification effect of the drill pipe loose state is better than that of the broken state.
作者
庄国昌
ZHUANG Guochang(Project Department,the Fourth Engineering Co.,Ltd.of China Railway 18th Bureau Group,Tianjin 300350,China)
出处
《上海电机学院学报》
2023年第2期69-73,共5页
Journal of Shanghai Dianji University
