Research on the diagnosis method of abnormal faults in the drum brake device of the tobacco production line belt conveyor

GUI Tengyue; SHAO Jianming; WAN Zhonghai; ZHANG Zhiming; WU Jiansheng; ZHANG Kuangshi

doi:10.16578/j.issn.1004.2539.2026.04.019

Abstract：[Objective]The vibration signal of the drum brake device may contain multiple frequency components,which may overlap with each other,making it difficult to separate the fault characteristics from the background noise.In response to the above issues,a diagnostic method for abnormal faults in the drum brake device of a tobacco production line belt conveyor was studied.[Methods]Firstly,the vibration signal of the drum brake device was obtained using a vibration sensor,and the impact fault and background noise in the signal were distinguished through spectral kurtosis analysis.The Wiener filter combined with spectral kurtosis optimization parameters was used to recover the original vibration signal from the noisy vibration signal,and the noise signal was removed by setting a threshold.Secondly,the variational mode decomposition(VMD)algorithm was used to decompose the filtered vibration signal and multiple intrinsic mode functions(IMFs)that describe the signal components in different frequency bands were obtained.Thirdly,complex vibration signals containing multiple frequency components were decomposed into several modal components,each corresponding to a specific frequency bandwidth,making fault features easier to identify and extract.Finally,by calculating the kurtosis value,power spectrum,and centroid frequency of each IMF component,the characteristics of the fault signal of the drum brake device were extracted and input into the self-organizing map(SOM)network to achieve abnormal fault diagnosis of the drum brake device.[Results]The test results show that this method can diagnose multiple types of faults in the drum brake device,which are the same as the preset fault types,and there is no misdiagnosis.