基于小波包分解的回转窑筒体故障提取研究

Research on Fault Extraction of Rotary Kiln Shell Based on Wavelet Packet Decomposition

为了诊断回转窑工作故障和评估窑运行状况,有效提取窑筒体故障的特征信号极为重要。通过分析故障状态下窑筒体与托轮之间受力关系,建立托轮振动模型,得出窑故障与托轮位移振动的关联关系。针对现有窑筒体故障特征信息提取方法的不足,提出基于小波包分解的特征频率提取方法,对实际采集的数据进行小波包分解和提取特征频段进行重构。对重构后的数据进行Hilbert分析表明,采用小波包分解方法在托轮位移信号中提取2个窑故障特征频率,即筒体工作频率(KH)与托轮工作频率(RH),并以KH和RH的能量密度作为评估参数来分别反映筒体弯曲和各托轮超载受力的故障程度。通过对实测回转窑托轮信号进行处理,表明所提出方法有效,从而为后续研究回转窑运行故障的在线监测提供了新思路。

In order to diagnose the working fault of rotary kiln and evaluate the operation condition of rotary kiln,it is very important to extract the characteristic signal of kiln shell fault effectively.By analyzing the stress relationship between kiln shell and supporting roller under fault condition,the supporting roller vibration model is established,and the relationship between kiln fault and supporting roller displacement vibration is obtained.Aiming at the shortcomings of the existing methods for extracting the fault feature information of kiln shell,this paper proposes a feature frequency extraction method based on wavelet packet decomposition,which decomposes the actual data and reconstructs the feature frequency band.The Hilbert analysis of the reconstructed data shows that the wavelet packet decomposition method is used to extract two kiln fault characteristic frequencies,namely,the kiln shell working frequency(KH)and the supporting roller working frequency(RH),and the energy density of KH and RH is used as the evaluation parameters to reflect the fault degree of the cylinder bending and the overload of each supporting roller.The results show that the proposed method is effective,which provides a new idea for the follow-up study of online monitoring of rotary kiln operation fault.