基于振动信号的中速磨煤机故障识别  

Fault Identification of Medium-speed Coal Mills Based on Vibration Signals

作　　者：陈雪飞 杨锴 徐文杰 谢丹 CHENG Xuefei;YANG Kai;XU Wenjie;XIE Dan(CHN Energy Changyuan Wuhan Qingshan Co-Gengeration Co.Ltd.,Wuhan 430080,China;School of Mechanical and Electronic Engineering,Wuhan University of Technology,Wuhan 430070,China)

机构地区：[1]国能长源武汉青山热电有限公司,湖北武汉430080 [2]武汉理工大学机电工程学院,湖北武汉430070

出　　处：《数字制造科学》2023年第4期287-292,共6页

基　　金：国家自然科学基金资助项目(51775400)

摘　　要：为了深入分析磨煤机在发生不同故障时的振动异常原因,构建了磨煤机振动监测系统,并基于所采集振动信号,提出了一种K值优化的VMD-HHT边际谱结合LSSVM模型的磨煤机故障识别方法。首先,构建了磨煤机振动监测系统,并采集了振动故障信号。其次,对故障样本集进行数据预处理,利用VMD-HHT边际谱对磨煤机的振动信号进行处理,提取了不同状态下的边际谱作为故障特征,然后利用LSSVM模型对各故障特征进行决策分类。最后,通过实验证明了所提出方法的有效性。结果显示,VMD-HHT边际谱能更清晰地表达故障信息,该方法的分类准确率高于基于EMD和EEMD的方法,准确率可达96%。To thoroughly analyze the causes of vibration abnormality in the coal mill during different faults,a coal mill vibration monitoring system was constructed.Based on the collected vibration signals,a method combining K-value optimized VMD-HHT marginal spectrum with the LSSVM model is proposed for coal mill fault identification.Initially,the coal mill vibration monitoring system was established,and vibration fault signals were gathered.Then,the fault sample set underwent data preprocessing.VMD-HHT marginal spectra processed the coal mill′s vibration signals.Marginal spectra from various states were extracted as fault features.Subsequently,the LSSVM model was employed to classify each fault feature.Through experimental validation,the proposed method′s effectiveness was confirmed.Results indicate that the VMD-HHT marginal spectrum provides a clearer representation of fault information.The classification accuracy of this method surpasses those based on EMD and EEMD,achieving up to 96%.

关 键 词：振动 磨煤机 VMD-HHT 故障识别 

分 类 号：TM621[电气工程—电力系统及自动化]