基于小波包变换预处理的铝电解槽针振信息元分析与诊断  

Analysis and diagnosis of information elements of noise in aluminum reduction cells based on wavelet packet preconditioning

作　　者：李贺松[1] 梅炽[1] 黄涌波 唐骞 周乃君[1] 姜昌伟[1] 

机构地区：[1]中南大学能源与动力工程学院 [2]中铝公司平果分公司

出　　处：《轻金属》2006年第2期36-39,共4页Light Metals

摘　　要：简述了小波包变换的基本原理及利用小波包对电压信号进行分解的方法。针对铝电解槽电压波动信号的频谱特点,采用小波包分析方法提取了电压信号的特征向量,将信号分解到8个频段内,进行预处理得到频段能量特征向量。应用BP神经网络建立了特征向量到振针信息元之间的映射。仿真结果表明,小波包分析能够有效地将隐藏在正常电压信号之中的早期弱故障信号提取出来,从而发现槽子的早期不良症状。The theory of wavelet packet transform and the method of voltage signal analysis with wavelet packet was presented. The wavelet packet analysis was used to abstract the characteristic of signal according to the frequency spectrum characteristics of voltage vibration signal of aluminum reduction cells. The signals were decomposed into eight frequency bands and the information preconditioned was used as an energy characteristic vector. BP neural network was used to realize the map between feature and information element of noises. The simulation results show that the initial fault information hidden in the voltage vibration signals can be extracted effectively by wavelet packet analysis, so the bearing initial fault can be detected in the early stage.

关 键 词：铝电解槽 针振 小波包分析 BP神经网络 变换预处理 

分 类 号：TF351[冶金工程—冶金机械及自动化]