基于EMD-ICA和HMM的风机故障分类方法

摘要

Hidden Markov Model (HMM) is a kind of method for patterns clustering and recognition,Independent component analysis (ICA) is a powerful tool for analyzing nongaussian data.In ICA,Four order cumulant is a kind of robust and steady algorithm,which is specially appropriate to feature extraction of vibration signal.But independent component analysis has certain premise condition,with the aid of empirical mode decomposition (EMD),remove high frequency IMF part,eliminate noise interference,make signal satisfy the ICA condition.In this paper,the ICA is proposed for feature extraction of different state patterns (including normal and rotor misalignment) of one iron mill blower,HMM realize the final classification.Contrast the results classification experiments showed that the compound EMD-ICA and HMM classifier can be constructed in simpler way,and classify various fault patterns at considerable accuracy,both of which imply its great potential in health condition monitoring of machines.%隐马尔可夫模型(HMM)是一种模式聚类和识别方法,独立分量分析(ICA)则是一种非常有效的非高斯数据分析工具.其中,四阶累积量算法是一种数值稳定且鲁棒的ICA方法,非常适合用于振动信号的特征抽取,但独立分量分析有一定的前提条件,借助经验模态分解(EMD),消除噪声干扰,去除高频IMF部分,满足ICA的条件.因此,利用ICA算法对某炼铁厂风机不同状态模式(包括正常和转子不对中)进行特征提取,HMM实现模式的最终分类.对照分类实验结果,表明基于EMD-ICA的HMM的故障分类方法不仅具有良好的模式分类能力,且实现简单,在风机健康状况监测中有较大的应用潜力.