摘要
在齿轮箱齿轮振动信号处理中,可将经验模式分解(Empirical Mode Decomposition,EMD)和分形技术结合在一起.先进行EMD分解,并针对EMD的端点效应问题提出了基于最小能量误差的端点效应抑制方法;计算各IMF分量的分形维数,根据同类故障具有相似特征,主要表现在模式空间上距离相近,从而利用未知故障网格维数与已知故障网格维数的距离远近来对故障类型进行判断.实测验证结果表明:通过比较网格维数,能有效判断故障的类型,该方法大大简化了故障诊断过程.
During gearbox gear vibration signal processing, combines the empirical mode de- composition with fractal method. First with EMD end effect based on the minimum energy error suppression method for end effects of EMD. According to the fact that similar failures have similar characteristics and their distance was near in the pattern space, calculate the gridding dimension of IMF component. Then the fault type was judged by using the distance from the unknown faults gridding to known faults grid gridding dimension. Experiments show that the method can greatly simplifies the troubleshooting process.
出处
《陕西科技大学学报(自然科学版)》
2013年第1期115-120,共6页
Journal of Shaanxi University of Science & Technology
基金
国家自然科学基金项目(50915028)