摘要
独立分量分析方法广泛应用于机械设备故障诊断领域。在稳健独立分量分析方法的基础上,结合故障特征频率先验信息,提出了一种约束稳健独立分量分析方法。该算法首先讨论了如何产生参考信号,然后定义了参考信号和期望信号的接近性度量函数,最后提出了改进的稳健独立分量对比函数。仿真和试验结果表明,该算法在收敛速度和计算精度方面都明显优于传统的Fast ICA算法。
The method of independent component analysis is widely used in mechanical equipment fault diagnosis domain.A novel method named as constrained robust independent component analysis（ cRobust ICA）based on Robust ICA algorithm is proposed which utilized prior information about the fault characteristic frequency.Firstly,how to create the reference signal is discussed.Then,the measurement function between reference signal and desired independent component（ IC） is defined.At last,an enhanced contrast function is acquired by modifying a generally used kurtosis contrast function with closeness measurement.A contrastive study on the conventional Fast ICA is adopted to demonstrate the effectiveness and accuracy of the cRobust ICA by numerical simulations and experiments.
出处
《机械传动》
CSCD
北大核心
2015年第11期154-160,共7页
Journal of Mechanical Transmission
基金
中央高校基本科研业务费专项资金(ZYGX2012J099)
