基于小波去噪的快速定点ICA及其在振源识别中的应用

Fast ICA Algorithm and Its Application to Recognition of Vibration Sources Based on Wavelet De-noising

在多源激励情况下,从传感器阵列输出的高维响应中分离、识别不同振源的性质,对于大型结构振动分析以及对其状态监测与健康评价有着重要的意义.本文结合大型结构振动的性质,介绍了小波阈值去噪原理,使用其改进的快速定点算法实现了多源激励情况下振源信息分离的方法.文中通过实例给出了这种方法的应用效果.

To separate and recognize characters under the excitation of multiple sources is a useful technique for infrastructure vibration analysis. This paper discusses an excitation sources separating approach using improved fast fixed point ICA and wavelet de-noising in consideration of peculiarities from the vibration of large scale structures. In additions, its merits are illustrated by real world vibration date.