基于DCT的微弱信号提取和识别

微弱信号提取一直是故障诊断领域的难点。结合离散余弦变换(DCT),将离散时间序列经过离散余弦变换处理成对应的系数向量,在阈值处理的基础上,重构信号提取出微弱故障信息。与小波降噪和低通滤波方法进行对比分析,该算法突出了信号的微弱故障特征信息,较好的再现了夹杂在信号中的微弱成分,参数设定简单,结果对参数不敏感。最后通过实验证实该方法的有效性。本算法速度快,简单易行,可用于实时故障监测。

Weak Fault Signal Extrction and Identification Based on DCT

Weak signal extraction is a traditional difficulty in fault diagnosis, In this paper, a new method based on discrete cosine transform （DCT） is introduced. In this method, the original time series is transferred to a corresponding coefficient vector, then the signal is reconstructed and the weak fault information is extracted. This method is compared with the wavelet noise-reduction method. In the simulation of an example, the weak fault components in mixed signals were strongly reappeared and the information of the weak fault signal was extracted efficiently from the strong powerful background noise. This method is less computer-time consuming. The process is easy to realize. It provides an efficient approach for real time monitoring.