基于负熵的转子混叠振动信号盲识别

Blind Identification of Rotors' Mixed Vibration Signals Based on Negative Entropy Arithmetic

介绍了盲源分离的基本概念,研究了基于负熵的盲源分离技术。在给出了基于负熵的快速定点算法后,成功地进行了仿真分析。首先将该技术应用于带有多个振源的转子的混叠振动信号盲源分离识别中,成功地实现了转子的复杂振动信号的盲分离,并发现了分离后的时域信号的独立性特征不明显,而在频域中能够体现满意的分离效果的特性。该研究为复杂振动信号的分析和故障诊断提供了一种新方法。

The basic conception of blind source separation was introduced and the research of blind source separation technique based on negative entropy was dealed with. The simulation was carried through successfully after the presentation of fast fixed--point calculation algorithms. It is the first research on the separation of rotors＇ vibration signals that include several mixed vibration sources. The rotors＇ complex vibration signals were separated successfully. It is discovered that the independence characteristics of signals in time domain is not clear, but the satisfactory separated results can be obtained in frequency domain. It provides a new method for analysis of the complex vibration signals and fault diagnosis.