基于盲源分离的齿轮箱状态检测与故障诊断

Condition monitoring and fault diagnosis of gear box based on blind source separation

针对高背景噪声的影响及混合故障诊断困难这一问题,采用盲源分离技术作为信号预处理与故障特征提取的主要工具,并应用这一新技术解决了故障曝气机WA5齿轮箱振动信号的故障诊断问题.通过盲源分离得到准确的故障信号,并根据故障振动信号频谱诊断出故障.利用盲源分离技术可以有效地去除外来干扰,提高故障诊断精度,解决了现实工作中的故障定位、早期故障诊断率低等难题.

In order to solve the effect of high background noise and the difficulty of mixed fault diagnosis, the blind source separation technique was considered as a main tool of signal pre-processing and feature extraction. This method has been applied in fault diagnosis and condition monitoring of the gear box of explode spirit machine （WAS）. Through blind source separation technique, the accurate failure signal can be extracted and then be used to diagnose faults. Using blind source separation technique, the useless interference can be eliminated, which will improve the diagnosis accuracy and solve such problems as fault positioning difficulty and low ratio of early fault diagnosis.