基于多方法结合应用的齿轮箱故障特征提取研究

The study in gearbox fault feature extraction based on multi-methods joint application

基于振动信号分析的齿轮箱故障诊断的关键是实现对信号中故障特征的提取.由于在工程实际中采集到的齿轮箱振动信号含有较强的噪声干扰,所以单一的信号分析方法难以实现对故障特征的提取.因此将两种或两种以上方法相结合应用于齿轮箱振动信号的处理成为当前的研究趋势.为研究将不同方法相结合应用于齿轮箱故障信号特征提取的优势,对大量文献的研究成果进行了归纳整理.综合分析发现:将多种方法结合应用于齿轮箱振动信号特征提取,可有效避免单一方法的局限性,充分发挥不同方法的优势.总结了在齿轮箱故障诊断领域中分别以频谱分析为基础和以非线性理论为基础的将不同信号处理方法结合应用于齿轮箱故障特征提取的现状,最后针对多种方法结合应用于齿轮箱故障诊断的发展趋势提出了建议.

The realization of fault feature extraction is the key point during the fault diagnosis in gearbox based on vibration signal analysis. Because the strong noise interference consist in the vi- bration signals which collected from the actual work gearbox, it is hard to extract the fault fea- ture using single signal analysis method. Two different or more methods jointly used in gearbox vibration signal processing have been the tendency in current research; In order to study the ad- vantages of different methods jointly used in gearbox fault signal feature extraction, we analyzed large amount of literature and found that the combining methods can avoid the limitations ap- peared in single method, and make full use of the advantages in different methods. The current situation of the application of combining methods separately based on frequency analysis and non- linear theory in gearbox fault feature extraction was summarized, and finally some suggestions were made for the development of the different signal processing methods used in gearbox fault diagnosis.