基于改进的自适应回归时序模型故障诊断

Fault Diagnosis Based on Improved Adaptive Regression Time Series Model

汽车变速箱的故障诊断工作比较复杂,由于传统的诊断方法已不能满足复杂的故障现象,文中提出了一种基于改进的自适应回归时序模型故障诊断方法;方法采用了基于时间序列故障诊断技术,首先测取工作环境下的振动信号,然后建立被诊断对象的时间序列数学模型,最后用信息距离判别法诊断出故障类型,提高了诊断效率;最后在变速箱进行了实验研究;选用型号为621B40型ICP加速度传感器测取变速箱的振动信号,通过设置模型参数(n,m)来模拟故障检测,实验分析表明,提出的算法可以有效地识别变速箱系统中不同严重程度的故障,且与传统的故障诊断算法相比,提出的算法对提高识别率和降低计算复杂度都有着明显的优势。

the fault diagnosis of automobile transmission is more complex, because the traditional diagnostic methods have been unable to meet the complex fault phenomena, this paper proposes an improved adaptive fault diagnosis method based on regression time series model. Methods by using the technology of fault diagnosis based on time series, firstly, the vibration signals are measured under working environ- ment, and then establish the diagnosis time series object mathematical model, the information of distance discriminant analysis method to di- agnose the fault type, improve the efficiency of fault diagnosis. Experiment research is carried out in the gearbox of mulberry. Experimental results show that, the algorithm can effectively identify the fault severity of gearbox system, and compared with the traditional fault diagno- sis algorithm, the proposed algorithm improves the recognition rate and reduce the computational complexity, has obvious advantages.