多参数与多测点信息融合的行星轮故障诊断

Planetary gear fault diagnosis based on information fusion of multi-parameters and multi-points

为了找到多级行星齿轮传动系统复杂故障诊断的合适方法,对三级齿轮传动系统进行故障模拟和振动信号测试。针对行星齿轮传动系统振动信号的非线性和非平稳性、故障特征信号难以提取等特点,采用关联维数、最大Lyapunov指数、样本熵3个混沌特征参数作为故障辨识特征量。用不同测点和不同混沌特征参数的信息融合,通过支持向量机分类方法建立信息融合故障诊断模型及6种不同故障状态的训练集,实现对三级齿轮传动系统复杂故障类型的识别与诊断。分析结果表明:多测点信息融合或不同混沌特征参数融合,均能不同程度提高故障分类准确率。而经多测点与多混沌特征参数的信息融合后,通过支持向量机的故障分类准确率最高。

In order to find the appropriate method of multistage planetary gear transmission system complex fault diagnosis,fault simulation and vibration signal test were carried out on three stage gear transmission system.Aiming at the nonlinear and non-sta-tionary characteristics of the vibration signals of the planetary gear transmission system and the difficultly in picking up the fault fea-ture signals,3 chaotic characteristic parameters,including the correlation dimension,maximum Lyapunov exponent and sample entropy were adopted as fault identification features.The information of different measuring points and different chaotic characteris-tic parameters was fused;and the information fusion fault diagnosis model and the training set for 6 kinds of different fault states were established based on the classification method of support vector machine,then the recognition and diagnosis of the three stage gear transmission system with complex fault types were achieved.Analysis results show that the multiple measurement information fusion or different chaotic characteristic parameter fusion both can improve the accuracy of fault classification at different degrees. After the information fusion of multipoint and multiple chaotic characteristic parameters,and using support vector machine,the ob-tained fault classification accuracy rate is the highest.