LMD和支持向量机相结合的齿轮毂故障诊断方法
Fault Diagnosis of Gear Hub Based on LMD and Support Vector Machine

局部均值分解(LMD)作为一种新的自适应时频分析方法,在故障诊断领域展现了良好的应用前景。根据某型航空发动机减速器一级齿轮毂出现裂纹故障时其振动信号会产生调制现象的特点,提出了基于LMD和支持向量机(SVM)的某型航空发动机减速器一级齿轮毂裂纹故障诊断方法。对某型航空发动机进行振动测试获取其振动样本数据,利用LMD提取故障样本数据的故障特征信息、构造特征向量,并将其作为SVM的输入特征参数,成功建立了针对目标故障的故障诊断模型。对一级齿轮毂工作状态的分析结果表明了该方法的有效性。
As a new kind of self-adaptation time-frequency analysis approach, the local mean decomposition(LMD) shows a good application prospect in fault detection. According to the fact that the characteristics of vibration signal can be modulated when the crack fault exists on the gear hub of the reducer in an aero engine, we spropose a fault detection method based on LMD and support vector machine(SVM). Firstly, vibration sample data acquired through vibration tests performed in the aero engine. Then the fault information of fault gear hub was extracted by LMD to construct feature vectors. Eventually, taking the feature vectors as the input parameter of SVM, a fault detection model towards the target fault was successfully established. The analysis results demonstrate the effectiveness of the proposed method