基于多维度排列熵与支持向量机的轴承早期故障诊断方法

Early fault diagnosis of bearing based on multi-dimension permutation entropy and SVM

针对许多现有方法无法有效诊断滚动轴承早期故障的问题,引入排列熵的方法对轴承振动信号进行早期故障分析。通过研究嵌入维数和延迟时间对信号排列熵的影响,提出多维度排列熵的特征提取方法。利用多维度排列熵方法所提取的特征,建立了基于支持向量机的轴承早期故障智能诊断模型。对轴承不同类型、不同程度的故障数据进行分析,证明了多维度排列熵方法可以有效提取轴承不同状态的特征信息,与支持向量机结合的智能诊断模型可以精确地诊断轴承不同类型的早期故障,具有很强的通用性;该模型在贫样本的情况下,依然具有很高的诊断精度,适用于滚动轴承早期故障状态的在线监测。

Aiming at the problem that the early fault diagnosis of rolling bearing could not be carried out effectively by most exiting methods, the permutation entropy algorithm was used to analyze the early faults of bearing vibration signal. Through researching the impact of embedding dimension and delaying time on signal permutation entropy, a feature extraction method based on multi-dimension permutation entropy algorithm was proposed. The intelligent di- agnosis model was established to diagnose bearing early faults according to the support vector machine theory and the features calculated by the proposed method. The fault data with different types and varying degrees of bearings were analyzed to prove the effectiveness of multi-dimension permutation entropy method for extracting the features information of beating＇s different states effectively and the universality of proposed intelligent model for diagnosing the bearing early faults accurately. Under the circumstance of poor samples, the proposed model had- high diagnos- tic capability which was suitable for on-line monitoring for bearing early faults.