论文部分内容阅读
针对旋转机械振动过程的复杂性和振动故障产生的随机性以及振动故障样本获取难的问题,在信息熵理论的基础上,融合了支持向量机(SVM)小样本、全局性和泛化性好的优点,提出了过程功率谱信息熵(功率谱熵)SVM的故障诊断方法。结合转子实验台,得到了4种典型振动故障在多测点多转速下的数据,通过计算提取了其功率谱熵特征值作为故障样本,即故障向量,并建立SVM诊断模型,对转子振动故障的类别、严重程度和部位识别诊断,验证了该方法在转子振动故障诊断方面效果良好。
Aiming at the complexity of the vibration process of rotating machinery and the randomness of the vibration fault and the difficulty of obtaining the sample of the vibration fault, based on the information entropy theory, a small sample of Support Vector Machine (SVM) , The fault diagnosis method of process power spectrum information entropy (power spectral entropy) SVM is proposed. Combining with the rotor test bench, the data of four kinds of typical vibration faults at multi-measuring point and multi-speed were obtained. The power spectral entropy eigenvalues were extracted as fault samples, ie fault vectors, and the SVM diagnosis model was established. The rotor vibration fault The classification, severity and part recognition diagnosis of this method are validated. The method is effective in the diagnosis of rotor vibration fault.