基于支持向量机的转子振动故障融合诊断技术

针对某些大型复杂旋转机械振动信号特征提取和故障样本获取难的问题，提出了一种基于小波包特征谱熵支持向量机（SVM）的转子振动故障融合诊断方法.通过转子实验台模拟了转子振动的4种典型故障，并采集其振动故障数据.用小波包对振动故障信号进行分解，提取故障信息含量大的频带并计算出其小波特征谱熵作为故障特征，建立故障诊断模型.通过对故障类别的区分和故障严重程度的判断，验证了该方法在解决转子振动故障信号的特征提取及小样本情况下的故障诊断问题等方面是有效的.

Rotor vibration fault fusion diagnosis based on support vector machine

For the difficulties in extracting vibration signal features of some large scale intricate rotating machinery and obtaining efficient fault samples, a rotor vibration fault fusion diagnosis method based on wavelet packet characteristic spectral entropy support vector machine （SVM） was proposed. Four typical rotor vibration faults were simulated through rotor experiment table, and some vibration fault data were collected. The vibration fault signals were disassembled. The frequency band containing a large amount of fault information was extracted and its wavelet characteristic spectral entropy was calculated to serve as fault feature, and then the rotor vibration fault diagnosis model was established. Through distinguishing the rotor 〖JP2〗fault type and determining the fault severity, it is proved that this method is effective in solving such problems 〖JP〗as extracting the rotor vibration fault signal feature and realizing the fault diagnosis under small samples.