基于改进模糊SVM的转子振动故障诊断技术研究

本文首先对常用的隶属度确定方法进行改进，提出了基于改进模糊支持向量机（FSVM）的融合故障诊断方法，并建立了改进FSVM故障诊断数学模型；然后，利用转子振动模拟实验台对四种典型的转子诊断故障进行模拟，并提取其故障信号特征；最后，通过实例计算分析，验证了该方法在转子振动故障诊断方面是可行和有效的，为转子振动故障准确诊断提供了一种新方法。

Research on Improved Fuzzy SVM Method for Rotor Vibration Fault Diagnosis

In this paper, a Fuzzy Support Vector Machine (FSVM) fault diagnosis method was proposed based on the improvement to the ordinary method used to determine memberships, and an improved diagnosis mathematical model was established at first. Secondly, four kinds of typical rotor vibration faults were simulated on a rotor vibration experimental rig, and effective samples of fault were collected. Finally, the validity and feasibility of this method using for rotor vibration fault diagnosis were validated by calculating and analyzing to the obtained examples, an accurate novel rotor vibration fault diagnosis method was provided.