Abstract: In order to monitor the vibration state of the aero-engine in real time more efficiently and intuitively, a method of assessing the rotor vibration state was proposed based on information entropy distance and fuzzy support vector machine (FSVM) membership. Firstly, information entropy distance, the indicator of the rotor vibration state, was put forward by studying information entropy feature of vibration signal. Then, a multi-parameter rotor vibration state assessment model was established, combining information entropy distance with the fuzzy membership matrix determined by FSVM.Finally, the model was applied to rotor vibration signal system analysis and quantitative calculation.The results show that the method for rotor vibration state assessment is effective and feasible.

Key words: Rotor vibration; Feature extraction; Information entropy distance; Fuzzy support vector machine; Fuzzy membership; State assessment

摘要： 为了更有效、直观地对航空发动机的振动状态进行实时监控,运用信息熵和模糊支持向量机(FSVM)方法,建立了基于信息熵距和FSVM隶属度的转子振动状态评估方法。研究了振动信号的信息熵特征,提出了可以表示转子振动状态的指标—信息熵距；通过模糊支持向量机(FSVM)确定模糊隶属度矩阵,将模糊隶属度矩阵与信息熵距相结合,建立了一个多参数的转子振动状态评估模型；应用此模型对转子振动信号进行系统分析和定量计算,验证了该方法用于转子振动状态评估是有效、可行的。 

关键词: 转子振动;特征提取;信息熵距;模糊支持向量机;模糊隶属度;状态评估 