摘 要：对模糊支持向量机中的传统隶属度确定函数进行了改进，得到了紧密度隶属函数的模糊隶属度确定方法；针对航空发动机整机振动中多类故障诊断的特点，引入模糊隶属度函数建立了更有效地FSVM融合诊断的数学模型，并将该模型应用到航空发动机整机振动故障诊断中。计算结果显示：该方法不但具有较高的正确诊断率，而且也具有很强的抗噪声能力，从而为航空发动机整机振动故障分析提供了一种新方法。

Abstract: A more effective fuzzy membership based on samples’ affinity function was investigated for the measurement of the samples inaccuracy through synthesizing the advantages between the fuzzy theory and SVM under applying membership to SVM. And the method was applied to some aeroengine vibration fault analysis, and a better diagnosis mathematical model was established based on FSVM. And then the model was used to fault diagnosis of aeroengine whole-body vibration. The result of analysis reveal that this method doesn’t has the better diagnosing accurate rate, but also takes on good ability to overcome the noise, and a novel analytical technique for aeroengine whole-body vibration fault to be determined rapidly is provided.