中文摘要：

为解决发动机所监控的健康指数不能多于测量参数的问题,采用平滑支持向量机方法（SSVM）,用4个参数对发动机九类衰退故障进行诊断,并与传统的支持向量机方法（采用LSSVM）进行对比。研究表明：在样本数量小、样本分布不平衡等条件的影响下,SSVM对各类部件性能衰退故障的诊断正确率均在90%以上。相对于LSSVM,SSVM无需优化参数,鲁棒性强,对样本集大小和样本集数目不平衡性的适应性良好,更适合航空发动机性能衰退故障的诊断。

英文摘要：

For solving the problem that the number of healthy parameters monitored by the engine is no more than the measurement parameters,and Smooth Support Vector Machines（SSVM） is adopted in this paper to diagnose nine kinds of deterioration fault by using four parameters.Meawhile a comparison between SSVM and traditional SSVM is given.The results show that though the amount of samples is small and their distribution are unbalanced,the accuracy of deterioration fault diagnosis by SSVM are always over 90%.Compared with LSSVM,it is not necessary for SSVM to optimize parameters and SSVM is more robust.SSVM can better adapt to the amount and unbalance of samples and is more suitable to diagnose the deterioration fault of aeroengine performance.