论文部分内容阅读
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。
In order to solve the problem that the health index monitored by the engine can not exceed the measured parameters, a smooth support vector machine (SSVM) method is used to diagnose nine types of engine failure with four parameters. Compared with traditional support vector machine (LSSVM) )comparing. The results show that under the condition of small number of samples and unbalanced sample distribution, SSVM can diagnose all kinds of components’ performance degradation faults by more than 90%. Compared with LSSVM, SSVM does not need to optimize the parameters, robustness, good adaptability to the sample size and sample number imbalance, more suitable for the diagnosis of aero engine performance degradation failure.