基于支持向量机的航空发动机故障诊断研究

Research on Fault Diagnosis of Aeroengines Based on Support Vector Machine

研究利用支持向量机对发动机的两类故障——失速和喘振进行识别。介绍了支持向量机理论,选取适当的学习算法、惩罚因子和核函数,建立了支持向量机,并采用4组已知故障模式的数据对其进行训练和测试,之后对另外两组数据进行仿真识别,仿真结果与实际故障模式一致。

Identification of two faults of aeroengines stall and surge is studied with a method of SVM （ support vector machine）. The theory of SVM is introduced. After choosing right study algorithm and punition gene and kernel function, the SVM is set up. Then it is trained and tested with four sets of data whose fault patterns are known. At last, another two sets of data are imported for emulation. The results are consonant with their real fault patterns.