基于单类支持向量机的航空发动机故障检测

通过监控航空发动机性能参数，准确判断发动机的工作状态，预知发动机的异常变化．为预防和排除故障提供充足的时间和决策依据。开发出利用QAR（quick access Feeorder．快速存取记录器）数据的发动机故障检测系统，该系统基于民航发动机的QAR数据，由于发动机正常运行数据容易获取．而故障样本难以获得．因而采用单类支持向量机（OCSVM），仅依靠发动机的健康数据建立其分类器。利用OCSVM分类器．监控后续航班参数是否出现异常，通过分析检测结果，实现发动机故障检测。采用该系统监控航空发动机性能参数．及时发现发动机运行状态异常，证明了系统的可行性和有效性。

Aero-engine fault detection based on one-class support vector machine

By monitoring the aero-engine performance parameters, the engine state can be accurately determined and the engine performance can be predicted in advance, providing adequate time and decision-making basis for the prevention and troubleshooting. Yielding to a health model by using one-class support vector machine （0CSVM）, meanwhile the data sample of normal running status is easier to gain, aero-engine fault detection system is designed based on the quick access recorder （QAR） data of the civil aviation engine. Furthermore, the follow-up flights parameters can be monitored by using OCSVM classification. Ultimately, the engine operation status and performance trend can be obtained. The monitoring results of aero-engine performance parameters show that the faults can be detected in advance and demonstrate the feasibility and effectiveness of this system.