摘要
航空发动机突发故障严重影响飞机的飞行安全。为了解决航空发动机故障诊断中因缺乏样本和突发故障信息难以提取的困难,提出了基于支持向量机、小波包分解和智能模块相结合的发动机突发故障识别与监控方法。该方法在强噪声和少样本条件下,用结构风险最小原理建立发动机故障特征与运行状态之间的对应关系,再根据该函数的输出来识别故障状态和调用相应的智能模块对故障进行监控。实验结果表明,该方法能有效地提高航空发动机突发故障的识别率,并能对突发故障进行监控修复。
Sudden fault of aero-engine seriously affects the flight safety of an aircraft.But there is a lack of data samples and it is difficult to extract information on sudden fault.Therefore,we propose what we believe to be an effective method for their identification and monitoring,which combines a support vector machine（SVM） and wavelet packet decomposition（WPD） with intelligent modules.We extract the features of the sudden faults with the WPD and identify the sudden faults with the SVM.Under the conditions of strong background noise and data sample lacking,we use the structural risk minimization principle to establish the function between the aero-engine＇s fault characteristics and its operation states.Then we use the output of the function to identify the sudden faults and then employ the intelligent modules to monitor and repair the sudden faults.Finally,we did experiments on their identification and monitoring.The experimental results,given in Figs.1 and 3 and Tables 1 and 2,and their analysis show preliminarily that our method can extract quickly the features of sudden faults of an aero-engine and identify them accurately and that it can also monitor and repair the sudden faults,thus effectively enhancing the safety and reliability of the aero-engine particularly at their early stage.
出处
《西北工业大学学报》
EI
CAS
CSCD
北大核心
2013年第3期401-405,共5页
Journal of Northwestern Polytechnical University
基金
国家自然科学基金(51075330
50675178)资助
关键词
飞机发动机
特征提取
突发故障
识别
监控
支持向量机
小波包分解
aircraft engines
feature extraction
monitoring
support vector machines
identification
sudden fault
wavelet packet decomposition
