摘要
由于传统建模技术有高成本、高复杂性的缺点,计算机仿真技术随之兴起,正逐渐取代传统建模技术。AMEsim仿真平台由于其完备的模型库,在建模中发挥了越来越重要的作用。论文在AMEsim建立的航空发动机燃油调节系统模型的基础上,人为设置故障,得到系统的故障数据,并在Matlab上对故障数据用主元分析法进行降维,得到特征数据,采取支持向量机法对特征数据进行训练,最后用测试数据验证结果的准确性。实验结果完全分辨出测试数据代表的故障,这验证了支持向量机法用于航空发动机燃油调节系统故障诊断的优越性和可靠性,可用于航空发动机燃油调节系统的故障诊断。
Because of the high cost and high complexity of traditional modeling technology, computer simulation technology is gradually replacing the traditional modeling technology. AMEsim simulation platform has comprehensive model library, so it plays an increasingly important role in modeling. Based on aviation engine fuel control system built on AMEsim, some faults are set to obtain failure data, and Principal Component Analysis is used to reduce the dimensions and obtain the characteristic data on Matlab, Then support vector machines is taken to train feature data. Finally test data is used to verify the accuracy of the result. The result distinguishs completely failure represented by test data, which demonstrate the superiority and reliability of SVM for aviation engine fuel control system's fault diagnosis and SVM can be used for fault diagnosis of aviation engine fuel control system.
出处
《计算机与数字工程》
2014年第9期1536-1541,1720,共7页
Computer & Digital Engineering
基金
远程宽体客机综合健康管理机载系统体系架构及先进健康管理算法研究项目资助