摘要
现代民航发动机大多使用VSV系统来提高发动机工作稳定性和避免发动机失速或喘振。为了诊断VSV系统故障,提出了一种基于PSOBP神经网络对VSV位置进行监控的方法,当PSOBP神经网络模型的预测值与实际值的偏差超过一定值时,则判断VSV系统故障。利用发动机健康状态的QAR数据,基于PSO算法优化的BP神经网络建立了发动机VSV在飞机下降段的调节规律模型,同时建立BP神经网络模型。经过对比分析,通过PSOBP神经网络建立的VSV调节规律模型的诊断精度,高于传统的BP神经网络模型,可为民航发动机状态监控和故障诊断提供依据,具有一定的工程实用价值。
Variable stator vane(VSV)system is widely used in modern civil aviation engines to improve engine stability and prevent engine stall or surge.In order to diagnose the fault of VSV system,a method of monitoring VSV position is proposed based on PSOBP neural network.When the deviation between the predicted value and the actual value of PSOBP neural network model exceeds a certain value,the VSV system fault is judged.Based on the QAR(Quick Access Recorder)data of engine health status and BP neural network optimized by PSO algorithm,the regulation rule model of engine VSV in aircraft descent phase is established,so is the BP neural network model.Through comparative analysis,the diagnosis accuracy of VSV regulation rule model based on PSOBP neural network is higher than that of traditional BP neural network model,which can provide basis for civil aviation engine condition monitoring and fault diagnosis,and is of certain engineering practical value.
作者
阚玉祥
KAN Yu-xiang(Flight College,Binzhou University,Binzhou 256603,China)
出处
《滨州学院学报》
2020年第6期5-10,共6页
Journal of Binzhou University
基金
滨州学院青年人才创新工程科研基金项目(BZXYQNLG201802)。
关键词
民航发动机
VSV
故障诊断
PSO算法
BP神经网络
civil aviation engine
VSV
fault diagnosis
particle swarm optimization algorithm
BP neural network
作者简介
阚玉祥(1990—),男,山东无棣人,助教,硕士,从事民航发动机状态监控与故障诊断研究。E-mail:1510366683@qq.com。
