• 首页期刊介绍编委会期刊订阅广告合作联系我们English
  • 选择皮肤:
来晨阳,郭迎清,于华锋.基于RF-SVR的燃油计量装置性能衰退检测和剩余寿命估计方法[J].航空动力学报,2019,34(7):1624~1632
基于RF-SVR的燃油计量装置性能衰退检测和剩余寿命估计方法
Fuel metering unit performance degradation detection and remaining useful life estimation method based on RF-SVR
投稿时间:2018-10-25  
DOI:10.13224/j.cnki.jasp.2019.07.024
中文关键词:  发动机燃油系统  健康管理  性能衰退  随机森林-支持向量回归  剩余使用寿命
英文关键词:engine fuel system  health management  performance degradation  random forest-support vector regression  remaining useful life
基金项目:
作者单位
来晨阳 西北工业大学 动力与能源学院,西安 710072 
郭迎清 西北工业大学 动力与能源学院,西安 710072 
于华锋 西北工业大学 动力与能源学院,西安 710072 
摘要点击次数: 263
全文下载次数: 229
中文摘要:
      为了实现航空发动机燃油系统的安全状态监测和健康管理,开展了燃油系统性能衰退检测和剩余使用寿命估计方面的研究。以燃油系统燃油计量装置为例,分析了其主要的性能衰退模式,设计了基于电流-速度数据的健康指标(HIs)选取方案,并考虑环境及模型参数不确定性,进行模型不确定性仿真,基于健康数据与性能衰退数据间的马氏距离对部件性能衰退进行检测。提出了基于随机森林-支持向量回归(RF-SVR)的剩余使用寿命(RUL)估计方法,利用通过RF特征选择优化的SVR模型实现部件RUL估计。最后基于某型民用涡扇发动机机械液压模型仿真数据对该方法进行了验证,结果表明:该方法的性能衰退检测虚警率及漏报率低于2%,RUL估计误差低于3%,可为航空发动机燃油系统的预测性维护提供参考。
英文摘要:
      In order to realize the safety state monitoring and health management of the aero-engine fuel system, the fuel system performance degradation detection and remaining useful life estimation was researched.Taking the fuel system fuel metering device as an example, the main performance degradation mode was analyzed.The health indicators (HIs) selection scheme based on current-speed data was designed.Considering the uncertainty of environment and model parameters, the model uncertainty simulation was carried out.The component performance degradation was detected based on the Mahalanobis distance between the healthy data and the performance degradative data.A remaining useful life (RUL) estimation method based on random forest-support vector regression (RF-SVR) was proposed.The component RUL estimation was realized by SVR model optimized by RF feature selection.Finally, the method was validated based on the simulation data of a certain type of civil turbofan engine mechanical hydraulic model.The results show that the performance of the method has a false alarm rate and a false negative rate of less than 2%, and the RUL estimation error is less than 3%. This provides a reference for predictive maintenance of aero engine fuel systems.
查看全文  查看/发表评论  下载PDF阅读器
关闭
.