基于空地数据链的航空发动机剩余寿命预测研究

发布时间：2019-06-04 04:04

【摘要】：为适应我国大飞机产业的飞速发展,保障飞机的安全性,控制飞机的维修成本,实现安全与效益的双赢,对航空发动机剩余寿命进行准确预测成为发动机健康管理中的核心工作。本文基于空地数据链获取了发动机健康监测参数,分析了发动机的性能衰退趋势,建立了衰退模型,设计了预测算法,实现了发动机剩余寿命的预测。首先,介绍了空地数据链传递监测数据的标准与协议,实现了监测数据的译码与预处理,选取了多个关键监测参数,为后续的预测工作提供了准确数据。其次,设计了基于Kalman滤波估计的发动机剩余寿命预测算法。该算法采用分阶段的处理方式对多监测参数进行了融合,基于状态空间方法建立了发动机的性能衰退趋势模型,借助Kalman滤波实现了模型参数的估计。对比了基于单阶段线性融合的Kalman滤波预测以及基于分阶段线性融合的Kalman滤波预测。实例分析表明,基于分阶段线性的Kalman滤波预测具有较准确的寿命分布以及较好的预测演化过程。再次,鉴于发动机健康状态与多个监测参数之间存在着非线性的映射关系,设计了基于粒子滤波的发动机剩余寿命预测算法。该算法将分阶段处理与非线性处理组合起来对多个监测参数进行融合,基于状态空间方法建立了发动机性能衰退模型,借助粒子滤波实现了模型的参数估计。对比了基于单阶段线性融合的粒子滤波预测、基于分阶段线性融合的粒子滤波预测以及基于分阶段非线性融合的粒子滤波预测三种预测算法。实例分析表明,基于分阶段非线性的粒子滤波预测具有较准确的寿命分布以及较好的预测演化过程。最后,本文基于MATLAB设计了可以实现数据融合、性能衰退建模以及预测实施功能的发动机剩余寿命预测系统。
[Abstract]:In order to adapt to the rapid development of large aircraft industry in our country, ensure the safety of aircraft, control the maintenance cost of aircraft, and realize the win-win situation of safety and benefit, Accurate prediction of the remaining life of aero-engine has become the core work in engine health management. In this paper, the engine health monitoring parameters are obtained based on the air-to-ground data link, the performance decline trend of the engine is analyzed, the recession model is established, the prediction algorithm is designed, and the prediction of the residual life of the engine is realized. Firstly, the standards and protocols of air-to-ground data link transmission monitoring data are introduced, the decoding and preprocessing of monitoring data are realized, and several key monitoring parameters are selected, which provides accurate data for subsequent prediction work. Secondly, a residual life prediction algorithm based on Kalman filter estimation is designed. In this algorithm, the multi-monitoring parameters are integrated in stages, the performance decline trend model of the engine is established based on the state space method, and the estimation of the model parameters is realized by Kalman filter. The Kalman filter prediction based on single-stage linear fusion and the Kalman filter prediction based on phased linear fusion are compared. The example analysis shows that the prediction based on phased linear Kalman filter has more accurate life distribution and better prediction evolution process. Thirdly, in view of the nonlinear mapping relationship between the health state of the engine and several monitoring parameters, an algorithm for predicting the residual life of the engine based on particle filter is designed. The algorithm combines phased processing with nonlinear processing to integrate multiple monitoring parameters. Based on the state space method, the engine performance decay model is established, and the parameter estimation of the model is realized by particle filter. Three prediction algorithms, particle filter prediction based on single-stage linear fusion, particle filter prediction based on phased linear fusion and particle filter prediction based on phased nonlinear fusion, are compared. The example analysis shows that the particle filter prediction based on phased nonlinear has more accurate life distribution and better prediction evolution process. Finally, based on MATLAB, an engine residual life prediction system is designed, which can realize data fusion, performance degradation modeling and prediction implementation function.
【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：V23

【相似文献】