解析模型/智能信息融合的航空发动机全寿气路健康自监测及诊断研究

Aiming at gas path health online analysis of aircraft engine during its lifetime, a novel methodology of gas path health self-monitoring and diagnosis is proposed based on analytical model/intelligent information fusion for aircraft engine in this project. It is developed from the latest research results of on-board dynamic modeling and intelligent fault diagnosis. The model mismatch due to component degradation and measurement uncertainty are negative to engine health performance on-line monitoring. Besides, the complex correlation of multi-source different fault characterizations pose a tough challenge to the reliable diagnosis implementation. Hence, new problems of aircraft engine gas path health online monitoring and fault diagnosis arise. The project focuses on the following key issues: lifecycle combined modeling approach to the engine real time adaptive dynamic model including the nonlinear and linear subsystems, gas path health self-monitoring mechanism with measurement uncertainty of the engine over time, and multiple source information intelligent fusion diagnosis. Attempt to the researches on this project, the mechanism of engine component natural degradation is revealed; lifecycle real-time combined dynamic modeling theory with satisfactory confidence is established; moreover, gas path anomaly on-line monitoring isolation logic and threshold self-adjustment strategy are explored to tune to the natural degradation and measurement uncertainty; a new intelligent information fusion diagnosis method is obtained using component characteristics knowledge from analytical model and real-time measurements, and the feasibility analysis and verification are achieved, which will provide the theory foundation for aircraft engine fault diagnosis.

针对航空发动机全寿气路健康状态在线监测及诊断问题，通过借鉴机载动态建模与智能诊断最新研究成果，首次提出一种基于解析模型/智能信息融合的发动机全寿气路健康自监测及诊断方法。发动机性能自然蜕化引起的模型失配和测量不确定是影响全寿气路健康在线监测的主因，同时多源异质气路故障信息的复杂关联性对可靠诊断实现提出了更高要求，因而蕴涵着新的健康监测及诊断问题。本项目旨在探索的核心问题包括：发动机全寿机载自适应非线性和线性组合动态建模，全寿测量不确定鲁棒性的气路健康状态自监测，及多源异质智能融合诊断。试图通过本项目研究，不仅揭示航空发动机寿命期内气路部件性能自然蜕化机理，建立高置信度全寿机载组合动态解析模型，而且掌握全寿测量不确定下气路健康状态在线监测及诊断逻辑与阈值自调整策略，获得综合解析模型部件特性变化信息和实测数据的智能融合诊断新方法，完成可行性分析与验证，以期为航空发动机故障诊断提供理论储备。

针对航空发动机全寿气路健康状态在线监测及诊断问题，通过借鉴机载动态建模与智能诊断最新研究成果，提出了一种基于解析模型/智能信息融合的发动机全寿气路健康自监测及诊断方法。考虑到发动机性能自然蜕化引起的模型失配和测量不确定是影响全寿气路健康在线监测的主因，同时多源异质气路故障信息的复杂关联性对可靠诊断实现提出了更高要求。本项目主要完成探讨了发动机全寿机载自适应非线性和线性组合动态建模，全寿测量不确定鲁棒性的气路健康状态自监测及多源异质智能融合诊断关键技术，通过本项目研究，揭示航空发动机寿命期内气路部件性能自然蜕化机理，建立了高置信度全寿机载组合动态解析模型，掌握全寿测量不确定下气路健康状态在线监测及诊断逻辑与阈值自调整策略，获得综合解析模型部件特性变化信息和实测数据的智能融合诊断新方法，最后通过真实数据分析与仿真，验证了本设计方法的优势，部分成果应用于某型航空发动机预测健康管理系统研制中，完成国防科技成果一项，发表学术论文12篇，其中SCI检索论文6篇，申请发明专利10项，获批软件著作权1项，培养研究生8名。