基于多维状态特征空间的飞机异构作动系统多故障诊断与主动容错控制

The distributed and more-electric hybrid actuation system with dissimilar redundancies is the development direction of the actuation system on large civil aircraft, and fault diagnosis and fault-tolerant control is an effective tool to guarantee the high reliability of the actuation system. However, some scientific issues still exist in fault diagnosis and fault-tolerance control of hybrid actuation system, such as the issues of multi-fault coupling and diagnosis, the issues of the effective combination of fault diagnosis and fault-tolerance control, etc. To address these issues, on the basis of the previous research on efficient control of hybrid actuation system, this project proposes a novel multi-dimensional state feature space-based multi-fault diagnosis and active fault-tolerant control theory. Multi-dimensional state feature space is constructed according the typical failure modes, the health state vector is formed by utilizing multi-dimensional fault state characterization, then the integrated design of fault diagnosis and fault-tolerant control is achieved by using the health state vector. Designated component energy-based multi-fault quick diagnosis method is proposed, in which the obviousness of the designated fault model is enhanced by space projection energy of the health state vector. The health index is designed according the norm of the health state vector, then an active fault-tolerant control method based on health index and intelligent model matching is also proposed to achieve reasonable performance degradation. The research achievements of this project can provide theoretical basis and effective support to design the more-electric hybrid actuation system of our new wide-body civil aircraft.

分布式多电化的非相似异构作动系统是大型飞机作动系统的发展方向，故障诊断与容错控制是确保其高可靠健康服役的必要手段。但是，异构作动系统的故障诊断与容错控制尚存在诸多难题，如多故障耦合与诊断问题、诊断与容错的集成设计问题等。针对这些问题，本项目将在前期异构作动系统高效控制方法研究的基础上，提出基于多维状态特征空间的多故障诊断与主动容错控制理论。根据系统的典型故障模式确立与之对应的空间坐标轴，形成多维状态特征空间；提取每一种故障状态的有效表征，综合多维故障状态在空间中形成系统的健康状态矢量，以实现故障诊断与容错控制的有效集成设计。将健康状态矢量进行空间投影，通过指定元能量分析，实现多故障快速诊断；利用健康状态矢量的模构建系统健康指数，通过健康指数在线匹配最优参考模型，实现合理性能降级的主动容错控制。本项目的研究成果将为我国新型宽体客机的多电化异构作动系统设计提供理论依据和有效支持。

针对大型飞机非相似余度作动系统的故障诊断与容错控制两个关键问题，本项目按研究计划、围绕研究目标开展五方面研究工作：（1）作动系统多故障耦合机理与多维故障状态表征；（2）基于扩展卡尔曼滤波和交互多模型的多故障诊断方法；（3）基于复合空间的非相似余度控制系统容错架构构建；（4）多作动器与传感器故障工况下的飞行控制系统主动容错控制；（5）大型飞机水平安定面损伤情况下的自适应主动容错控制。.本项目在执行期间围绕这五方面研究内容开展相关研究工作，在作动系统级，运用状态方程分析方法建立非相似余度作动系统的耦合模型，对不同作动器间的耦合机理进行分析；提出基于扩展卡尔曼滤波和交互多模型结合的多故障诊断方法，能够在参数偏离正常值时确定实际系统的健康状态，从而采取适当的措施提高系统的可靠性和安全性。在飞行控制级，提出一种基于复合空间的非相似余度控制系统容错架构，有利于确立层级容错控制律分配体系；提出了多作动器与传感器故障工况下的飞行控制系统主动容错控制，使得在多执行器和传感器作动性能下降或者完全失效情形下保证整机控制增益；针对大型飞机水平安定面的损伤故障开展自适应主动容错控制研究，从而保证了在安定面损伤的情况下系统的稳定性。.本项目通过四年的研究，突破了飞机非相似余度作动系统多故障诊断与容错控制关键技术；发篇学术论文35篇，授权国家发明专利7项，培养研究生8名。本项目在飞机容错控制方面的研究成果已经成功应用于我国新一代大型飞机电液动力控制与作动系统的研制和我国某型反潜直升机等国家重大型号中伺服作动系统的设计和能力提升，显现出重要的应用价值和良好的社会效益，荣获中国航空学会科学技术二等奖1项。