数据驱动视角下飞机故障预测与健康管理系统设计及验证

Design and Verification of Aircraft PHM System Under Data-Driven Viewpoint

摘要:

由于故障预测与健康管理系统(PHM)是涉及多层次、多领域、耦合度高的复杂系统,很难从整体对PHM系统进行设计和验证,所以始终没有成熟的工具和方法；从数据驱动的视角,研究了飞机PHM系统设计及验证；以数据为主线,采用了4种方法实现PHM系统设计及验证:1)采用系统建模语言(SysML)进行PHM系统的架构设计；2)采用基于事件的方法将被观测系统的行为状态化,实现PHM系统中资源的充分利用,降低局部的资源压力；3)采用基于模型的推理方法,实现了系统状态模型的应用；4)采用多线程、异构能量耦合等方法定义了一个PHM仿真框架,并在该框架指导下开发了一个仿真环境,并针对机载配电系统设置了几种测试用例进行诊断推理,对提到的各种方法进行了验证；研究结果表明,上述方法是合理有效的,可以构建描述完整、资源均衡的PHM系统架构。

Abstract:

PHM system is a multi-layers, multi-disciplines, high-coupled complex system and hard to be designed and verified as a whole, so there are no mature tools and methods for PHM design and verification. Design and verification of aircraft PHM system is researched from data-driven viewpoint. Four methods are used:1)System Modeling Language(SysML) could be used for the design of PHM system under Data-Driven Viewpoint; 2)Event-based method could convert the monitored system’s behavior to some states which will utilize resource of PHM system adequately so as to reduce the local resource pressure;3)Model-based reasoning method is introduced and system’s state model is implemented. 4)Eventually, a multi-thread method and a heterogeneous energy-coupled method are used to define a PHM simulation framework under which develops a simulation environment and verifies mentioned methods, and some test cases of power distributing system are developed for reasoning. The result of research shows that methods mentioned above are valid and could be used for construction of PHM system architecture completely and resource-balanced.