故障检测与隔离
实时计算
计算机科学
健康状况
状态监测
特征提取
模糊逻辑
断层(地质)
遥控水下航行器
工程类
健康管理体系
汽车工程
分类器(UML)
人工智能
能量操作员
控制工程
能源管理
能量(信号处理)
可靠性工程
推进
无人水下航行器
作者
Dan Zeng,Xiaoyu Guo,Jiabin Shen,Zhen Dong,Chenliang Wang,Lu Liu
标识
DOI:10.1109/tie.2025.3603055
摘要
The hydrogen-powered unmanned aerial vehicle (UAV) is a pivotal part of zero-carbon aviation, yet the online health monitoring and management of fuel cells (FCs) during flight remain a predominant challenge. In response to the critical concern, this article initiates an online systematic approach of health monitoring and management (SAHMM) for hydrogen fuel cells in aerial applications consisting of four stages. First, a voltage-based adaptive health monitor continuously monitors the state of health of the fuel cell, detecting faults in real time. Second, to classify the faults, a hybrid excitation operator (HEO) including a converter regulation signal and an unmanned aerial vehicle maneuver signal is injected to extract fault features when a fault is detected. The HEO, as a control-for-diagnosis design, exploits unmanned aerial vehicle maneuvers for fault feature extraction while maintaining position tracking performance, relieving the burden of the energy subsystem during excitation. A fuzzy logic classifier then categorizes the detected fault. Finally, recovery measures according to the fault type will be adopted to restore system health. The proposed SAHMM is the first to tackle the health monitoring and recovery of aviation hydrogen fuel cell systems in a systematic approach. The effectiveness of SAHMM is validated through real flights and ground experiments, where online fault detection and 97.22% fault classifier accuracy are demonstrated, and successful health recovery is achieved based on the diagnosis result. Moreover, the monitor reduces false detection rate by integrating online model estimation to track the fuel cell performance drifts, and a more than 37% reduction on energy subsystem stress is observed thanks to the HEO design.
科研通智能强力驱动
Strongly Powered by AbleSci AI