PEMFC是能有效解决环境污染和能源危机的发电设备，具有高功率密度、高效、零排放等特点，但其可靠性和耐久性限制了它的商业应用，对PEMFC进行状态监测和故障诊断是保证其安全运行的有效方法。本项目以PEMFC为对象，针对其状态监测和故障诊断的难题，提出利用FBG传感器对PEMFC多工况参数进行分布式实时原位检测，并基于深度学习实现PEMFC故障特征的自适应提取和健康状态准确识别。重点围绕PEMFC系统故障与FBG传感信号特征之间的关联规律、PEMFC非线性多源异构传感信号的信息融合方法、基于多退化量和随机阈值的PEMFC故障识别方法等科学问题展开研究，解决PEMFC健康状态的监测、表征和识别等关键技术问题，为PEMFC的状态监测和故障诊断提供一条新的解决思路。本研究对进一步丰富和发展PEMFC的状态监测和故障诊断理论和方法具有重要意义，可为PEMFC安全、稳定、高效且长久地运行提供科学依据。

With the advantages of high power density，high efficiency and environmental friendly property，the proton exchange membrane fuel cell （PEMFC）as a alternative power generator，has been considered to solve the crucial issues in environment and energy. However，reliability and durability are still two barriers which block its wide application. Thus, fault diagnosis is an efficient solution to guarantee the safe operation of the fuel cell. Developing a prognostics methodology for PEMFC has become a complex and complicated process. The distributed real-time in-situ monitoring of PEMFC parameters using fiber Bragg grating（FBG）sensors is proposed. The adaptive extraction of fault features and the accurate identification of health status for PEMFC are realized based on deep learning. Focusing on the mapping relationship between fault features of PEMFC and FBG sensing signals, the information fusion method of non-linear multi-source heterogeneous sensing signals, and the fault identification method of PEMFC based on multi-degradation and multi-source information fusion, the key technical problems such as on-line monitoring, identification and recognition of PEMFC health status are solved, which provides a new solution for real-time in-situ monitoring and fault diagnosis of PEMFC. This research is of great significance to further enrich and develop the real-time monitoring and fault diagnosis technology of PEMFC，and can provide scientific basis for the safe，stable，efficient and long-term operation of PEMFC.