Search for a command to run...
Fuel cell durability being one of the technical bolts regarding the technology industrialization in the automotive sector, durability improvement methods are particularly relevant. Fault tolerant control process enables to increase fuel cell durability by detecting and correcting fuel cell faults in real time. Fuel cells are prone to faults because they are very sensitive to operating conditions. In vehicle application, fault risk is exacerbated as dynamic conditions are often encountered. Dynamic conditions make the fuel cell control harder because it impacts reactants supply, thermal management, water management… If not corrected, those faults degrade the fuel cell and reduce its remaining useful lifetime. Fault tolerant control consists in diagnosing faults, then taking corrective actions to resolve those faults. This article treats the diagnosis part, which consists in detecting and identifying faults, in vehicle application. Vehicle applications engender several constraints as the reduced cost, the hydrogen usage and computation limitations or the safety regulations for algorithms implementation. Three steps are necessary for diagnosis: real time measurements, useful information extraction, and classification. In this article, a state-of-the-art of methods for each of these steps independently is presented. In the last section, useful explanations to convert offline diagnosis algorithm into an embedded diagnosis tool are provided.