PEM fuel cell performance and lifetime strongly depend on the polymer membrane and MEA hydration. As the internal moisture is very sensitive to the operating conditions (temperature, stoichiometry, load current, water management…), keeping the optimal working point is complex and requires real time monitoring. This article is centered in the PEM fuel cell stack health diagnosis and more precisely on stack fault detection monitoring. This paper intends to define new, simple and effective methods to get relevant information on usual faults or malfunctions occurring in the fuel cell stack. For this purpose, the authors present a fault detection method using simple and non-intrusive on-line technique based on the space signature of the cell voltages. The authors have the objective to minimize the number of embedded sensors and instrumentation in order to get a precise, reliable and economic solution in a mass market application. A very low number of sensors are indeed needed for this monitoring and the associated algorithm can be implemented on-line. This technique is validated on a 20-cell PEMFC stack. It demonstrates that the developed method is particularly efficient in flooding case. As a matter of fact, it uses directly the stack as a sensor which enables to get a quick feedback on its state of health.
This paper scope is centered in the PEM fuel cell stack health diagnosis. For this purpose, authors present fault detection and identification methods using simple and non-intrusive on-line monitoring techniques. The approach is gradual based on detection and identification methods applied to a single cell up to multi-cells stacks used for power applications like transportation. A very low number of sensors are needed for the monitoring and the technique can be implemented on-line. Numerical simulation results illustrate the advantages of the different techniques.
In transport application, long high power PEM fuel cell stacks could suffer from voltage discrepancy between cells due to severe constraints or appearance of localized faults in case of bad water management, moreover the output power of the stack is limited by the weakest group of cells. This article proposes a three-part segmented fuel cell associated with an isolated four-source DC/DC converter which makes possible a power sharing between the fuel cell segments according to their state-of-health. This topology allows an enhanced utilization of the stack as well as fault tolerance. ZVS operation is achieved to improve global converter efficiency.
To fulfill the transport applications, either for traction or on-board auxiliaries systems, a power generator based on fuel cell needs significant power. For this purpose, long fuel cell stacks either mono or multi-stack systems are already implemented as technological solutions. Long stacks may though be affected by spatial discrepancies (fluidics, temperature) causing possible failures. The latter often occur on localized stack sections. A corrective action has to be taken to quickly restore the fuel cell state-of-health. As an alternative to fluidic action, segmented electrical action is explored in this paper. First, an "All or Nothing" solution achieved with electrical bypass circuits is analyzed: it proved simple to implement but restrictive to exploit. Consequently, a "gradual" action is proposed by using the power electronics converter associated to the fuel cell. Hence, the present work investigates the approach consisting in individually driving the electric power delivered by each segment of a long Polymer Electrolyte Membrane Fuel Cell stack. Each segment is controlled independently according to its state of health. To achieve this objective, the article provides an extended multi-criteria analysis of several power converter topologies. The converter topology has to be in agreement with transportation specifications: simple, compact, having a high efficiency and should be adapted to manage fuel cell degraded modes. Among several studied topologies, resonant isolated boost stands out as a candidate topology. The related multi-port architecture and algorithm structure are analyzed by numerical simulations, taking into account degraded modes and technology considerations.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.