Hydrogen fuel cells have the potential to dramatically reduce emissions from the energy sector, particularly when integrated into an automotive application. However, there are three main hurdles to the commercialisation of this promising technology; one of which is reliability. Cur- rent standards require an automotive fuel cell to last around 5000 h of operation (equivalent to around 150,000 miles), which has proven difficult to achieve to date. This hurdle can be overcome through in-depth reliability analysis including techniques such as Failure Mode and Effect Analysis (FMEA), Fault Tree Analysis (FTA) and Petri-net simulation. This research has found that the reliability field regarding hydrogen fuel cells is still in its infancy, and needs development, if the current standards are to be achieved. In this research, a detailed reliability study of a Polymer Electrolyte Membrane Fuel Cell (PEMFC) is undertaken. The results of which are a qualitative and quantitative analysis of a PEMFC. The FMEA and FTA are the most up to date assessments of failure in fuel cells developed using a comprehensive literature review and expert opinion. Advanced modelling of fuel cell degradation logic was developed using Petri-net modelling techniques. 20 failure modules were identfied that represented the interactions of all failure modes and operational parameters in a PEMFC. Petri-net simulation was used to overcome key pitfalls observed in FTA to provide a verfied degradation model of a PEMFC in an automotive application, undergoing a specific drive cycle, however any drive cycle can be input to this model. Overall results show that the modeled fuel cell's lifetime would reach 34 hours before falling below the industry standard degradation rate of more than 5%. The degradation model has the capability to simulate fuel cell degradation under any drive cycle and with any operating parameters. A fuel cell test rig was also developed that was used to verify the simulated degradation. The rig is capable of testing single cells or stacks from 0-470W power. The results from the verification experimentation agreed strongly with the degradation model, giving confidence in the accuracy of the developed Petri-net degradation model. This research contributes greatly to the field of reliability of PEMFCs through the most up-to-date and comprehensive FMEA and FTA presented. Additionally, a degradation model based upon Petri-nets is the first degradation model to encompass a 1D performance model to predict fuel cell life time under specific drive cycles.

La membrane conductrice protonique constitue un rouage essentiel du fonctionnement des piles à combustible PEMFC. Les travaux de recherche présentés dans ce document consistent à développer une membrane non perfluorée basée sur une technologie nanocomposite hybride susceptible d’être produite à faible coût. Cette membrane est composée d’une matrice poly(VDF-co-HFP) dans laquelle sont dispersées des nanoparticules de silice fonctionnalisée par de l’acide poly(styrène sulfonique) (PSSA). Ce travail a porté sur l’étude de la mise en oeuvre de la membrane afin d’obtenir une membrane homogène et dense avec des caractéristiques physico-chimiques et électrochimiques intéressantes. Les performances en pile après rodage à 60 °C sont extrêmement satisfaisantes avec un gain en densité de puissance de 40 % à 0,7 V par rapport au Nafion® NRE211. Les études de durabilité de la membrane ont mis en évidence un phénomène d’élution de la silice fonctionnalisée ayant pour conséquence un fort déclin de tension. Différentes stratégies de modification de la membrane ont été proposées pour améliorer la stabilité de la membrane. Les plus intéressantes consistent à modifier la morphologie de la matrice (grades de PVDF plus rigides ou réticulation du poly(VDF-co-HFP) par irradiation) pour mieux confiner les charges ou à greffer la silice fonctionnalisée sur la matrice. Cette dernière stratégie conduit à une diminution par trois du gonflement et par 2,5 de la vitesse de déclin à 80°C. / The proton conductive membrane is an essential part of the operation of PEMFC. This document presents the development of a non-perfluorinated membrane based on a hybrid nanocomposite technology likely to be produced at low cost. This membrane is composed of a poly(VDF-co-HFP) matrix in which are dispersed poly(styrene sulphonic acid) (PSSA) functionalized silica nanoparticles. This work focuses on the study of the implementation of the membrane to obtain a homogeneous and dense membrane with good physicochemical and electrochemical characteristics. Fuel cell performances after running at 60 °C are extremely satisfactory with a gain, compared to Nafion NRE211, of 40% for the power density at 0.7 V. However, the durability studies showed an elution phenomenon of the functionalized silica particles which results in a high voltage decline. Different membrane modification strategies have been proposed to improve the stability of the membrane. The most interesting involve modifying the morphology of the matrix (more rigid grades of PVDF or poly(VDF-co-HFP) crosslinking by radiation) to better confine the particles or grafting functionalized silica to the matrix. This last strategy leads to a threefold decrease of the swelling and 2.5 factor of the decay rate at 80 °C.