Search

A proton-exchange membrane fuel cell (PEMFC) constitutes today one of the preferred technologies to promote hydrogen-based alternative energies. However, the large-scale deployment of PEMFCs is still hampered by insufficient durability and reliability. In particular, the degradation of the polyelectrolyte membrane, caused by harsh mechanical and chemical stresses experienced during fuel cell operation, has been identified as one of the main factors restricting the PEMFC lifetime. An innovative chemical-mechanical ex situ aging device was developed to simultaneously expose the membrane to mechanical fatigue and an oxidizing environment (i.e., free radicals) in order to reproduce conditions close to those encountered in fuel cell systems. A cyclic compressive stress of 5 or 10 MPa was applied during several hours while a degrading solution (H 2 O 2 or a Fenton solution) was circulated in contact with the membrane. The results demonstrated that both composite Nafion ™ XL and non-reinforced Nafion ™ NR211 membranes are significantly degraded by the conjoint mechanical and chemical stress exposure. The fluoride emission rate (FER) was generally slightly lower with XL than with NR211, which could be attributed to the degradation mitigation strategies developed for composite XL, except when the pressure level or the aging duration were increased, suggesting a limitation of the improved durability of XL.

Sequential processes can hardly be modeled with Colored Petri Nets (CPN). Their standard functions (such as "Succ" and "Prec") can only describe simple cases whereas modeling a complex sequence requires the definition of a function via a cumbersome and static table. In order to overcome these limitations, we first introduce a mixed structure based upon CPN and FIFO queues; then, we define an isomorphism between the set of colors and a finite field Z/pZ enabling symbolic calculation using polynomials associated with arcs instead of linear functions mapping sets into sets. Finally, several examples illustrate their modeling capabilities including failure recovery.