This thesis addresses the mechanical degradation of the proton exchange membrane and gas diffusion layer in a low-temperature fuel cell. A computationally efficient two-dimensional model based on a discrete spring lattice is developed, enabling the evaluation of spatially and temporally varying stress–strain states. The model accounts for the effects of temperature, relative humidity, material anisotropy, and the nonlinear mechanical response of the membrane through a simplified representation of time-dependent material properties. Dynamic loading is treated using the rainflow counting method to identify and count individual loading cycles. The identified cycles are subsequently coupled with a constitutive description of the mechanical response and a linear damage accumulation model, enabling the identification of critical loading cycles, the assessment of their contribution to the total accumulated damage, and the estimation of the number of cycles to failure. The results demonstrate that the model is capable of determining spatially and temporally varying stress–strain states, identifying critical loading cycles, and estimating their contribution to accumulated damage. The numerical behaviour of the model is evaluated through convergence analyses performed at different levels of discretisation refinement, while comparison with related research from the literature indicates agreement with key physical trends associated with constrained hygrothermal membrane expansion and cyclic mechanical loading. Owing to its reduced formulation and low computational cost, the developed model has the potential to be applied for the real-time evaluation of the mechanical state of the membrane and, in the future, to be integrated with monitoring and control systems for low-temperature proton exchange membrane fuel cells.
|
