No abstract

The existing models of proton-exchange membrane fuel cells, such as computational fluid dynamics and lumped-parameter models, cannot provide a comprehensive description of mass-transport mechanisms without incurring a high computational load. This paper proposes a reduced-dimension dynamic model of fuel cells that considers main mass-transport mechanisms, including two-phase flow in the gas diffusion layer. Under the quasi-steady-state assumption, the distributions of liquid water and gas species in the through-plane direction are calculated through analytical expressions. Thus, the reduceddimension fuel-cell model can obtain a similar output to a one-dimensional model without needing to solve partial differential equations. The reduceddimension model was validated on steady state and transient experimental data. The computational load of the reduced-dimension model was only 1/60 that of a one-dimensional model. To simulate the transient behavior of the fuel-cell system, the reduced-dimension fuel-cell model was integrated with a comprehensive balance-of-plant model. The system model could execute longtime-scale transient simulations and evaluate the internal states of the fuel cell while minimizing the computing resources. Finally, the contradiction between the power response rate and oxygen starvation under the load-delay strategy was analyzed with the system model, and the delay duration was optimized.Results verify the potential of the reduced-dimension model in controloriented applications. Novelty statementThe existing models of proton-exchange membrane fuel cells, such as computational fluid dynamics and lumped-parameter models, cannot provide a comprehensive description of mass-transport mechanisms without incurring a high computational load. This work proposes a reduced-dimension dynamic model of fuel cells that considers main mass-transport mechanisms, including two-phase flow in the gas diffusion layer. Under the quasi-steady-state assumption, the distributions of liquid water and gas species in the through-plane direction are calculated through analytical expressions. Thus, the reduceddimension fuel-cell model can obtain a similar output to a one-dimensional model without needing to solve partial differential equations. The

A new concept offaulty component detection via robust observation is proposed. An algorithm is developed and computer coded for failure-diagnosis-system design using a robust observation approach. Such diagnosis systems can locate system failures to first-order components without a priori information on possible failure types.

No abstract

No abstract