(345f) The Evaluation of the Feeding Effect on Liquid-Feed Dmfc Using Rigorous Dynamic Simulation

The objective of this research is to evaluate the feeding effect on liquid-feed DMFC (Direct Methanol Fuel Cell) using the combined discrete dynamic simulation technique. A one-dimensional, two-phase rigorous dynamic model for a single-cell liquid-feed DMFC is developed and simulated dynamically. To simulate the feeding effect of methanol on the system, several feasible feeding strategies are selected and mathematically modeled including discrete events that make the simulation difficult to solve. The simulation results of various feeding strategies show that feeding methods have an effect on the dynamic behavior of the system and the dynamic response of the system changes according to the feed flow rate as well as the feed concentration of methanol. The feeding effect on liquid-feed DMFC system can be fully understood to reduce undesired phenomena such methanol crossover. The liquid-feed DMFC system recently gains a lot of interest in fuel cell research area due to its possibility of miniaturization and commercialization, and lots of researches have focused especially on the modeling and simulation of liquid-feed DMFC to prevail the internal phenomena that experiments cannot show. Wang and Wang (Journal of the Electrochemical Society, 150(4), A508-A519, 2003) developed a two-dimensional, two-phase model of liquid-feed DMFC and simulated the model using computational fluid dynamics to show the mixed potential effect of methanol oxidation at the cathode as a result of methanol crossover. Though their model is rigorous enough to describe the internal phenomena like methanol crossover, their results cannot explain the dynamic behavior which arises from the feeding effect on the DMFC system due to their steady-state model. The feed flow rate and concentration of methanol plays a key role in the operation of the DMFC system. If the feed concentration of methanol is too low then the maximum power density cannot be obtained, and if it is too high then the methanol crossover occurs and the performance of the fuel cell decreases significantly. It is therefore very important to find the optimal feed concentration and flow rate of methanol to get maximum power density. Sundmacher et al. (Chemical Engineering Science, 56, 333-341, 2001) have developed a dynamic model and have simulated the model to show that methanol crossover can be reduced by periodically pulsed methanol feeding. Although this periodical feeding reduces the undesired methanol crossover significantly, their research focuses only on the periodical feeding concentration and neglects other feeding method. Xu et al. (Computers & Chemical Engineering, 29, 1849-1860, 2005) performed a dynamic optimization based on the model of Sundmacher et al. to obtain the optimal feeding concentration of methanol for maximizing the total cell voltage. According to their research, the periodically pulsed feeding strategy of Sundmacher et al. may be unnecessary to achieve the optimal cell voltage, and they suggest the optimal feed concentration profile as a function of time that changes slightly as time changes. Nevertheless higher technical depth and accuracy, their research ignores the feasibility and operability of the miniaturized and commercialized DMFC system. In this research, we provide a one-dimensional, two-phase rigorous dynamic model of a single-cell liquid-feed DMFC. The model is rigorous enough not only to describe the dynamics of the DMFC system but also to express the feeding strategy that contains discrete events making the simulation more difficult to reach the convergence. The single-cell of this research divides into seven modules: anode flow field, anode diffusion layer, anode catalyst, membrane, cathode catalyst, cathode diffusion layer, and cathode flow field. The mathematical models of two diffusion layers and membrane have the first order dimensionality resulting in PDE system and the others have no dimensionality resulting in DAE system. Each model contains momentum, mass and energy balances, electrochemical correlations, and many other equations that describe the DMFC rigorously. This mixed PDE and DAE system is solved for each methanol feeding strategy using a conventional dynamic simulator. The simulation results show that the dynamic response of the liquid-feed DMFC system changes as the flow rate and concentration of methanol change. This research is significant in that the feeding effect on the DMFC system is fully evaluated using a rigorous dynamic simulation.