PEM fuel cell systems in automotive application face challenging power requirements, as they must provide full load as well as part load efficiently. Especially in systems without an external humidifier, relatively low cathode stoichiometries and increased operating pressures are the key to prevent membrane dry-out by increasing the local relative humidity. During low load operation at increased temperature and pressure, low gas velocities within the flow field occur. Too low gas velocities are known to result in a transition of the two-phase flow from film flow to slug flow regime [1]. This transition within the flow field channel structure can result in a local or global blockage of the reactant paths by liquid water droplets.

While the flow regime transition within simple channel geometries is well known in literature, there is a lack of experimental data for full stack situations in fuel cell systems. Some work regarding the stationary minimal flow limitation exists [2], but countermeasures enabled by dynamic operation of the whole system are seldom in focus of systematic studies. However, well guided dynamic operation enables robust low load capability. The negative impact of flooding on stack power and lifetime can be minimized and a possible dry-out can be prevented, while maintaining a relatively low required compressor energy consumption.

In the work to be presented, targeted experiments conducted with an automotive fuel cell system confirm the existence of a minimum cathode gas velocity for a stable operation and its relevance in application-like operation. It is shown that during low load operation of a non-humidified system, the targets of sufficient membrane humidification and liquid water drainage are incompatible. Because of a response time of multiple seconds until flooding or dry-out occurs, an intermittent operation with periodic switching between two operating points becomes feasible. During the first phase, the membrane is humidified due to a relatively low cathode stoichiometry and an increased cathode inlet pressure. As the liquid water accumulation becomes critical, the pressure is decreased and the stoichiometry is increased for the second phase, where droplets are blown out by utilizing a high gas velocity, as shown in figure 1.

The exemplary results suggest that dynamic water management analysis can widen the operating range of a fuel cell system towards lower loads due to dynamic control. It is expected that a deeper understanding of the underlying processes will allow the development of optimized control strategies in the future.

[1] M. Grimm, E.J. See, S.G. Kandlikar. Modeling gas flow in PEMFC channels: Part I – Flow pattern transitions and pressure drop in a simulated ex situ channel with uniform water injection through the GDL. Int. Journal of Hydrogen Energy 37, 12489-12503 (2012).
[2] D. Jenssen, O. Berger, U. Krewer. Anode flooding characteristics as design boundary for a hydrogen supply system for automotive polymer electrolyte membrane fuel cells. Journal of Power Sources 298, 249-258 (2015).

Figure 1: Normalized current, high frequency resistance, cell voltages, cathode stoichiometry, cathode inlet pressure and cathode inlet velocity, measured on a system test rig during intermittent low load operation. The operating point is periodically switched between membrane humidification (green) and liquid water drainage (orange).