UC Davis Institute of Transportation Studies

For passenger fuel cell vehicles (FCVs), customers will expect to start the vehicle and drive almost immediately, implying a very short system warmup to full power. While hybridization strategies may fulfill this expectation, the extent of hybridization will be dictated by the time required for the fuel cell system to reach normal operating temperatures. Quick-starting fuel cell systems are impeded by two problems: 1) the freezing of residual water or water generated by starting the stack at below freezing temperatures and 2) temperature-dependent fuel cell performance, improving as the temperature reaches the normal range. Cold start models exist in the literature; however, there does not appear to be a model that fully captures the thermal characteristics of the stack during sub-freezing startup conditions. Existing models do not include stack internal heating methods or endplate thermal mass effect on end cells.

The focus of this research is the development and use of a sub-freezing thermal model for a polymer electrolyte fuel cell stack and system designed for integration within a direct hydrogen hybrid FCV. The stack is separated into individual cell layers to determine an accurate temperature distribution within the stack. Unlike a lumped model, which may use a single temperature as an indicator of the stack's thermal condition, a layered model can reveal the effect of the endplate thermal mass on the end cells, and accommodate the evaluation of internal heating methods that may mitigate this effect.

This research is designed to answer the following motivating questions:

* What detailed thermal model design will accurately characterize the fuel cell stack and system during the sub-freezing startup operation? * What are the effects of different startup strategies on energy consumption and time to normal operation?

These questions are addressed in this dissertation. Major research findings include the following recommendations for the best startup strategies based on model parameter values and assumptions: 1) use internal heating methods (other than stack reactions) below 0ºC, 2) circulate coolant for uniform heat distribution, 3) minimize coolant loop thermal mass, 4) heat the endplates, and 5) use metal such as stainless steel for the bipolar plates.