Xinyi Liu

Publication Date


Document Type

Honors Thesis




Proton exchange membrane fuel cells, Thermodynamics, Chemical models, Miniature PEM fuel cell, Modeling, Dynamic


Energy plays a crucial role in supporting human activities and is essential to daily life. For the past century, people relied heavily on fossil fuels that are non-renewable. There are a number of environmental concerns resulting from consuming fossil fuels, including the release of carbondioxide and other greenhouse gases into the atmosphere. There is an urgent need to find a clean energy source to replace coal, oil, and natural gas. A fuel cell is an electrochemical engine that produces work through the conversion of hydrogen and oxygen to water. It is a clean, efficient power system that can be applied in many fields. A fuel cell's performance depends on many factors. Two important factors are temperature and humidity and they have a coupled effect on fuel cell performance. When the fuel cell systems operate at extreme ambient temperature, pressure and humidity conditions, monitoring and control become critical for successful operation. With a good model of fuel cell response, we are able to monitor and control the cell in order to increase power and energy density. A fuel cell response depends on the range of operating conditions experienced as well as the system architecture. For miniature fuel cells, the application of interest in this work, the fuel cell operates with a dry reactant supply, one of the most challenging operational conditions. However, it results in mass reductions that lead to very promising specific energy densities(Wh/g). In order to achieve stable performance under these conditions, thermal and mass dynamics must be carefully regulated. In this paper, we extend an existing one-control-volume model, first to a two-control-volume thermal mode and then to a multi-control-volume mass and thermal model, using the experimental data generated from a miniature fuel cell. Those models are tuned and experimentally validated. The one-control-volume model is not able to decouple gas transients from load transients, and thus can not be used for temperature regulation in this application. Adding a cathode control volume results in an accurate temperature estimation, except in the region where water evaporation or condensation takes place. The multi-CV-model combines mass dynamics with thermal dynamics by adding several mass and temperature states. Because of its added complexity and increased number of model states, the multi-CV-model is computationally intensive. The multi-CV-model is currently hand-tuned and is able to predict the direction of the temperature response. Future work is recommended to address computational stability relating to the evaporation and condensation dynamics to enable standard automated parameter identification techniques. Model order reduction techniques could then be performed to reduce the number of modeled states as well as the computational time




xiv, 57 pages : illustrations (some color). Honors Project-Smith College, 2014. Includes bibliographical references (pages 56-57)