Inscrição na biblioteca: Guest

ISSN Online: 2379-1748

7th Thermal and Fluids Engineering Conference (TFEC)
SJR: 0.152 SNIP: 0.14 CiteScore™:: 0.5

Indexed in

Clarivate CPCI (Proceedings) Scopus
May, 15-18, 2022 , Las Vegas, NV, USA

ANALYSIS OF DYNAMIC WATER TRANSPORT IN PEM FUEL CELL GAS DIFFUSION LAYERS USING LONG SHORT TERM MEMORY

Get access (open in a dialog) pages 1327-1343
DOI: 10.1615/TFEC2022.ees.040804

Resumo

The Long Short Term Memory model, a recurrent neural network model is utilized to predict the state of liquid water in the flow channel of a PEM fuel cell. Ex-situ experimental data is used for training and testing. The input features of the model are based on the measured pressure drop across the channel, water injection rate, and air flow rates. The water content is estimated using the wetted area ratio obtained from high speed images of water in a transparent test channel. The model is designed to predict the flooding risk and the water accumulation over time within the flow channel. The goal of this study is to establish a modeling framework that uses pressure drop, which can be practically measured as the main input and is capable of estimating flow field liquid content as an output.