应用相变材料与预测模型优化质子交换膜燃料电池

Abstract Rising global energy demands and environmental concerns necessitate significant advancements in efficient and sustainable energy technologies. Proton exchange membrane fuel cells represent a promising technology for clean energy generation. However, performance inconsistencies and thermal management challenges during downtime limit their practical application. Therefore, this study employs a comprehensive approach to enhance the performance and efficiency of proton exchange membrane fuel cells. It leverages response surface methodology and artificial neural networks for predictive modeling and optimization, as well as phase change materials for maintaining optimal conditions during downtime. A central composite design was implemented to evaluate the influence of critical operational parameters, including temperature, pressure, and inlet flow rates, on the power density. The developed response surface methodology and artificial neural networks models demonstrated high predictive accuracy, with coefficient of determination values of 98.66 % and 99.11 %, respectively. Optimization results revealed that a temperature of 79.1 °C, a pressure of 200 kPa, and anode and cathode inlet flow rates of 5 L per minute yielded a maximum power density of 1.71 W cm −2 . Furthermore, the innovative thermal management solution using phase change materials with insulation extended the operating temperature range duration to 6.43 h at 25 °C ambient, nearly 5 times longer than using insulation alone. Additionally, in cold environments at −20 °C, this approach increased the operating temperature range and above freezing point duration by 3.5 and 2.7 times, respectively. These findings contribute to the advancement of fuel cell technology by improving performance and thermal management strategies.