Design of the Local Three-phase Interface and Machine Learning Assisted Life Prediction of Low-Platinum PEMFCs

With the increasing energy demand, fuel cells have drawn extensive attention as highly efficient and clean energy conversion devices. Nevertheless, the high cost and poor durability of membrane electrode assemblies (MEAs) in fuel cells have severely hindered their large-scale application. This makes the reduction of platinum usage and the improvement of durability become the two important research areas. The three-phase interface plays a pivotal role in the electrochemical reaction process of fuel cells. Accurate design of its structure is the core strategy for enhancing the performance and durability of low-platinum fuel cells. Meanwhile, the durability test requires a lot of power and material resources. The simplification of the life assessment method is the key to improve the development efficiency. In this study, the three-phase interface was precisely regulated by adjusting the distribution of ionomers and optimizing the structural parameters of catalyst layer, thereby improving the performance of MEA (2A/cm2@0.71V; 0.5A/cm2@0.82V) at a low-platinum loading. In addition, the durability of MEAs was accurately predicted by machine learning. Based on the fuel cell durability data set, the attenuation prediction model is verified that the life prediction errors are less than 5%. Related achievements have effectively enhanced the performance of MEAs and reduced the amount of platinum used in fuel cells. This research provides an important theoretical foundation and technical support for the further development and commercial application of low-platinum fuel cells. It is expected to promote the wide-spread application of fuel cells in the energy sector.