Fuel cells major obstacles in its wide spread adoption are high costs, use of expensive materials & coatings, lack of mass-production technologies for achieving economies of scale, and difficulties in predicting the performance and durability of individual cells, or integrated stacks, and complete fuel cell systems with all its sub-systems. The underlying behaviour in cell performance and its manufacturing process is usually described by the principles of electrochemistry and thermodynamics, which are often difficult to model mathematically, resulting in labor-driven, error- prone process with high rate of defects, sub-optimal materials, and less-than-desirable electrochemical performance. In this white paper, we have attempted to look at several key aspects of fuel cell design, operational control, material development, and manufacturing process control, with an aim to find cost-effective solutions through machine learning, artificial intelligence and data analysis techniques, which have received increasing attention recently. Machine learning models have unique advantages in running simulations, as they require no prior knowledge, especially of the complex coupled transport and electrochemical processes occurring in fuel cell operation.

Updated On: June 2021