Surrogate modeling of ammonia decomposition reactor and catalyst loading optimization with genetic algorithm

초록

Ammonia as a hydrogen carrier is an emerging route to the hydrogen economy for cleaner and environmentally friendly energy systems. Recovering hydrogen from ammonia by decomposition in ammonia-cracking reactors is an established technology. With ammonia gaining significance as a green hydrogen carrier, it must be studied as a part of larger integrated energy systems. Surrogate modeling of the ammonia decomposition reactor aids real-time optimization and digital twins, yet research work on the topic is scarce, if available at all. To address this gap, a robust data-driven model of an ammonia cracker was developed from CFD-generated data of a reactor producing 200 kg/day of hydrogen. A segment-wise surrogate modeling approach was followed to enable the surrogate model to predict the reactor performance at varying parameter values not only at the reactor inlet but also inside the reactor. This method was successfully applied to vary the kinetic reaction rate along the length of the reactor and find the optimum catalyst loading distribution through the application of the genetic algorithm. The optimization results predicted about 20% savings in catalyst load with less than 1% decrease in ammonia conversion. Valuable insights are gained in the surrogate modeling of the ammonia cracker system.