Predicting the Ammonia Synthesis Performance of Plasma Catalysis Using an Artificial Neural Network Model
Xing Wang, Xuesen Du, Kunlu Chen, Ziwen Zheng, Yanggu Liu, Xiaoqiang Shen, Chenlong Hu
Abstract
Ammonia synthesis in a plasma catalysis system coupling dielectric barrier discharge and an alumina-loaded ruthenium catalyst was investigated. The discharge energy, the N 2 to H 2 ratio, and the reactant flow rate greatly affect the production of NH 3 and the efficiency of energy. Higher ammonia synthesis rates and higher energy yields were obtained at high discharge powers and flow rates. An artificial neural network (ANN) was built to describe the influence of operating parameters on the NH 3 synthesis performance, including ammonia synthesis rate and energy yield. The proposed ANN model was trained using experimental data. The results showed the N 2 to H 2 ratio was the most impactful parameter with a relative importance of 41.7% on the model, followed by the flow rate and discharge power of 32 and 26.3%, respectively. This ANN model can effectively help to optimize the operating parameters of the plasma catalysis system for NH 3 synthesis and predict the catalysis performance under specific situations.