Development of NOx reduction system utilizing artificial neural network (ANN) and genetic algorithm (GA)

Abstract

Abstracts A submerged combustion vaporizer (SCV) is normally used at a liquefied natural gas (LNG) terminal to vaporize LNG and supply natural gas to domestic and industrial markets. In this study, a new methodology was developed to reduce NOx efficiently in the flue gas of the SCV. The experimental setting of the SCV was prepared, and the impact of various process variables (e.g., excess O2 concentration, temperature, pH of water, and H2O2 concentration) on NOx reduction was investigated. For this, a surrogate model was developed by using an artificial neural network (ANN) algorithm based on the experimental data. The contributions of the input variables on NOx removal were assessed using sensitivity analysis. A genetic algorithm (GA) was integrated with the model to optimize the variables to maximize the removal of NOx in the flue gas. Finally, the ANN-GA model was validated through experimental verification, performed under the identified optimum condition.