Investigation of Sarin Gas Dispersion in an Indoor Environment: A CFD-ANN Study

Abstract

This study investigated sarin gas dispersion in an indoor environment using transient three-dimensional Computational Fluid Dynamics (CFD) and Artificial Neural Network (ANN) approaches. To achieve this, the CFD model was first verified and validated. Then, random locations in the indoor environment were considered as inlets of airflow with sarin gas, and the dangerous times were calculated using the CFD model. Finally, these results of the CFD model were used as inputs to train the ANN model. The results of this study demonstrated that the present model exhibited strong agreement with experimental data. Also, the results of training the ANN showed that for all sections, the training, validation, and testing data and model results were consistent with a high R-squared value. Moreover, the results of different air inlet locations showed that if the air inlet was placed in the corner sections of the indoor environment, the danger time increased. Additionally, if the air inlet was placed near the open region, the danger time also increased, which is an important result for designing indoor environments.