Development of Artificial Neural Network based Metamodels for the Inactivation of anthrax spores in ventilated spaces using Computational Fluid Dynamics

SHAMIA HOQUE (1) Bakhtier Farouk (2) Charles N. Haas (1)

(1) Department of Civil, Architectural and Environmental Engineering, Drexel University (2) Department of Mechanical Engineering and Mechanics, Drexel University

     Abstract Number: 425
     Last modified: May 12, 2010

Abstract
Decontamination of large indoor spaces and buildings following release of biological agents is challenging, as the response to the fall 2001 anthrax-release events indicate. The ability to efficiently and rapidly decontaminate rooms/buildings is limited by the lack of quantitative understanding of the behavior of biological agents and decontaminants in relation to building geometry, airflow pattern and surface properties. In response to any potential new releases it would be necessary to rapidly determine the optimal way to decontaminate the enclosed spaces.
In the present study, artificial neural network based metamodels have been developed for predicting the number of viable bioaerosols remaining in an arbitrary enclosed space after sometime when a specific disinfectant is being applied. Computational fluid dynamics was applied to numerically simulate the transport of the bioaerosol (e.g., spores of Bacillus anthracis) and inactivation by a decontaminant (chlorine dioxide). A comprehensive CFD model was first developed to generate accurate data for developing a metamodel. Large eddy simulation (LES) technique was applied to compute the airflow. Anthrax spores were modeled using the Lagrangian treatment. Disinfectant mass fraction was calculated by solving a convective/diffusive mass transport equation. Kinetic decay constants were included for spontaneous decay of the disinfectant and for the reaction of the disinfectant with room surfaces. An inactivation rate equation accounted for the reaction between the spores and the disinfectant. To obtain better mixing a momentum source was included in the CFD model. The significant dimensionless groups influencing bioaerosol inactivation was identified and multiple scenarios were simulated. The study showed successfully the ability to apply accurate CFD results to develop neural network based metamodels for quick and accurate predictions for bioaerosol inactivation.