This work presents the development of a multi-input, multi-output neural network structure to predict the time dependent concentration of chemical agents as they participate in chemical reaction with environmental substrates or moisture content within these substrates. The neural network prediction is based on a computationally or experimentally produced database that includes the concentration of all chemicals presents (reactants and products) as a function of the chemical agent droplet size, wind speed, temperature, and turbulence. The utilization of this prediction structure is made userfriendly via an easy-to-use graphical user interface. Furthermore, upon the knowledge of the time-varying environmental parameters (wind speed and temperature that are usually recorded and available), the time varying concentration of all chemicals can be predicted almost instantaneously by recalling the previously trained network. The network prediction was compared with actual open air test data and the results were found to match.
A general-purpose multi-phase and multi-component computer model capable of solving the complex problems encountered in the agent substrate interaction is developed. The model solves the transient and time-accurate mass and momentum governing equations in a three dimensional space. The provisions for considering all the inter-phase activities (solidification, evaporation, condensation, etc.) are included in the model. The chemical reactions among all phases are allowed and the products of the existing chemical reactions in all three phases are possible. The impact of chemical reaction products on the transport properties in porous media such as porosity, capillary pressure, and permeability is considered. Numerous validations for simulants, agents, and pesticides with laboratory and open air data are presented. Results for chemical reactions in the presence of pre-existing water in porous materials such as moisture, or separated agent and water droplets on porous substrates are presented. The model will greatly enhance the capabilities in predicting the level of threat after any chemical such as Toxic Industrial Chemicals (TICs) and Toxic Industrial Materials (TIMs) release on environmental substrates. The model’s generality makes it suitable for both defense and pharmaceutical applications.