School of Computer and Information Science, Edith Cowan University, Perth, Western Australia
A novel methodology for solving the chemical plume tracing problem that utilizes data from a network of stationary sensors has been developed in this study. During a toxic chemical release and dispersion incident, the imperative need of first responders is to determine the physical location of the source of chemical release in the shortest possible time. However, the chemical plume that develops from the source of release may evolve into a highly complex distribution over the entire contaminated region, making chemical plume tracing one of the most challenging problems known to date. In this study, the discrete Fourier series method was applied for re-construction of the contour map representing the concentration distribution of chemical over the contaminated region based on point measurements by sensors in a pre-installed network. Particle Swarm Optimization was then applied to the re-constructed contour map to locate the position of maximal concentration. Such a methodology was found to be highly successful in solving the chemical plume tracing problem via the sensor network approach and thus closes a long-standing gap in the literature. Furthermore, the nature of the methodology is such that a visual of the entire chemical dispersion process is made available during the solution process and this can be beneficial for warning purposes and evacuation planning. In the context of such chemical release scenarios, the algorithm developed in this study is believed to be able to play an instrumental role towards national defense for any country in the world that is subjected to such threats.