Plume Source Detection using a Process Query System

A Process Query System (PQS) has the capability of filtering large volumes of real time data originating from a field of networked Physical Sensors. Modern air quality monitoring techniques such as Fourier Transform Infra-Red (FTIR) spectroscopy will eventually provide massively distributed real time contamination data at high fidelity. As large networks of these sensors are deployed, improved techniques of data retrieval and assimilation will be required. The case of detecting a diffusion event such as a hazardous chemical plume is considered. In this scenario, a plume model based on an Ensemble Kalman Filter (EnKF)is submitted to the PQS which manages multiple hypotheses explaining the current observations. The feasibility of such an application is demonstrated and results from preliminary simulations are presented.