Parameter Estimation of Large Scale Systems using Dynamic Data Monitoring : Application to Plume Dispersion Phenomenon

In this research, the effects of data measurement on source parameter estimation are studied. The concept of mutual information is applied to locate the optimal location for each sensor to improve the accuracy of the overall estimation process. For validation purposes, an advection diffusion simulation code, SCIPUFF (Second-order Closure Integrated PUFF) is used as a modeling testbed to study the effects of using dynamic data measurement. Bayesian inference framework is utilized to perform source parameter estimation using stationary and mobile sensor networks, where in mobile sensors, the proposed approach of Dynamic Data Monitoring is used to locate mobile data observation sensors. As our numerical simulations show, using dynamic data monitoring leads to a considerably better estimate of the source parameters, while just using fewer sensors, than the stationary sensors case, or other alternative approaches.