Synthetic Event Reconstruction Experiments for Defining Sensor Network Characteristics
An event reconstruction technology system has been designed and implemented at Lawrence Livermore National Laboratory (LLNL). This system integrates sensor observations, which may be sparse and/or conflicting, with transport and dispersion models via Bayesian stochastic sampling methodologies to characterize the sources of atmospheric releases of hazardous materials. We demonstrate the application of this event reconstruction technology system to designing sensor networks for detecting and responding to atmospheric releases of hazardous materials. The quantitative measure of the reduction in uncertainty, or benefit of a given network, can be utilized by policy makers to determine the cost/benefit of certain networks. Herein we present two numerical experiments demonstrating the utility of the event reconstruction methodology for sensor network design. In the first set of experiments, only the time resolution of the sensors varies between three candidate networks. The most ''expensive'' sensor network offers few advantages over the moderately-priced network for reconstructing the release examined here. The second set of experiments explores the significance of the sensors detection limit, which can have a significant impact on sensor cost. In this experiment, the expensive network can most clearly define the source location and source release rate. The other networks provide data insufficient for distinguishing between two possible clusters of source locations. When the reconstructions from all networks are aggregated into a composite plume, a decision-maker can distinguish the utility of the expensive sensor network.
- Research Organization:
- Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
- Sponsoring Organization:
- USDOE
- DOE Contract Number:
- W-7405-ENG-48
- OSTI ID:
- 894010
- Report Number(s):
- UCRL-TR-217762; TRN: US200701%%124
- Country of Publication:
- United States
- Language:
- English
Similar Records
Dynamic Data-Driven Event Reconstruction for Atmospheric Releases
Dynamic Data-Driven Event Reconstruction for Atmospheric Releases