ERI: Efficient identification of passive scalar source in turbulence using the Probabilistic Domain of Dependence (P-DoD)

San Diego State University Foundation, San Diego CA

Investigators

Abstract

This project addresses the complex challenge of tracking and locating sources of scalar fields like chemical leaks, oil spills, or unusual heat emissions, often obscured or altered by the unpredictable nature of atmospheric and aquatic currents. The research is driven by the need for efficient, accurate tools that can interpret complex data and pinpoint the origin of these pollutants, even without complete knowledge of the environment. To address such needs, the project proposes a revolutionary approach using the domain of dependence, a concept derived from the governing equations of fluid dynamics, for a network of sensors. The project also aims to understand the intricate dynamics between pollutants and turbulent flows, leveraging the probability distribution of these domains of dependence to effectively interpret sensor data and give uncertainty bounds for the prediction. The implications of this work extend beyond pollution detection, offering new insights and tools for weather forecasting and disaster response, thereby contributing to a safer, more resilient society. The core of this research is to develop and utilize a novel framework for computing and utilizing the probability distribution of the domain of dependence in turbulent, uncertain flows. These dependence fields can be obtained from the adjoint simulations of the scalar transport equations, where flow physics is integrated into scientific search strategies of localized scalar sources. By aggregating multiple domains of dependence, the project designs a geometric approach to locate the scalar source, enhancing accuracy and reliability. A pivotal aspect of this research involves adapting to the inherent uncertainties in the flow fields by utilizing the probabilistic nature of the probabilistic domain of dependence, which allows for a more nuanced and informed interpretation of sensor data. Uncertainties in the prediction can also be evaluated and mitigated by optimizing sensor networks to maximize information extraction. The project will use advanced sensitivity analysis techniques to exploit the probabilistic domain of dependence for sensor measurements, providing a new method for optimizing sensor placement and trajectory in pollution detection. This research is not only transformative for environmental science and urban planning but also holds the potential for significant advancements in the detection of extreme weather events and other anomalies in fluid dynamics. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.