Scope: DART Experiments: CPS: Medium: Data-Driven Adaptive Real-Time (DART) Flow-Field Estimation Using Deployable UAVs

The Fukushima Daiichi nuclear disaster and the Aliso Canyon natural gas leak are recent high- profile examples of emergency situations that resulted from the unplanned release of an airborne chemical, biological, radiological, or nuclear (CBRN) contaminant. In such cases, accurate real-time prediction of contaminant movement is invaluable for planning emergency response, protecting emergency workers, and assessing environmental impact. The objective of this project is to develop and demonstrate a new data-driven adaptive real-time (DART) CPS that is capable of producing accurate real-time micrometeorological estimates and forecasting contaminant dispersion near the source. Several new CPS technologies must be developed to realize the potential of DART flow field estimation. Thus, this proposal targets Technology for Cyber-Physical Systems. Predicting contaminant dispersion is challenging because the atmospheric boundary layer contains unsteady 3-dimensional turbulent flow, which is difficult to forecast. DART CPS aims to produce accurate flow field and CBRN-dispersion estimates by using physical measurements to continually improve a computational fluid dynamic (CFD) model in real time. This project will use autonomous unmanned aerial vehicles (UAVs) to obtain sparse flow and contaminant-concentration measurements, which are used to adapt the CFD. Thus, DART flow field estimation incorporates a DART-CFD cyber system and a UAV-based physical system. The novel DART flow field estimation method will be validated and demonstrated using fixed-wing UAVs, which will provide real-time data for DART-CFD. In turn, DART-CFD will use this data to improve flow field and contaminant-concentration estimates. Then, the UAVs will be autonomously re-routed based on cyber feedback of the DART-CFD-predicted flow and contaminant concentration. This synergy project relies on an interdisciplinary team with expertise in atmospheric CFD, model adaptation, control systems, UAVs, and experimental turbulence. This project addresses the difficulty of predicting atmospheric contaminant (or pollutant) dispersion in real time. Tackling this challenge requires: i) new techniques for real-time data-driven model adaption (i.e., DART), ii) advances in CFD turbulence modeling, iii) improvements in UAV-based sensing and data processing, and iv) new cyber-feedback-based UAV guidance methods. DART is a new approach to data-driven model adaptation, which takes advantage of recent advances in adaptive control theory and addresses the CPS focus area of Real-Time Control and Adaptation. Although preliminary DART-CFD results using Reynolds averaged Navier-Stokes (RANS) models (e.g., k.-w) are promising, new high-speed high-fidelity turbulence models will be developed to enable accurate real time predictions. Novel UAV guidance methods, which take advantage of cyber feedback from DART-CFD, will also be developed. This project will make transformational progress towards the real-time prediction of contaminant dispersion in an atmospheric flow field. Fukushima Daiichi and Aliso Canyon are only two recent examples of emergency situations that would benefit from DART flow field estimation. Other emergency response applications include forest fires, oil spills, fracking accidents, and train derailments, where the severity of the disaster and the immediate risk to emergency responders needs to be quickly assessed. DART- CFD also has application in wind energy and aviation safety, where predicting the atmospheric flow in a wind farm or near an airport can provide critical information for optimizing operations and safety. In these examples, the UAV-based measurement system could be replaced by a spatially xed measurement system, which is not possible in emergency response. This project will create an exciting opportunity for high-school students to learn about STEM. Over 300 high-school students across Kentucky (a rural EPSCoR state) will be exposed to UAV design and airborne measurement systems. The project will also create unique educational opportunities for undergraduate and graduate students to participate in the exciting research field of CPS. In collaboration with the University of Kentucky's AMSTEMM program, this project will create research opportunities for underrepresented minorities. These activities address NSF's goal of preparing and engaging a diverse STEM workforce.