Real-time data-driven flow diagnostics with event-based cameras

Objective

Experiments in fluid dynamics remain an expensive bottleneck for designing, improving and benchmarking new solutions. Pointwise probes provide either sparse information, or only statistics for the full field. Optical measurement techniques deliver instantaneous flow fields, but at the expense of high data volumes, complex and costly instrumentation, and delayed feedback. RT-FLOW will demonstrate a compact, event-based imaging velocimetry system enhanced by data-driven techniques, enabling real-time flow field measurements that can be used for integration in design and active control loops. Event-based sensors bring forth reduced data throughput and higher sensitivity than conventional imaging systems, allowing low-latency spatial coverage with inexpensive illumination sources. These features unlock access to high-frequency measurements with cost-effective hardware, lowering the entry barrier to advanced aerodynamic testing, and laying the ground for widespread use. Furthermore, this solution naturally integrates into the process of flow control and optimization, thus paving the way for groundbreaking innovations in many industrial sectors. RT-FLOW will define the requirements for a fast quantitative flow visualization platform tailored to research and industrial stakeholders, and will tackle the scaling of data-driven spatio/temporal resolution enhancement techniques to wind tunnel operating regimes. The project will draft the path to commercialization and exploitation of the solution. RT-FLOW will thus demonstrate the potential of real-time flow field feedback as the cornerstone of a transformative change in aerodynamic testing, enabling on-the-fly design validation and optimization.