Our most recent publication “Complete flow characterization from snapshot PIV, fast probes and physics-informed neural networks”, introduces a groundbreaking methodology for reconstructing fluid behavior using physics-informed neural networks (PINNs) and field data.
PINNs are capable of incorporating physics laws during training, and the proposed methodology addresses the lack of time resolution in experimental applications (such as snapshot particle image velocimetry) by feeding PINNs with time-resolved fields estimated from non-time-resolved field measurements and time-resolved point-probe measurements. This innovative approach not only enhances the precision of data but also allows for the extraction of additional derived quantities, such as pressure distribution. We validate the method on a synthetic test case -fluidic pinball- and a experimental scenario (wake of a wing profile), demonstrating its effectiveness in characterizing fluid flow. The methodology’s potential for broader applications beyond the specific cases studied is evident, making it a significant contribution to the field of fluid mechanics and data-driven modeling. Overall, this article presents a powerful and versatile framework that bridges the gap between experimental limitations and the need for high-fidelity flow characterization.