Journal article
Mohamed Elrefaie, Steffen Hüttig, Mariia Gladkova, Timo Gericke, Daniel Cremers, Christian Breitsamter
Experiments in Fluids, 2024
Experiments in Fluids, 2024
View PDF
Cite
Cite
APA
Click to copy
Elrefaie, M., Hüttig, S., Gladkova, M., Gericke, T., Cremers, D., & Breitsamter, C. (2024). Real-time and on-site aerodynamics using stereoscopic piv and deep optical flow learning. Experiments in Fluids. https://doi.org/10.48550/arXiv.2401.09932
Chicago/Turabian
Click to copy
Elrefaie, Mohamed, Steffen Hüttig, Mariia Gladkova, Timo Gericke, Daniel Cremers, and Christian Breitsamter. “Real-Time and on-Site Aerodynamics Using Stereoscopic Piv and Deep Optical Flow Learning.” Experiments in Fluids (2024).
MLA
Click to copy
Elrefaie, Mohamed, et al. “Real-Time and on-Site Aerodynamics Using Stereoscopic Piv and Deep Optical Flow Learning.” Experiments in Fluids, 2024, doi:10.48550/arXiv.2401.09932.
BibTeX Click to copy
@article{mohamed2024a,
title = {Real-time and on-site aerodynamics using stereoscopic piv and deep optical flow learning},
year = {2024},
journal = {Experiments in Fluids},
doi = {10.48550/arXiv.2401.09932},
author = {Elrefaie, Mohamed and Hüttig, Steffen and Gladkova, Mariia and Gericke, Timo and Cremers, Daniel and Breitsamter, Christian}
}
We introduce Recurrent All-Pairs Field Transforms for Stereoscopic Particle Image Velocimetry (RAFT-StereoPIV). Our approach leverages deep optical flow learning to analyze time-resolved and double-frame particle images from on-site measurements, particularly from the 'Ring of Fire,' as well as from wind tunnel measurements for real-time aerodynamic analysis. A multi-fidelity dataset comprising both Reynolds-Averaged Navier-Stokes (RANS) and Direct Numerical Simulation (DNS) was used to train our model. RAFT-StereoPIV outperforms all PIV state-of-the-art deep learning models on benchmark datasets, with a 68% error reduction on the validation dataset, Problem Class 2, and a 47% error reduction on the unseen test dataset, Problem Class 1, demonstrating its robustness and generalizability. In comparison to the most recent works in the field of deep learning for PIV, where the main focus was the methodology development and the application was limited to either 2D flow cases or simple experimental data, we extend deep learning-based PIV for industrial applications and 3D flow field estimation. As we apply the trained network to three-dimensional highly turbulent PIV data, we are able to obtain flow estimates that maintain spatial resolution of the input image sequence. In contrast, the traditional methods produce the flow field of 16x lower resolution. We believe that this study brings the field of experimental fluid dynamics one step closer to the long-term goal of having experimental measurement systems that can be used for real-time flow field estimation.
Overview of our approach for flow field estimation using RAFT-StereoPIV.
- a) Stereoscopic configuration of Ring of Fire where the car crosses the measurement section and particle images are processed on-site by RAFT-StereoPIV.
- b) The Mapping function used for 3D reconstruction from 2D Stereoscopic PIV Data.
- c) Inputs to RAFT-StereoPIV and its main operations.
- d) Processing pipeline RAFT-StereoPIV (adapted from \cite{RAFT}), which takes a video image sequence from the stereoscopic configuration and estimates the corresponding optical flows of the in-plane components (in x- and y-directions) for each camera. The 3D velocity field (U, V, W) can be reconstructed from the 2D optical flow pair.
Ring of Fire experimental data - Development of the normalized out-of-plane streamwise velocity component in the wake of the car at x = 0.25, 0.50, 0.75, and 1.00 m. (Top) we show the results obtained from LaVision DaVis 10.2 after postprocessing and (bottom) - from the deep learning model RAFT-StereoPIV.
