Image-Based Prediction of Lift and Drag Coefficients of a Circular Cylinder Using Convolutional Neural Networks and Flow Visualization
Open Access
Abstract
Force measurement is a fundamental diagnostic technique in experimental fluid mechanics. However, conventional instrumentation, such as mechanical balances and pressure taps, is often intrusive, geometry-dependent, and mechanically constrained. This study proposes a non-contact, image-driven force-estimation framework wherein convolutional neural networks (CNNs) infer the instantaneous lift and drag coefficients of a circular cylinder directly from wake streakline flow-visualization images. A comprehensive hybrid numerical–experimental dataset was established across laminar and turbulent regimes. Navier–Stokes simulations (DNS at ; LES at and 3900) were post-processed into synchronized streakline–force pairs, complemented by water-tunnel dye-visualization experiments at , 8000, and . Five CNN architectures, including lightweight custom networks and vision backbones (AlexNet, ResNet18), were trained and validated. For numerical datasets, all models show high prediction accuracy in lift and drag coefficients. In simulation-to-experiment transfer at , the models capture the dominant vortex-shedding Strouhal number within ( vs. ), while CNNModel_v5 predicts the mean drag coefficient within . On experimental datasets, ResNet18 achieved the best predictive performance among the models tested under random splitting. These results indicate that wake flow-visualization images contain rich physical representations for non-intrusive fluid force estimation.
Keywords
Funding Information
- Shenzhen Science and Technology Program (GXWD20220811152110003)
- Shenzhen Science and Technology Program (SYSPG20241211173609005)
This publication has 24 references indexed in Scilit:
- Convolutional neural networks for fluid flow analysis: toward effective metamodeling and low dimensionalizationTheoretical and Computational Fluid Dynamics, 2021
- Assessment of supervised machine learning methods for fluid flowsTheoretical and Computational Fluid Dynamics, 2020
- An improved smoke-wire flow visualization technique using capacitor as power sourceTheoretical and Applied Mechanics Letters, 2018
- A CNN-based vortex identification methodJournal of Visualization, 2018
- An Immersed-Boundary Finite-Volume Method for Simulations of Flow in Complex GeometriesJournal of Computational Physics, 2001
- A COMPARISON OF METHODS FOR EVALUATING TIME-DEPENDENT FLUID DYNAMIC FORCES ON BODIES, USING ONLY VELOCITY FIELDS AND THEIR DERIVATIVESJournal of Fluids and Structures, 1999
- Passive control of the aerodynamic forces acting on a circular cylinderExperimental Thermal and Fluid Science, 1998
- FORCE PREDICTION BY PIV IMAGING: A MOMENTUM-BASED APPROACHJournal of Fluids and Structures, 1997
- FLUCTUATING LIFT AND DRAG FORCES ON FINITE LENGTHS OF A CIRCULAR CYLINDER IN THE SUBCRITICAL REYNOLDS NUMBER RANGEJournal of Fluids and Structures, 1997
- The velocity field of the turbulent very near wake of a circular cylinderExperiments in Fluids, 1996