Estimation of passive drag in swimming via experimental and computational means

Discussed is a comparison of computational and experimental evaluations of passive drag during human swimming. Experimentally, ten trials were conducted per athlete at five chosen velocities, using a commercial resistance trainer to record the tension force in a rope during a streamline position tow test. The resistive force recorded was assumed equal to the passive drag force and an average value of passive drag was found across each tow test. Mean passive drag values measured during the tow test were agreed well with existing experimental data across the range of velocities used, varying between 20 N at 1 ms⁻¹ up to 100 N at 2 ms⁻¹. Computationally, using the immersed boundary method in OpenFOAM, basic geometry validation cases and streamline passive drag cases were simulated. Validation cases were completed on 2D cylinders and 3D spheres with the drag coefficient found at low and high Reynolds numbers, using the simpleFoam solver within OpenFOAM. Results tended to be slightly over predictive when compared with existing simulation and experimental data in literature. The accuracy of results could potentially be improved using a finer mesh and better quality geometries. The passive drag was also computed using OpenFOAM over a range of velocities, similar to the experiments, varying from 30 N at 1 ms⁻¹ to 120 N at 2 ms⁻¹. Drag forces computed using simpleFoam were over predictive when compared to existing literature and the completed experiments, likely due to the inaccuracy of the geometry used in the simulations. When results were compared to existing literature for swimmers not in a perfect streamline position, more similar to the geometry used in this study, results were in better agreement. The accuracy of the results could be improved using a better quality geometry in the correct position.