Computational analysis of dive phase in competitive swimming
Date of Issue2017
School of Mechanical and Aerospace Engineering
Analysis of starting phases of swimming like the diving phase has not been studied much due to the difficulty in setting up accurate experimental setup. The aim of this study is to understand the performance impact of diving at different angles of entry using a 2D geometry of a swimmer by employing Computational Fluid Dynamics. Dynamic simulations were conducted using OpenFOAM with the help of an overset meshing algorithm called OPErA and six degree of motion solver. A previously validated 2D geometry was taken from literature and meshed, and was then validated against existing data in passive drag analysis. A short study on the significance of turbulence model choice in predicting drag forces was also done which concluded that it is necessary to use wall refined meshes with y+ < I and employ low Reynolds number turbulence models to accurately predict near wall physics. A brief 30 CFD analysis of passive drag was also conducted using a previously validated geometry and this study identified the regions where strong three dimensional flow features prevailed like arms, neck and legs. Using the validated computational mesh, multi-phase diving simulations were performed in 3 set of studies (a) using no motion constraint (b) setting an angular velocity constraint (c) motion analysis. With no motion constraint, simulations exhibited nonphysical rotational moments which were attributed to the unrealistic forces exerted by the fluid due to the inherent 2D approximation. Thus an angular velocity constraint was introduced to the second set of simulations which resulted in realistic trajectory of motion of the swimmer. The force predictions between the three angle of entries (35°, 40°, 45°) showed similar behavior and exhibited 2 prominent peaks and only differed in the location and the amplitude of the peaks. The third analysis compared the motion analysis data extracted from video footages of diving trials against the simulation data. This study reported that the peaks in force predictions in the 2D CFD results were not prominent in the motion analysis and were found to be generated by high pressure regions around the neck, arms and legs. These regions exhibited prominent 3D flow features and deviate the most from the human anatomy in a 2D approximation. Thus the results suggest that 2D simulations are insufficient to predict realistic dynamics of the swimmer and will require further study using 3D simulations is necessary to predict the forces accurately.