CFD Investigation on Sliding Mesh as a Method to Model Wheel Rotation - Implementation and Analysis on Different Rims

Abstract

Energy efficiency has become an incentive to further investigate sustainable alternatives for passenger mobility. Many engineering disciplines work continuously on minimizing the energy consumption of today’s vehicles, being automotive aerodynamics one of them. In a moving car, aerodynamic drag becomes a dominant force after, approximately, 60 km/h and thus, it must be reduced to improve the vehicle’s efficiency. In a passenger vehicle, the wheels and wheelhousings are a major contributor to the total drag of a vehicle. Hence, capturing the complex flow around the wheels is essential to improve on current solutions. Computational Fluid Dynamics (CFD) simulations are widely used in the aerodynamic improvements of a vehicle. Traditionally, wheel rotation is modelled as a moving wall for the tyres and a Multiple Reference Frame (MRF) region for the rims. The MRF approach, is a simplified method that aims to model wheel rotation while keeping the computational resources low, and thus it is the common practice among carmakers. However, it is an approximation of the real motion of a wheel and it has a number of drawbacks. Alternatively, sliding mesh is an improved model that captures the true motion of a wheel and therefore it is considered to deliver an enhanced numerical solution compared to MRF. This thesis aims at investigating the use of sliding mesh as a wheel rotation method and its correlation with wind tunnel tests. The method is evaluated on four different rims using the same vehicle as a baseline. All cases were solved in ANSYS Fluent 19.0 using a hybrid RANS-LES turbulence model. The investigation is carried out to enable a comparison against previously obtained wind tunnel force measurements. As part of the analysis, the results of interest are the drag, front-lift and rear-lift deltas (CD, CFL and CRL). The implementation of the sliding mesh yielded comparable results to the ones obtained from the MRF simulations. However, the computation time increased significantly. Therefore restricting the number of cases performed in this study. In addition to the increased computational cost, the sliding mesh implementation failed to improve significantly over the MRF and thus putting in question a number of set parameters and boundary conditions applied. However, sliding mesh has shown better results in similar studies [1] [2] [3] [4], and thereby motivates for further studies regarding its implementation and applications.