Steady and Unsteady Numerical Analysis of the DrivAer Model

Abstract

This work consists of steady and unsteady numerical analysis of the DrivAer vehicle model. The CAD model is free for anyone to use and is setting the new standard for benchmarking numerical analysis against experimental data in academia and industry. The model consists of a standardised car with interchangeable rear ends, resulting in a fastback, notchback, and squareback model. The models can be analysed with at or detailed underbody and with or without closed wheelhouses, resulting in no wheels. A comparison of the numerical results and experimental results has been performed using test data from scale model testing at TU Munchen and TU Berlin, and full scale testing at Volvo Cars performed by Ford of Europe and made available to the author. The numerical analysis consists of three di erent methods, first a steady-state method using the realisable k turbulence model with standard wall treatment and a mesh consisting of 50 million cells for a full car simulation. Second, a similar method as the first one except for a refined mesh using six prism layers, resulting in 200 million cells and combining this with enhanced wall treatment. Finally a unsteady delayed detached eddy simulation, DDES, method with Spalart Allmaras, SA, turbulence modelling is used combined with the same mesh in the second method mentioned. The results of this work show that the unsteady method has a better capability of predicting drag force trends, surface pressure distribution at the base, and ow field behaviour in the wake. While the steady-state method with enhanced wall treatment shows good ability to predict absolute drag values for the detailed underbody, the unsteady method show similar abilities but excels in many different configurations when in comparison.