Search Results

The Ahmed body is a simplified vehicle geometry that results in flow features representative of those found at the rear of most passenger vehicles. By adjusting the rear slant angle, separation can take place at the sharp corner, on the rear slant panel, or not at all. Accurate prediction of the separation and reattachment of the flow is essential in predicting the correct drag trends. This separation and reattachment is known to be a highly unsteady phenomenon. The objective of this study is to evaluate the ability of a lattice Boltzmann based CFD code to predict the correct drag trends and flow structures for the Ahmed body at varying rear slant angles. Component and total drag values show excellent agreement with the original experiments of Ahmed over a wide range of rear slant angles (5 to 35 degrees).

The lattice Boltzmann (LB) method has been successfully used in conjunction with a Very Large-Eddy Simulation (VLES) turbulence modeling approach for over a decade for the accurate prediction of automotive fluid dynamics. Its success lies in the unique underlying physics that is significantly different from traditional computational fluid dynamics methods. In this paper, we provide a complete description of the method followed by a set of examples which show its use in the automotive industry. We will first provide a review of the physics and numerical methods. Here the LB method and its relationship to kinetic theory and the Navier-Stokes equations will be briefly discussed. We will summarize the strengths of LB method, especially for the solution of transient flows in extremely complex geometries. The VLES turbulence modeling method will be presented next, as well as how VLES neatly fits into the LB framework.

The purpose of this paper is to describe some of the technology behind the Lattice Boltzmann approach to exterior airflow simulations as incorporated into the commercial CFD code, PowerFLOW®. The fundamental approach used is the Lattice Boltzmann Method (LBM) coupled with both a turbulence model to recover the dissipation of sub-grid eddy scales and a wall model to allow reduced resolution in the near-wall region. A description of LBM and both models is given. Comparisons to methods that directly solve the Navier-Stokes equations, such as finite volume or finite element methods (hereafter, collectively referred to as RANS methods) are also presented. A demonstration of the technology is presented by comparing numerical simulations with extensive experimental test data on Ford's standard calibration models. These models were originally described in SAE paper 940323 [1].

Presented is a study of the air flow over the Morel body [1] in ground proximity which was obtained using a discrete particle method, referred to as DIGITAL PHYSICS. The results were computed at several back-light angles and will be compared to experimental observations. Separation and reattachment along the angled section at a back-light angle of 30 degrees, and complete separation at 35 degrees, were both accurately predicted.