Search Results

This paper discusses an emerging area of applying machine learning (ML) methods to augment traditional Computational Fluid Dynamics (CFD) simulations of road vehicle aerodynamics. ML methods have the potential to both reduce the computational effort to predict a new geometry or car condition and to explore a greater number of design parameters with the same computational budget. Similar to traditional CFD methods, there exists a broad range of approaches. In particular, the accuracy and computational efficiency of a CFD simulation vary greatly depending on the choice of turbulence model (DNS, LES, RANS) and the underlying spatial and temporal numerical discretizations. Similarly, the end-user must select the correct ML method depending on the use-case, the available input data, and the trade-off between accuracy and computational cost. In this paper, we showcase several case studies using various data-driven ML methods to highlight the promise of these approaches.

The 2nd Automotive CFD Prediction workshop (AutoCFD2) was organized to improve the state-of-the-art in automotive aerodynamic prediction. It is the mission of the workshop organizing committee to drive the development and validation of enhanced CFD methods by establishing publicly available standard test cases for which high quality on- and off-body wind tunnel test data is available. This paper reports on the AutoCFD2 workshop for the Ford DrivAer test case. Since its introduction, the DrivAer quickly became the quasi-standard for CFD method development and correlation. The Ford DrivAer has been chosen due to the proven, high-quality experimental data available, which includes integral aerodynamic forces, 209 surface pressures, 11 velocity profiles and 4 flow field planes. For the workshop, the notchback version of the DrivAer in a closed cooling, static floor test condition has been selected.

This paper presents the results from an investigation into the performance of OpenFOAM v1806 on the Amazon Web Services (AWS) Elastic Compute Cloud (EC2) service for a realistic racing vehicle using a high-fidelity hybrid RANS-LES CFD approach. It is shown that AWS can provide the HPC environment to enable greater use of high-fidelity CFD methods by allowing higher core counts to reduce turn-around time. With the correct instance type - which potentially differs between meshing and solving - AWS was competitive against a high-performance Cray XC30 supercomputer, up to 1920 cores and meshes up to 280 million cells. However it is recognised that this Cray XC30 displayed superior scaling whilst containing older generation processors (Intel Ivybridge) compared to the AWS Instances (Intel Skylake).

A wide-ranging investigation into the sensitivity of the hybrid RANS-LES based OpenFOAM CFD process at Audi was undertaken. For a range of cars (A1, TT, Q3 & A4) the influence of the computational grid resolution, turbulence model formulation and spatial & temporal discretization is assessed. It is shown that SnappyHexMesh, the Cartesian-prismatic built-in OpenFOAM mesher is unable to generate low y+ grids of sufficient quality for the production Audi car geometries. For high y+ grids there was not a consistent trend of additional refinement leading to improved correlation between CFD and experimental data. Similar conclusions were found for the turbulence models and numerical schemes, where consistent improvements over the baseline setup for all aerodynamic force coefficients were in general not possible. The A1 vehicle exhibited the greatest sensitivity to methodology changes, with the TT showing the least sensitivity.

Computational Fluid Dynamics (CFD) is now one of the most important design tools for the automotive industry. Reliable CFD simulations of the complex separated turbulent flow around vehicles is becoming an ever more crucial goal to increase fuel efficiency and reduce noise emissions. In this study Reynolds Averaged Navier-Stokes (RANS) models (both at eddy-viscosity and second-moment closure levels) are compared to hybrid RANS-LES methods (Detached-Eddy Simulation). The application is the DrivAer model; a new open-source realistic car model which aims to bridge the gap between simple Ahmed body and MIRA/SAE Reference car models and actual car geometries in use by the major car manufacturers. To date, many hybrid RANS-LES studies on complex geometries have been under-resolved compared to more academic cases, due to a limit on computational resources available.