Image
In the automotive industry, aerodynamic design relies more and more on numerical simulation. Computational Fluid Dynamics (CFD) is used to study how air flows around a vehicle. It reduces the need for physical prototypes and works alongside wind tunnel testing. It is a key tool for optimising body shape, reducing aerodynamic drag, and improving fuel consumption, electric vehicle range and acoustic comfort.
Simulations that keep getting more accurate need advanced models and a lot of computing power. That is where High Performance Computing (HPC) comes in, and increasingly artificial intelligence too. Both are becoming strategic technologies for speeding up analysis, shortening development times and supporting sounder design decisions. This field fully reflects the mission of EuroCC Italy, which works to connect companies and research centres with supercomputing expertise and infrastructure. The 5th edition of the Automotive CFD Prediction Workshop (AutoCFD5) takes place in this context. It is organised in collaboration with EuroCC Italy and will be held on 8–9 October 2026 at the DAMA Tecnopolo in Bologna.
An international comparison of CFD’s predictive capability
When: 8–9 October 2026
Where: DAMA Tecnopolo, Via Stalingrado 84/3, Bologna
Language: English
Registration deadline: 1 October 2026
AutoCFD5 is organised by Upstream CFD, EVCM and CINECA, in collaboration with EuroCC Italy. It features a series of international workshops where industry, academia and research centres tackle one specific question: how reliable are Computational Fluid Dynamics (CFD) codes at predicting the aerodynamics of a car?
AutoCFD is built on comparison. Different organisations apply different codes, models and methods to the same test cases and then compare their results. This work produces practical insights on several topics:
- turbulence modelling
- mesh generation
- numerical schemes
- simulation uncertainty
- computational performance
AutoCFD5 builds on the work of previous editions through two benchmarks based on the configurations used for AutoCFD4. The activities on artificial intelligence and machine learning also use the open-source DrivAerML dataset.
HPC, GPUs and AI for ever more advanced CFD
CFD is one of the fields where supercomputing can have a direct impact on industrial processes. Access to HPC infrastructure makes it possible to run simulations at ever higher resolution, explore more configurations and get results useful for design in less time. GPU-accelerated architectures are also profoundly changing the performance of simulation codes.
Artificial intelligence now adds to these technologies, with the development of surrogate models and geometric deep learning techniques that can work alongside traditional numerical simulation to predict aerodynamic quantities. The collaboration with AutoCFD5 is part of EuroCC Italy‘s mission to connect Italian companies and organisations with the infrastructure, expertise and opportunities that supercomputing offers, with a particular focus on the manufacturing and automotive sectors.
Why register
AutoCFD5 is a unique chance to exchange ideas with people who develop and use CFD simulation at the highest level, in industry and in research. Taking part means you can:
- Learn about the state of the art in aerodynamic simulation for automotive, through a direct comparison of results from different codes and approaches on the same test cases.
- Take home practical guidance on turbulence modelling, mesh generation and numerical schemes that you can apply right away in your own workflows.
- Explore the frontiers of innovation, from GPU-accelerated simulations to the use of artificial intelligence and surrogate models for aerodynamic prediction.
- Connect with an international community of car manufacturers, software vendors, supercomputing centres and universities. The workshop committee includes representatives from Ford, NVIDIA, Barcelona Supercomputing Center and Loughborough University, among others.
- Network in an informal setting, thanks to a social programme that includes a visit to the Museo Enzo Ferrari in Modena.
The programme
Across the two days, five Technology Focus Groups (TFGs) present the work they have done during the year. Each group covers one of the main areas where automotive CFD is evolving:
- Meshing: turbulence models, mesh refinement and automatic mesh generation
- Scale Resolving Simulations: high-order spectral methods, DDES and hybrid RANS–LES approaches
- High Performance Computing: GPU-accelerated simulations and performance comparison
- Noise Factors: uncertainty quantification and ensemble averaging methods
- AI/ML: geometric deep learning, neural surrogates and physics-based models for aerodynamic prediction
The programme also includes a visit to the Museo Enzo Ferrari in Modena.
Click here for more information on the event programme.
Register
To attend the event, click here to register. Registration is open until 1 October 2026.