AI in the Automotive Industry — Digitalization and Recent Improvements
Over the past few years, we have seen Neural Concept’s solution being deployed more and more within the automotive industry, by industrial leaders such as Plastic Omnium, Bosch and Stellantis. Here is what we have observed, and our thoughts on why this industry is rapidly adopting AI based design and engineering workflow.
As of today, the automotive industry is the largest industry within the global CAE market. The development cycles of new vehicles are becoming shorter, putting higher pressure on automotive suppliers and the whole automotive supply chain. This is also why companies have invested massively in this digitalization journey, and are relying heavily on CAD/CAE software.
The growth of high-performance computing infrastructure and the progress in high-fidelity simulation methods have contributed to numerical simulation being key in the reduction of physical testing and in product performance improvement. However, the time required to run a simulation is a major bottleneck in the engineer’s design loop, and hence in the whole development cycle of the vehicle.
This is where Artificial Intelligence has a major impact, bringing real-time, simulation-driven design workflows to engineering companies. By using the large amount of data generated on past vehicles' development by engineering teams, AI has allowed future design campaigns to become much shorter, with reduced development effort, and a smarter usage of the CAE tools.
Let’s use a practical example, on the development of a Heating, Ventilation, and Air Conditioning (HVAC) system.
The HVAC of a SUV, although similar to the one of a small compact car, has different specifications and requirements to be met. Therefore, for each new vehicle being developed, the climate control computational fluid dynamics (CFD) team in the company needs to iterate with the corresponding computer-aided engineering (CAD) team. This is to make sure that the airflow at each of the registers matches the KPIs, and that there is no undesirable behavior for the whole range of operating conditions of the HVAC. This is a lengthy process, where the CAD and CAE teams iterate together, to converge towards a design that fulfills all the requirements. Time-consuming CFD runs are needed at each design loop, which is a major bottleneck in the iterations between the teams. Moreover, for each new design campaign, the engineering teams need to start from scratch again, and they do not leverage on data generated from previous projects, which is a huge waste of data considering the high pace at which new vehicles are being developed.
On the contrary, Artificial Intelligence (AI) can help train models using this historical data coming from past programs. The first designs iterations can be done using simulation results from the predictive AI model, that are instantaneous (in the order of seconds from a given CAD geometry). This enables the CAD teams to evaluate some issues by themselves during the first HVAC design iterations(detachment of the flow at specific regions, recirculation, large pressure drops, etc.). Iterations with the Climate Control team are limited to the ones which have much more added value in the design chain, or where further analysis or fine-tuning are needed.
This results in:
1. Further frontloading for design exploration
2. Smarter CFD usage to higher value applications e.g., acoustics
3. Smarter usage of existing high value software tools (move to high fidelity usage of licenses – fewer higher quality runs
= reduction in overall license spend)
Therefore, AI is revolutionizing the automotive sector. If you think that NC Shape can help shorten your development cycles and increase your product performance, feel free to get in touch with us.
.jpg)
