In This Article:
Luminary Cloud, in partnership with automotive manufacturer Honda and technology company NVIDIA, has launched what it claims to be the world’s first Physics artificial intelligence (AI) foundation model for analysis of SUV aerodynamics.
The SHIFT-SUV model, which forms part of Luminary Cloud's SHIFT Models suite, enables real-time interactive aerodynamic analysis, a critical advancement for early-stage automotive development.
The collaboration addresses a significant challenge in automotive design by combining Luminary's virtual wind tunnel with the Physics AI model.
This synergy allows for the seamless integration of design aesthetics and engineering performance, which has traditionally been a bottleneck due to the slow pace of performance-based decision-making in the initial stages of development.
Luminary Cloud CEO Pete Schlampp said: “This project represents a fundamental shift in automotive design methodology.
“By providing instantaneous, physics-informed aerodynamic feedback, our open-source foundation model enables designers and engineers to collaborate more effectively at the earliest stages of development, saving time and money—while also advancing innovation.”
Luminary said the SHIFT-SUV aerodynamic dataset has been crafted by simulating thousands of AeroSUV geometry variants using its “high-fidelity” Computational Fluid Dynamics platform.
NVIDIA's PhysicsNeMo, with its DoMINO architecture, was instrumental in training the foundational model, which utilises Physics AI-based simulation for inference.
This modern approach outperforms traditional design methods by providing instantaneous and precise aerodynamic performance insights.
At its inception, the SHIFT-SUV model has been trained on around 1,000 simulations, with ambitions to expand to 25,000 by year's end.
Luminary Cloud said it plans to release new data monthly and will offer pre-trained models at various dataset milestones. Both model and dataset will be available as open-source.
The foundation model presents numerous advantages for automotive players, including access to scarce open-source, industry-relevant data, and the creation of the most extensive open-source Physics AI database for SUV aerodynamics.
Honda's contribution of design variations ensures a training set with industrial relevance, while NVIDIA provides accelerated GPU computing and the NVIDIA Omniverse Blueprint for real-time digital twins CAE resources to support model training.
Honda principal aero design lead Fong Loon Pan said: “The automotive industry has long needed a solution that accelerates styling and engineering development cycles.