Siemens and PhysicsX collaborate to transform engineering with AI-Driven physics simulations

PhysicsX has harnessed Siemens’ Xcelerator portfolio to train its latest pre-trained deep physics model, LGM-Aero, for aerodynamics. This innovative model leverages high-fidelity simulation data to enable engineers to fine-tune AI models with minimal effort, significantly accelerating design processes.

Jean-Claude Ercolanelli, Siemens’ Senior Vice President of Simulation and Test Solutions, highlighted the transformative potential of AI in simulation. “To scale AI adoption, open collaboration is key,” he said, emphasising Siemens’ commitment to advancing AI-driven engineering alongside PhysicsX.

Robin Tuluie, founder and chairman of PhysicsX, underscored the importance of high-quality synthetic data and robust integrations between Computer-Aided Engineering (CAE) and AI. The release of LGM-Aero and Ai.rplane, an open-access reference application, represents a significant milestone in their collaboration.

LGM-Aero, trained on over 25 million geometries and billions of data points, is designed to generalise across a wide range of aerodynamic applications. Using Siemens tools like Simcenter™ STAR-CCM+™ and Simcenter™ Nastran, the model reduces the need for extensive simulations, allowing engineers to achieve insights rapidly.

Ai.rplane, a free-to-use technology demonstrator, enables users to explore aerodynamic designs and predict performance characteristics in mere seconds. This zero-shot model delivers instant results, offering unparalleled efficiency compared to traditional simulation methods.

newsroom.sw.siemens.com/en-US/siemens-physicsx/