#London #United Kingdom #PhysicsX #LGM-Aero #Ai.rplane

PhysicsX Launches AI-Powered LGM-Aero for Aerospace Engineering

London, December 4, 2024 – PhysicsX, a pioneering startup based in London, has unveiled a groundbreaking Large Geometry Model (LGM) specifically designed for aerospace engineering. Dubbed LGM-Aero, the model is empowered by artificial intelligence and built using Amazon Web Services (AWS) cloud technologies, making it a formidable tool for aerospace developers.

This innovative model has been developed using a colossal dataset of over 25 million different geometric shapes that together represent more than 10 billion vertices. By leveraging such extensive training data, LGM-Aero dramatically transforms the development timeline of aerospace concepts, reducing the previous months-long processes to just hours.

One of the standout features of this launch is Ai.rplane, a reference application created to demonstrate the capabilities of LGM-Aero. This platform allows engineers to explore an infinite design space for aircraft, producing fresh design ideas and quickly evaluating their performance. Ai.rplane adds an interactive layer to the engineering process, positioning it as an essential tool for innovation in aerospace design.

The underlying technology incorporates a wealth of high-fidelity physics simulations, expertly generated in collaboration with Siemens’ Xcelerator portfolio. This component significantly enhances the model's reliability in predicting various aerodynamic qualities including lift, drag, and stability. Notably, LGM-Aero can deliver results in less than a second, a stark contrast to the traditional numerical simulations which can take hours.

Accelerating Engineering with Advanced Technology

Ozgur Tohumcu, the General Manager for Automotive and Manufacturing at AWS, emphasized the collaboration's significance. He stated, "We are delighted to work with PhysicsX as they develop their first Large Geometry Model and release its showcase application Ai.rplane. This technology will accelerate the transformation of engineering in Advanced Industries for AWS customers, enabling them to bring their products to the market faster while increasing product performance."

PhysicsX's co-founder and CEO, Jacomo Corbo, further elucidated on the innovation behind Ai.rplane, likening its geometric comprehension to that of large language models in understanding textual data. He noted, "The technology can optimize across multiple types of physics in seconds, many orders of magnitude faster than numerical simulation, and at the same level of accuracy."

The LGM-Aero model is essentially a zero-shot learning architecture, allowing it to infer performance metrics for a wide array of flying shapes without requiring extensive retraining. This capability signifies a monumental shift in how aerospace applications approach design and development phases.

Future Aspirations and Goals

PhysicsX envisions this as merely the beginning. As they continue to collaborate with AWS, they plan to introduce further functionalities within both LGM-Aero and Ai.rplane, including the ability to select different powertrains and controls, potentially evolving the platform into a comprehensive solution for aircraft design.

In an industry where every second counts and optimization is key to safety and performance, LGM-Aero and Ai.rplane position themselves at the forefront of aerospace engineering technology. By making these tools publicly accessible at airplane.physicsx.ai, PhysicsX aims to democratize access to advanced engineering capabilities, encouraging innovation and reducing barriers for engineers.

In conclusion, the launch of LGM-Aero and Ai.rplane is set to redefine how aerospace engineering conceptualizes and iterates designs, paving the way for faster, more efficient development processes and ultimately transforming the aerospace landscape.