Free-to-Use Aerospace-Targeted AI Model Opens New Doors for Plane Optimization

You have heard of large language models, as exemplified by chatbots like ChatGPT, but have you heard of large geometry models? This month, London-based PhysicsX released a free-to-use large geometry model tailored to aerospace (LGM-Aero), which serves as a harbinger to how we might simulate and analyze the aerodynamics of airplane structures in the future.

“In the same way that large language models understand text, PhysicsX has a vast knowledge of the shapes and structures that are important to aerospace engineering” said Jacomo Corbo, cofounder and CEO of PhysicsX. “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. We’re excited about what LGM-Aero brings as capabilities to our customers while recognizing that it is also an important stepping stone towards developing physics foundation models.”

Delivered as a browser-based application, PhysicsX's LGM-Aero gives you the ability to use the target payload as the starting point, then gradually shape the optimal geometry by considering thrust, cruising speed, wingspan, lift, and other parameters. There is also an upload button that let users upload CAD geometry as STL. 

The application responds to parameter changes in real-time because it's not running full physics-based CFD analysis. Instead, it's relying on an AI model, trained on a Siemens dataset involving 25 million shapes, according to Neil Cameron, Principal Delivery Engineer, PhysicsX. 

“The application is constrained to modest aerodynamics, so it's meant for hobbyists and students building a drone-size plane, for example,” Cameron clarified. “

If the user attempts to analyze a commercial passenger jet with a payload that exceeds the application's intended range, the application will issue an error message. Similarly, if a user uploads a model that does not resemble an airplane (for example, a vehicle model), the application will recognize it as topology it cannot confidently make predictions on. Therefore, it will prompt the user to upload a more acceptable model. These safeguards and others prevent the users from using the application beyond it's intended purpose, Cameron said.

The morph-the-design command allows you to specify target points and reshape the plane's topology. This command, too, has built-in safeguards that prevent you to morphing the design beyond the acceptable scope.

“We have a platform for hosting these types of models using different architectures. We can work with customers to help them train and develop a model specific to their engineering problem, using a mix of our foundational model and the customers' data,” explained Cameron.

The use of AI-trained models that can act as surrogate for full physics-based CFD or FEA is gaining momentum, in part due to products like Ansys SimAI. 

“We're looking forward to having conversations with enthusiasts and students about this, just to prepare them for where we think the industry is going,” said Cameron. 

“To scale AI adoption, open collaboration is key and Siemens is delighted to see PhysicsX emerge out of stealth and unveil to the world what it is working on—together, we’re exploring an AI-enhanced simulation industry that has the potential to change how products are ideated and engineered,” said Jean-Claude Ercolanelli, senior vice president, Simulation and Test Solutions, Siemens Digital Industries Software.

To access Ai.rplane go to airplane.physicsx.ai.

More PhysicsX Coverage