SMART: Scalable Mesh‑free Aerodynamic Simulations from Raw Geometries using a Transformer‑based Surrogate Model

Machine learning–based surrogate models have emerged as more efficient alternatives to numerical solvers for physical simulations over complex geometries, such as car bodies. Many existing models incorporate the simulation mesh as an additional input, thereby reducing prediction errors. However, generating a simulation mesh for new geometries is computationally costly. In contrast, mesh-free methods, which do not rely on the simulation mesh, typically incur higher errors. Motivated by these considerations, we introduce SMART, a neural surrogate model that predicts physical quantities at arbitrary query locations using only a point-cloud representation of the geometry, without requiring access to the simulation mesh. The geometry and simulation parameters are encoded into a shared latent space that captures both structural and parametric characteristics of the physical field. A physics decoder then attends to the encoder's intermediate latent representations to map spatial queries to physical quantities. Through this cross-layer interaction, the model jointly updates latent geometric features and the evolving physical field. Extensive experiments show that SMART is competitive with and often outperforms existing methods that rely on the simulation mesh as input, demonstrating its capabilities for industry-level simulations.

Physical simulations are essential for the evaluation and optimization of technical designs during product development. For instance, improving the efficiency of vehicles such as cars or airplanes often depends on aerodynamic simulations of the airflow around their complex shapes. These simulations typically rely on numerical solvers that are computationally expensive and time-consuming. Recent advances in machine learning offer a faster alternative by replacing the numerical solver with a model that approximates the solution. However, many of these models still rely on the simulation mesh as an additional input, which means they continue to depend on a preprocessing step that is itself costly and time-consuming. Our work introduces SMART, a new machine learning model that removes the need for a simulation mesh while still achieving high accuracy. Instead of relying on a simulation mesh, SMART only requires a simple point cloud that describes the object's shape. The model learns a shared internal representation that captures the structural and parametric characteristics of the solution, and uses this representation to predict physical quantities, such as pressure or velocity, at arbitrary spatial locations. Across a wide range of test cases, SMART performs as well as or better than methods that still require a simulation mesh. This demonstrates that SMART can deliver fast, accurate predictions suitable for real engineering applications while eliminating another bottleneck in the simulation pipeline.