AutoDIDACTS: Automated Discovery for Design and Control of Turbulent Systems

Despite the growing sophistication of computational tools for aeronautical design, increasingly detailed simulation models suffer from diminishing returns – the number of variables involved makes the problems intractable, no matter how precise the calculations or how much computing power is used. Further, these models are often inaccurate and not generalizable across parameter space, limiting their use for design innovation.

The AutoDIDACTS program envisions a different way forward, revolutionizing aeronautical design by pathfinding accurate, predictive, and generalizable principles to guide design exploration and innovation. By accumulating precision ground-truth data from compact, tabletop turbulence simulators, the program will demonstrate data-informed development of macroscale principles and generalizable models. These models aim to provide rapid design capabilities with broad applicability across aeronautical regimes from tiny drones to hypersonic vehicles.