- Journal
- Année
- 2023
Abstract
To answer the needs presented by DARPA's symbiotic design for cyber physical systems challenge, a fast-running, analytical aerodynamic model is developed to guide artificial intelligence systems in the design of unmanned aerial vehicles. As part of the program, aircraft are assembled using a corpus of parts, including commercial off-the-shelf and a few parametric components. Aerodynamic information of the design is required for simulation in a six degree-of-freedom flight dynamics code. In the analytical model, each corpus part is assigned a primitive shape which resembles the part geometry, including cylinders, streamline bodies, and plates/boxes. Appropriate transforms for each primitive are applied such that a surrogate aircraft is constructed, which is then aerodynamically analyzed. The model computes coefficients of lift and drag over a range of pitch angles and freestream velocities individually for each part. Aerodynamic interference between pairs of parts is considered for every pitch angle using analytical relationships and the surrogate aircraft geometry. The model's estimation of drag enables the simulation of the artificial intelligence created designs and provides a source of feedback to encourage logical aerodynamic design choices.