Sequential design of UAV fuselage pods using bounding aerodynamic models

In this work, a Sequential Decision Process (SDP) is applied to perform fuselage design using Computational Fluid Dynamics (CFD). The SDP uses models to provide two-sided estimates that attempt to bound the exact solution, ultimately converging to an optimal design space to be analyzed with models of increased fidelity. The present work proposes the use of laminar and turbulent physics in CFD models to form lower and upper bounds on drag calculations, respectively. These bounding models are then used in a formal SDP to cull the design space, focusing the region of interest for increased fidelity modeling and analysis. Increasing mesh resolution is used to increase fidelity, creating a multi-fidelity approach to aerodynamic shape design. In this work the SDP-CFD design approach is applied to two design problems: (1) drag minimization of a fairing with a defined thickness and (2) drag per unit volume minimization of a fairing. The results of this study demonstrate that the SDP-CFD approach can accurately and quickly improve the fuselage design.