Comparative Evaluation of Finite Element-Driven Optimization Methods for Additively Manufactured UAV Primary Structures

Abstract:
Unmanned aerial vehicle (UAV) primary structures require high specific strength and stiffness, traditionally necessitating expensive carbon fiber composites. This study evaluates simulation-driven, additively manufactured polymer alternatives fabricated from PLA and computationally optimized via macroscopic Topology Optimization (TO), mesoscopic Variable-Thickness Lattices (VTL), and uniform Triply Periodic Minimal Surfaces (TPMS). Evaluations were conducted under a superimposed, multi-axial flight envelope. Physical testing demonstrated that VTL architectures maximized the Structural Efficiency Index (SEI) by pushing mass to the extreme geometric fibers and increasing global flexural rigidity. In contrast, mass-constrained TO yielded misleading specific strength due to volumetric starvation and elevated compliance, while the uniform TPMS baseline exhibited favorable specific stiffness but lacked targeted root robustness, resulting in reduced specific strength. Off-axis testing further showed that Diamond lattices dominated vertical bending and inverted impulse loading, whereas Octet and Kelvin geometries more efficiently resolved transverse shear. Experimental data identified a performance-optimized efficiency asymptote in the 73-79 g VTL specimens, which achieved a 19-24% mass reduction relative to a 97 g carbon fiber baseline. To assess assembled-vehicle relevance, the selected fully 3D-printed replacement arms were installed on the baseline quadcopter and subjected to nine dynamic ground tests comprising staircase and cyclic propulsive loading under freestream conditions of 0, 10, and 20 knots. The optimized arms completed the full test matrix without fracture, mount failure, screw loosening, visible yielding, or permanent deformation, demonstrating structural viability in a realistic multi-part UAV assembly without carbon fiber reinforcement.

Did you attend Forum 82? Click the preview below to access the full paper.