On 29 September 2026, Esther Rocío Valladolid Ortíz presented “Towards UVL-based variability modelling of drone propellers” at MODEVAR 2026, the International Workshop on Languages for Modelling Variability, held as part of VARIABILITY 2026 in Limassol, Cyprus. The work is co-authored with David Romero-Organvidez, José Antonio Pérez Castellanos and David Benavides, from the University of Seville.
Feature models are usually applied to software, where a feature is a choice that is either included in a configuration or not. Physical engineering domains are just as variable, but their variability is often continuous and bound by physics. The design of a drone propeller depends on numeric parameters such as radius and pitch, together with discrete choices such as the number of blades or the tip shape, and these parameters are linked by physical relations that decide whether a design is feasible at all. Without a model of this space, the number of combinations quickly exceeds manual exploration, designs are hard to reproduce, and nothing prevents physically unfeasible propellers.
The authors model the variability of fixed-pitch drone propellers as a UVL feature model that combines discrete features with the numeric parameters of the propeller, covering 432 discrete configurations before any numeric value is assigned. An external validator enforces the physical feasibility relations of the domain, such as the minimum pitch-to-diameter ratio or the proportion between hub and propeller radius. A generator then turns every valid configuration into an executable parametric model for Autodesk Fusion 360, with a unique identifier that links each generated design back to the configuration that produced it.
Because feasibility is checked on the variability model before any geometry is built, every derived propeller is physically valid by construction and fully traceable, and a large design space can be explored in little time. To the best of the authors’ knowledge, this is the first feature-model account of drone propeller variability, and it shows that UVL can serve as a common representation for physical domains, not only for software.
The artefacts are openly available:
UVL model on UVLHub: https://www.uvlhub.io/doi/10.5281/zenodo.20148676
Generator and validator source code: https://doi.org/10.5281/zenodo.20148834
Live platform: https://spldroneprop.diversolab.net
This work was partially supported by FEDER, the Ministry of Science, Innovation and Universities, Junta de Andalucía and the State Research Agency through the projects Data-pl (PID2022-138486OB-I00), PREMISE (PID2025-171313OB-I00), PLANT (DGP_PIDI_2024_01144) and SENSOLIVE (PLSQ_00162).