PoETS: A Proposed Powertrain Topology Encoding Scheme
Laminn McLay, James P. Scanlan, András Sóbester · AIAA SCITECH 2023 Forum · 2023
View Video Presentation: https://doi.org/10.2514/6.2023-1363.vid With the rise of hybrid powertrains (particularly on Uncrewed Aerial Vehicles (UAV)), there is an increased need for the systematic, automated exploration of the space of powertrain topologies, with the goal of identifying the layout that optimizes mission performance. The first building block of such an optimizer is a scheme capable of the flexible, future-proof, yet economical encoding of powertrain topologies, and we propose such a scheme in this paper. Our starting point is an idea first formulated in computational chemistry, where the construction of new molecules from a dictionary of atoms and bonds has parallels with the ‘automated invention’ of drivetrains through the systematic search through the space of component catalogues and their connectivities.