In the thesis, we present the development process of a backdrivable planetary gear train of Wolfrom topology. The starting point is an analytical gear model that integrates geometry, kinematics, forces, losses and load-carrying capacity verification, according to ISO 6336 in regards to tooth bending strength and pitting. The system is optimized in regards to multiple objectives using a genetic algorithm. The optimization is carried out with a Mixed-Variable Genetic Algorithm, with early constraints implemented, so that computationally expensive load-capacity calculations are performed only on promising individuals. Material, heat treatment, bearing, shaft seals and lubrication type are all chosen. The chosen Pareto-optimal solution is validated in KISSsoft and a 3D model is designed.
|
