Gearbox Design by means of Genetic Algorithm and CAD/CAE Methodologies 2010-01-0895

The paper discusses a gearbox design method based on an optimization algorithm coupled to a fully integrated tool to draw 3D virtual models, in order to verify both functionality and design. The aim of this activity is to explain how the state of the art of the gear design may be implemented through an optimization software for the geometrical parameters selection of helical gears of a manual transmission, starting from torque and speed time histories, the required set of gear ratios and the material properties. This approach can be useful in order to use either the experimental acquisitions or the simulation results to verify or design all of the single gear pairs that compose a gearbox. Genetic algorithm methods are applied to solve the optimization problems of gears design, due to their capabilities of managing objective functions discontinuous, non-differentiable, stochastic, or highly non-linear. The final design tool, implemented in Matlab® environment, is based on the calculation of load capacity of helical gears, including the computation of tooth bending strength, of surface durability (pitting) and the estimation of service life under variable load, as suggested by International Standards. An automated macro procedure for Solidworks® interacts with the Matlab® environment to get the dimensional parameters of each gear and produces the models of each gear and their assembly.