CAD-Integrated Automatic Gearbox Design with Evolutionary Algorithm Gear-Pair Dimensioning and Multi-Objective Genetic Algorithm-Driven Bearing Selection

Abstract

This paper investigates global optimization methods applied to the design of a one-stage gearbox, aiming to partially automate the design using artificial intelligence. The developed software autonomously determines the gearbox parameters (number of teeth, gear width, modulus, etc.), optimizes them, and then models the assembly in Siemens NX CAD (computer-aided design). The direct connection between optimization and CAD leads to a faster designing process. The literature review reveals that the field of machine design is quite conservative, and only a few articles with some similarities to our research have been found. The paper describes gear dimensioning and the application of the Ipopt algorithm to the optimization of gear-pair parameters. Then, it addresses shaft design and bearing selection through multi-objective optimization using the NSGA-II algorithm, balancing cost, weight, and volume while meeting strength and durability constraints. The paper also describes the transfer of the optimized parameters and the creation of a CAD model. The last part is dedicated to the problems encountered, their potential solutions, and the advantages of the new approach. The proposed approach delivers a functional, optimized CAD model in about 10 min, providing a notable speed advantage over typical manual workflows.