SHAPE AND SIZE OPTIMIZATION OF A GEARBOX BRACKET USING GENETIC ALGORITHMS

?This paper represents a case study regarding the structural optimization of a typical gear bracket used in automobiles. The functional specifications mimic those found in real gear brackets, both in terms of geometry and boundary conditions, including the loading, thus, a direct comparison with such existing brackets can be made. The optimization procedure uses an original implementation of genetic algorithms, as found in the Object-Oriented Genetic Algorithm framework for MATLAB. For the evaluation of solutions, the FEA software Abaqus was used, in conjunction with the scripting language Python. The results show the methodology can lead to lighter parts while keeping the same strength and rigidity requirements. It also allows the automation of a task that would otherwise require important design efforts, without guaranteeing the same quality of solutions.