Sizing Optimization of a Exoskeleton Structure Utilizing Finite Element Analysis and Multi-Objective Search

João Carlos Pereira Passos¹; Carlos Eduardo da Silva Santos²; Renato Coral Sampaio³; Leandro dos Santos Coelho⁴; Carlos Humberto Llanos Quintero¹

¹ University of Brasília - Department of Mechanical Engineering - Brasília - DF; ² Federal Institute of Education, Science and Technology of Tocantins - Campus Palmas - Palmas - TO; ³ University of Brasilia - Faculty of Gama, Software Engineering Group - Gama - DF; ⁴ Federal University of Paraná - Electrical Engineering Graduate Program - Curitiba - PR

Baixar PDF doi:10.20906/CPS/CBA2018-1375

Resumo

Nowadays there are many elderly people who need assistance to walk. For that purpose, an exoskeleton is a type of wearable mechanism that can help them in both rehabilitation and locomotion tasks. In its autonomous form, an exoskeleton is required to be both lightweight and sturdy. However, there is a trade-off between these two characteristics. The mechanical design of its parts usually require simulations and stress tests to determine the best shapes that reduce weight and obey a safety factor, which are contradictory criteria. This mechanical design demands highly skilled professionals and is time and cost consuming. In this context, this work proposes the use of multi-objective optimization algorithm to automate mechanical simulations and assist mechanical designers. A case study of the modelling of an exoskeleton part is presented. The multi-objective optimization is performed using finite element analysis by simulating the part under external forces while altering its dimensional parameters. At each iteration, a cost function evaluates the solutions based on the safety factor and the mass of the resulting model. Finally, multiple results in the Pareto Front are presented and discussed.

Palavras-chave: exoskeleton; sizing optimization; multi-objective search; finite element analysis; differential evolution