Pareto front generation for integrated drive-train and structural optimisation of a robot manipulator conceptual design via NSGA-II

Gulec M. O., Ertugrul S.

Advances in Mechanical Engineering, vol.15, no.3, 2023 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 15 Issue: 3
  • Publication Date: 2023
  • Doi Number: 10.1177/16878132231163051
  • Journal Name: Advances in Mechanical Engineering
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Applied Science & Technology Source, Computer & Applied Sciences, INSPEC, Directory of Open Access Journals, Civil Engineering Abstracts
  • Keywords: Robot design, integrated conceptual design optimisation, dynamic simulation of flexible bodies, the lumped parameter estimation, drive-train optimisation, non-dominated sorting genetic algorithm
  • Istanbul Technical University Affiliated: Yes

Abstract

Due to the complexity of the process, there is no single solution for determining the motors, gearboxes and structures of a robot manipulator according to the desired dynamic performance while minimising both the deflections in the structure during the dynamic motion and total robot weight. The solution of this integrated drive-train and dynamic structural optimisation problem is generalised for three degrees of freedom (DOF) robot manipulator via Non-Dominated Sorting Genetic Algorithm II (NSGA-II) to obtain the Pareto front of any desired robot manipulator overall conceptual design, including motors, gearboxes and thicknesses of the links. A flexible body dynamic simulation model was created in the MATLAB Simmechanics environment. The flexible bodies were defined via lumped parameter estimation method, which allows observation of the deflections in links during the dynamic motion. A library containing technical data related to motors and gearboxes was created to be utilised in the optimisation algorithm. The method accelerates the time-consuming iterative process for obtaining optimum conceptual design solutions for a dynamic system and allows for easy modification of design parameters and constraints. It also makes the algorithm suitable for different types of dynamic system designs.