Structural optimization of robotic manipulators with fiber-reinforced composite materials.

摘要

The design of structurally advanced robotic arms with improved stiffness-to-weight and strength-to-weight ratios is investigated. Methodologies are presented enabling the design of manipulators with maximum structural stiffness and load carrying capacity based on (1) shape optimization of isotropic links without increasing the total weight of the arm, and (2) optimal prediction of ply angles and ply thicknesses of composite links. Both approaches incorporate finite element based objective functions. A new element capable of modeling tubular composite links is developed. Numerical solutions are obtained through a nonlinear programming algorithm adapted to the requirements of the analysis, the recursive surrogate optimization (RSO) method, which has proven efficient for computationally expensive objective functions. Uniform stiffness and strength improvement within the workspace is obtained by multi-postural static performance criteria. Such design criteria based on multiple postures also ensure the reliability of the optimal designs.;Application studies on a three-link robotic arm have been conducted. Experimental characterizations of the manipulator with static and dynamic modal analysis tests have been also performed. All application cases display the effectiveness of the method, and illustrate significant improvement margins in specific stiffness and strength. Redistribution of the material with optimal design has doubled both specific stiffness and strength. Optimal tailoring of composite links, without altering the geometric dimensions, has also doubled the specific stiffness and strength with respect to unidirectional plies. The specific stiffness and strength of the optimal composite design have been improved by factors of three and thirty, respectively, compared to the original aluminum design. The suitability and superiority of laminated fiber composite materials for the production of high-performance manipulators is conclusively illustrated.