ESTIMATION OF NATURAL FREQUENCIES FOR ROBOTICS AT DIFFERENT POSTURES USING PARTICLE SWARM OPTIMIZATION AND RAYLEIGH QUOTIENT
Journal
Proceedings of the International Congress on Sound and Vibration
ISBN
9788011034238
Date Issued
2023-01-01
Author(s)
Abstract
The paper develops a nonlinear dynamic model to predict the natural frequencies of a robotic manipulator at different postures. The model with inertial and stiffness matrix is established by using the Lagrange approach. The linear and nonlinear parameters of the stiffness matrix are identified by using modal testing and particle swarm optimization (PSO) method. The contribution of the inertial and stiffness matrix at different postures are further evaluated by using Rayleigh quotient. Experimental results show that the variation of natural frequency of the arm at different postures is mainly caused by the changing inertia at different postures. Moreover, the nonlinear stiffness term could affect the natural frequency significantly at the extreme locations of the robotics. The averaged difference of the first mode natural frequency of the arm at different postures was less than 0.6% between the experiment and that estimated using the proposed dynamic model. The averaged difference was less than 2.6% if the first and second mode natural frequencies were combined with weightings of 0.7 and 0.3 using in the PSO. It is sufficient to provide time-varying input shaping commands for arm vibration suppression; however, further improvements of the dynamic model will be conducted in the near future.
Subjects
Dynamic model | Modal testing | Natural frequencies | Particle swarm optimization | Six-axis robotic arm
Type
conference paper