Dynamic Model Identification of Axial Force in Robotic Friction Stir WeldingÕ

摘要

This paper presents the dynamic model identification of axial force during the robotic Friction Stir Welding (FSW) process. Due to the large external force exerted on the robot, the natural stiffness of the industrial robot is not sufficient and the deformation is nonnegligible. Thus, two dynamic models of axial force are proposed by using the positioning error along the spindle axis and its changing rate. By means of a nonlinear observer and data filtering, experimental studies have been carried out on a robot-based FSW system in order to identify the proposed models based on the least squares (LS) method. The results especially for errors of identified models have been systematically analyzed, showing good performance of the identification.