Compared to conventional industrial robots, robots that are used in machining require greater stiffness. When designing the Innoflex robot at the IFW, the stiffness was increased by using WITTENSTEIN Galaxie® drives. The robot is equipped with three rotary axes and one linear axis. A key element of the robot kinematics is the second rotary axis, which is actively stiffened by a hybrid drive concept consisting of a servo and torque motor. This combination increases the static stiffness by a factor of ten compared to conventional drives. The torque motor is used to actively stiffen the axle and compensate for joint angle errors at higher frequencies. The third axis with direct drive enables high axis dynamics on the TCP, but leads to vibrations in the robot structure. This was not considered in previous work, as no active compensation of structural vibrations was implemented. Furthermore, the pose-dependent vibration behaviour of the robot structure has not been considered so far. Accordingly, individual vibration modes cannot be specifically damped.
As part of the research project "Active vibration damping of a machining robot with hybrid drive", a model-based control method for active vibration damping of the robot structure is now being developed. The non-linear, position-dependent and therefore time-variant vibration behaviour represents the central challenge of the damping system for machining robots.
The control approach for the hybrid drive is based on a real-time capable multi-body model. This enables precise control of the varying vibration modes and the changing system properties of the robot structure. The dynamic stiffness of the robot, which depends on the robot pose in the workspace, can thus be effectively considered. During the machining process, dynamic forces act on the robot, which lead to vibrations and thus to pose deviations. The aim of the model-based control of the hybrid drive is to actively reduce these deviations in order to improve the surface quality of the machined component. Robust control approaches are being investigated in order to compensate for parameter uncertainties from the modelling.
The control of the hybrid drive consists of two controllers: the rigid body controller for the axle drive by the servomotor and a vibration controller for the torque motor. As the torque motor is in a power flow with the servomotor, there is an interaction between the two drives, which has an effect on the respective control loop. Therefore, the investigation of the coupling of both control loops is an important aspect of this research project.
Contact:
For further information, please contact Taha Araoud, Institute of Production Engineering and Machine Tools at Leibniz Universität Hannover, on +49 511 762 18340 or by e-mail (araoud@ifw.uni-hannover.de).
