Model predictive motion/force control in robotic grinding system for turbine blade

The nonlinear time-varying contact state and low robot motion accuracy in the robot grinding system cause unstable contact forces and position errors in tool paths, reducing the grinding quality. This is particularly problematic when grinding complex curved workpieces such as turbine blades. Therefore, a model predictive motion/force control (MPMFC) algorithm is proposed to achieve precise control of the robot's trajectory and contact force. In this algorithm, an online trajectory interpolation method between adjacent cutter-contact (CC) points is first implemented to acquire the ideal trajectory information of the robotic end-effector within one control cycle. Furthermore, the control system of the MPMFC is developed based on optimum control, incorporating an admittance model for contact force prediction between the workpiece and grinding tool, as well as a robot kinematics model for motion prediction. Simulation and experiments are conducted to verify the superiority of the control algorithm. The force control accuracy and position control accuracy with the MPMFC method during the tracking process of the blade are respectively 1.934 N and 0.132 mm, which are improved by over 38 % and 37 % than that with some conventional motion/force controllers.