Force/Position Control in Robotic Grinding System for Point Clouds of Turbine Blade

Ziling Wang, Jiantao Li, Lai Zou

Abstract: With the development of optical measurement technology, 3D points are gradually used to characterize actual geometry of turbine blade. Moving trajectories can be generated by slicing the point clouds of the turbine blade. However, due to the low precision of robotic motion and the complex curvature changes of turbine blade surfaces, fluctuations in contact forces and trajectory errors can easily occur, ultimately affecting the profile accuracy during the grinding process of the blade. Therefore, to address the issues, a force/position hybrid control method is proposed. In this approach, the position/attitude in the contact environment between the grinding tool and the turbine blade is firstly determined, then, the position/posture controller is constructed based on PID method. Furthermore, an admittance controller illustrates the contact characteristic of the contact environment is established. Based on this, an online neural network is used to adjust the key admittance parameter based on the control errors to further improve force control accuracy.