Nonlinear-Disturbance-Observer-Based Model-Predictive Control for Servo Press Drive

The servo press is a nonlinear system with built-in force amplifiers, and the moment of inertia varies across positions. Due to the existence of assembly clearance, a conventional mini-closed loop of electrical machine position employing linear control methods cannot follow the curve very well. Moreover, the error of the cascaded position loop under heavy load conditions gives electromagnetic torque that counteracts the disturbance torque. In this article, we propose an effective semi-closed loop for crank position control. The modified nonlinear disturbance observer based on computed torque (NDOCT) in a servo drive is designed to improve the transient performance during the stamping process. The Lagrange equation allows a nonlinear disturbance observer to select suitable parameters and get a real-time estimate of the torque load in a servo press, which is beneficial to improving the dynamic behavior under heavy load. The model-predictive control offers fast current responses and reduces the switching frequency for an electrical drive. An experimental study is tested in an industrial 1600-kN servo press. As far as the authors know, this is the first time the NDOCT has been implemented on an actual servo press to solve problems with nonlinear position control under heavy load conditions and improve dynamic performance.