Model predictive trajectory planning with fallback-strategy for an active Heave Compensation system

Active Heave Compensation (AHC) systems need an active control of the heaving winch in order to decouple the offshore crane's lift operation from the vertical motion of the vessel. Two-degree-of-freedom control (2DOF) can give good tracking control performance for such applications. However, every 2DOF control system requires a reference trajectory which has to fulfill requirements regarding continuity and differentiability. In case of AHC applications, the drive system of the winch is usually limited by velocity and acceleration constraints. Thus, the trajectory planning needs to take these constraints into account. Furthermore, the trajectory needs to be generated in real-time which narrows down the number of possible algorithms. In this work, a real-time model predictive trajectory planner is presented. By means of predictions of the vertical position and velocity of the crane tip, an optimal control problem is formulated and solved online for every time step. Furthermore, state constraints are taken into account. For the case when no optimal control solution can be found within the available time, a repetitive polynomial-based trajectory planner is used as fallback-strategy. The proposed approach is validated with simulations as well as experiments on the Liebherr AHC test bench.