| dc.description.abstract | Control of Remotely Operated Vehicles (ROV) with manipulator arms, require control of all Degrees of Freedom (DOF) of the vehicle. Merlin WR200 is a work-class ROV currently equipped with automatic depth and heading control, which is the industry standard. It is of interest to equip Merlin WR200 with local Dynamic Positioning (DP) capabilities to allow for more efficient operations close to the sea floor. This includes systems for station keeping, trajectory tracking, path following and low-speed maneuvering. The system is local in the sense that there are no global reference systems, and the position is thus found relative to the sea floor. Robustness is essential to account for uncertain modeling and suppression of disturbances from ocean currents and tether-induced motions.
Previous work on DP for Merlin WR200 lays the foundation for this thesis. The previously proposed control system has been extended for full-DOF control, including compensation of the actuator dynamics and tether disturbances in the control loop. The proposed changes have been documented in a simulation study using the high-fidelity Merlin WR200 simulator at the headquarters of IKM Subsea Solutions AS. Using the simulator, it was possible to develop a controller that was further verified in full-scale experiments. As such, the simulator proved to be a highly valuable tool for verification of control design.
Systems for station keeping, trajectory tracking, path following and low-speed maneuvering was implemented in Matlab and verified using the simulator. The proposed controller was able to meet the control objective in all scenarios considered in this thesis, even when disturbed by ocean current. Successful station keeping was possible with current speed up to 2.5 knots, at which point two thrusters were running at full speed. The modules for low-speed maneuvering and station keeping was implemented on a PLC, and used in successful positioning and stabilization of a Merlin WR200 ROV in a full-scale experiment.
The robustness of the system can be credited to the disturbance observer, which allows for fast and effective integral action by balancing the equations of motion. The extension to compensate for actuator dynamics, although effective, does not pull its weight in terms of the additional implementation complexity. It is suggested to replace the parameter adaptation scheme with the proposed disturbance observer, and consider omitting compensation of actuator dynamics for a simpler implementation. | |
