Automation of underwater operations on wave energy converters using remotely operated vehicles
- Plats: Häggsalen, Ångströmlaboratoriet, Lägerhyddsvägen 1, Uppsala
- Doktorand: Remouit, Flore
- Om avhandlingen
- Arrangör: Elektricitetslära
- Kontaktperson: Remouit, Flore
In the last fifteen years, the Division of Electricity at Uppsala University has been developing a wave energy converter (WEC) concept. The concept is based on a point-absorbing buoy with a directly driven linear generator placed on the seabed.
Several units are connected to a marine substation, whose role is to collect and smooth the power absorbed from the waves and then bring it to the shore through one single cable.
A big challenge in the project is to reduce the costs related to the deployment and maintenance of the WECs and substation. Currently, those operations are performed by divers, which is costly and entail considerable risks. A possibility is to replace divers with automated solutions using small robots called remotely operated vehicles (ROVs). This PhD thesis proposes and analyses a method for deployment and maintenance of underwater devices with no use of diving operations.
Existing ROVs need additional modules and equipment in order to carry out operations with the required force and precision. Typical missions are inspection, shackles or slings removal, valve closing, and cable connection. The latter demands especially high precision in the positioning: 5 mm in distance and 5◦ in heading angle. In addition, this operation involves forces up to 200 N. This combination power-precision is not reached by existing ROVs. This PhD thesis presents a positioning system for underwater robot to enable autonomous positioning of the vehicle before cable connection.
The positioning system is composed of two green lasers and a monocular camera, and is based on image processing. Experimental results from laboratory testing show that the mean absolute error in distance measurement is as low as 6 mm at 0.7 m from the target, whereas the heading is minimized to 2◦. The computational time for the image processing is 13.6 ms per image, meaning the possibility of a 30 Hz measurement system. Used together with a closed-loop path-following unit, this positioning system can support autonomous docking. This PhD thesis presents the model of an autopilot and results from docking simulations, showing the performance of the positioning system used in closed-loop.