Imagine standing under a load with a weight equivalent to more than 5,000 Tesla’s model X. Panicking is not an option and you need to be fully focused on completing a crucial task. There is no option to leave this precarious situation and your contribution is vital to the success of a multi-million dollar project. Sounds frightening, right? Oddly enough, it’s a common event during offshore topside installations at Heerema Marine Contractors. During these kind of operations a rigger foremen and an assistant superintendent take place on the jacket. With the role to communicate topside positioning information to the superintendent, on the crane vessel. They have to provide clear instructions until the installation is completed. Although the installation of offshore topsides have always been carried out with people on the jacket, without any major incident to date, having people operating under suspended loads of up to 10,000 tonnes is considered unwanted. Therefore, HMC is looking for a robust system to replace the presence of people on a jacket during topside installations. Existing techniques developed by HMC consist of robotic total stations and the use of augmented reality. However, these techniques are limited by the view from the stern of an SSCV.
In this thesis, a novel motion tracking algorithm is developed based on drones, fixed cameras and visual object tracking. Drones are already starting to change how businesses operate – and this is happening today. Companies across industries are using them for inspection, monitoring, repair work and onsite security. They are also being used for real-time data collection. Drones are able to take any position with respect to the topside or jacket and can mimic the view from people on the jacket. They are therefore not limited by the view from the SSCV. The developed algorithm is able to localise a pair of Aruco markers in an image captured by the vision system. Aruco was only recently introduced which makes this solution unique in the offshore sector. If a marker pair is recognized successfully, relative distance calculations can be made. By conveniently placing these markers on the topside stabbing cones and jacket legs, the topside relative motions can be estimated. A minimum of two locations need to be monitored in order to perform a successful estimation.
Two configurations have been proposed to test the algorithm. In the first configuration use is made of four fixed cameras on the stern of the SSCV. The four fixed cameras will need to track the marker pairs. Also one drone is available to provide visual confirmation from every desired position. In this configuration, the stabbing cones which are closest to the SSCV are monitored. In the second configuration three drones and one fixed camera on the stern of the SSCV are used. In this configuration the drones are used to track the marker pairs. Unlike configuration 1, the stabbing cones diagonally opposite to each other are monitored. Both configurations are able to provide relative positioning information during a topside installation. The first configuration is limited by the view from the SSCV while the latter configuration is not - since drones are used.
The motion tracking algorithm was tested in a virtual simulation experiment. Experimental 3-DOF results demonstrate the accuracy of the proposed method compared with the simulation log of the virtual environment. For configuration 1 the mean absolute error was found to be 0.048m with a standard deviation of 0.040m. For configuration 2 the mean absolute error was found to be 0.034m with a standard deviation of 0.020m. Both configurations are therefore within the 0.15m acceptable error margin. It can be concluded that the configuration using drones seems to perform better than fixed cameras from the stern of the SSCV. An obstacle of using drones for this purpose is the need of certified operators and the limited power supply. Autonomous drones can be a solution for the first obstacle. The second obstacle might be tackled in the future with the continuous battery improvements fostered by the automotive and electronic consumer goods industry. Nevertheless, the motion tracking algorithm using Aruco markers looks very promising. Taking into account the steep developments of drones over the past year, the future use of drones during offshore topside installations looks promising.