AbstractsComputer Science

This Msc graduation study covers the installation of topsides on a jacket by an offshore heavy lift crane vessel, and its related problems. One of the problems take place after the topside is hoisted on the jacket, during the weight shift from the hoist cable onto the jacket. The topside then exerts large horizontal forces upon the crane and vessel. The problem is that the DP system is not designed to cope with the changed force characteristics. Due to the changing force characteristics the DP system may build up oscillations and even instable behavior. The ultimate consequence may inflict great (economic) damage, human injury or even loss of lives. The current industry solution to this problem is found to be unsatisfying. The goal of this thesis is to propose one final conceptual solution to this problem. This final conceptual solution is selected from four candidate solutions after simulation performance tests. All candidate solutions use the same basic idea of estimating the large horizontal force and feed this into the DP controller. The performance of the candidate solutions is tested on a simulated model of a crane vessel hoisting a topside on a jacket. This simulation model is based on the Imtech Marine DP and vessel model which is extended with a heavy lift crane. With the use of this simulation model the problem of instable behavior is reproduced in a simulation environment using realistic parameters. One of the candidate solutions, denoted the Feed Forward solution candidate, translates the estimated force into compensation thrust by the vessel's own actuators. As a result, the horizontal forces are eliminated, whereby the DP controller will not be disturbed by the changed force characteristics anymore. In literature, this feed forward control law performed very well during scale model tests. In this study however, the performance proved to be very poor. It is shown that the performance is mainly poor because of the acceleration and deceleration rates of the thrusters are too low. The second candidate solution, denoted the Kalman solution candidate, is a more simple candidate and only feed the estimated forces to the Kalman filter. This candidate is performing good, but DP stability is not guaranteed under all conditions. To determine the stability, with or without the Kalman solution candidate, an theoretical analysis method is explained. It is concluded that of all candidate solutions the Kalman solution candidate proved the best performance. Consequently, it is recommended to develop the conceptual candidate to an industrial solution. During this development it is recommended to take extra measures to increase the robustness against differences between the heavy lift vessel model and the true heavy lift vessel.