Dynamic analysis of the stinger hang-off system cable forces using ship motion data of Audacia.

Abstract

Dynamic analysis of the stinger hang-off system cable forces using ship motion data of Audacia. Already for some time the desire exists to acquire insight into the phenomenon of stinger uplift. This is the situation where external forces and motions acting on the ship, stinger cause the hang-off cables to have no or barely any cable tension. Consequently, when the stinger “comes down” snap loads occur in the cables. This situation should be avoided. To gain quantitative insight into uplift occurrences, four load cells have been installed on the vessel Audacia to record stinger cable forces. However, the usability of these recordings was unclear. It was therefore decided to develop two dynamic models to predict measured cable forces on the basis of measured ship motions. This is the subject of the present study: A computational model is presented to predict time-varying cable forces for the operational setting that the stinger is lifted out of the water. Before the models were created, a direct comparison was performed between the measured cable forces and the motions at the centre of gravity of the ship. It appeared that, that the pitch motion is the most influential motion for the cable forces. However, a type of prediction of the cable forces directly by the time traces of the motions in the centre of gravity is not considered accurate enough. Additionally, it became evident that the roll motion showed no correlation at all. Therefore, maintaining the 2-D analysis in further stages of the thesis is considered justified. This thesis considers two dynamic models which are performance compared, i.e. the rigid links model (RL model) and the flexible links model (FL model). The RL model assumes the complete system of the ship, stinger and hang-off wires to be rigid. The FL model assumes the ship and the stinger rigid but it accounts for a hinged stinger connection to the bow of the ship and stretching of the cables. Additionally, the flexible links model is based on multiple approaches. The first approach is a direct time-integration method for solving the equations of motion. The second approach is using the modal analysis to solve the equations of motion. Audacia’s on board measured motion data is used as input for the computational models. For verification of the model results, the measured cable forces are used. The information of the measured displacement data is read-out, resampled, filtered and processed for reviewing. The RL model and the FL model were reviewed for their correlations and/or similarities to the measured cable forces using eight reviewing methods. Reviewing these results it can be concluded that the most accurate model to predict the cable forces in the hang-off system based on ship motions is the FL model. This model can predict the standard deviation of the cable forces with an maximum error of 3,7%, which is essentially the worst case scenario uncertainty due to input errors and the uncertainty due to errors in the measured cable forces. This accuracy of on a total predicted mean cable force with an error of 0.08% can be regarded satisfactory for the data time trace considered. When studying the FL model sensitivity the stiffness values appeared to play a dominant role. The first and closest eigenfrequency, considering the FL model, is 1.9 Hz (or 0.52 seconds period). Logically, the most sensitive excitation frequencies are the frequencies closest to this first eigenfrequency. Noise in the measured data used as model input may adversely affect the utility of model results. After research multiple causes for the high frequency noise are found and mitigation strategies are proposed. Reviewing the models it is apparent that the performance of the FL model is better than the performance of the RL model in prediction of the cable forces based on the ship motions at the centre of gravity, for the cases considered in this thesis.