Dynamic Positioning in Extreme Sea States: Improving Operability Using Hybrid Design Methods
Master thesis
Permanent lenke
http://hdl.handle.net/11250/238924Utgivelsesdato
2014Metadata
Vis full innførselSamlinger
- Institutt for marin teknikk [3472]
Sammendrag
This thesis investigates the performance of high-level hybrid dynamic positioning (DP) algorithms in extreme environmental conditions. A vessel in DP uses the thrusters as the sole means of keeping position in wind, waves and current, giving both flexible and accurate position keeping. Therefore vessels with DP capabilities are high in demand in industries like for instance the offshore, aquaculture, renewable energy industries and emerging fields like offshore mining. Due to high day rates the focus today is on developing DP systems for extending the operational window to even harsher environments, while keeping the solutions safe and environmentally friendly.When a sea state transitions from calm to extreme, the wind velocities increase generating higher and longer incident waves. This makes both horizontal and vertical relative motions of the vessel larger with longer periods of oscillation, which has consequences for the DP system onboard. Given the nature of a transitioning sea state, hybrid design methods are used to design a controller and an observer concept.A hybrid controller, for a vessel in a varying sea state is designed, and global asymptotic stability is established. Simulations in a sea state varying from calm to extreme are conducted with the hybrid controller, consisting of four candidate controllers, and a single controller with adaptive wave filtering for comparison. The single controller becomes unstable in extreme seas whereas the hybrid controller shows good performance. Candidate controllers are selected based on spectral analysis of the vessel wave frequency motions.A simplified sensor-based hybrid observer concept is investigated for noise robust position estimation. The concept assumes that acceleration measurements are readily available, and can be integrated to obtain position estimates. Position measurements are taken occasionally, and at these instances the position estimate is updated. Stability of the concept is analyzed giving uniform global asymptotic stability, and the simulation of two one degree of freedom sensor-based hybrid observers which rely on acceleration, velocity and position measurements is conducted.