| With the development of science and technology, offshore ocean exploration,especially oil and gas exploration in deep sea has been experiencing a continuous boom.Better positioning performances are needed for deep sea structures and offshore ships, asthe sea condition becomes harsher with the increase of water depth and distance from theshore. MP (Mooring Positioning) is not suitable for deep sea operation, due to its ownlimitations, such as lack of accuracy and mobility, high cost for construction andespecially its bad functionality with large water depth and obstructed seabed, etc. DP(Dynamic Positioning) has replaced the traditional MP, as a better option to maintain theposition of a floating structure in deep sea or to make it follow a pre determined operationtrajectory by means of active thrusters. DP is becoming more and more important to allkinds of offshore operations, such as oil gas exploration, scientific drilling, pipe laying,mineral drilling, etc.Ship DP system consists of reference system, control system and power andpropulsion system, among which control system is of great importance. Most of theprevious researches on ship DP control system are based on a certain sea condition,utilizing fixed control algorithms and models. Seldom are there publications dealing withsystem operating under complex sea conditions. In this thesis, different sea states areconsidered, and various observers and controllers with different structures andfunctionalities are designed to meet diversified needs. The final hybrid control systemcapable of auto switching among control algorithms and models is made possible by aswitching logic. The performance of the control strategies are tested and verified throughcomputer simulations. Firstly, the mathematical model of ship DP control system including wind, wave,current and ship motion model are built up. In the control system, feedforward wind loadis approximated as Gaussian white noise. High Frequency (HF) first order wave modelwith large amplitude and slowly varying low frequency (LF) second order wave modelwith small amplitude are treated separately. To reduce the energy consumption andthruster wear, first order wave load is especially modeled and filtered. Current issimplified as slowly varying bias, since its complex physical characters make it hard tomodel.Secondly, four different feedback control loops with proper observers and controllersare constructed in four typical sea states. By comparing three filtering algorithms, namelythe traditional low pass filter, Kalman filter and nonlinear passive filter, a more suitablemodel is chosen to eliminate the first order wave disturbance.Thirdly, in an attempt to get a smooth response of the control system from themoderate sea state to the extreme sea state, a separate controller is built by assigningweights to controllers of the second and fourth control loops. For the fourth control loopwith extreme sea state, the relatively low frequency of the first order wave renders thenotch frequency of the filter designed, within the control bandwidth of ship's LF motion.For this reason, the observer for the fourth sea state is modified so that the first orderwave load enters the control process.Finally, a hybrid control system is constructed based on scale independent hysteresisswitching logic, which allows the system to automatically switch among different controlstrategies in four different sea states.In this thesis, a ship dynamic positioning hybrid control system is designed, whichcan operate in different sea states. It provides an effective solution for the control systemdesign of DP, thus contributes to the design and wider application of ship DP system. |