Optimal control, self-tuning techniques and their application to dynamically positioned vessels
This thesis consists of two parts. The development of a self-adaptive stochastic control system for dynamically positioned vessels is described in Part One. Part Two is the investigation and the development of self-tuning control techniques. In Part One, the dynamic ship positioning control problems and basic components are described. The modelling techniques of low frequency ship motions and wave motions are given. The various Kalman filtering methods are appraised. An optimal state feedback control with integral action for the ship positioning system is proposed, followed by the simplification of the complex control structure to allow easy implementation. A self-tuning Kalman filter is proposed for systems which have low frequency outputs corrupted by high frequency disturbances. This filter is used in the ship positioning system. Simulation results of scalar, multivariable and non-linear cases are given. Part Two begins with the development of an adaptive tracking technique for slowly varying processes with coloured noise disturbances. Estimated results for various wave signals are given. The self-tuning control techniques are overviewed, followed by the development of an explicit multivariable weighted minimum variance controller. Simulation results including the estimation of system time delay are given. Finally, an implicit weighted minimum variance controller for single input-single output system is developed.