Application Of Model Predictive Control To Dynamic Positioning System

Dynamic Positioning (DP) is a technique of automatically maintaining the position and heading of the floating vessel within the specified limits by controlling onboard thrusters which generate thrust vectors to counter the wind, wave and current forces. The research in the paper is a simplified design based on the main idea of the GreenDP, which is a new DP product made in Kongsberg Company in Norway. The purpose of the paper is to apply the advanced control technology of Model Predictive Control to the DP controller design.A simple motion mathematical model of the vessel is established and educed according to some reasonable supposition. A bias model is established for the slowly varying environmental forces and moments, such as second-order wave drift, sea current. The modeling of environmental disturbances is simplified.A Vessel Parallel Reference Frame is introduced. Small angle theory is used to get the linear model of the vessel's low frequency motion, that is, model used in the controller design. This work avoids the complex Taylor linearized method.A Kalman filtering method is used to separate the low frequency signals and high frequency signals as well as to get the optimal state estimate of the vessel motion. Combining with the wind force feedforward control, a LQG-DP system based on Kalman filtering is designed.The paper is the first study for introducing the MPC technolgy to DP system design. Because of the particular advantage of MPC in dealing with the constraints, the working point of the designed DP system can approach the edge of the allowable operation area and can achieve higher performance and more economic benefits. Moreover, due to the decoupling function of the arithmetic itself, a lot of work is avoided. It shows that the control effect of MPC is rather good.Lots of simulation results show that MPC is suitable for DP system. It can well deal with constraints such as the thruster limits, operation limits. It also can satisfy the requirements of the position control accuracy and can attain better rapidity and stability than LQG control.