Research On Controller Of Dynamic Positioning System By Using Model Predictive Control

With the development of deep sea space exploration and the popularization of ships and marine work platforms which have specific features,there is a growing need for ships to maintain a set position or to set a track for navigation.Considering the feasibility,cost and accuracy,the traditional methods of ship control can not meet the requirement of marine operations in new period while dynamic positioning system(DPS)can satisfy this demand.The DPS consists of a measuring system,a control system and a thrust system,controller is the key of the DPS.In this paper,focusing on the design of controller based on model predictive control,the main contributions can be summarized as follows.Firstly,two coordinate system and low frequency simplification model for ship motion are set to describe the ship movement.Also,the high frequency model of the ship motion is simulated by harmonic oscillation.Besides,the interference caused by wind,second-order sea wave and sea current are resolved into the force and torque from ship's three degrees of freedom direction.Secondly,a state observer for dynamic positioning ship based on the fusion of UKF and PF is designed.The fusion algorithm has a good state estimation accuracy compared to conventional observers which use EKF or UKF,whenever the ship model parameters match or not.Thirdly,a self-correcting model predicative controller is designed.This controller also has a good control effect while ship model parameters doesn't match better.In this design,the mode which directly sets the upper speed limit is replaced by the ship position change item in rolling optimization.This change allows the ship to approach quickly and slowly reach the set position so as to achieve the purpose of reducing or eliminating overshoot.Finally,the control effect of the model predictive controller is verified by a near sea navigation model in the pool.In the experiment,the influence of the position deviation weight and the position change weight is consistent with the design.Besides,the control effect of the controller in ship model matching and model mismatch is verified respectively in still water,the results show that the controller effect of the algorithm can meet the setting requirement in the two conditions.