EKF based model identification for a relaxed dynamic positioning ship using NMPC method

The extended Kalman filtering (EKF) based identification is an effective identification technology. It has made successful applications in ship model identification fields to cruise-speed ships. This paper introduces it to the dynamic positioning (DP) ship model identifications. The DP ship is of different motion characteristics with that at the cruise-speed. Because the ship speed is close to zero, DP ships are easily affected by the sea environment disturbances, which makes the identification difficult. In order to assure the correctness and improve the accuracy of the DP ship model identification, the batched identifications regarding to different degree of freedom (DOF) is necessary and must. The main hydrodynamic parameters which affect the ship motion characteristics primarily are need to successfully identify firstly, and then, other hydrodynamic parameters are identified in 3 DOF coupling ship motions. After the identification, the identified ship model is tested, to verify whether the motion characteristics of the identified model is basically consistent with the actual ship. Also a model based control method - nonlinear model predictive control (NMPC) is designed to test the controlling performance which reflects the correctness and accuracy of the identified ship model meeting requirement from aside.