Intelligent Modeling And Fuzzy Generalized Predictive Control Research In Ship Anti-pitching Control

Ships influenced by the wave will produce six degree-of-freedom movements in the sea. The serious pitch movement and vertical swing movement may cause first bang and seas over deck, which is great influence of seaworthiness, security, weapon attack's accuracy in the war, comfortableness for people's work and life to ships, especially naval vessel. It is possible that serious pitch movement results in naval vessel to lose the operational capacity, even to be capsized due to swing. Ship rolling motion control technology makes the rich progress after the widespread research. At the same time, longitudinal movement's control research is also obtained widely taken seriously.First of all, the reasonable ship movement model is the key factor to ship control technology application. Therefore this paper used two different methods to build ship motion model. The first method was that using the cascade combination neural network built ship longitudinal motion hydrodynamic force parameters adaptive model, because the ship movement hydrodynamic force parameter is nonlinear function of the sea sentiment, the speed, the heading angle. The other method was using generalized dynamic fuzzy neural network based on empirical datum of ships model to establish respectively ship longitudinal motion adaptive model of ship vertical swing and pitching. Then they was carried on the partial dynamic linearization in each sampling point to the system, and the ship control neural network adaptive models were obtained. Simulation results proved that neural network adaptive model established by above two methods was reasonable.The second, according to randomness and uncertainty of ship swaying motion model, generalized forecast control law was deduced under the CARMA model(single variable) and the state space model(multivariable) based on the generalized predictive control(GPC) theory, which was applied to the ship longitudinal motion adaptive model.The simulation result indicated: Both can realize the effective control to the ships pitching. Control effect of using GPC based on(Generalized Dynamic Fuzzy Neural Network) intelligent adaptive CARMA model was better than that based on(Cascade Combination Neural Network) adaptive state space model.