Study On Ship Maneuvering Motion Simulation With Intelligent Technology Applications

Started with the first International Symposium on Ship Maneuverability held in 1960 in Washington, D.C., U.S.A, the research on ship maneuverability has been progressively improved. (International organizations-such as the IMO (International Maritime Organization), the ITTC (International Towing Tank Conference) and the SNAME (Society of Naval Architects and Marine Engineers) have become more active in the stimulation of research on those aspects of ship maneuverability which are vital for navigation safety and oceans protection. During last several years, the requirement to improve the maneuverability of ship has been increased among many international societies such as IMO, IMPA (International Maritime Pilot Association) and IACS (International Association of Classification Societies).By analyzing ship maneuvering motion, we can build ship maneuvering motion mathematic model and perform ship maneuvering motion simulation. Thus not only ship maneuverability and running status can be predicted, but also port and channel design can be optimized.Generally, the parameters of ship maneuvering motion are subjected to vary with ship running status and circumstances. Therefore, ship maneuvering motion is time-variant, non-linear and noise disturbing process. The mathematical model of ship maneuvering and control system is extremely complicated and it is very difficult to use the traditional mathematical tools to describe.In the thesis, based on the traditional ship maneuvering motion equations, some simulations of ship maneuvering motion are carried out using different mathematical models, among which some intelligent technology such as Neural Network.The most important and difficult work in ship maneuvering simulation andprediction is to accurately estimate its hydrodynamic derivatives. In this thesis, Linear Neural Network is introduced for estimating the hydrodynamic derivatives using the data accumulated in full-scale tests.To apply Neural Networks to the simulation of ship maneuvering motion, an NNRM (Neural Network Recursive Model) is designed and used to simulate a serial full-scale tests conducted in Yangtze River and the comparison between simulated results and the measured ones is satisfactory.Ship trajectory tracking is a well-known maneuvering problem with an increasing practical and theoretical interest. But the real-world tracking applications encounter a number of difficulties caused by the presence of different kinds of uncertainty due to the unknown or not precisely known system model and environmental effects. A preliminary study is tempted with regard to the trajectory tracking problem by using NNRM, CTE and LOS hybrid controller.The problem of marine simulation and ship maneuvering motion is mainly and usually studied in traditional two right-handed coordinate systems, i.e., the space-fixed and body-fixed. In fact, it is shown that the novel kinematic frame, Serret-Frenet frame is the newest approach to follow trajectory path and to control the motion of surface vessel. Some research on ship maneuvering problem in Serret-Frenet frame is also discussed.Above all, ship maneuvering motion prediction and simulation is a developed science that provides highly useful tools for study ship maneuverability. From a technology perspective, the future is promising for improving the accuracy, flexibility and extent of prediction, simulation and simulator-based training applications. Consequently, further study on ship maneuvering motion prediction and simulation is badly needed.