Simulation Study On Motion Control Of Hybrid-drive Underwater Glider For Mesoscale Process Detectio

As a new type of underwater vehicles,the hybrid-driven underwater gliders combine the structure and operation mode of traditional underwater gliders and autonomous underwater vehicles.The hybrid-driven underwater gliders are characterized by large range,long endurance,high concealment and mobility which can perform various oceanographic monitoring missions.The hybrid-driven underwater gilders can realize two different motion modes: The first is gliding mode,which realizes the“sawtooth” motion by means of an active buoyancy regulating unit and an attitude regulating unit.The second is propulsion mode,which completes the sailing and yaw motions efficiently in virtue of the propeller propulsion unit and steering mechanism.The hybrid-driven underwater glider,“Petrel-II”,is regarded as the object in this thesis,the dynamic model and motion control are studied.Furthermore,the intelligent PID control and the motion simulation for the spatial sampling of mesoscale process is realized.The details are listed as follows:1.The establishment of the dynamic model of Petrel-II.Taking the buoyancy unit,the attitude regulating unit,the propeller propulsion unit and the rudder regulating unit into account,the force analysis of Petrel-II is provided,and the hydrodynamic characteristics are synthesized to construct the dynamic model in three-dimensional space.2.The analysis,simplification and motion characteristic simulation of the dynamic model.Based on the motion features of Petrel-II,the dynamic model pre-constructed is simplified,and then,the simulation of basic motion property under different modes are studied.The results show that the established and simplified dynamic model can describe the motion behavior of underwater glider accurately.3.Intelligent PID control and simulation.The simplified dynamic equations in the vertical plane and the horizon plane is analyzed,based on the simplified dynamic equations,the control schemes of velocity,pitch angle and yaw angle are provided,respectively.The PID control is synthesized with neural network and fuzzy control to realized the faster convergence speed and better performance.And then,the comparison simulation is presented for the proposed control schemes.4.The motion simulation for the spatial sampling of mesoscale process.Based on the simplified dynamic equations,for the observation of the front in Sea of Japan,the sub-mesoscale anticyclone eddy in South China Sea and the large-scale anticyclonic vortex in Irminger Sea,the simulation motion schemes are presented based on their detection requirements,so as to provide reference and optimization parameters for the future observation and application.