Research On Control Of Ship Shaft-Driven Doubly Fed Induction Generator

In the face of rising oil prices and increasingly stringent environmental regulations,the major shipping companies in the world are paying more attention to energy saving and emission reduction of ships.In recent years,the shaft-driven power generation system has been widely used on ocean and river vessels.The shaft-driven power generation system utilizes the reserve power of the ship main engine to generate electricity,which can save fuel costs and reduce exhaust emissions.The doubly fed induction generator(DFIG)has a good application prospect in the field of the ship shaft-driven power generation due to its advantages such as variable speed constant frequency operation,small converter capacity and flexible power regulation.In this thesis,the control techniques of the grid-side and rotor-side converters of the ship shaft-driven DFIG are investigated.The main work of the thesis is as follows:First,a neural network dynamic surface control(NNDSC)method is proposed for the grid-side converter of the ship shaft-driven DFIG.The neural networks are applied to approximate the model uncertainties of the grid-side converter,which eliminates the dependence of the controller design on the line resistance and the load equivalent resistance values.Moreover,by introducing the tracking differentiator,the analytic derivation of the virtual control law is avoided,and the complexity of the controller design is reduced.The NNDSC method can regulate the DC bus voltage accurately and realize the unit power factor operation of the grid-side converter.Meanwhile,it has strong load adaptability.In order to further improve the control performance of the grid-side converter,a predictor-based neural network dynamic surface control(P-NNDSC)method is proposed.The tracking error of the predictor is employed to update the weight vector of the neural network,which reduces the peak of the neural network output in the initial moment and expands the selection range of the adaptive gain.The P-NNDSC method can suppress the DC bus voltage overshoot and shorten the DC bus voltage adjustment time effectively.The stability of the closed-loop control systems is analyzed by using Lyapunov stability theory.Simulation and experimental results are given to verify the effectiveness and feasibility of the proposed control methods.Then,a neural network integral sliding mode rotor current control method is proposed for the grid-connected ship shaft-driven DFIG.The neural networks are applied to approximate the equivalent control items of the sliding mode control laws,which eliminates the dependence of the controller design on the DFIG parameters.The presented control scheme can regulate the rotor currents quickly and accurately.Meanwhile,it has strong robustness to the rotor rotating speed variation.As a consequence,the grid synchronization and the stator active and reactive power decopupling control of the DFIG can be achieved.The stability of the closed-loop control system is analyzed by using Lyapunov stability theory.Simulation and experimental results are given to verify the effectiveness and feasibility of the proposed control method.Finally,a neural network integral sliding mode stator voltage direct control method is proposed for the stand-alone ship shaft-driven DFIG.The stator voltages are controlled by the rotor voltages directly,such that the rotor current control loops are not required,the controller structure is simplified,and the response speed of the stator voltages is improved.The presented control scheme can maintain the stability of the stator voltage amplitude and frequency.Meanwhile,it has strong robustness to the load and the rotor rotating speed variations.As a consequence,the variable speed constant frequency stable voltage operation of the DFIG can be achieved.The stability of the closed-loop control system is analyzed by using Lyapunov stability theory.Simulation and experimental results are given to verify the effectiveness and feasibility of the proposed control method.