Research On DTC Control System Based On Fuzzy Space Vector Modulation

The direct torque control takes motors and inverters as a whole, builds the model of torque and flux directly in the stator coordinate system, and selects the appropriate switch vector to control the operation of the motor using hysteretic control. The traditional hysterics control transforms into the logic control signals to implement the appropriate voltage vector based on the size of the torque and flux error. The characteristics of this control method are monotonic logic signal, with less compensation voltage vector and complex switching table. At the same time, it produces the torque and flux ripple, without fixed switching frequency. As a result, the use and development of the direct torque control technology is limited severely.According to the induction motor, inverter principle and mathematical model, this thesis analyzes the influence of the voltage space vector on the torque and flux, studies the application of modern intelligent control in direct torque control, and proposes a method of combining the fuzzy and space vector to control the direct torque control. This method uses the intelligent fuzzy theory and neural network theory to establish the corresponding controller to replace the traditional hysterics and switching table, according to the deviation of the flux, the torque error and stator flux space position. The corresponding error compensation voltage vector is obtained by combined intelligent controller and voltage space vector. Eventually, the fixed frequency is transformed to improve the flux and torque ripple as well as the dynamic performance of the system.In order to verify the feasibility of DTC control system based on fuzzy space vector modulation, firstly, this thesis establishes conventional direct torque control system simulation model based on the principle of direct torque control, and then uses the basic principles of fuzzy control to design fuzzy controller. Besides, neural networks, which has the advantages of self-learning ability, is put into the fuzzy control to design the fuzzy neural network controller, the purpose of which is to further optimize the control structure and to improve the control effect. What’s more, combine the two controllers and voltage space vector, and then establish the direct torque control system simulation model based on these two controllers and the space vector modulation. Finally, MATLAB/SIMULINK toolbox is used to simulate and verity these two control methods. The results show that compared with conventional direct torque control, fuzzy neural network direct torque control has a good dynamic response and adaptive capacity. At the same time, it can reduce the flux and torque ripple effectively and improve the robustness of the system. Finally, after numerous and extensive theoretical and practical research, this thesis chooses TMS320LF2808as the core controller, which is produced by TI Corporation, and uses C programming language to design the system.