Study On Modeling And Intelligent Control Strategy For Brushless Doubly Fed Machine

Brushless Doubly Fed Machine (BDFM) is a new kind of machine developed in recent years, which consists of two sets of stator control winding and stator power winding unattached one another, and the rotor winding of special structure. It can be used in AC adjustable-speed system as a motor, using an inverter to adjust the power frequency of stator control winding and releasing the speed control of BDFM. On the other hand, BDFM can be applied in a variable velocity constant frequency and voltage generating system as a generator. The BDFM system has the advantages, such as brushless, simple structure, less capacity inverter, controlled power factor, and so on. So it will have extensive application prospects in saving energy adjustable speed system of pumps and wind machines and in generating system of water power and wind power.The major objectives of this research project are to propose an intelligent control strategy used in BDFM adjustable speed system, and enhance the robustness and static and dynamic performances, because of the indiscernible problems in the process of building the mathematical models of BDFM and the models behaving nonlinear characteristics in the three phase static coordinate system and two phase rotating frame.The research methods in this thesis are that based on deeply understanding the structural analysis, electromagnetism relationship analysis and running principle of BDFM, the theory studies have been done in the BDFM models establishing method, motor parameters identification method, and intelligent control methods. Because of the uncertainty of models establishing and the time variation of the part motor parameters, in order to get the better control performances of BDFM adjustable speed system, some intelligent control algorithms, including the fuzzy control, neural network control, sliding mode control, predictive control, support vector machine, and so on, have been employed in the applications research by the computer simulation.The dissertation is organized as follows:In Chapter 1, the developing statues, structure, operating principle, and application prospects of BDFM at first are introduced. And then, the modeling, motor parameters computing methods and control strategies of BDFM are summarized. It points out the disadvantages and problems existing in the design and running of BDFM. The jobs, purposes and significance of the project in this dissertation are given at the end of the chapter.In Chapter 2, the inductance computing methods of stator power winding and control winding, and rotor winding of BDFM have been introduced. Aiming at the information of the rotor parameters, rotor flux linkage, and so on, which are difficult to measure, it presents that the parameter identification of motor can be achieved by making the best of neural network (NN). The basic theory of NN and the problems identification are introduced. The technique of NN has turned into a rising method in identification of system. This chapter also proposes a method of structuring the network based on mathematics model of system. The parameters waiting for identification are acted as the weighted values of network. In practice of BDFM adjustable speed system, a general law and algorithm of structuring the network are given, which are proved by the mathematics model of motor parameters identification, and a simulation is carried out. All of results show the validity of the method.In Chapter 3, based on the idealization physics model and suppose of BDFM, the mathematics model in three phase static coordinate system has been built according to the electromagnetism relationship of BDFM. And using coordinate conversion theory, the BDFM mathematics models in the rotor speed d-q coordinate frame and in the double synchronism coordinate frame. Then, based on the model in the double synchronism coordinate frame, the rotor field orientation control equations have been deduced, and the dynamic torque control of BDFM has been released also and the simulation has been done. Simulation results show that the control strategy is the same with the BDFM adjustable speed system, and can reach the same effects with the induction machines, which has better dynamic characteristics.In Chapter 4, aiming at the indiscernible problems in the process of building the dynamic model of BDFM and improving the control performance of BDFM, an adaptive fuzzy sliding mode variable structure control algorithm is applied to control the speed in the BDFM adjustable speed system, and simulation has been done using the MATLAB/SIMULINK tool and software. By introducing the adaptive fuzzy control method,and genetic algorithm, the switch rate in the sliding mode switch control and feedback gains in the sliding mode equivalent control can be adaptively adjusted, which can reduce the chattering effects caused by the sliding mode variable structure control, and keep the stability and dynamic properties. The simulation and experimental results show that the employed control algorithm can not only keep the robustness of sliding mode variable structure control for DBFM, but also is of the advantages of fast responding speed, no overshoot, no steady-state error, etc.In Chapter 5, using the fuzzy predictive control algorithm based on the RBF neural network, the simulation study for the BDFM speed control has been done. RBF neural network is used to identify the model of BDFM speed performance, and the multiple inference and performance measuring method of adaptive fuzzy are applied to calculate the performance evaluation index, and then the fuzzy predictive control algorithm is employed to realize the BDFM speed control. Through predicting the output of future system, the effective control for the complex process can be realized, and the nonlinear plants like BDFM can obtain better control effect. Using the MATLAB/SIMULINK tool, the simulation results indicate that the control method applied can effectively realize the optimal control for the BDFM nonlinear system.In Chapter 6, the support vector machine (SVM) in the field of machine learing has been used in the modeling and control of BDFM, and the modeling method of SVM, which is employed in the modeling and control of BDFM, has been explored. By combining the SVM with the fuzzy control technology, the fuzzy inference controller based on the SVM has been designed, and the good dynamic response characteristics have been released in the BDFM adjustable speed system. The designed SVM identifier can accurately describe the nonlinear relationship of the parameters between the speed and the torque and flux linkage in the BDFM speed control system. The simulation results indicate that the control method applied can effectively realize the real-time control for the BDFM nonlinear system.In the last chapter, the advantages, progresses, and the lack of the dissertation are summarized, and the dissertation ends with presenting the keystones, directions in the future research and problems waiting to be solved.