Research On Dynamic Character & Control Strategy Of Brushless DC Motor For HEV

By way of the core parts of electrical vehicle, motor, driver and controller play important roles, and researches on them have important theoretical and practical meanings. Permanent magnet BLDC motor has been the best choice for HEV with its merits of high efficiency, high power density and superior speed regulation performance. In this thesis, the BLDC motor for HEV and its driver are taken as the research object with elaborate theory analysis and experiments. The thesis addresses eight basic issues constituting the main work as follows:1. By using Matlab/Simulink, the brushless DC motor is modeled. The simulation of open loop control and PI-PWM control are acquired and results are analyzed. With use vector control of induction motor for reference, a model of BLDC motor based on rotor position has been derived and simulation researches on the conventional time-based-variable model and rotor-position-based-variable model have been carried out.2. The dynamic characteristics of BLDC motor are analyzed in detail. By solving dynamic differential equation of BLDC motor and using first-order Taylor expression, the hyper-equations for the motor operation process has changed to the general differential equations and the expressions of current and torque are derived. Comparing the simulation results with the computation results, the computation results are precise enough to the engineering applications.3. The research on the effect of different conduction modes to the torque ripple is carried out. Based on the analysis of two-two conduction and three-three conduction modes, a 6 conduction mode has been presented to compensate the torque ripple caused by commutation, using leading conduction and delay conduction method. Simulation results demonstrate that this conduction mode well restrain the torque ripple.4. A network weights modification algorithm including neural network MRAC and auto adaptive control is presented combining different modern control strategies and signal analysis technologies. The simulation results show that the algorithm has the preferable varying speed signals tracking performance and the convergence speed is accelerated.5. A PWM predictive control expression is acquired by the analysis of the dynamic model ofBLDC motor. In order to solve the lag problem of conventional PI adjuster, two neural networks are used, one for the identification of the model of BLDC motor, the other for the control of motor. The comparison between PI control and predictive control simulation results shows that the predictive control is superior to the tracking of abrupt speed variation signal and time varying signal.6. A fuzzy PI control strategy of BLDC motor is presented. With reference to the neural network weights modification algorithm, the connection weights of fuzzy controller and PI controller are modified. When the error is relative higher, the weight of fuzzy control is strengthened; when the error is relative lower, the weight of PI control is strengthened. So the overshoot in the speed tracking process and the higher constant state error by using single fuzzy control problems are well solved. Due to the less sensitivity of system variables, fuzzy PI controller can achieve better performance in speed tracking system.7. The obvious current ripple of BLDC motor by using PWM control leads to the difficulty in using conventional sample and signal analysis technology in the speed tracking system. In this thesis, the wavelet analysis method is applied to the sample and processing of three -phase current, filtering out the high-frequency components and basic components being kept. Thus the PWM control of BLDC motor is realized and the current tracking goal is achieved.8. A BLDC motor for HEV experimental test-bed has been constructed. We use the DSP as the central controller and IPM modules as the driver device. The currents waveforms under half nominal load and nominal load conditions by using five different modulatio...