基于RBF神经网络的开关磁阻电机定子电流的研究-控制科学与工程专业论文.docxVIP

基于 RBF 神经网络的开关磁阻电机定子电流的研究 Research on Stator Current of the Switched Reluctance Motor based on RBF Neural Network ABSTRACT Switched reluctance motor drive system (SRD) is another promising new speed regulation system after AC and DC motor system. However, the non-linear characteristics of Switched Reluctance Motor (SRM) ,which is the core of SRD, especially its apparent special structure and mode of operation has led to large torque ripple and prominent noise. Optimization of Switched Reluctance Motor on structure design and novel stator current control strategies has become the focus of experts and scholars. Due to the occurance of new nonlinear control methods and the ease of control the switched reluctance motor stator current, an increasing number of new control theories are applied to optimal control of stator current. In this paper, RBF neural network was proposed to control the stator current of switched reluctance motor by comparison and analysis of several typical nonlinear control strategy used in controlling the motor current ,as well as torque-angle and flux characteristics of switched reluctance motor: in order to derived an optimal control of stator current of switched reluctance motor, RBF neural network is used for flux modeling, which is used to achieve dynamic adjust of current PWM combining with mechanical and electrical characteristics of switched reluctance motor. In addition, two-phase excitation model of switched reluctance motor is running, the mutual inductance effect is more obvious and make serious impact for the motor magnetic characteristic curve. In this paper, on the basis of analysis on the impact of electric self-inductance and mutual inductances on the current, the RBF neural network is improved on network topology to solve the problem of complex magnetic circuit and decouple. Finally, the control system of switched reluctance motor current was constructed using MATLAB simulation software based on RBF neural network to achieve the r