Position Sensorless Control and Adaptive Speed Control of Permanent Magnet Synchronous Motor

Position Sensorless Control and Adaptive Speed Control of Permanent Magnet Synchronous Motor
　　Permanent Magnet Synchronous Motor (PMSM) has the advantages of small volume, light in weight, small torque ripple, simplicity in torque control, steady rotate speed, fast dynamical speed response and strong overload ability, so it is widely used in AC speed regulating system. However, the existing sensors increase the cost, reduce the reliability of the system and are difficult to mount and maintain, so it can’t be used in bad environment and high demand Neodymium magnets conditions. Therefore, position sensorless control has become an important topic of PMSM research. Moreover, because PMSM is a multi-variable, nonlinear and strong coupling system, the conventional PID controller is easily subjected to the influence of motor parameter variation and load disturbance.
　　So it is hard to satisfy the need in the situations when load variation is big and the requirement of speed accuracy and torque control precision is high.The neural network (NN) is well known for its strong nonlinear mapping capability, and can be applied to the complex speed control system. Conventional BP neural network has the disadvantages such as complex structure, slow training speed, and easily landing into local minimum points. All of these disqualify it from filling the need of control of the PMSM. However, RBF neural network has many outstanding merits, for example, quick learning speed and strong generalization ability. Satisfying control results can be obtained by combining RBF neural network with quick training arithmetic and appropriate control mode.At first, this paper establishes the dynamic simulation models of the 5-phase PMSM and its driving system based on MATLAB/SIMULINK.
　　Then, the vector control speed regulating system of PMSM is simulated and analyzed. Meanwhile, the system hardware and software design based on TMS320F2812 DSP is also introduced in the following chapter. The rotor position detection and speed regulation, as well as overcurrent and undervoltage protection are realized. And then the research on intelligent control strategies is processed on it.Firstly, this paper presents a new control method for PMSM, which is based on radial basis function (RBF) neural network. By mapping the voltages and currents of stationaryα-βreference frame to the angle and speed of rotor, the network can replace the traditional position sensors.
　　The theory in this paper is verified by http://www.chinamagnets.biz/ simulation results. Then, the adaptive speed control system for permanent magnet synchronous motor based on RBF neural network is presented. It constructs speed controller by the network that has been trained off-line and regulates network’s parameters in the course of controlling in order to adapt to the change of the environment. In addition, the system constructs another wavelet network to offer gradient parameters by on-line identification, which are used by the control network for its on-line study. The simulation results show that with this method, quick response speed, high control precision and good adaptability can be achieved.