Application Of Artificial Fish Swarm Algorithm In Parameter Identification Of Permanent Magnet Synchronous Machines

With the development of science and technology, the degree of automation and complexity of electrical equipment is increasing continuously whilst. Large complex system is applied in industrial production more and more widely. The voltage source inverter based permanent magnet synchronous machine (PMSM) drive system has been widely used in industrial control and civilian facilities. However, since the inverter utilized in actual application is not ideally linear element and contains nonlinearities such as dead time effect, voltage drop due to device, switch on/off time delay, etc., these nonlinearities will introduce waveform distortion of voltages and currents into the actual system and affect the performance of whole drive system, especially in high accuracy servo control. Therefore, it is necessary to identify and compensate these nonlinearities accurately in application. Since the voltage source inverter is part of the permanent magnet synchronous machine drive system, the identification of its nonlinearities can be regarded as part of the technology of machine parameter identification. Thus, the estimation which combines the estimation of machine parameters and the estimation of inverter nonlinearities has become one of the popular research trends in power electronics technology.In this thesis, the artificial fish swarm algorithm is employed to identify the motor winding resistance and the equivalent voltage of inverter nonlinearities simultaneously and the estimated inverter nonlinearities will be used for compensation, which enables a good control performance in drive system. Compared with the traditional compensation methods, the proposed scheme can identify both the inverter nonlinearities and motor winding resistance. In addition, the proposed estimation scheme has been applied on a permanent magnet synchronous machine drive platform and its effectiveness is verified by the comparison between the estimated parameter values and the nominal values in datasheet, which shows the estimation is of high accuracy.The main content of this thesis are as follows:(1) The permanent magnet synchronous machine and inverter nonlinearities are analyzed and modeled. The structure of permanent magnet synchronous machine and its mathematic model are investigated and the state equations including the inverter nonlinearities, in the static reference frame and rotating reference frame, are modeled, respectively. From the analysis of PMSM state equation, it is evident that the permanent magnet synchronous motor model in the rotating reference frame is similar to the model of a dc motor and a normal vector control can be used for the drive control of PMSM. In addition, based on the modeling of inverter nonlinearities, the model for identifying the stator winding resistance and inverter nonlinearities simultaneously is derived, in which the stator winding resistance consists of wire resistance and the on-state resistance of the inverter..(2)The artificial fish swarm algorithm is introduced. Through the analysis of the typical behavior of artificial fish, it shows that the fish swarm algorithm is suitable for the identification of inverter nonlinearities and PMSM parameters.(3)The equivalent voltage of inverter nonlinearities and stator winding resistance are simultaneously identified by artificial fish swarm algorithms, which is verified on a commercial PMSM drive system. From the experimental results, it shows that the identified parameter values are close to the nominal values in datasheet, which verifies the effectiveness of proposed method...