Application Of Articicial Immune Clonal Algorithm In Vector Control Parameter Identification

With the development of power electronics technology, various new types of voltage source inverter play important roles in motor drive and control systems. Among them, the insulated gate bipolar transistor (IGBT) has been widely used in the fields of power converter and electric drive system due to its high current density, low power cost in switch on/off, etc. However, since IGBT is a typical non-linear power device, its nonlinearity factors need to be compensated in actual application to improve the control performance of drive system. Futher, the key point for compensating the inverter nonlinearity is to obtain the IGBT and machine parameters accurately. Therefore, the modeling of inverter and corresponding model parameter identification are regarded to be important research issumes in nowaday power electronics technology.A novel strategy based on immune clonal algorithm for identifying the parameters of IGBT and motor is proposed in this paper and the inverter nonlinearity is compensated through the identified parameters. Compared with the traditional compensation methods, the proposed strategy can intelligently identify the equivalent voltage caused by inverter nonlinearity and motor winding resistance, respectively. The proposed strategy is employed in a permanent magnet synchronous machine drive system and its accuracy is verified through the comparison between the identified results and the parameter values listed in the inverter datasheet.The main research work in this paper focuses on the following aspects:Firstly, the Artificial Immune Clone Selection Algorithm is introduced. Artificial Immune Clone Selection Algorithm is a novel algorithm which was inspired by the biological immune system. And it has become a new member of the family of intelligent algorithms. Immune clonal selection algorithm has evolved from the standard clonal selection algorithm to the senior clonal selection algorithm, namely adaptive dynamic clonal selection algorithm. Comparing the two kinds of clonal selection algorithm, the adaptive dynamic clonal selection algorithm based on the standard clonal selection algorithm divided the antibody into memory cell and ordinary cell dynamically, which have different mutation probability and the latter is much greater than the former. The mutation probability and mutation probability of the normal cells is much greater than memory cells, so the algorithm can guarantee inheritance parent can ensure that the diversity of their antibodies.Secondly, the mathematical model of permanent magnet synchronous motor and the non-linear factor model of IGBT are analyzed. The structure and permanent magnet synchronous motor model is presented, and then the mathematical model and state equation are structured based on the stationary and rotating coordinate system. Through mathematical models and equations of state can see, rotating reference frame permanent magnet synchronous motor model and DC model is very similar to vector control strategies can be used on the permanent magnet synchronous motor control. In addition, the IGBT inverter to model the nonlinear factors, the model derived by the equivalent motor stator winding resistance (the stator winding resistance and equivalent resistance only and inverter), and non inverter linear factors in the amount equivalent voltage.Finally, this thesis distinguishes between nonlinear equivalent voltage IGBT volume and motor winding resistance using clonal selection algorithm, and use of a commercial permanent magnet synchronous motor drive system on the performance of the algorithm was validated. With the experimental data analysised, the experimental identification of parameters and the device out of manual parameter values are very consistent, which proves the clonal selection algorithm on the drive system parameter identification is accurate.