Parameter Identification Of Permanent Magnet Synchronous Motor Based On Genetic Algorithm

With the progress of society,the performance requirements of motor in various fields have been further improved.Reliability and energy saving are the development trend of motor control in the future.However,previous control strategies are mostly based on linear theory.It is difficult to maintain the controller performance for a long time under the condition that the motor is a high-order,non-linear and time-varying system.In order to inherit the theoretical research achievements of the predecessors and control the motor better,parameter identification becomes indispensable.In order to illustrate the importance of parameter identification to the mainstream control methods,on the basis of the analysis of the mathematical model of permanent magnet synchronous motor,this thesis studies the process of changing the internal state of the motor by space vector control and direct torque control,and draws the conclusion that the change of parameters affects the control accuracy.So the problem of under-rank of parameter identification equation and its solution are analyzed.Finally,the full-rank equations are established by searching two unknowns which lead to the inferior rank:of the identification equations with genetic algorithm,.and the four parameters identified are solved.In order to get rid of the influence of the nonlinearity of the converter on the accuracy of the identification results,the state equation of the motor is transformed equivalently,and the differential types are eliminated,the identification process can be sampled directly from the motor,thus improving the accuracy of the sampling.Different from the previous methods of judging the quality of identification results,the results are not directly compared with the motor design values,but based on the control requirements,through the construction of evaluation function,to achieve the scientific evaluation of the parameters identification results.The results of simulation and experiment prove that the identification method can identify the parameters of the motor quickly and effectively.This thesis applied the difference between the old and new methods in theory and practice.In theory,based on the population Markov chain model of genetic algorithm,the reason that the new method is better in accuracy and speed is qualitatively analyzed.