A Optimal Design For Vector Control Of PMSM Based On Improved Genetic Algorithm

PMSM has been widely used in industrial ac servo, because of its characteristic of small volume, simple structure, high power factor, energy saving and high reliability, However, the application of PMSM depends on its control technology, therefore, the research of PMSM has both theoretical significance and practical application value.There are a lot of control strategy for PMSM, the vector control is one of the most popular control method. There are three PI controller in the vector control system, the three PI controller parameters will directly affect the performance of this system. In order to find the optimal parameters of PI controller, the optimization problem can be regarded as to solve a problem which has multivariable, nonlinear, multi-peak and multi performance index.Genetic algorithm as a kind of evolutionary computation technology, is often used to solve many complex optimization problems in the field of application. Especially on the controller parameter optimization problem, it can achieve a good optimization effect, therefore, this article design the PMSM vector control system based on the genetic algorithm. However using common genetic algorithm to optimize this problem, often exists the phenomenon of premature convergence and slow convergence speed. Therefore, this paper introduces an improved genetic algorithm which is improved mainly includes:select operation using fitness value and a combination of small group competition, in the choice to join the elite strategy at the same time, the initial population is not only to maintain optimal can keep its diversity; Crossover and mutation probability of crossover and mutation in the operation and the way to improve, make relevant crossover probability and evolution algebra, mutation probability is related to individual fitness and population can better and faster to get the global optimal solution.In order to proving the result in this paper, linking the Mat files and SimPowerSystem toolbox in Matlab, the results show that the improved genetic algorithm is compared with the simple genetic algorithm has the optimal time is short, fast convergence speed and global optimization ability, etc.