基于改进竞技粒子群算法的DFIG参数辨识

Improved Competitive Particle Swarm Optimization for Parameter Identification of DFIG

为了获得高性能的双馈感应发电机（Doubly-Fed Induction Generator, DFIG） 的控制，必须确保发电机参数的准确性。针对DFIG多参数辨识问题，提出了基于改进竞技粒子群算法的参数辨识方法。首先在两相同步旋转坐标系下将DFIG数学模型转化为DFIG参数辨识的模型。然后针对竞技粒子群算法收敛速度较慢的问题，对所引入的竞技机制进行了改进。在粒子迭代过程中，优胜的粒子仍需向个体最优和全局最优的粒子学习，从而提高了算法的收敛速度和收敛精度。最后在Matlab／Simulink中将改进竞技粒子群算法用于DFIG的参数辨识，并与粒子群算法、量子粒子群算法和竞技粒子群算法进行了对比验证。仿真结果表明改进竞技粒子群算法能提高定子电阻、定子电感、转子电阻、转子电感以及定转子互感5个参数的辨识精度。

In order to obtain the high performance control of doubly-fed induction generator （DFIG）, the accuracy of generator parameters must be ensured. In view of the problem of DFIG multiple parameters identification, an improved competitive particle swarm optimization （ICPSO） for parameter identification is proposed. Firstly, the mathematical model of DFIG is converted into the parameter identification model of DFIG at the two-phase synchronous rotating coordinate. Secondly, the competition mechanism introduced in the competitive particle swarm optimization （CPSO） is improved to overcome the slow convergence problem. In the process of particle iteration, the superior particle still needs to learn from the personal best position and the global best position. Therefore, the convergence rate and precision of the converged solution of CPSO is improved. Finally, ICPSO is compared with particle swarm optimization （PSO）, quantum-behaved particle swarm optimization （QPSO） and CPSO on the parameter identification of DFIG in Matlab/Simulink. Simulation results show that ICPSO can improve the identification accuracy of the five parameters including the stator resistance, stator inductance, rotor resistance, rotor inductance, and mutual inductance of stators and rotors.