论文部分内容阅读
针对标准遗传算法在励磁系统参数辨识时存在的收敛速度慢、易早熟的缺点,本文提出将混沌遗传算法应用到励磁系统参数的辨识。该算法克服了其缺点并优化了其算法。文中在两个方面将混沌序列与标准遗传算法相结合,一方面是用混沌序列对辨识参数进行初始化,这样可以保证初始参数的遍历性;另一方面是在种群交叉和变异后,用混沌序列重构新的个体,这样可以防止辨识参数早熟而得到局部最优解。为了验证提出方法的有效性,分别采用标准遗传算法和混沌遗传算法对MEC3300T型励磁系统的参数进行辨识。结果证明,基于混沌遗传算法的励磁系统参数辨识方法计算效率和精度均高于标准遗传算法,所得到的辨识模型能够更好地反映出实际励磁系统的物理特性,也能够更好地应用于电力系统分析研究。
Aiming at the shortcomings of standard genetic algorithm, such as slow convergence and premature convergence in the identification of excitation system parameters, this paper proposes to apply chaos genetic algorithm to the identification of excitation system parameters. The algorithm overcomes its shortcomings and optimizes its algorithm. In this paper, chaos sequences are combined with standard genetic algorithm in two aspects. On the one hand, chaotic sequences are used to initialize the identification parameters so as to ensure the ergodicity of the initial parameters; on the other hand, after the crossover and mutation of the populations, chaotic sequences Reconstruction of new individuals, so as to prevent premature identification parameters and get the local optimal solution. In order to verify the effectiveness of the proposed method, the parameters of MEC3300T excitation system were identified by using standard genetic algorithm and chaos genetic algorithm respectively. The results show that the efficiency and accuracy of the identification method of excitation system based on chaos genetic algorithm are higher than those of standard genetic algorithm. The obtained identification model can better reflect the physical characteristics of the actual excitation system and can also be better applied to electric power System Analysis and Research.