改进遗传算法和神经网络在电能质量扰动识别中的应用

Improved genetic algorithm and neural network's application in the recognition of power quality disturbances

提出了一种改进遗传算法(IGA)。该算法采取了有效的选择和交叉策略,设计了自适应变化的交叉概率和变异概率,以及随进化代数自适应变化的移民算法。该改进遗传算法稳定性和收敛性能较好,将其应用到基于BP网络的电能质量扰动识别中,提高了识别的稳定性和收敛性能,并使基于BP网络的电能质量扰动识别系统更具实用性。文中最后通过仿真计算验证了本文方法的性能。

A new Improved Genetic Algorithm (IGA) was proposed in this paper. In the IGA, the effective selection and cross strategy was adopted, and the new adaptive probability algorithm of crossover depending on the number generations, and a new adaptive probability algorithm of mutation depending on the fitness value were designed. Moreover, a crossover strategy for preventing inbreeding is also proposed. All these methods can help to enhance the capability of the genetic algorithm. IGA is also used in the recognition system based on the BP neural network, to improve its stability and convergence capability. Finally, the validity of the proposed method is verified by the results of the simulation.