基于遗传算法的电力系统自适应卡尔曼滤波动态状态估计

摘要

针对卡尔曼滤波动态状态估计中 Holts′两参数均为常数，在电力系统运行状态变化时易产生较大的预测误差的不足，提出采用指数平滑法对参数进行动态调整。该方法在预测步中利用遗传算法来动态确定参数大小，实现了预测参数的自适应优化。最后，对 IEEE 14节点系统进行仿真计算，与传统方法进行比较，结果表明本文方法具有明显的优势。%Aiming at problem of Kalman filter dynamic state estimation of which both Holts′parameters are constants and may produce bigger forecast error when power system running state changing,this paper proposes to use exponential smoot-hing method to carry on dynamic adjustment for parameters.By applying genetic algorithm in forecast steps,this method was able to dynamically confirm parameter size and realize self-adaptive optimization of forecast parameters.At last,simulation calculation on IEEE 1 4 node system was conducted.Compared with traditional method,the result indicated that this method was provided with obvious advantages.