摘要
将遗传算法应用于电力系统综合负荷建模.以三阶感应电动机为综合负荷模型,以待辨识参数为未知向量,以系统实测与模型响应误差平方和为目标函数;以随机初始种群为基础进行交叉—变异—选择运算并产生下一代种群;通过若干代进化即可获得具有足够精度的辨识结果.通过实验数据将遗传算法与传统模式搜索算法的建模结果比较,表明遗传算法所得模型的描述精度比模式搜索法高10倍,其模型参数呈现很好的稳健性,从而有效地克服了传统优化方法的模型参数分散性.
Genetic algorithm was applied to the comprehensive load modeling of the power system. To be specific, this paper used the 3-order induction motor as the comprehensive load model, the parameters to be identified as the unknown vector, and the square sum of errors between the system measurement data and the model response as the objective function. Under the condition of random original colony and by crossover-mutation-selection operation, we obtained the next era colony and acquired satisfactory identification result after some eras' evolution. The identification results showed that the calculation precision based on genetic algorithm improved 10 times compared with the mode-detection method, and at the same time, the model parameters were more stable, thus effectively overcoming the diffusiveness of the model parameters of traditional optimization methods.
出处
《湖南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2005年第2期29-32,共4页
Journal of Hunan University:Natural Sciences
基金
高等学校骨干教师资助计划(教技司[2002]65号)
湖南省教育厅重点资助项目(湘教通[2001]197号)
关键词
遗传算法
电力系统
综合负荷
负荷建模
参数辨识
genetic algorithms
power system
aggregate load
load modeling
parameter identification
