针对现有无功分区中缺乏分区结果定量评价标准的问题,提出一种可以对分区结果进行量化评估的大电网双阶段无功电压分区方法。基于对数电压灵敏度计算各母线间的电气距离,利用K-均值对互联电网进行初始分区。提出表征无功电压分区标准的指标体系,通过权重设置构建适用于多目标优化的适应度函数,满足不同运行方式下的量化评估要求,最后,利用改进遗传算法求解帕累托最优分区。IEEE118节点算例和某实际系统算例表明:所提出的双阶段无功电压分区方法在不同运行方式下都具有较好的分区效果,且计算结果可以量化评估,具有理论和工程实用价值。
In view of few quantitative evaluation standards on reactive power partition results, this paper proposed a dual-stage partitioning method to evaluate reactive power and voltage partition results quantitatively. Electric distance was calculated based on the logarithmic voltage magnitude sensitivity and then the power system was partitioned originally by utilizing the K-mean cluster analysis. The index system was established to evaluate the partitioning standards. The fitness fimction applied for multi-objective optimization was constructed to meet the quantitative assessment requirements of different operation schemes. The improved genetic algorithms was used to seek the Pareto optimal result. Examples of IEEE 118-bus and a practical area network show that the partitioning method is effective for different operation schemes, the partitioning results can be quantitatively evaluated, and the method has significance in theory sense and engineering practice.