摘要
提出了一种基于区域划分的500 kV区域电网动态无功电压优化控制策略.该策略首先通过模糊聚类方法分析各节点母线电压之间的依存度,确定相互之间的影响因子,以此作为分区控制的基础.然后建立基于人工神经网络的智能母线电压预测模型,预测各节点电压的越限情况,以此实现无功设备投切的事前控制.最后建立基于全网的动态无功电压智能优化控制模型,综合考虑分区内各节点电压的分布情况、影响因子、无功设备投切次数,选取最优调节节点,确定无功设备的投切数量,最终实现500 kV区域电网的无功电压优化控制.采用IEEE 14模型,应用该策略进行仿真试验,结果表明该策略有效地预测了各节点母线电压的越限情况,并提出了最优的控制方案.
A 500 kV regional dynamic reactive-power voltage optimal control strategy based on zoning is proposed. First, using fuzzy clustering analysis, the dependence of the bus voltages on each node is analyzed, and the mutual influence factor is determined as basis for the zoning control. Second, a smart bus-voltage prediction model is built based on the artificial neural network, which predicts the off-limit circumstances of each node, to achieve control of the reactive device switching beforehand. Finally, the smart optimal control model of a dynamic reactive-power voltage is established based on the entire network by considering the distribution, effect factors, and reactive device switching times of each node voltage in a sub-region. The optimal regulating node is pinpointed to determine the reactive device switching number; ultimately, optimal control of the reac- tive-power voltage of a 500 kV regional power grid is achieved. This strategy is applied to the simulation test under the IEEE 14 model, which indicates that the bus voltage off-limit circumstances of each node are effec- tively predicted, and the optimal control approach is proposed.
出处
《信息与控制》
CSCD
北大核心
2014年第1期88-95,共8页
Information and Control
关键词
无功优化
分区控制
电压预测
遗传算法
模糊聚类分析
人工神经网络
reactive power optimizationzoning-control
voltage prediction
genetic algorithm
fuzzy clustering analysis
artificial neural network