节能发电调度中的有功安全校正策略  

Correction Strategy of Active Power Security in Energy Conservation Generation Dispatching

作　　者：朱足君[1] 李扬[1] 陈文浩 杨梓俊 姚颖蓓[1] 

机构地区：[1]东南大学电气工程学院,南京210096 [2]江苏电力调度通信中心,南京210024

出　　处：《电力系统及其自动化学报》2008年第6期47-50,共4页Proceedings of the CSU-EPSA

基　　金：江苏省电力公司节能发电调度策略研究(J2007011)

摘　　要：有功安全校正是节能发电调度安全控制的一项重要内容,调整量越少则发电计划偏离节能目标越少。文中对直流灵敏度法中灵敏度取值方法作了改进。第一次调整使用基于直流潮流的灵敏度值,下一次调整使用精确的灵敏度值。用线路有功变化量除以发电机调整量得到实际灵敏度值。将记录的线路有功变化量和实际灵敏度值作为一个三层的反向传播(BP)神经网络的训练用样本集合。训练完成后对应某过载量可求得精确的灵敏度值。对IEEE14节点系统计算的结果表明,采用该方法可以充分使用历史数据,精确地进行有功安全校正。Active power security correction is an important part of security control methods in energy conservation generation dispatching. The power generation plan can be less changed from energy conservation target if the adjustment is little. In this paper, the method based on DC sensitivity has been improved on how to get the sensitivity value. The sensitivity value based on DC power flow is used in the first adjustment, and the accurate sensitivity value is used in the next adiustment. The actual sensitivity value can be got by dividing active power change of lines by adjustment of generators. A three-tier backpropagation（BP） neural network is trained by samples of active power change of lines and actual sensitivity value in recoads. After training, accurate sensitivity value which is corresponding to overload in some cases can be got. The calculating results on IEEE 14 buses system show that this method can make full use of historical data,and make active power security correction more accurate.

关 键 词：有功安全校正 反向传播神经网络 历史数据 精确 

分 类 号：TM731[电气工程—电力系统及自动化]