基于径向基神经网络的输电线路动态容量在线预测  

Online Prediction of Transmission Dynamic Line Rating Based on Radial Basis Function Neural Network

作　　者：王孔森 盛戈皞[2] 孙旭日 王威[4] 王世强 江秀臣[2] 

机构地区：[1]国网物资有限公司,北京市西城区100120 [2]上海交通大学电气工程系,上海市闵行区200240 [3]聊城供电公司,山东省聊城市252000 [4]国电南瑞科技股份有限公司,江苏省南京市210061

出　　处：《电网技术》2013年第6期1719-1725,共7页Power System Technology

基　　金：国家863高技术基金项目(SS2012AA050803);国家自然科学基金项目(50977057)~~

摘　　要：在线预测输电线路的动态热容量,合理安排负荷高峰时期运行方式和调度管理,对输电线路的安全和经济运行具有重要意义,同时也对确定风电等间歇式可再生能源的接入容量提供技术支持。为此,提出了利用径向基神经网络实现线路动态容量的在线预测方法。该方法首先利用径向基神经网络进行风速和日照辐射温度的在线学习和预测,基于IEEE 738标准进行输电线路动态容量的在线短期预测。利用典型的夏季和冬季实测数据进行动态容量预测后发现,预测未来1、2、4 h的动态容量的最大相对误差分别在10%、20%、40%以内。将短期的负荷预测与该方法结合起来,可为电力紧张地区和负荷高峰时期的智能调度提供决策支持。It is significant for secure and economic operation of transmission lines to predict dynamic line rating （DLR） in realtime mode and reasonably arrange the operation mode during peak load period and scheduling management, meanwhile it can provide technical support for determining grid-connected capacity of intermittent renewable energy resources such as wind power and so on. For this reason, using redial base function neural network （RBFNN） an online prediction method to implement the prediction of DLR is proposed. Firstly, the online learning and prediction of wind and sunshine radiation temperature is performed by RBFNN, then based on IEEE 738 standard the online short-term DLR is predicted. In the prediction of DLR by use of typical measured data in summer season and winter season, it is found that the maximum relative error in the prediction of dynamic capacity for next one, two and four hours are less than 10%, 20% and 40% respectively, Combining the short-term load forecasting method with the proposed method, the decision-making can be provided to intelligent dispatching during the peak load period in the region lacking of power supply.

关 键 词：输电线路 动态容量 径向基神经网络 在线预测 

分 类 号：TM726.3[电气工程—电力系统及自动化]