改进K近邻算法在风功率预测及风水协同运行中的应用  

Wind Power Prediction Based on Improved K-Nearest Neighbor Algorithm and Its Application in Co-Operation of Wind and Hydro Powers

作　　者：杨秀媛[1] 裘微江 金鑫城 陈勇 邹卫美 郑志伟 郭中华 秦泽阳 

机构地区：[1]北京信息科技大学自动化学院,北京市海淀区100192 [2]电网安全与节能国家重点实验室(中国电力科学研究院有限公司),北京市海淀区100192 [3]国网北京电力工程公司,北京市丰台区100070

出　　处：《电网技术》2018年第3期772-778,共7页Power System Technology

基　　金：国家自然科学基金项目(51377011);中国电力科学研究院有限公司电网安全与节能国家重点实验室开放基金项目~~

摘　　要：风电输出功率的不确定性和不可控性成为了制约风电发展的根本问题。研究风功率预测技术,为电网运行提供准确的风电输出功率预测数据和信息,是解决风力发电发展的根本途径。提出基于改进K最邻近算法的风功率预测模型,并将模型应用到了风水协同运行中,在风水协同运行计划的基础上增加了数据实时修正。通过Python语言实现仿真,通过实际仿真结果表明该方法具有较好的预测精度,提高了协同运行系统的精度和准确性。验证了该方法的有效性。Uncertainty and uncontrollability of wind power output become a fundamental problem restricting wind power development.It is the fundamental way to solve this problem by studying wind power forecasting technology and providing accurate wind power output data and information for power grid operation.In this paper,a wind power prediction model based on improved K-nearest neighbor algorithm is proposed and applied to co-operation of wind and hydro powers based on collaborative operation plan of wind power and hydropower.Real-time correction is added.Simulation results shows that the method has better prediction accuracy and improves accuracy of collaborative operation system.

关 键 词：PYTHON 风功率预测 改进K近邻算法 风水协同 实时修正 

分 类 号：TM721[电气工程—电力系统及自动化]