摘要
针对风电场输出功率预测受气象因素不确定性和异常历史数据的影响而出现的预测结果精度不高的问题,提出基于关联规则及BP(back propagation)神经网络的风电场输出功率预测方法.对异常和缺失数据进行处理,采用改进K-means聚类算法对温度/风速气象数据进行聚类分析,使用Apriori算法挖掘风电场输出功率与气象因素间的关联规则,将关联规则应用于BP神经网络.将4种方法的预测误差进行对比,结果表明:相对其他3种方法,该文方法的最大相对误差、最小相对误差、平均相对误差均最小;其最大相对误差不超过5.78%,最小相对误差仅为0.01%.因此,该文方法能提高风电场输出功率预测的准确度,具有有效性.
In order to solve the problem of low accuracy of wind farm output power prediction due to the influence of uncertainty of meteorological factors and abnormal historical data,a wind farm output power prediction method based on association rules and BP neural network was proposed.The abnormal and missing data were processed.The improved K-means clustering algorithm was used to cluster the temperature/wind speed meteorological data.The Apriori algorithm was used to mine the association rules between the wind farm output power and meteorological factors.The association rules were applied to the BP neural network.The results showed that the maximum relative error,minimum relative error and average relative error of the proposed method were all the least compared with the other three methods.The maximum relative error was no more than 5.78%,and the minimum relative error was only 0.01%.Therefore,the proposed method could improve the accuracy of wind farm output power prediction and was effective.
作者
雷蕾潇
张新燕
孙珂
LEI Leixiao;ZHANG Xinyan;SUN Ke(College of Electrical Engineering, Xinjiang University, Urumqi 830046, China)
出处
《安徽大学学报(自然科学版)》
CAS
北大核心
2021年第5期72-76,共5页
Journal of Anhui University(Natural Science Edition)
基金
国家自然科学基金资助项目(51667018)。
关键词
数据挖掘
聚类分析
关联规则
输出功率预测
data mining
clustering analysis
association rules
output power prediction