摘要
光伏发电功率预测有助于提升电网的稳定性和纳电能力,但在利用神经网络方法预测中,若输入变量和输出结果相关性差,则会导致多数据的预测模型存在预测精度差、收敛速度慢的问题,因此,提出一种基于关联规则与狼群算法优化BP神经网络(WPA-BPNN)算法的短期光伏功率预测模型.通过关联规则算法(Apriori)对光伏功率与气象因素进行关联规则挖掘确定影响光伏发电功率的关键因素,减少数据冗余,优化BP算法神经结构,再利用WPA优化BPNN初始权值和阈值,提高其拟合能力.试验表明,所提预测模型误差最高减少34%,有较好的实际应用价值.
Photovoltaic power prediction is helpful to improve the stability and capacity of power grid,If the correlation between input variables and output results is poor,it will lead to poor prediction accuracy and slow convergence speed of multi-data prediction model.Therefore,a short-term photovoltaic power prediction model based on association rules and wolf swarm optimization BP neural network(WPA-BPNN)algorithm is proposed.Through association rule mining of photovoltaic power and meteorological factors by association rule algorithm(Apriori),the key factors affecting photovoltaic power generation power are determined,data redundancy is reduced,neural structure of BP algorithm is optimized,and BPNN initial weight and threshold are optimized by WPA to improve its fitting ability.The experimental results show that the accuracy of the proposed prediction model is significantly improved and the error is smaller.
作者
谷鹏
肖建于
宋香鹏
徐成振
GU Peng;XIAO Jianyu;SONG Xiangpeng;XU Chengzhen(School of Computer Science and Technology,Huaibei Normal University,Huaibei 235000,China)
出处
《湖北民族大学学报（自然科学版）》
CAS
2020年第3期322-327,共6页
Journal of Hubei Minzu University:Natural Science Edition
基金
安徽省自然科学基金项目(1908085QF286).
关键词
狼群算法
BP神经网络
光伏发电功率
预测
关联规则
Wolf Pack Algorithm
BP neural network
photovoltaic power
prediction
association rule
