摘要
在光伏发电预测中,一般都采用与发电功率正相关因素作为输入变量,这样操作很容易陷入局部最优。文章改变输入变量的选取范围,选取与光伏发电功率正、负相关性较大的因素作为光伏发电预测模型的输入变量,利用模糊系统具有收敛速度较快和神经网络具有自学习和调整参数容易等优点,提出Takagi-Sugeno模糊神经网络模型应用于光伏发电功率短期预测中,并与BP神经网络预测进行比较,其结果显示,所述预测模型预测精度比BP神经网络预测精度提高了10%。
In photovoltaic power forecasting,the positive correlation factors with power are generally used as input variables,which makes the operation easy to fall into local optimum.In this paper,the selection range of input variables is changed,and the factors that have greater positive and negative correlation with photovoltaic power are selected as input variables of photovoltaic power generation prediction model.Fuzzy system has the advantages of fast convergence speed,easy self-learning and parameter adjustment.Takagi-Sugeno fuzzy neural network model is proposed for short-term prediction of photovoltaic power,and compared with BP neural network prediction.The results show that the predictian accuracy of the forecasting model is 10% higher than BP neural network.
作者
文立
WEN Li(Hunan Vocational Institute of Technology,Xiangtan Hunan 411104,China;Hunan Provincial Photovoltaic Power Generation System Control and Optimization Engineering Laboratory,Xiangtan Hunan 411104,China))
出处
《智能计算机与应用》
2019年第3期118-121,125,共5页
Intelligent Computer and Applications
基金
湖南省教育厅资助科研项目(17C0739)
关键词
正、负相关因素
功率预测
模糊神经网络
减法聚类
positive and negative related factors
power prediction
fuzzy neural network
subtraction clustering
