基于深度模糊神经网络的太阳总辐射预测研究
RESEARCH ON GLOBAL SOLAR RADIATION FORECAST BASED ON DEEP FUZZY NEURAL NETWORK
乔楠 1蒋波涛 1郑雨 1刘燕东 1王锦1
作者信息
- 1. 西安工程大学电子信息学院,西安 710600;西安市电气设备互联感知与智能诊断重点实验室,西安 710600
- 折叠
摘要
提出一种基于深度模糊神经网络的太阳总辐射预测模型.首先利用Pearson相关系数分析太阳总辐射关键影响因素,其次利用深度学习多隐含层所具有的特征提取优势将模糊神经网络模块重复连接,构建深度模糊神经网络模型,并使用蝗虫优化算法对其中心值和宽度进行优化.利用所提太阳总辐射预测模型对5个气象站点的相关数据进行仿真实验,并对结果进行分析.仿真结果表明:所提预测模型较其他模型具有较高的预测精度,验证了模型的有效性,可满足无辐射监测站点太阳总辐射预测的需要.
Abstract
This paper proposes a global solar radiation forecast model based on deep fuzzy neural network.Firstly,Pearson correlation coefficient is used to analyze key influence factors of global solar radiation.Then,the fuzzy neural network modules are repeatedly connected to construct a deep fuzzy neural network model by using the feature extraction advantage of deep learning multiple hidden layers.Moreover,the width and center value of the membership function in this model are optimized by the grasshopper optimization algorithm.Finally,simulation experiments are conducted by using the proposed global solar radiation forecast model based on related data of five meteorological sites.The simulation results show that the proposed model has higher forecast accuracy than other models,and verifies the validity of the model,which meets the requirements of global solar radiation forecast at some sites without radiation monitoring.
关键词
太阳能/太阳辐射/预测/深度模糊神经网络/蝗虫优化算法Key words
solar energy/solar radiation/forecasting/deep fuzzy neural network/grasshopper optimization algorithm引用本文复制引用
基金项目
国家自然科学基金青年项目(11705135)
陕西省自然科学基础研究计划项目(2020JM-573)
西安工程大学博士科研启动基金(BS1339)
出版年
2024
NETL
NSTL
维普
万方数据