摘要
为进一步提高光伏/光热(photovoltaic/thermal,PV/T)综合利用系统中PV/T组件温度预测精度,该文采用主成分分析法对原始输入样本数据进行预处理,提取主成分,并利用遗传算法优化BP神经网络结构,对组件温度数据建立预测模型。仿真结果表明,相对于未经主成分提取的神经网络,该方法使得网络模型在晴天和多云天气条件下的预测精度分别提高了7.68%和4.97%,使得网络模型预测精度更高,泛化性能更强。
In order to further improve the prediction accuracy of PV/T module temperature in photovoltaic-thermal( PV/T) system,a principal component analysis method is used to preprocess the original input sample data to extract the main component. Then combined with the BP neural network theory and Genetic optimization algorithm,the prediction model of component temperature data is built. The simulation results show that the proposed method makes the prediction accuracy of the network model higher while the generalization performance stronger than that of the neural network without the main component extraction. The prediction accuracy is increased by 7. 68% and4. 97% in sunny and cloudy weather.
引文
[1]肖文波,吴华明,傅建平,等.光强和温度对硅光伏电池输出特性的影响[J].华中科技大学学报(自然科学版),2017,45(1):108-112.
[2]李光明,刘祖明,李景天,等.新型PV/T太阳能利用复合系统的实验研究[J].中国电机工程学报,2013,33(17):83-89.
[3]KHELIFA A,TOUAFEK K,MOUSSA H,et al.Modeling and detailed study of hybrid photovoltaic thermal(PV/T)solar collector[J].Solar Energy,2016,135(3):169-176.
[4]YANG T,ATHIENITIS A.A review of research and developments of building-integrated photovoltaic/thermal(BIPV/T)systems[J].Renewable&Sustainable Energy Reviews,2016,66(8):886-912.
[5]HASSANI S,TAYLOR R,MEKHILEF S.A cascade nanofluid-based PV/T system with optimized optical and thermal properties[J].Energy,2016,112(1):963-975.
[6]龚莺飞,鲁宗相,乔颖,等.光伏功率预测技术[J].电力系统自动化,2016,40(4):140-151.
[7]潘进军,申彦波,边泽强,等.气象要素对太阳能电池板温度的影响[J].应用气象学报,2014,25(2):150-157.
[8]丁坤,刘振飞,高列,等.基于主成分分析和马氏距离的光伏系统健康状态研究[J].可再生能源,2017,35(1):1-7.
[9]刘俊,王旭,郝旭东,等.基于多维气象数据和PCA-BP神经网络的光伏发电功率预测[J].电网与清洁能源,2017,33(1):122-129.
[10]李芬,宋启军,蔡涛,等.基于PCA-BPNN的并网光伏电站发电量预测模型研究[J].可再生能源,2017,35(5):689-695.
[11]李芬,刘迪,胡超,等.基于PCA-LMBP神经网络的北京地区直散分离预测[J].水电能源科学,2017,35(4):208-212.
[12]杜子芳.多元统计分析[M].北京:清华大学出版社,2016:239-269.
[13]张云龙,袁建华,吴仕玄,等.基于GA-PSO的分布式光伏发电最大功率跟踪控制[J].水电能源科学,2017,35(1):212-214,84.
[14]刘沛汉,袁铁江,梅生伟,等.基于遗传算法优化神经网络的光伏电站短期功率预测[J].水电能源科学,2016,34(1):211-214.
[15]王江荣,赵睿,袁维红,等.基于遗传算法因素筛选的BP神经网络在软土路基沉降数据处理中的应用[J].矿山测量,2016,44(5):87-90.
[16]张敬尧,潘雅清,陈彦瑞,等.基于Matlab遗传算法工具箱的同心式磁力齿轮优化设计[J].机械传动,2015,39(6):135-138.
[17]王勃,冯双磊,刘纯.基于天气分型的风电功率预测方法[J].电网技术,2014,38(1):93-98.