基于Transformer-LSTM及误差校正的太阳辐照度预测  

Solar irradiance prediction based on Transformer-LSTM and error correction

作　　者：唐志伟 高慧敏[2] Tang Zhiwei;Gao Huimin(School of Information Science and Engineering,Zhejiang University of Technology,Hangzhou,Zhejiang 310018,China;School of Information Science and Engineering,Jiaxing University)

机构地区：[1]浙江理工大学信息科学与工程学院,浙江杭州310018 [2]嘉兴学院信息科学与工程学院

出　　处：《计算机时代》2023年第7期123-126,132,共5页Computer Era

基　　金：嘉兴市公益性研究计划项目(2020AY10012)。

摘　　要：为提高太阳辐照度的长序列数据预测精度,提出一种基于Transformer-LSTM及误差校正的太阳辐照度预测模型。将长距离依赖学习中更有优势的Transformer模型与能够提取数据位置信息的LSTM网络结合,并且引入误差校正机制来提高模型预测精度。对三种不同模型在不同时间步长时的预测性能进行仿真实验,结果表明,Transformer-LSTM模型在太阳辐照度预测中具有更高的预测精度,并且在长序列数据预测中具有一定优势;引入误差校正机制后的仿真实验也表明了该机制的有效性。In order to improve the prediction accuracy of long series solar irradiance data,a solar irradiance prediction model based on Transformer-LSTM and error correction is proposed.The Transformer model,which is more advantageous in long-distance dependent learning,is combined with LSTM network,which can extract data location information,and an error correction mechanism is introduced to improve the prediction accuracy of the model.The prediction performance of three different models at different time steps is simulated.The results show that Transformer-LSTM model has higher prediction accuracy in solar irradiance prediction,and has certain advantages in long series data prediction.The simulation experiment after introducing the error correction mechanism also shows the effectiveness of this mechanism.

关 键 词：太阳辐照度预测 TRANSFORMER LSTM 误差校正 

分 类 号：TM615[电气工程—电力系统及自动化]