山东科学 ›› 2023, Vol. 36 ›› Issue (1): 115-123.doi: 10.3976/j.issn.1002-4026.2023.01.015

• 能源与动力 • 上一篇    下一篇

基于时间序列疏系数模型的太阳辐射年际变化趋势预测

贾兴斌(), 宫响*()   

  1. 青岛科技大学 数理学院,山东 青岛 266061
  • 收稿日期:2022-02-16 出版日期:2023-02-20 发布日期:2023-02-08
  • 通信作者: *宫响(1977—),女,博士,副教授,研究方向为海洋大气大数据分析。Tel: 13210086655, E-mail:gongxiang@qust.edu.cn
  • 作者简介:贾兴斌(1995—),男,硕士研究生,研究方向为海洋大气大数据。E-mail:m18409351359@163.com
  • 基金资助:
    国家自然科学基金-山东省联合基金(U1906215)

Predicting interannual variation of global solar radiation trends in Jinan City based on time series sparse coefficient model

JIA Xingbin(), GONG Xiang*()   

  1. School of Mathematics and Science, Qingdao University of Science and Technology, Qingdao 266010, China
  • Received:2022-02-16 Online:2023-02-20 Published:2023-02-08

摘要:

利用1961—2016年山东省济南市太阳年总辐射量观测数据,通过模型识别和统计检验,对比分析时间序列模型AR(5)和ARIMA((1,2,4),1,0)的拟合结果。残差检验结果表明,疏系数模型ARIMA ((1,2,4),1,0)可用于预测地表太阳年总辐射量,预测结果显示2017—2025年济南市地表太阳辐射的年际变化整体呈增长趋势。对比多元线性回归模型结果,时间序列疏系数模型误差较小,预测准确度相对较高。

关键词: 太阳年总辐射量, 时间序列分析, ARIMA疏系数模型, 年际变化, 趋势预测, 模型对比

Abstract:

In this paper, we have used the observed data of annual total solar radiation from 1961 to 2016 in Jinan, Shandong Province, and compared and analyzed the fitting results of time series models AR(5) and ARIMA((1,2,4),1,0) via model identification and statistical tests. As per the residual test results, the sparse coefficient model ARIMA((1,2,4),1,0) can be used to predict the total annual surface solar radiation. The prediction results show that the overall interannual variation of surface solar radiation in Jinan from 2017 to 2025 follows an increasing trend and the utilization of solar energy resources can be further explored. Compared to the results of the multiple linear regression model, the time series sparse coefficient model has less error and higher prediction accuracy.

Key words: total annual solar radiation, time series analysis, ARIMA sparsity coefficient model, interannual variability, trend prediction, model comparison

中图分类号: 

  • TK511