Forecasting Research on Long-term Solar Irradiance with An Improved Prophet Algorithm

被引：6

作者：

Yang Xinpei ^{[1
]}

Li Yiguo ^{[1
]}

Shen Jiong ^{[1
]}

机构：

[1] Southeast Univ, Natl Engn Res Ctr Power Generat Control & Safety, Sch Energy & Environm, Nanjing 210096, Peoples R China

来源：

IFAC PAPERSONLINE | 2022年 / 55卷 / 09期

基金：

中国国家自然科学基金;

关键词：

integrated energy systems (IES); solar irradiance; Planning; Prophet Algorithm; Long-term forecasting forecasting; GLOBAL HORIZONTAL IRRADIANCE; RADIATION;

D O I：

10.1016/j.ifacol.2022.07.085

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

To solve the problem of predicting solar irradiance in integrated energy systems (IES), this paper improves the Prophet algorithm by removing the trend and holidays terms, changing to the multiplicative form, and adding the monthly regressors. The improved Prophet algorithm is compared with the original and autoregressive integrated moving average (ARIMA) algorithms respectively, and the forecast accuracy is significantly improved, and the RMSE of the long-term forecast is 108.30% lower than that of ARIMA. The improved algorithm can be applied to forecast solar irradiance on long time scales with single information, providing a solution to the problem of PV array planning for IES. At the same time, the forecast data obtained by the algorithm can also be used as a reference value to optimize the day-ahead dispatch. Copyright (C) 2022 The Authors.

引用

页码：491 / 494

页数：4

共 13 条

[1]

[Anonymous], 2018, SOLAR RAD

[2]

Bolyen E., 2018, PREPRINT, DOI [DOI 10.7287/PEERJ.PREPRINTS, 10.7287/peerj.preprints.]

[3] Stan: A Probabilistic Programming Language [J].