Forecasting the Distortion in Solar Radiation during Midday Hours by Analyzing Solar Radiation during Early Morning Hours

被引:2
|
作者
Noman, Abdullah M. [1 ]
Haidar, Zeyad A. [2 ,3 ]
Aljumah, Ali S. [1 ]
Almutairi, Sulaiman Z. [1 ]
Alqahtani, Mohammed H. [1 ]
机构
[1] Prince Sattam Bin Abdulaziz Univ, Coll Engn, Elect Engn Dept, Al Kharj 11942, Saudi Arabia
[2] Aden Univ, Coll Engn, Elect Engn Dept, Aden 6312, Yemen
[3] King Saud Univ, Sustainable Energy Technol Ctr, Riyadh 11421, Saudi Arabia
来源
APPLIED SCIENCES-BASEL | 2023年 / 13卷 / 10期
关键词
forecasting; PV plants; distortion index; real solar radiation data; ARTIFICIAL NEURAL-NETWORK; EMPIRICAL-MODELS; IRRADIANCE; PREDICTION; SUNSHINE;
D O I
10.3390/app13106049
中图分类号
O6 [化学];
学科分类号
0703 ;
摘要
Knowing the fluctuation of solar radiation is essential for reliable and safe operation of power systems with a high share of solar PV power plants. This paper introduces a novel method for forecasting the distortion in global solar radiation during the midday time (during the peak of solar radiation at which the generation of the PV plants is maximum), by analyzing the solar radiation during the very early hours. This paper shows that there is a strong relation between the condition of the solar radiation during the very early minutes of the day (when the generation of the PV plants is low) and the condition of the solar radiation during the midday hours (when the output of the PV plants is enormous). This relation can be utilized to help power systems operators in determining the safe share of PV plants that can be fed to power systems. We analyzed real data of one complete year for two sites in Saudi Arabia to approve this approach. The difference between each two readings was calculated for the hours 7 a.m. and 8 a.m. Then, the negative, zero, and amplitude of the difference was used to formulate a distortion index (DI) that predicts the distortion/fluctuation in the global radiation. It was concluded that the DI could classify the days into three categories: clean, distorted, and medium level according to the value of the DI. The accuracy of this approach was 85.2% and the error was 14.8%.
引用
收藏
页数:24
相关论文
共 50 条