Perhitungan Intensitas Radiasi Matahari Berdasarkan Pola Sebaran Awan Menggunakan Metode Support Vector Regression (svr)

Abstract

Abstrak
Intensitas radiasi matahari yang diterima oleh permukaan bumi dapat diketahui melalui lintasan
matahari. Tingkat intensitas radiasi matahari dipengaruhi oleh banyak faktor, yang terpenting adalah
posisi, pola, serta sebaran awan. Penelitian ini menganalisis hubungan antara awan dengan intensitas
radiasi matahari menggunakan metode Support Vector Regression (SVR). Data awan diperoleh dari
METARs dan data intesitas radiasi matahari dari PySolar dan University of Oregon. Hasil perhitungan
model menunjukan nilai koefisien determinasi (R²) yang dihasilkan oleh model perhitungan adalah
sebesar 0,80022, dimana model mampu menghitung nilai global solar pada kondisi clear sky dan cloudy
sky dengan nilai persentase error dinyatakan dalam NMBE sebesar 10,38 %, serta CVRMSE sebesar
21,03%. Data hasil penelitian ini dapat diperlukan untuk membuat desain bangunan agar didapat kondisi
termal yang baik.

Kata kunci: machine learning, intensitas radiasi matahari, awan, support vector regression (SVR)

Abstract
The intensity of solar radiation received by the surface of the earth can be known through the path of the
sun. The level of radiation intensity is influenced by many factors, the most important is the potition,
pattern, and distributon of clouds. This research analyzes the relationship between clouds and the
intensity of solar radiation using the Support Vector Regression (SVR) method. Cloud data were obtained
from METARs and solar radiation intensity data from PySolar and the University of Oregon. The model
calculation results show the coefficient of determination (R²) generated by the calculation model is
0.80022, where the model is able to calculate the global solar value in clear sky and cloudy sky conditions
with the percentage error value expressed in NMBE of 10.38%, and CVRMSE of 21.03%. The data from
the results of this study are needed to create a building design to obtain good thermal conditions.

Keywords: machine learning, radiation intensity, cloud, support vector regression (SVR)