loading page

Prediction of Poly-Crystalline PV Performance by Machine Learning and Analytical Modeling in Jordan
  • Sinan S. Faouri,
  • Salah Abdallah,
  • Dana Helmi Salameh
Sinan S. Faouri
Applied Science Private University

Corresponding Author:[email protected]

Author Profile
Salah Abdallah
Applied Science Private University
Author Profile
Dana Helmi Salameh
Applied Science Private University
Author Profile

Abstract

In this work, different prediction models of poly-crystalline PV performance in Jordan are studied based on experimental results, analytical, and machine learning models. Two types of PV grid connected systems working at Applied Science Private University in Amman-Jordan, having a power capacity of 5 KWp each were involved in study. These two PV systems are fixed with tilt angle of 11 which are: poly-crystalline south directed, and poly-crystalline EW directed (which means that half of these PV modules directed to the east and the other half is directed to the west). Both PV systems have been trained via linear regression, LASSO, ElasticNet in addition to ANN algorithms. Linear regression model surpasses all the other three models when compared by different performance measures. The results for poly-crystalline EW directed PV system show that the yearly analytical electrical power generation is 1433.9 KWh/KWp, where the value of error equals 3.12% as compared with experimental value. Whereas the yearly electrical power generation by prediction using linear regression is 1510.45 KWh/KWp, where the value of error equals 2.1% as compared with experimental value. In the case of poly-crystalline south directed PV system, the yearly analytical electrical power generation is 1772.9 KWh/KWp, where the value of error equals 7.89% as compared with experimental value. Whereas the yearly electrical power generation by prediction using linear regression is 1658.15 KWh/KWp, where the value of error equals 1.5%. Also, the results show that the productivity of poly-crystalline south directed PV system is better than the poly-crystalline EW directed system with power gain equals 23.64% using analytical method, equals 10.43% using experimental method, and equals 9.77% using prediction by linear regression.