Estimation of solar radiation and photovoltaic power potential of Türkiye using ANFIS

Abstract

Abstract In this study, we estimated global horizontal irradiance (GHI) and photovoltaic (PV) power potential in Türkiye using fan Adaptive Neuro-Fuzzy Inference System (ANFIS) with inputs from Meteonorm 8.0. We considered latitude and longitude for GHI estimation and added tilt and azimuth angles for PV power potential. Training and testing data yielded Root Mean Square Errors (RMSE) of 10.429 kWh/m 2 /year and 35.395 kWh/m 2 /year for GHI, and 3.984 kWh/kWp/year and 39.624 kWh/kWp/year for PV power potential, respectively. The coefficient of determination ( $${R}^{2}$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:msup> <mml:mrow> <mml:mi>R</mml:mi> </mml:mrow> <mml:mn>2</mml:mn> </mml:msup> </mml:math> ) was 0.995 for GHI training data, and $${R}^{2}$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:msup> <mml:mrow> <mml:mi>R</mml:mi> </mml:mrow> <mml:mn>2</mml:mn> </mml:msup> </mml:math> = 0.946 for GHI testing data, and $${R}^{2}$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:msup> <mml:mrow> <mml:mi>R</mml:mi> </mml:mrow> <mml:mn>2</mml:mn> </mml:msup> </mml:math> = 0.993 for PV power potential training data, and $${R}^{2}$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:msup> <mml:mrow> <mml:mi>R</mml:mi> </mml:mrow> <mml:mn>2</mml:mn> </mml:msup> </mml:math> = 0.951 for PV power potential testing data. Results showed that the southeast of Türkiye had the highest GHI, around 1900 kWh/m 2 /year, while the north had the lowest, around 1400 kWh/m 2 /year. The highest electricity production occurred in the south, approximately 1900 kWh/kWp/year, while the north yielded the least, about 1450 kWh/kWp/year. Tilt angle increase and azimuth angles of -90° or + 90° led to decreased electricity production. This study provides precise GHI and PV power potential estimates for Türkiye, using ANFIS with diverse meteorological inputs.

Topics & Concepts