A Novel Long Short-Term Memory (LSTM)-Based Degradation Detection in Photovoltaic Systems using Performance Ratio

Photovoltaic (PV) systems deteriorate over time as a result of prolonged exposure to UV light and harsh weather conditions. This degradation can have a major impact on the performance of the PV systems and increase their levelized cost of energy (LCOE). To deal with this problem, in this paper, a method has been proposed based on predicting The PV systems' performance ratio, which is a measure of the actual output compared to the expected output. To estimate the performance ratio, a dataset including a PV array of normal and degraded states under different environmental conditions has been created. The main model uses an architecture with long short-term memory (LSTM). The model results show high accuracy in predicting the performance ratio of systems. The model achieved the outstanding <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{R}^{2}$</tex> scores of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{9 9. 4 7 \%}$</tex> and <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{9 9. 4 6 \%}$</tex> and the root mean squared error (RMSE) values of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{0. 0 1 6 3}$</tex> and 0.0169 for the training and testing processes, respectively.