Evaluation of PV Panel Power Loss Using Gabor Filter Bank

Abstract

Photovoltaic panels are subject to thermomechanical stresses during their production and subsequent life stages. These conditions give rise to cracks and other defects in panels that can affect power output. Cell cracking is one of the most important causes of power loss in photovoltaic panels. Therefore, photovoltaic panels and cells need to be monitored to achieve the maximum output during production and further downstream. Electroluminescent imaging is a powerful and established technique consisting of many electrically connected solar cells arranged on a grid, and it is employed in order for assessing the quality of photovoltaic panels. In this study, the detection of photovoltaic panel defects in electroluminescent images was examined through image processing methods. PV panels can consist of different numbers of cells. Performance evaluation is made on a cell-based and whole module basis. In PV panel production, unlike EL images taken in standard environments in the factory environment, EL images taken under field conditions require preprocessing before actually being processed. Features were extracted from the preprocessed EL images by exploiting Gabor filter. The obtained features were evaluated as cell-based and the stability of the cells was determined. The performance of the panel was calculated according to the power loss of the cells of the panel. When the calculated performance values were compared with the power values obtained by I-V measurement, the highest error was found to be 0.059, the lowest error was 0.004, and the average error was 0.0213. As a result, the highest success rate was 99.99%.