Comparative Analysis of Methods for Estimating the Average Photon Energy: the Case Study of a Sunny Site
Year: 2014
Type: article
Abstract: The average photon energy is a useful and meaningful index which many have assumed to be unique to a given solar spectral distribution. Many contributions in literature use this factor to model the impact of the solar spectrum on thin film photovoltaic modules. However, spectral irradiance measurements needed to obtain this index have been scarcely addressed in meteorological stations worldwide. This work is mainly aimed at calculating the average photon energy by means of different methods using only astronomical and geographical parameters together with meteorological variables commonly available in meteorological databases. A novel contribution to model the average photon energy was based on an artificial neural network. A comparative study among some methods proposed by other authors or a multiple linear regression model was carried out based on the same input dataset. The results presenting a lower error between modeled and measured values of the average photon energy are obtained using the artificial neural network, with tiny discrepancies between modeled and measured values. Specifically, the mean absolute percentage error and root mean square percentage error were found to lie below 0.3% and 0.45%, respectively.
Source: EU PVSEC
Authors: Torres-Ramírez, M., García-Domingo, B., Elizondo, D., Nofuentes Garrido, G., Lopez Talavera, D.
Language: English
Field: Energy
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6144276