Using artificial neural networks and multiple regression models to generate economic solar thermal energy
Using artificial neural networks and multiple regression models to generate economic solar thermal energy lead image
As the demand for energy has increased, many energy experts are focusing on developing innovative sources of renewable energy, such as parabolic trough solar collector (PTSC) systems, a mature, reliable, and economical solar thermal technology. The energy produced by PTSC systems has been widely used in various applications, like solar cooking, cooling systems and power plants.
Ajbar et al. simulate and optimize the PTSC system to investigate a simpler and less expensive alternative model by identifying the most influential input variables on the PTSC outlet temperature. The researchers propose using an artificial neural network and four multiple linear regression models to determine which model proved a preferable alternative to existing models.
PTSC systems are composed of a concave sheet made of reflective materials. This sheet reflects and concentrates the incident solar radiation in its focal line, where an absorber tube is located. The concentrated solar energy heats a working fluid inside the tube, and that fluid could be water, air, oil, or nanofluids
“The results showed that the inlet temperature is the parameter that has the most significant relative importance on the PTSC outlet temperature, among others,” said author Wassila Ajbar.
The findings may help predict the PTRC system’s outlet temperature for water heating of the industrial processes or domestic use. This procedure also can be beneficial for optimizing the transfer of energy in heat exchangers used in thermal absorption transformers, among other systems.
Going forward, the researchers will apply other more advanced optimization algorithms based on “swarm” behaviour or communities, such as ant colony optimization, shuffled frog leaping and lion optimization algorithm.
Source: “Identification of the relevant input variables for predicting the parabolic trough solar collector’s outlet temperature using an artificial neural network and a multiple linear regression model,” by Wassila Ajbar, A. Parrales, S. Silva-Martínez, A. Bassam, O. A. Jaramillo, and J. A. Hernández, Journal of Renewable and Sustainable Energy (2021). The article can be accessed at https://doi.org/10.1063/5.0055992 .
