Abstract

Artificial neural networks (ANNs), inspired by human learning, have allowed an optimal solution to problems in different fields of knowledge. The precise representations of the connections between each of the elements of a complex system determine the analysis and efficient processing of the information. Experimental data from systems involving renewable energy are good candidates to be modeled with ANN due to their variation over time, environmental factors, and complexity. This was exemplified by the development of a neural network model capable of predicting the exit temperature in a low-cost solar parabolic trough collector. The inlet temperature of the fluid, month, day, hour, ambient temperature, glass surface temperature, and reflective surface temperature were the variables considered for the input layer. The best adjustment of training data was acquired with an architecture 7–19–1 considering a sigmoidal hyperbolic tangential transfer function in the hidden layer and a linear transfer function in the output. The coefficient of determination R2 obtained was 0.9894.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call