Use of empirical regression and artificial neural network models for estimation of global solar radiation in Dubai, UAE

Hassan A.N. Hejase, Ali H. Assi, Maitha H. Al Shamisi

Research output: Chapter in Book/Report/Conference proceedingChapter

2 Citations (Scopus)

Abstract

The geographical location of the United Arab Emirates (UAE) (latitude between 26° and 32° North and longitude between 51° and 56° East) favors the development and utilization of solar energy. This chapter presents estimation models for the global solar radiation (GSR) in Dubai, UAE. It compares between six empirical regression models and the best of 11 different configurations of artificial neural network (ANN) models. The models have been developed using measured average daily GSR data for 7 years (2002–2008) while the measured data for the years 2009–2010 are used for testing the models. Results of monthly average daily GSR data of all the empirical models for the test period 2009–2010 yield low statistical error parameters and coefficients of determination (R2) better than 96 %. Comparison with ANN models and Solar Radiation (SoDa) Web site data shows that the optimal multilayer perceptron (MLP) ANN model is the best with R2 98 %, and with the lowest statistical error parameters. The results also confirm that a simple linear regression model provides a very good estimation for monthly and daily average GSR data.

Original languageEnglish
Title of host publicationCauses, Impacts and Solutions to Global Warming
PublisherSpringer New York
Pages61-86
Number of pages26
ISBN (Electronic)9781461475880
ISBN (Print)9781461475873
DOIs
Publication statusPublished - Jan 1 2013

Keywords

  • Artificial neural networks
  • Empirical regression
  • Global solar radiation

ASJC Scopus subject areas

  • General Energy

Fingerprint

Dive into the research topics of 'Use of empirical regression and artificial neural network models for estimation of global solar radiation in Dubai, UAE'. Together they form a unique fingerprint.

Cite this