Evolving ICA and HGAPSO algorithms for prediction of outlet temperatures of constructed solar chimney

Abstract

A small-scale experimental solar chimney consisted of a 3 m diameter air collector and a 2 m tall chimney was built. The absorber and fluid temperatures were recorded based on the practical weather conditions, and the experimental data were evaluated by a hybrid of genetic algorithm with particle swarm optimization (HGAPSO). In next step, imperialist competitive algorithm (ICA) method was applied as an artificial intelligence approach to forecast the temperature changes due to radiation variations. In this case, a data set of 1000 condition parameters for 30 days operation of solar chimney was used, which include eight and four sensors for input and output variables, respectively. Finally, according to the value correlation coefficient (R2) and the mean square error, the results of the trained networks were reported and the temperature prediction was done with high accuracy. The results showed that the solar chimney's experimental data were qualified with no noise.