ANN and ANFIS models to predict the performance of solar chimney power plants

A precise model of the behavior of complex systems such as solar chimney power plants (SCPP) would be much beneficial. Also, such a model would be quite contributing to the control of solar chimney operation. In this paper, the identification and modeling of SCPP utilizing ANN and Adaptive Neuro Fuzzy Inference System (ANFIS) are discussed. The modeling is based on the data of three working days which were taken of a built pilot in University of Zanjan, Iran. The input parameters are time, radiation and ambient temperature, while the output is the air velocity at the inlet of the chimney. The results of ANN model and ANFIS model were compared; it was found that ANFIS model exhibited better performance than ANN. The R-Square error of testing in ANFIS is about 0.91, therefore there is good agreement between the ANFIS model and experimental data. Therefore the ANFIS model used to predict the SCPP performance for coming days. A numerical simulation of the problem is conducted to provide a comparison between the conventional method and the presented approach. The results indicated that the performance of solar chimney power plants will be accurately predictable via such a method providing less computational cost. (c) 2015 Elsevier Ltd. All rights reserved.