Title: Prediction of solar pond performance parameters using artificial neural network

Authors: K. Karunamurthy; R. Manimaran; M. Chandrasekar

Addresses: School of Mechanical and Building Sciences, VIT University, Chennai 600127, India ' School of Mechanical and Building Sciences, VIT University, Chennai 600127, India ' Anna University (BIT Campus), Tiruchirappalli 620024, India

Abstract: In this paper, artificial neural networks (ANNs) model was used to predict the performance parameters of a laboratory model salinity gradient solar pond (SGSP), which is used for supplying hot water. Experiments were conducted on three different solar ponds provided with and without twisted tapes in the flow passage of the in-pond heat exchanger during the month of May 2015 at Chennai weather conditions in India. The performance parameters of solar pond such as outlet water temperature, efficiency of solar pond and effectiveness of in-pond heat exchanger were determined experimentally for two different flow rates of Reynolds numbers 1,746 and 8,729. The experimental data obtained from the observations were utilised for training, validating and testing the proposed artificial neural network model. The parameters like incident solar radiation, inlet water temperature, lower convective zone (LCZ) temperature and flow rate are responsible for the outlet water temperature of the solar pond. Based on the experimental readings as inputs a computational program was developed in Python. This program was trained using artificial neural network with back propagation algorithm to predict the outlet water temperature of the in-pond heat exchanger. The results predicted using the model developed is in good agreement with the experimental results.

Keywords: solar pond; performance parameters; artificial neural network; ANN; twisted tapes.

DOI: 10.1504/IJCAET.2019.098133

International Journal of Computer Aided Engineering and Technology, 2019 Vol.11 No.2, pp.141 - 150

Received: 22 Apr 2016
Accepted: 21 Oct 2016

Published online: 26 Dec 2018 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article