Utilization of Radial Basis Function Neural Network model for Water production forecasting in Seawater Greenhouse units

The solar seawater greenhouse desalination is a plant that simulates the natural water cycle through seawater evaporation and condensation into freshwater. In the present work, the radial basis function neural network integrated with different optimization algorithms is presented to predict the water production capacity considering the effect of various parameters on the performance seawater greenhouse system. Different statistical metrics are employed to examine the performance of the proposed models. Also, in order to obtain a more satisfactory performance of water production forecasting in a seawater greenhouse, the effect of neurons' numbers in the hidden layer of the proposed neural network is studied. According to the obtained results, the forecasting accuracy of the proposed radial basis function neural network optimized with the hybrid particle swarm optimization-gray wolf optimizer algorithm with nine neurons in the hidden layer with the correlation coefficient and coefficient of determination of 0.998 and 0.996 in training phase is much better than those of the other models. Also, the best values for the front greenhouse dimension are obtained as Width = 125 m, Length = 200 m, and Height = 4 m. Also, the roof transparency is obtained by 0.6. (C) 2021 Published by Elsevier Ltd.

Automated summary

The study introduces a neural network model for predicting water production capacity in solar seawater greenhouse desalination, integrating different optimization algorithms. By investigating the impact of hidden layer neuron numbers, accurate prediction results were obtained. The optimized neural network model showed superior forecasting accuracy under specific parameters.