Artificial neural network approach for modeling and prediction of energy consumption in a seawater greenhouse

Seawater greenhouse using humidification-dehumidification method can desalinate saline water and utilize fresh water for the greenhouse and drinking. Many parameters affect the performance of the seawater greenhouse. In this study, the effect of the width and length of the greenhouse, the height of the first evaporator and the roof transparency parameters on the energy consumption in the seawater greenhouse were investigated with the artificial neural network method. Artificial neural networks of the multi-layer perceptron (MLP) have been used for modeling. An appropriate structure for this method was obtained and the mathematical statistics of %AARE, RMSE and R2 were used to evaluate the network performance. The existing method is in good agreement with experimental data. Using this optimized network, the effect of each parameter on the energy consumption was evaluated. Finally, a greenhouse with a width of 125 meters, a length of 200 meters, an evaporator height of 4 meters, and a roof transparency of 0.6, which produces 161.6 m3/day of fresh water and 1.558 kWh /m3 of energy consumption, was introduced as an optimal seawater greenhouse.