Multi-objective optimization of a solar chimney for power generation and water desalination using neural network

•A hybrid solar chimney is investigated for power generation and water desalination.•Neural networks are trained to predict turbine velocity and water evaporation rate.•Multi-objective optimization is used to maximize the production of power and water.•Geometrical parameters of the system including curvature parameters are optimized.•The most optimal point provides 719 kW of power and 14.28 kg.s-1 of water. Given that the initial cost of constructing a solar chimney system is high, the multi-purpose use of this system for power generation and freshwater production from seawater is vital to make the system economically feasible. To achieve the best balance between the turbine power generation and freshwater production, the optimization of design parameters such as collector’s height and diameter, chimney’s height and diameter, and the curvature of the outer wall of the chimney is necessary. Also, due to the high volume of calculations required for the numerical simulation of any arrangement of parameters, a neural network for the prediction of the output quantities is advantageous. Therefore, a perceptron neural network with two hidden layers has been implemented to predict the average temperature on the collector’s surface and the average air velocity at the turbine inlet to calculate total power and condensed water. Finally, the genetic algorithm is implemented for optimization, and the Pareto frontier is obtained for this problem. The total power generation values and freshwater production corresponding to the most optimal point are 719 kW and 14.28 kg.s-1, respectively.