A new approach to performance analysis of a seawater desalination system by an artificial neural network

An integrative system of air-conditioning and desalination driven by heat pumps is presented. Recently some analytical methods for the desalination process have been developed. The analytical methods use experimental function of reliability to study the performance of desalination. The numerical methods use some differential equations coupled with heat and mass transfer to simulate the desalination process, but in these methods, some correlative factors are neglected and some hypotheses are ideal, all of these affecting the accuracy and validity of the model. The artificial neural network (ANN) is widely used as technology offering an alternative way to deal with the complex and ill-defined problems. This paper analyzes the seawater desalination process and presents a new approach to simulate the water production ratio of the system using ANN technology. The ANN model of a seawater desalination system for performance prediction has been proposed. Based on the trained ANN model, it can predict the influence of the dry and damp bubble temperature of the air, the inlet and outlet cooling water temperature, and the sprinkler temperature of seawater on the water production ratio for the desalination system. The water production ratio of the ANN model was compared with experimental value; error was small and within an acceptable range. The simulative results show that the application of ANN to seawater desalination is feasible and has the distinctive characteristics of convenient operation, high efficiency and precision.