Neural Network Modeling and Experimental Study of Radon Separation from Water

Abstract

Against the improvement of science and technology, earthquake is one of natural disasters that human cannot predict. Researchers indicated that there are many earthquake precursors. One of these precursors is a change in radon concentration in thermal waters about of active faults. Most of radon monitors cannot detect radon concentration in water directly, and they can detect radon concentration in air. Therefore, radon molecules should be separated from water and transferred to air. In this study, a bubbling system was used for the transfer of radon from water to air. Mathematical and neural network modeling of this system has been performed. The time constant parameter as an indicator of the separation rate was less than 20 min in most of experimental conditions. After validation of the models with experimental data, the effects of water and air flow rates and water temperature on the speed response of this system have been studied.