Estimation of volumetric water content during imbibition in porous building material using real time neutron radiography and artificial neural network

Movement of fluids through porous media is of great significance in different sciences such as civil engineering and chemical technology. When a fluid is introduced to a porous media, the fluid starts moving and diffusing through that. In this condition, the amount of fluid in each position of the porous material is a function of the distance between the desired position and the fluid movement starting point as well as the elapsed time. In this research, a novel methodology is proposed for estimating the volumetric water content in different times and positions during the water imbibition inside the porous building materials using a combination of real time neutron radiography technique and artificial neural network (ANN). For this purpose, a brick as a porous construction sample was positioned in contact with a water container and 145 neutron radiographs were recorded during the water imbibition inside the brick. To extract the required data for training, testing and validating the ANN, a line of pixels in the center of brick's image along with the water movement direction inside the sample was considered. The calculated volumetric water content in each position of the mentioned line was used as the output of the ANN. In addition, the elapsed time and position were utilized as the two inputs of the ANN. After training, the proposed ANN model could estimate the volumetric water content in different times and positions in direction of water movement inside the brick with a mean square error (MSE) of less than 8.4$\times$10⁻⁴.