Modeling of pollutant transmission in rivers By using Finite volume method and ANFIS model

Abstract

Study in surface water quality is important. Rivers are one of the main sources of water supply for<br />drinking, agriculture and industry. Unfortunately, sometime Rivers where wastewater discharges<br />are considered. For this reason,the pollutant transmission in river is one of the most important<br />problems in Environmental Engineering. the Advection Dispersion Equation (ADE) is governed<br />on the pollutant transport in river. this equation is partial differential equation (PDE). This<br />equation is used most widely in fluid engineering spatially water engineering and in general<br />named convection equation.to Preparate computer programs that can simulate the pollutant<br />transmission, in addition to the used numerical method to solution of the convection equation ,the<br />calculation of Longitudinal Dispersion coefficient (LDC) is needed. Many empirical formulas are<br />presented for calculating the LDC. the assessment of this equations shows that Most of their<br />accuracy is not satisfactory. Thus the use of Artificial Intelligent techniques is inevitable. In this<br />paper, the finite volume method is used discretiz of ADE and to estimate the diffusion coefficient,<br />adaptive neural network (ANFIS) has been developed. the result of ANFIS model shows that The<br />accuracy and performance of the model in training and testing process is very suitable. After the<br />development of an artificial intelligence model general model was developed. the result of the<br />final model tested with analytical solution of ADE and observation data of the Severn River in the<br />UK. the final result show that performance of model is suitable.<br />Keywords: transmission of pollutant, finite volume method, Severn river