Abstract
The finite element method is one of the typical methods that are used for numerical water quality modeling of the topographically complicated river. In this paper, based on the principle of probability theory the probability density of pollutants is introduced. A new model for the grid size optimization based on the finite element method is developed with the incorporation of the maximum information entropy theory when the length of the grid is given. Combined with the experiential evaluation approach of the flow discharge per unit river width, this model can be used to determine the grid size of the finite element method applied to water quality modeling of the topographically complicated river when the velocity field of the river is not given. The calculating results of the application of the model to an ideal river testified the correctness of the model. In a practical case-the application of the model to the Xingjian River (the Hengyang section of the Xiangjiang River), the optimized width of the grid of the finite element method was gained and the influence of parameters was studied, which demonstrated that the model reflected the real situation of the pollutants in the river, and that the model had many excellent characteristics such as stabilization, credibility and high applicability in practical applications.