Two-dimensional modeling of groundwater flow using a Meshless Local Petrov-Galerkin method based on RBF radial function (Case study: Rafsanjan plain)

In this investigation, the Meshless Local Petrov-Galerkin (MLPG) method is used to model flow in the porous media of Rafsanjan plain. The discretization of the governing equations was done using the partial residual method and the radial basis function, and the model was implemented in two dimensions. A circular support domain was used for shape functions. The research focuses on the Rafsanjan Plain in southeastern Iran as a real case study, where no studies of meshless methods have been done. Computational models were implemented using MATLAB, integrating Radial Basis Functions (RBF) as the interpolation technique. The comparative analysis between observed experimental data and modeled values in flow dynamics showed a significant alignment. The sensitivity analysis of the parameters showed that the value of the shape parameter (αc) can have the greatest effect in achieving a more accurate prediction value. The accuracy of the calibrated parameters was evaluated by the coefficient of determination (R2=0.97). The high coefficient of determination shows a strong correlation between the observed and simulated data, which indicates that the model parameters are within an acceptable range.