Aquifer parameter and zonation structure estimation using meshless local Petrov-Galerkin method and particle swarm optimization

Abstract

The simulation-optimization (SO) modeling approach can be effectively used for aquifer parameter estimation. In this study, a numerical approach based on meshless local Petrov-Galerkin (MLPG) method is used for groundwater flow simulation and coupled with particle swarm optimization model for optimization. The study deals with the identification of the most suitable model structure for a hypothetical heterogeneous confined aquifer from a number of alternate models using zonation method of parameter estimation. A range of alternate models starting from homogeneous to more complex model structures are considered for the zonation. Inverse modeling of different model structures is carried out based on weighted least square performance criterion. The suitable models are selected and reliability analysis ascertained by computing three parameters of composite scaled sensitivity, coefficient of variation, and confidence interval, and the best model is selected. Sensitivity of estimated parameters is investigated by considering different sets of head data involving possible measurement errors. The solutions are compared with another inverse model using the MLPG and Levenberg-Marquardt algorithm. Based on the results, it is found that the proposed methodology can be utilized in the estimation of different unknown parameters in a regional groundwater system.