Groundwater Flow and Contaminant Transport Simulation with Imprecise Parameters

Abstract

A methodology has been developed in this study wherein a genetic algorithm (GA) is used to find a global optimal solution to a groundwater flow and contaminant problem by incorporating an artificial neural network (ANN) to evaluate the objective function within the genetic algorithm. The study shows that an ANN-GA technique can be used to find the uncertainties in output parameters due to imprecision in input parameters. The ANN-GA methodology is applied to five case studies involving radial flow in a well, one-dimensional solute transport in steady uniform flow, a two-dimensional heterogeneous steady flow, a two-dimensional solute transport, and a two-dimensional unsteady groundwater flow to demonstrate the efficiency and effectiveness of the developed algorithm. The results show that, with this approach, one can successfully measure the uncertainty in groundwater flow and contaminant transport simulations and achieve a considerable reduction in computational effort when compared to the vertex method that has been widely used in the past.