A simulation-optimization approach for optimal design of dewatering systems in unconfined strata

Visual Modflow software was adopted to numerically simulate the groundwater flow in unconfined strata for six case studies of previously executed projects in Egypt that were in need of dewatering systems. This software was then integrated with a modular groundwater optimizer (MGO) as an optimization model, where the genetic algorithm (GA) was the optimization solver. The models run an unmanaged nonoptimal simulation based on the results of a pumping test and the executed pumping wells to reach the observed drawdown values in the piezometric wells at the construction sites. The models run once more for pumping well optimization using the MGO with the objective of minimizing the number of wells, the total amount of extracted water and, consequently, the cost of dewatering systems. Comparing the drawdown depth (D.D) from the two consecutive runs for each project, the optimization results for the unconfined strata proved that the required D.D can be achieved with only a percentage of the implemented wells ranging from 58 to 75%. The results identify the importance of applying the MGO as a robust method in the design of future dewatering projects.