Groundwater contaminant source identification by a hybrid heuristic approach

Abstract

This study proposes an approach, SATS-GWT, that combines simulated annealing (SA), tabu search (TS), and the three-dimensional groundwater flow and solute transport model (MODFLOW-GWT). Our approach is used to estimate the source information: source location, release concentration, and release period. The sampling concentrations at monitoring points are simulated by the MODFLOW-GWT with an assumed release concentration and release period at a known source location. In the source estimation process, the source location is selected by TS within the suspected source area, and the trials for release concentrations and release periods are generated by SA. The MODFLOW-GWT is employed to compute the simulated concentrations at the monitoring points with the trial solution. The above mentioned procedures are repeated until the stopping criterion regarding the differences of objective function value (OFV) is met. The last trial source information which yields the best OFV is the final solution. Six studies on the homogeneous site, two studies on the heterogeneous site, and one study on the transient flow problem are conducted in this study. A suggestion regarding the optimal number of monitoring points and the condition for estimating the source information is given on the basis of the studies for homogeneous and heterogeneous aquifers. All results indicate that the proposed SATS-GWT can give good estimations, even when the sampling concentrations contain measurement errors.