OPTIMIZATION OF INTERMITTENT PUMPING SCHEDULES FOR AQUIFER REMEDIATION USING A GENETIC ALGORITHM1

Authors

  • Wei-Han Liu,

  • Miguel A. Medina Jr.,

  • Wayne Thomann,

  • Warren T. Piver,

  • Timothy L. Jacobs

    Search for more papers by this author
    • 2

      Respectively, Engineer, Geophex, Ltd., 605 Mercury St., Raleigh, North Carolina 27603; Professor, Dept. of Civil and Environmental Engineering, Duke University, Box 90287, Durham, North Carolina 27708; Director, Occupational and Environmental Safety, Box 3914, Med Center, Durham, North Carolina 27710; Engineer, National Inst. of Environmental Health Science, P.O. Box 12233, Research Triangle Park, North Carolina 27709; and Engineer, Sabre, Inc., Mail Drop 7390 TSG, One East Kirkwood Blvd., Southiake, Texas 76092 (E-Mail/Medina: miguel.medina@duke.edu).


  • 1

    Paper No. 99113 of the Journal of the American Water Resources Association.Discussions are open until August 1, 2001.

Abstract

ABSTRACT: Using a genetic algorithm (GA), optimal intermittent pumping schedules were established to simulate pump-and-treat remediation of a contaminated aquifer with known hydraulic limitations and a water miscible contaminant, located within the Duke Forest in Durham, North Carolina. The objectives of the optimization model were to minimize total costs, minimize health risks, and maximize the amount of contaminant removed from the aquifer. Stochastic ground water and contaminant transport models were required to provide estimates of contaminant concentrations at pumping wells. Optimization model simulations defined a tradeoff curve between the pumping cost and the amount of contaminant extracted from the aquifer. For this specific aquifer/miscible contaminant combination, the model simulations indicated that pump-and-treat remediation using intermittent pumping schedules for each pumping well produced significant reductions in predicted contaminant concentrations and associated health risks at a reasonable cost, after a remediation time of two years.

Ancillary