DESIGN OF OPTIMAL, RELIABLE PLUME CAPTURE SCHEMES - APPLICATION TO THE GLOUCESTER LANDFILL GROUNDWATER CONTAMINATION PROBLEM

A ground-water quality management model is applied to the Gloucester Landfill site, located near Ottawa, Canada, to examine the effectiveness of various single-well pumping schemes for the capture of dissolved contaminants. Deterministic and stochastic design analyses are conducted through ground-water solute transport modeling of the site. The purpose of the modeling analysis is to develop contaminant capture designs that both require minimum pumping rates and possess high probabilities of success. Optimization based upon deterministic simulation indicates that a well located at the front of the plume would effect plume capture and require the lowest pumping rate. However, a smaller total volume of water could be pumped and still effect plume capture if the well were located at the center of the plume and pumped at a higher rate for a shorter time. Stochastic optimization analyses are used to overdesign the pumping rates so that possible design error is overcome. The analyses indicate that design reliability may be increased from 50 to 90 percent by pumping an additional 18 percent at the front or 27 percent at the center of the plume. These pumping overdesign factors are the first such values calculated using the stochastic optimization approach applied to a field site.