Reliability and Remediation Cost of Optimal Remediation Design Considering Uncertainty in Aquifer Parameters

Abstract

Reliability of a “preassumed” optimal remediation design using a pump and treat method is evaluated. A “true” hydraulic conductivity field with a certain spatial correlation is supposed. Hydraulic conductivity values are sampled at several locations from the true field. The preassumed remediation design is optimized in the hydraulic conductivity field composed by interpolation of the sampled data. This design is applied to random conductivity fields generated by using conditional simulations in which the conductivity values sampled from the true field are used as the conditioning points. The remediation costs for each generated conductivity field are computed when a preassumed optimal remediation design is applied. The cumulative distribution functions (CDFs) of the required remediation costs are obtained. The CDFs of the remediation cost show that the density of sampling and concentrated sampling in a highly contaminated zone have a great influence on the reliability and uncertainty of the preassumed optimal remediation design.