Inexact Multiphase Modeling System for the Management of Subsurface Contamination under Uncertainty

This study presents an integrated approach with multiphase modeling, factorial design, and Monte Carlo simulation by which to deal with uncertainties in the fate and transport process of petroleum contaminants in a subsurface environment. A set of 24 factorial experiments is conducted for 4 factors, including the permeability and porosity of sandy loam and clay soils, in order to examine the effects of uncertain input characteristics on the modeling outputs. For each experimental set, a number of Monte Carlo simulation runs are undertaken through a well-developed multiphase compositional simulator. Thus, the problem of multiphase flow with uncertain information can be solved. The proposed methodology is applied to a petroleum-contaminated site in western Canada to illustrate its applicability. The simulation results reveal that the uncertainties in input parameters impose considerable influence on the predicted output: variations in the permeability of sandy loam soil especially will have significant impact on the modeling output. The proposed method offers an effective tool for evaluating uncertainty in multiphase flow simulation.