Document Type

Conference Paper

Journal/Book Title/Conference

Petroleum Hydrocarbons and Organic Chemicals in Ground Water: Prevention, Detection, and Restoration

Publisher

The American Petroleum Institute

Publication Date

11-1993

First Page

285

Last Page

299

Abstract

An innovative approach is presented to minimize pumping for immobilizing a floating plume of a light non-aqueous phase liquid (LNAPL). The best pumping strategy is determined to contain the free oil product and provide for gradient control of the water table. This approach combined detailed simulation, statistical analysis, and optimization. This modeling technique uses regression equations that describe system response to variable pumping stimuli. The regression equations were developed from analysis of systematically performed simulations of multiphase flow in an areal region of an unconfined aquifer. Simulations were performed using ARMOS, a finite element model. ARMOS can be used simulate a spill, leakage from subsurface storage facilities and recovery of hydrocarbons from trenches or pumping wells to design remediation schemes. Two gradient control points were located inside the area of the symmetric floating plume. Air-oil interface drawdowns with respect to water pumping rates were taken from ARMOS simulations at the two locations. These drawdowns were used to calculate elevation changes in air-oil table elevations (Zao) between the control points. These elevation changes of Zao between Points #1 and #2 versus pumping were plotted and fitted by statistical regression analysis for a pumping range of 150m3/day to 240m3/day. The resulting regression equation was used to represent system response to pumping in the simulation/optimization (S/0) model called Utah State Model for Optimizing Management of Stream/ Aquifer Systems Using the Response Matrix Method (US!REMAX). The containment problem was then optimized by US/REMAX to determine the minimum pumping rate required to reverse the water table gradient and immobilize the floating plume. Once regression equations are developed the optimal pumping state for alternative containment goals and scenarios can be quickly determined. A range of gradient control values can be easily evaluated to determine minimized pumping rates. Then, their impacts can be compared between alternatives for minimum pumping versus time to containment, residual or trapped oil volumes, and free oil area at containment time.

Download

Share

COinS